PLACE IN RETURN BOX to remove this checkout from your record.
To AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE

DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5/08 K:/Proj/Acc&Pres/CIRC/DaIeDue Indd

 

AQUEOUS PHASE CATALYTIC HYDROGENATION AND DEUTERIUM
EXCHANGE OF a-SUBSTITUTED ORGANIC ACIDS USING Ru/C
By

Jennifer Elizabeth Jacobs

A DISSERTATION
Submitted to
Michigan State University

in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Chemistry

2009 ’

ABSTRACT

AQUEOUS PHASE CATALYTIC HYDROGENATION AND DEUTERIUM
EXCHANGE OF a-SUBSTITUTED ORGANIC ACIDS USING Ru/C

By

Jennifer Elizabeth Jacobs
The biomass refinery of the future will require a wide range of catalytic conversions,
largely aqueous phase reductions. This dissertation explores the use of a 5% ruthenium
on carbon catalyst in hydrogenation and C-H activation studies of carboxylic acids, a
significant class of biogenic feedstocks. Several two- and three-carbon organic acid
substrates were studied to probe (a) the effects of a- substituents on hydrogenation and
H/D exchange reactions and (b) the relationship between these two processes. Substrates
included the following groups: -H, -CH3, -OH, -N+H3, -OCH3, -N+H2CH3, -N+H(CH3)2,
and -N+(CH3)3. Most reactions were performed at 100 °C or 130 °C in D2O solution but
ranng from 50 °C to 150 oC and 1000 psi H2 or D2. As expected, both catalyst and H2
gas are required to convert organic acids to the corresponding alcohols. Simple alkanoic
acids (acetic, propanoic, and isobutyric) and betaine, ((CH3)3N+CH2COOH-Cl'), a
quatemized glycine, are very unreactive. For the other methylated glycines, the rate of
hydrogenation falls off by a factor of three with each additional methyl, and reduction
requires acid, as found earlier by Jere. Use of water as reaction medium appears optimal;
addition of organic co-solvents such as ethanol, ethylene glycol, THF, or ethyl lactate led
to reduced hydrogenation rates. This apparent inhibition by organic additives may reflect
their tendency to tie up catalytic sites via metal insertion into C-H bonds, especially those
neighboring heteroatoms. Another feature of both the solvents and the substituents is

their capacity for hydrogen bonding, which may play a role in modulating hydrogenation

rates. Similarly, steric crowding may also affect the substrates’ ability to access the
surface in productive orientations. To explore this issue, activation of OH bonds next
door to amine N sites was probed via H/D exchange studies performed on the N-

methylated glycine series.

Reactions run in D20 with either H2 or D2 and catalyst were effective at exchanging
deuterium into the CH2 and CH3 sites of the amino acids. The Ru catalyst also exchanges
H (from H2) for D (from D20) to form HOD and HD, independent of substrate. Analysis
by internal standard-calibrated 1H NMR allowed complete isotopomer-specific inventory
of H concentration in substrates and water. Simple sequential replacements of H by D
were observed in the amino acids. Two kinetic models were developed to determine the
rate constants for H/D exchange at both the methylene and methyl positions. Both

included possible time delays due gas/liquid isotopic equilibration (kH) and/or a catalyst

induction (ktum_on) period; ultimately the latter was discarded as redundant. The first

model treated each isotope replacement step as a process unrelated to the others. Thus,
between forward and reverse replacements in CH2 and CH3 sites and k”, eleven rate
constants were freely varied in the fitting. The second model used a single rate constant
for each site, further modified by primary and secondary isotope effect values, resulting
in six values + kg for this fitting. The latter, more economical and interpretive model
gave fits nearly as good as those from the eleven-variable one. Somewhat surprisingly,
these analyses found H/D exchange rates to decrease in the following order: D2O >

sarcosine (CH3NH2+CH2CO2') > glycine (CH2 site) > N,N-dimethy1glycine.

TABLE OF CONTENTS

LIST OF TABLES ............................................................................................................. vi
LIST OF FIGURES .......................................................................................................... vii
LIST OF SCHEMES ........................................................................................................ xiv
LIST OF EQUATIONS ................................................................................................... xvi
Chapter 1. INTRODUCTION ............................................................................................ 1
1.1 Background ............................................................................................................... 1
1.2 Significance .............................................................................................................. 1
1.3 Literature Review ..................................................................................................... 4
1.3.1 Hydrogenation/Dehydrogenation ...................................................................... 4
1.3.2 H/D Exchange ................................................................................................. 18
1.3.3 Alkynylation of sp3 C-H Bonds Adjacent to a Nitrogen Atom ...................... 33
1.3.4 Secondary Deuterium Isotope Effects for Enolization Reactions ................... 35
1.3.5 Catalytic Hydrogenation of on-Amino and a-Hydroxy Esters ........................ 37
1.3.6 Catalytic Properties of Ru/C ........................................................................... 39
1.3.7 Hydrogenolysis of Glycerol to Propylene Glycol ........................................... 40
1.3.8 Deuterium NMR ............................................................................................. 43

1.4 Summary of Previous Work ............................................................................... 45
Chapter 2. EXPERIMENTAL ......................................................................................... 68
2.1 Hydrogenation Studies ............................................................................................ 68
2.1.1 Materials .......................................................................................................... 68
2.1.2 Batch Reactor System ..................................................................................... 69
2.1.3 Reaction Procedure ......................................................................................... 70
2.1.4 Organic Acid Analysis .................................................................................... 71
2.1.4.1 HPLC Analysis ........................................................................................ 71
2.1.4.2 pH Analysis .............................................................................................. 72

2.2 H/D Exchange Studies ........................................................................................... 72
2.2.1 Materials ......................................................................................................... 72
2.2.2 Batch Reactor System ..................................................................................... 73
2.2.3 Reaction Procedure ......................................................................................... 73
2.2.4 Amino Acid Analysis ...................................................................................... 74
2.2.4.1 HPLC Analysis ........................................................................................ 75
2.2.4.2 'H NMR ................................................................................................... 75
Chapter 3. CALCULATIONS ......................................................................................... 76
3.1 1H NMR Peak Fitting and Deconvolution ............................................................. 76
3.1.1 NMR Calibration Calculations ....................................................................... 76
3.1.2 Peak Fitting Equations .................................................................................... 77
3.1.3 Peak Integration Calculations ......................................................................... 80

3.2 H-D Content Inventory .......................................................................................... 82

iv

3.2.] Percent H in Water .......................................................................................... 82

3.2.2 Percent H in Sample ........................................................................................ 83
3.3 Kinetic Rate Data Determination ........................................................................... 88
3.3.1 Concentration Calculations ............................................................................. 88
3.3.2 Kinetic Model A: 10-variables ....................................................................... 91
3.3.3 Kinetic Model B -— 6 variables ........................................................................ 93
Chapter 4. RESULTS AND DISCUSSION .................................................................... 98
4.1 Control Experiments .............................................................................................. 98
4.1.1 Hydrogenation Studies .................................................................................... 98
4.1.2 H/D Exchange Studies .................................................................................. 102
4.2 Substrate Adsorption on the Catalyst/Carbon Support ........................................ 113
4.2.1 Adsorption on Ru/C ...................................................................................... 113
4.2.2 Adsorption on Carbon Support ..................................................................... 115
4.3 Effect of HCl ........................................................................................................ 116
4.3.1 Hydrogenation Studies ................................................................................... 116
4.3.2 H/D Exchange Studies .................................................................................. 122
4.4 Effect of H3PO4 .................................................................................................... 125
4.4.1 Hydrogenation Studies .................................................................................. 125
4.4.2 H/D Exchange Studies .................................................................................. 129
4.5 Solvent Effects on the Hydrogenation of Organic Acids ..................................... 143
4.6 Organic Acid Hydrogenation Studies .................................................................. 156
4.6.1 Effect of the a-Substituent on the Hydrogenation Reaction ......................... 156
4.6.2 Amino Acid Substrates ................................................................................. 162
4.7 H/D Exchange Studies ......................................................................................... 166
4.7.1 Amino Acid Substrates ................................................................................. 166
4.7.1.1 H/D Exchange at the Methylene Position .............................................. ' 166
4.7.1.2 I-I/D Exchange at the N-Methyl Position ............................................... 216
4.7.2 Other Substrates ............................................................................................ 237
4.8 Competition Experiments .................................................................................... 242
4.8.1 Hydrogenation Studies .................................................................................. 242
4.8.2 H/D Exchange Studies .................................................................................. 246
Chapter 5. SUMMARY ................................................................................................. 251
Chapter 6. REFERENCES ............................................................................................. 261

Table I.
Table 2.

Table 3.

Table 4.

Table 5.

Table 6.

Table 7.

Table 8.
Table 9.

LIST OF TABLES

Carbon characterization by N2 adsorption at 78 K ............................................ 53

Hydrogenation of C3 and C2 organic acids to alcohols after six hours (0.3 M
H3PO4, H20, 130 °C, 1000 psi H2, 1 g dry 5% Ru/C) ..................................... 161

Hydrogenation of glycine, sarcosine, N,N-dmg, and betaine at 100, 130, and
150 °C after six hours (0.3 M H3PO4, H2O, X °C, 1000 psi H2, 1g dry 5% Ru/C)

......................................................................................................................... 162
Rates of reaction for hydrogenation of glycine, sarcosine, and N,N-dmg (0.3 M
H3PO4, H20, 100, 130, and 150 °C, 1000 psi H2, 1g dry 5% Ru/C) ............... 163
Rate constants for glycine (100 °C, 1000 psi H2 or D2) using the original kinetic
model, k10 + kH, k6 + kH ................................................................................. 190
Rate constants for sarcosine (100 °C, 1000 psi H2 or D2) using the original
kinetic model, k10 + k”, k6 + kH ..................................................................... 193
Rate constants for N,N-dimethylglycine (100 °C, 1000 psi H2 or D2) using the
original kinetic model, k10 + k“, k6 + kH ....................................................... 196
pKa values for glycine, sarcosine, and N,N-dimethylglycine (25 °C) ............ 210

Summary of the rate kinetic rate constants and primary and secondary isotope
effects for glycine, sarcosine, and N,N-dimethylglycine ................................ 236

Table 10. Summary of rate constants for glycine, sarcosine, and N,N-dimethylglycine

......................................................................................................................... 258

vi

LIST OF FIGURES

Figure 1. Hydrogenation of 0.25 M refluxed lactic acid (0.3 M H3PO4, H20, 130 0C,
1000 psi He, 1 g dry 5% Ru/C) ......................................................................... 98

Figure 2. Hydrogenation of 0.25 M refluxed lactic acid (0.3 M H3PO4, H20, 130 °C,
1000 psi H2, no catalyst) ................................................................................... 99

Figure 3. Hydrogenation of refluxed lactic acid (0.3 M H3PO4, H20, 130 °C, 1000 psi
H2, 1 g dry 5% Ru/C) ...................................................................................... 101

Figure 4. Hydrogenation of 0.25 M glycolic acid in the absence and presence of 0.3 M
H3PO4 (H20, 130 °C, 1000 psi H2, 1 g dry 5% Ru/C) ..................................... 102

Figure 5. 1H NMR of H/D exchange at the methylene position of 0.25 M glycine (D2O,
130 °C, 1000 psi H2, no catalyst) .................................................................... 103

Figure 6. 1H NMR of H/D exchange at the methylene position of 0.25 M sarcosine (D20,
130 °C, 1000 psi H2, no catalyst) .................................................................... 104

Figure 7. 1H NMR of H/D exchange at the methyl position of 0.25 M sarcosine (D2O,
130 °C, 1000 psi H2, no catalyst) .................................................................... 104

Figure 8. Fit of the kinetic rate data for 0.25 M sarcosine (D2O, 130 °C, 1000 psi H2, no
catalyst) ........................................................................................................... 106

Figure 9. 1H NMR of H/D exchange at the methylene position next to the —OH of 0.25 M
choline bicarbonate (D20, 130 0C, 1000 psi H2, no catalyst) ......................... 108

Figure 10. lH NMR of H/D exchange at the methylene position next to the —N+(CH3)3 of
0.25 M choline bicarbonate (D2O, 130 °C, 1000 psi H2, no catalyst) ............. 108

Figure l 1. 1H NMR of H/D exchange at the methylene position of 0.25 M betaine (D2O,
130 °C, 1000 psi H2, no catalyst) .................................................................... 110

Figure 12. 1H NMR of H/D exchange at the methylene position of 0.25 M betaine (D20,
130 °C, 1000 psi H2, 1g dry 5% Ru/C) ........................................................... 110

Figure 13. 1H NMR of H/D exchange at the methyl position of 0.25 M betaine (D20, 130
°C, 1000 psi H2, no catalyst) ........................................................................... 111

Figure 14. 1H NMR of H/D exchange at the methyl position of 0.25 M betaine (D2O, 130
°C, 1000 psi H2, 1g dry 5% Ru/C) .................................................................. 111

Figure 15. Concentrations of CH2, CHD, and CD2 of 0.25 M betaine with and without 1
g dry 5% Ru/C catalyst (D2O, 130 °C, 1000 psi H2) ....................................... 113

Figure 16. 0.25 M Refluxed lactic acid concentrations measured (bench and reactor). 114

Figure 17. Substrate adsorption on the carbon support .................................................. 116

Figure 18. Hydrogenation of 0.25 M betaine hydrochloride (0.3 M H3PO4, H20, 100 °C
and 130 °C, 1000 psi H2, 1 g dry 5% Ru/C) .................................................... 118

Figure 19. Hydrogenation of 0.25 M betaine after 6 hours (0.3 M H3PO4, H20, 100 °C,
130 °C, and 150 °C, 1000 psi H2, 1 g dry 5% Ru/C) ...................................... 118

vii

Figure 20. Hydrogenation of 0.25 M glycine in the absence and presence of 0.3 M HCl

(0.3 M H3PO4, H20, 130 °C, 1000 psi H2, 1 g dry 5% Ru/C) ......................... 120
Figure 21. Hydrogenation of glycine alone and in the presence of NaCl or HCl (H3PO4,
H20, 130 °C, 1000 psi H2, 1 g dry 5% Ru/C) ................................................. 121

Figure 22. 1H NMR of H/D exchange at the methyl (3.2 ppm) and methylene position
(4.1 ppm) on 0.25 M betaine hydrochloride (D2O, 100 °C, 1000 psi H2, 1 g dry

5% Ru/C) ......................................................................................................... 123
Figure 23. Fit of the kinetic rate data for 0.25 M betaine (D20, 130 °C, 1000 psi H2. 1 g
dry 5% Ru/C) .................................................................................................. 124
Figure 24. Hydrogenation of 0.25 M glycolic acid in the absence and presence of 0.3 M
H3PO4 (H20, 130 °C, 1000 psi H2, 1 g dry 5% Ru/C) .................................... 125
Figure 25. Hydrogenation of 0.25 M glycine (no H3PO4, H20, 130 °C, 1000 psi H2, 1 g
dry 5% Ru/C) .................................................................................................. 126
Figure 26. Hydrogenation of 0.25 M glycine (0.3 M H3PO4, H20, 130 °C, 1000 psi H2, 1
g dry 5% Ru/C) ............................................................................................... 127
Figure 27. Hydrogenation of 0.25 M glycine in the absence and presence of 0.3 M
H3PO4 (H20, 130 °C, 1000 psi H2, 1 g dry 5% Ru/C) ..................................... 128
Figure 28. 1H NMR of H/D exchange at the methylene position of 0.25 M glycine (0.3 M
H3PO4; D20; 50 °C; 1000 psi D2, 1 g dry 5% Ru/C) ...................................... 130
Figure 29. 1H NMR of H/D exchange at the methylene position of 0.25 M glycine (D20;
50 °C; 1000 psi D2, 1 g dry 5% Ru/C) ............................................................ 130
Figure 30. Concentration data for 0.25 M glycine (0.3 M H3PO4, D20, 50 °C, 1000 psi
D2, 1 g dry 5% Ru/C) ...................................................................................... 131
Figure 31. 1H NMR of H/D exchange at the methylene position of 0.25 M sarcosine (0.3
M H3PO4, D20, 50 °C, 1000 psi D2, 1 g dry 5% Ru/C) .................................. 133
Figure 32. 1H NMR of H/D exchange at the methylene position of 0.25 M sarcosine
(D20, 50 °C, 1000 psi D2, 1 g dry 5% Ru/C) .................................................. 133
Figure 33. Concentration data for 0.25 M sarcosine (0.3 M H3PO4, D20, 50 °C, 1000 psi
D2, 1 g dry 5% Ru/C) ...................................................................................... 135
Figure 34. 1H NMR of H/D exchange at the methyl position of 0.25 M sarcosine (0.3 M
H3PO4, D20, 50 °C, 1000 psi D2, 1 g dry 5% Ru/C) ....................................... 135
Figure 35. 1H NMR of H/D exchange at the methylene position of 0.25 M sarcosine (0.3
M H3PO4, D20, 70 °C, 1000 psi D2, 1 g dry 5% Ru/C) .................................. 137
Figure 36. 1H NMR of H/D exchange at the methyl position of 0.25 M sarcosine (0.3 M
H3PO4, D20, 70 °C, 1000 psi D2, 1 g dry 5% Ru/C) ....................................... 137
Figure 37. Fit of the kinetic rate data for 0.25 M sarcosine (0.3 M H3PO4, D20, 70 °C,
1000 psi D2, 1 g dry 5% Ru/C) ........................................................................ 138
Figure 38. Fit of the kinetic rate data for 0.25 M sarcosine (0.3 M H3PO4, D20, 70 °C,
1000 psi D2, 1 g dry 5% Ru/C) ........................................................................ 138

viii

Figure 39. Rate constants for 0.25 M sarcosine in the absence and presence of 0.3 M

H3PO4 (D20, 70 °C, 1000 psi D2, 1 g dry 5% Ru/C) ....................................... 140
Figure 40. 1H NMR of H/D exchange at the methylene position of 0.25 M betaine (0.3 M
H3PO4, D2O, 100 °C, 1000 psi H2, 1 g dry 5% Ru/C) ..................................... 141
Figure 41. 1H NMR of I-I/D exchange at the methyl position of 0.25 M betaine (0.3 M
H3PO4, D2O, 100 °C, 1000 psi H2, 1 g dry 5% Ru/C) ..................................... 141
Figure 42. Fit of the kinetic rate data for 0.25 M betaine (0.3 M H3PO4. D20, 100 °C,
1000 psi H2, 1 g dry 5% Ru/C) ........................................................................ 143
Figure 43. .Hydrogenation of 0.25 M glycolic acid after 6 hours (THF/H2O, 130 °C, 1000
psi H2, 1 g dry 5% Ru/C) ................................................................................. 145
Figure 44. Hydrogenation of 0.25 M glycolic acid (0.3 M H3PO4, CH3CH2OH, 130 °C,
1000 psi H2, 1 g dry 5% Ru/C) ........................................................................ 148
Figure 45. Hydrogenation of 0.25 M glycolic acid (CH3CH2OH, 130 °C, 1000 psi H2, 1 g
dry 5% Ru/C) .................................................................................................. 148
Figure 46. Hydrogenation of 0.25 M glycolic acid in the absence and presence of 0.3 M
H3PO4 (CH3CH2OH, 130 °C, 1000 psi H2, 1 g dry 5% Ru/C) ........................ 149
Figure 47. Hydrogenation of 0.25 M ethyl lactate (H20, 130 °C, 1000 psi H2, 1 g dry 5%
Ru/C) ............................................................................................................... 150
Figure 48. Hydrogenation of 0.25 M glycolic acid after 6 hours (CH3CH2OH/H2O, 130
°C, 1000 psi H2, 1 g dry 5% Ru/C) ................................................................. 152
Figure 49. Hydrogenation of C3 organic acids to alcohols (0.3 M H3PO4, H20, 130 °C,
1000 psi H2, 1 g dry 5% Ru/C) ........................................................................ 159
Figure 50. Hydrogenation of mostly C2 organic acids to alcohols (0.3 M H3PO4, H20,
130 0C, 1000 psi H2, 1 g dry 5% Ru/C) .......................................................... 160
Figure 51. 1H NMR of H/D exchange at the methylene position of 0.25 M glycine (D2O,
130 °C, 1000 psi H2, 1 g dry 5% Ru/C) .......................................................... 167
Figure 52. Stacked lH NMR of H/D exchange at the methylene position on 0.25 M
glycine (D20, 100 °C, 1000 psi H2, 1 g dry 5% Ru/C) ................................... 168
Figure 53. Fit of the kinetic rate data for 0.25 M glycine (D20, 100 °C, 1000 psi H2, 1 g
dry 5% Ru/C) .................................................................................................. 169
Figure 54. 1H NMR of H/D exchange at the methylene position of 0.25 M sarcosine
(D20, 100 °C, 1000 psi H2, 1 g dry 5% Ru/C) ................................................ 171
Figure 55. Fit of the kinetic rate data for 0.25 M sarcosine (D20, 100 °C, 1000 psi H2, 1
g dry 5% Ru/C) ............................................................................................... 171
Figure 56. lH NMR of H/D exchange at the methylene position of 0.25 M N,N-dmg
(D2O, 100 °C, 1000 psi H2, 1 g dry 5% Ru/C) ................................................ 173
Figure 57. Fit of the kinetic rate data for 0.25 M N,N-dmg (D2O, 100 °C, 1000 psi H2, 1
g dry 5% Ru/C) ............................................................................................... 173

ix

Figure 58. 1H NMR of H/D exchange at the methylene position of 0.25 M glycine (D20,

100 °C, 1000 psi D2, 1 g dry 5% Ru/C) .......................................................... 178
Figure 59. Fit of the kinetic rate data for 0.25 M glycine (D20, 100 °C, 1000 psi D2, 1 g
dry 5% Ru/C) .................................................................................................. 178
Figure 60. 1H NMR of H/D exchange at the methylene position of 0.25 M sarcosine
(D20, 100 °C, 1000 psi D2, 1 g dry 5% Ru/C) ................................................ 180
Figure 61. Fit of the kinetic rate data for 0.25 M sarcosine (D2O, 100 0C, 1000 psi D2, 1
g dry 5% Ru/C) ............................................................................................... 180
Figure 62. 1H NMR of H/D exchange at the methylene position of 0.25 M N,N-dmg
(D20, 100 °C, 1000 psi D2, 1 g dry 5% Ru/C) ................................................ 182
Figure 63. Fit of the kinetic rate data for 0.25 M N,N-dmg (D20, 100 °C, 1000 psi D2, 1
g dry 5% Ru/C) ............................................................................................... 182
Figure 64. Rate constants for H/D exchange at the methylene carbon (D2O, 100 °C, 1000
psi D2, 1 g dry 5% Ru/C) ................................................................................. 183

Figure 65. Kinetic rate constant and primary and secondary isotope effects for glycine,
sarcosine, and N,N-dmg (D20, 100 °C, 1000 psi D2, 1 g dry 5% Ru/C) ........ 185

Figure 66. Percent H in D2O with either 1000 psi H2 or D2 .......................................... 186
Figure 67. Fit of the kinetic rate data for 0.25 M glycine (D2O, 100 °C, 1000 psi H2 or
linked H2 and D2, 1 g dry 5 % Ru/C) .............................................................. 191
Figure 68. Fit of the kinetic rate data for 0.25 M glycine (D20, 100 0C, 1000 psi D2 or
linked H2 and D2, 1 g dry 5 % Ru/C) .............................................................. 191
Figure 69. Kinetic rate constant and primary and secondary isotope effects for glycine
with 1000 psi H2 or 1000 psi D2 or H2 and D2 ................................................ 192
Figure 70. Fit of the kinetic rate data for 0.25 M sarcosine (D2O, 100 0C, 1000 psi H2 or
linked H2 and D2, 1 g dry 5 % Ru/C) .............................................................. 194
Figure 71. Fit of the kinetic rate data for 0.25 M sarcosine (D20, 100 °C, 1000 psi D2 or
linked H2 and D2, 1 g dry 5 % Ru/C) .............................................................. 194
Figure 72. Kinetic rate constant and primary and secondary isotope effects for sarcosine
with 1000 psi H2 or 1000 psi D2 or H2 and D2 ................................................ 195
Figure 73. Fit of the kinetic rate data for 0.25 M N,N-dmg (D2O, 100 °C, 1000 psi H2 or
linked H2 and D2, 1 g dry 5 % Ru/C) .............................................................. 197
Figure 74. Fit of the kinetic rate data for 0.25 M N,N-dmg (D2O, 100 °C, 1000 psi D2 or
linked H2 and D2, 1 g dry 5 % Ru/C) .............................................................. 197
Figure 75. Kinetic rate constant and primary and secondary isotope effects for N,N-dmg
with 1000 psi H2 or 1000 psi D2 or H2 and D2 ................................................ 198
Figure 76. 1H NMR of H/D exchange at the methylene position of 0.25 M glycine (D20,
70 °C, 1000 psi D2, 1 g dry 5% Ru/C) ............................................................ 199

Figure 77. Fit of the kinetic rate data for 0.25 M glycine (D20, 70 °C, 1000 psi D2, 1 g

dry 5% Ru/C) .................................................................................................. 199
Figure 78. 1H NMR of H/D exchange at the methylene position of 0.25 M sarcosine
(D20, 70 °C, 1000 psi D2, 1 g dry 5% Ru/C) .................................................. 201
Figure 79. Fit of the kinetic rate data for 0.25 M sarcosine (D20, 70 °C, 1000 psi D2, 1 g
dry 5% Ru/C) .................................................................................................. 201
Figure 80. ‘H NMR of H/D exchange at the methylene position of 0.25 N,N-dmg (D20,
70 °C, 1000 psi D2, 1 g dry 5% Ru/C) ............................................................ 202
Figure 81. Fit of the kinetic rate data for 0.25 M N,N-dmg (D20, 70 °C, 1000 psi D2, 1 g
dry 5% Ru/C) .................................................................................................. 202
Figure 82. Rate constants for H/D exchange at the methylene carbon (D20, 70 °C, 1000
psi D2, 1 g dry 5% Ru/C) ................................................................................. 203
Figure 83. Kinetic rate constant and primary and secondary isotope effects for glycine,
sarcosine, and N,N-dmg (D20, 70 °C, 1000 psi D2, 1 g dry 5% Ru/C) .......... 205
Figure 84. Rate constants for H/D exchange at the methylene carbon at 100 °C or 70 °C
(D2O, X °C, 1000 psi D2, 1 g dry 5% Ru/C) ................................................... 206
Figure 85. Fit of the kinetic rate data for 0.25 M glycine (D20, 50 0C, 1000 psi D2. 1 g
dry 5% Ru/C) .................................................................................................. 207
Figure 86. Fit of the kinetic rate data for 0.25 M sarcosine (D20, 50 °C, 1000 psi D2, 1 g
dry 5% Ru/C) .................................................................................................. 208
Figure 87. Kinetic rate constant for glycine, sarcosine, and N,N-dmg (D2O, X °C, 1000
psi D2, 1 g dry 5% Ru/C) ................................................................................. 209
Figure 88. Fit of the kinetic rate data for 0.25 M glycine (D20, 90 °C, 1000 psi D2, 1 g
dry 5% Ru/C) .................................................................................................. 213
Figure 89. Fit of the kinetic rate data for 0.25 M betaine (D20, 130 °C, 1000 psi H2, 1 g
dry 5% Ru/C) .................................................................................................. 214
Figure 90. 1H NMR of H/D exchange at the methylene positions of 0.25 glycylglycine
(D20, 100 °C, 1000 psi H2, 1 g dry 5% Ru/C ................................................. 215
Figure 91. IH NMR of H/D exchange at the methylene position of hydrolyzed glycine
from 0.25 M glyclglycine (D2O, 100 °C, 1000 psi H2, 1 g dry 5% Ru/C) ...... 216
Figure 92. 1H NMR of H/D exchange at the methyl position of 0.25 M sarcosine (D2O,
100 °C, 1000 psi H2, 1 g dry 5% Ru/C) .......................................................... 218
Figure 93. Fit of the kinetic rate data for 0.25 M sarcosine (D20, 100 °C, 1000 psi H2, 1
g dry 5% Ru/C) ............................................................................................... 218
Figure 94. 1H NMR of H/D exchange at the methyl position of 0.25 M N,N-dmg (D2O,
100 °C, 1000 psi H2, 1 g dry 5% Ru/C) .......................................................... 219
Figure 95. Fit of the kinetic rate data for 0.25 M N,N-dmg (D2O, 100 °C, 1000 psi H2, 1
g dry 5% Ru/C) ............................................................................................... 219

xi

Figure 96. 1H NMR of H/D exchange at the methyl position of 0.25 M sarcosine (D2O,

100 °C, 1000 psi D2, 1 g dry 5% Ru/C) .......................................................... 221
Figure 97. Fit of the kinetic rate data for 0.25 M sarcosine (D20, 100 °C, 1000 psi D2, 1
g dry 5% Ru/C) ............................................................................................... 221
Figure 98. 1H NMR of H/D exchange at the methyl position of 0.25 M N,N-dmg (D20,
100 °C, 1000 psi D2, 1 g dry 5% Ru/C) .......................................................... 223
Figure 99. Fit of the kinetic rate data for 0.25 M N,N-dmg (D20, 100 °C, 1000 psi D2, 1
g dry 5% Ru/C) ............................................................................................... 223
Figure 100. Fit of the kinetic rate data for 0.25 M sarcosine (D20, 100 °C, 1000 psi H2 or
linked H2 and D2, 1 g dry 5 % Ru/C) .............................................................. 226
Figure 101. Fit of the kinetic rate data for 0.25 M sarcosine (D20, 100 °C, 1000 psi D2 or
linked H2 and D2, 1 g dry 5 % Ru/C) .............................................................. 226
Figure 102. Kinetic rate constant and primary and secondary isotope effects for sarcosine
with 1000 psi H2 or 1000 psi D2 or H2 and D2 ................................................ 228
Figure 103. Fit of the kinetic rate data for 0.25 M N,N-dmg (D20, 100 °C, 1000 psi H2
or linked H2 and D2, 1 g dry 5 % Ru/C) .......................................................... 228
Figure 104. Fit of the kinetic rate data for 0.25 M N,N-dmg (D2O, 100 °C, 1000 psi D2
or linked H2 and D2, 1 g dry 5 % Ru/C) .......................................................... 229
Figure 105. Kinetic rate constant and primary and secondary isotope effects for N,N-
drn with 1000 si H2 or 1000 si D2 or H2 and D2 ........................................ 229
8 P P
Figure 106. Fit of the kinetic rate data for 0.25 M sarcosine (D20, 70 °C, 1000 psi D2, 1
g dry 5% Ru/C) ............................................................................................... 230
Figure 107. Fit of the kinetic rate data for 0.25 M N,N-dimethylglycine (D20, 70 °C,
1000 psi D2, 1 g dry 5% Ru/C) ........................................................................ 231
Figure 108. Kinetic rate constants and primary and secondary isotope effects for
sarcosine and N,N-dimethylglycine; 70 °C, 1000 psi D2 ................................ 232
Figure 109. Rate constants for H/D exchange at 100 °C or 70 °C (D20, X °C, 1000 psi
D2, 1 g dry 5% Ru/C) ...................................................................................... 233
Figure 110. Fit of the kinetic rate data for 0.25 M sarcosine (D20, 50 °C, 1000 psi D2, 1
g dry 5% Ru/C) ............................................................................................... 234
Figure 111. Rate constants for H/D exchange of sarcosine (D20, 50, 70, or 100 °C, 1000
psi H2 or D2, 1 g dry 5% Ru/C) ....................................................................... 235
Figure 112. lH NMR of H/D exchange at the methyl position of 0.25 M betaine (D20,
130 °C, 1000 psi H2, 1 g dry 5% Ru/C) .......................................................... 237
Figure 113. 1H NMR of H/D exchange at the methylene position next to the —OH of 0.25
M ethanolamine (D2O, 100 °C, 1000 psi H2, 1 g dry 5% Ru/C) ..................... 239
Figure 114. 1H NMR of H/D exchange at the methylene position next to the —NH2 of
0.25 M ethanolamine (D2O, 100 °C, 1000 psi H2, 1 g dry 5% Ru/C) ............. 239

xii

Figure 115. Concentration data for 0.25 M ethanolamine (D2O, 100 °C, 1000 psi H2, 1 g
dry 5% Ru/C) .................................................................................................. 240

Figure 116. 1H NMR of H/D exchange at the methylene position of 0.25 M glycolic acid
(D20, 100 °C, 1000 psi H2, 1 g dry 5% Ru/C) ................................................ 241

Figure 117. 1H NMR of H/D exchange at the methyl positions of 0.25 M tert—butanol
(D20, 100 °C, 1000 psi H2, 1 g dry 5% Ru/C) ................................................ 242

Figure 118. Hydrogenation of 0.25 M betaine hydrochloride and 0.25 M glycine (0.5 M
H3PO4, H20, 130 °C, 1000 psi H2, 1 g dry 5% Ru/C) ..................................... 243

Figure 119. Hydrogenation of 0.25 M glycine alone and 0.125 M glycine and 0.125 M
betaine (0.3 M H3PO4, H20, 130 °C, 1000 psi H2, 1 g dry 5% Ru/C) ............ 245

Figure 120. Hydrogenation of 0.25 M betaine alone and 0.125 M glycine and 0.125 M
betaine (0.3 M H3PO4, H20, 130 °C, 1000 psi H2, 1 g dry 5% Ru/C) ............ 246

Figure 121. Fit of the kinetic rate data for 0.25 M glycine alone and 0.25 M glycine and
0.25 M sarcosine (D2O, 100 °C, 1000 psi H2, 1 g dry 5% Ru/C) ................... 247

Figure 122. Fit of the kinetic rate data for 0.25 M sarcosine alone and 0.25 M glycine
and 0.25 M sarcosine (D20, 100 °C, 1000 psi H2, 1 g dry 5% Ru/C) ............. 247

Figure 123. Rate constants for H/D exchange of 0.25 M glycine alone, 0.25 M sarcosine
alone, 0.25 M glycine in the competition experiment, and 0.25 M sarcosine in
the competition experiment (D2O, 100 °C, 1000 psi H2, 1 g dry 5% Ru/C)... 249

xiii

LIST OF SCHEMES

Scheme 1. Biomass based organic acids in the production of commodity chemicals ....... 3
Scheme 2. Dehydrogenation of l-methylcyclohexene onto Pt(111) surface ..................... 5
Scheme 3. Suggested molecular structure of methylcyclohexenyl species (C2Hlo) .......... 6
Scheme 4. Hydrogenation of glycolic acid ........................................................................ 9
Scheme 5. Hydrogenation of 2-ethylheptanoic acid ........................................................ 10
Scheme 6. Hydrogenation of 2-ethylheptan-5-olide ........................................................ 11
Scheme 7. Hydrogenation of ethyl lactate ....................................................................... 15
Scheme 8. Hydrogenation of n-valeraldehyde ................................................................. 16
Scheme 9. Reaction scheme for generation of D2 gas ..................................................... 19
Scheme 10. 10% Pd/C H/D exchange on 5-phenyl-l-valeric acid .................................. 21
Scheme 11. Hydrogenation of L-alanine .......................................................................... 22
Scheme 12. Generalized amino acid hydrogenation sequence ........................................ 22
Scheme 13. Ion exchange with water of 2-acetylthiamin pyrophosphate and 2-acetyl-3,4-
dimethylthiazolium ............................................................................................ 23
Scheme 14. Calculated acidities for the zwitterions of glycine and its derivatives ......... 26
Scheme 15. H/D exchange at the a-carbon on glycine .................................................... 28
Scheme 16. Hydrolysis of N-protonated glycine methyl ester ........................................ 29
Scheme 17. Sodiated glycine in the charge-solvated (CS) and zwitterionic (ZW) forms 31
Scheme 18. H/D exchange on cyclododecene ................................................................. 32
Scheme 19. Proposed model for adsorption of ethylene glycol ....................................... 41
Scheme 20. Langmuir-Hinshelwood kinetic expression for competitive degradation of
ethylene glycol or propylene glycol .................................................................. 42
Scheme 21. Langmuir-Hinshelwood kinetic expression for hydrogenolysis of glycerol 43
Scheme 22. Langmuir-Hinshelwood model of hydrogenation of lactic acid .................. 48
Scheme 23. Langmuir-Hinshelwood kinetic expression of hydrogenation of lactic acid 48
Scheme 24. Deuterium incorporation in D2/D2O using Ru/C .......................................... 50
Scheme 25. Langmuir-Hinshelwood model of hydrogenation of protonated alanine ..... 52
Scheme 26. Heats of formation (kcal/mol) for lactic acid hydrogenation ....................... 57
S cheme 27. Catalytic hydrogenation of lactic acid and propanoic acid .......................... 59
Scheme 28. Rate expressions for single acid hydrogenation ........................................... 61
Scheme 29. Structure of choline bicarbonate ................................................................ 106

xiv

Scheme 30. Proposed mechanism for Ru/C catalyzed HD generation from H2 and D20

......................................................................................................................... 187
Scheme 31. 1H NMR positions for the methylene hydrogens on glycyl glycine ........... 214
Scheme 32. 1H NMR positions for the methylene hydrogens on ethanolamine ............ 238

XV

Equation 1.

Equation 2.

Equation 3.

Equation 4.

Equation 5.

Equation 6.

Equation 7.

Equation 8.

Equation 9.

Equation 10.
Equation 11.
Equation 12.
Equation 13.
Equation 14.
Equation 15.
Equation 16.
Equation 17.
Equation 18.
Equation 19.
Equation 20.
Equation 21 .
Equation 22.
Equation 23.
Equation 24.
Equation 25.
Equation 26.
Equation 27.
Equation 28.
Equation 29.
Equation 30.
Equation 31.

LIST OF EQUATIONS

NMR tube ratio: outer-to-insert volume ...................................................... 76
Lorentzian lineshape ..................................................................................... 77
Singlet peak fit .............................................................................................. 78
Triplet peak fit .............................................................................................. 78
Quintet peak fit ............................................................................................. 79
Total H concentration for each spectrum at time t ....................................... 80
Singlet area fraction ...................................................................................... 81
Triplet area fraction ...................................................................................... 81
Quintet area fraction ..................................................................................... 81

CH2 concentration ...................................................................................... 81
CHD concentration ..................................................................................... 81
CD2 concentration ...................................................................................... 81
Peak area of individual isotopomer ............................................................ 82

Conversion .................................................................................................. 82

Yield ........................................................................................................... 82
Selectivity ................................................................................................... 82
Peak area of water in sample ...................................................................... 83
Peak area of water in internal standard ....................................................... 83
Moles of deuterium in D20 (1) .................................................................... 84
Moles of hydrogen in H2 (g) ....................................................................... 84
Substrate — NHx contribution ...................................................................... 84
Substrate - methylene (CH2) and methyl contribution y*(CH3) ................. 85
Time zero H—D inventory ........................................................................... 85
Molarity of H at time zero .......................................................................... 85
Molarity of D at time zero .......................................................................... 85
Moles of D (l) at time zero ......................................................................... 86
Moles of H (1) at time zero ......................................................................... 86
Moles of H (g) at time zero ........................................................................ 87
Moles of D (l) at time t ............................................................................... 87
Total moles of H (l) and H (g) .................................................................... 87
%H at time t ................................................................................................ 87

xvi

Equation 32. Molarity of H at time t ................................................................................ 87
Equation 33. Molarity of D at time t ................................................................................ 87
Equation 34. AH2 concentration at time zero .................................................................. 88
Equation 35. AH2 concentration at time t ........................................................................ 88
Equation 36. AHD concentration at time t ....................................................................... 89
Equation 37. AD2 concentration at time t ........................................................................ 89
Equation 38. CH3 concentration at time zero ................................................................... 89
Equation 39. CH3 concentration at time t ........................................................................ 89
Equation 40. CH2D concentration at time t ..................................................................... 89
Equation 41. CHD2 concentration at time t ..................................................................... 89
Equation 42. CD3 concentration at time t ........................................................................ 89
Equation 43. Aqueous hydrogen concentration in H20 and HOD .................................. 90
Equation 44. Aqueous deuterium concentration in D2090
Equation 45. Sum of the H area .......................................................... ............................ 90
Equation 46. Aqueous hydrogen concentration equation of fit ....................................... 90
Equation 47. Rate equations for H/D exchange at the methylene position ..................... 92
Equation 48. CH2 concentration at time t ........................................................................ 92
Equation 49. CHD concentration at time t ....................................................................... 92
Equation 50. CD2 concentration at time t ........................................................................ 92
Equation 51. Rate equations for H/D exchange at the methyl position ........................... 92
Equation 52. CH3 concentration at timet ........................................................................ 93
Equation 53. CH2D concentration .................................................................................... 93
Equation 54. CHD2 concentration at time t ..................................................................... 93
Equation 55. CD3 concentration at time t ........................................................................ 93
Equation 56. Forward and reverse rate constants for H/D exchange at the methylene
position .............................................................................................................. 94
Equation 57. Forward and reverse rate constants for H/D exchange at the methyl position
........................................................................................................................... 95
Equation 58. Expression for H on the surface at time t ................................................... 96
Equation 59. Expression for D on the surface at time t ................................................... 96
Equation 60. Equilibrium expression for hydrogen and deuterium in the liquid and on the
surface ............................................................................................................... 97
Equation 61. Rate of deuterium on the surface versus time ............................................ 97

xvii

Equation 62. Conversion of substrate in the hydrogenation reaction ............................ 158
Equation 63. Yield of desired product in the hydrogenation reaction ........................... 158
Equation 64. Selectivity of desired product in the hydrogenation reaction ................... 158

xviii

Chapter 1. INTRODUCTION
1.1 Background
The Jackson/Miller group at Michigan State University is focused on emerging
plant/crop-based renewables technology, where these renewables can serve to replace or
complement conventional feedstocks to meet the ever-growing need for chemicals and
other materials. The group is working in an interdisciplinary and collaborative effort
using mechanistic organic chemistry and chemical reaction engineering. The studies
described in this dissertation explore the aqueous phase catalytic hydrogenation reaction
of organic acids and their derivatives to alcohols. In addition a detailed H/D exchange
study was conducted on glycine and its N-methylated derivatives. Though several
effective procedures have been developed, the mechanistic details of such reactions are
far from understood. By examining the effects of structure and medium on organic acid
reductions and H/D exchange, an improved understanding of some of the mechanisms
involved in heterogeneous catalysis can be obtained. This knowledge will prove to be far
reaching in that it helps map out the chemical transformations that will be needed for the

renewables-based “refinery of the future”.

1.2 Significance

Green chemistry, also known as sustainable chemistry, is an umbrella concept that has
grown substantially since it fully emerged in 1990. By definition, green chemistry is the
design, development, and implementation of chemical products and processes to reduce
or eliminate the use and generation of substances hazardous to human health and the

environment. Many of the concerns green chemistry is trying to address are not

laboratory curiosities or individual research projects,1 but global issues such as climate
change, energy consumption, pollution, and management of water resources. Green

chemistries can be used to help reduce our dependence on petroleum and natural gas.

The supply of petroleum is limited and finding an alternative to crude oil is increasingly
important.2 Petroleum has been the most used energy source due to automobiles since
the 1950’s due to its energy content, ease of transport, and abundance (albeit finite). It is
used as a raw material for nearly all organic chemical products, including
pharmaceuticals, solvents, fertilizers, etc. Identifying renewable feedstocks and green
pathways to producing the same high-valued chemicals is required to ensure a sustainable
future. Use of renewable feedstocks also has the potential to dramatically reduce U.S.
dependence on foreign oil and allow for domestic growth of a bio-based chemical

industry.

Research at Michigan State University seeks non-fossil based alternative pathways for
the production of a variety of commodity and specialty chemicals. The research
presented in this dissertation is aimed at the development and characterization of novel
pathways to produce high-valued chemicals using a “green chemistry” approach. Central
to this objective is the need for insight into the reactions’ mechanisms as well as a broad
descriptive knowledge of their variability as a function of input variables such as
substituents, solvent, pH, temperature and reaction pressure. In this work, the main tools
employed to probe the hydrogenation pathways of organic acids have been isotopic

deuterium exchange and kinetic modeling. The ability to successfully and selectively

hydrogenate biomass derived organic acids will open pathways to the use of renewable
starting materials in polymers, foods, and pharmaceutical applications, all at costs

competitive with existing petroleum based routes.

In evaluating the possible renewable “green” feedstocks, it is evident that carbohydrate
polymers (starch, cellulose) and monomers (glucose, fructose, mannose, etc.) and sugar
alcohols will be major contributors. Currently, bioconversion (fermentation) of
carbohydrates can selectively produce a large variety of products. The direct chemical
conversion of these. downstream fermentation products can in turn result in a plethora of
high-valued products. The combination of chemical- and bio-processing, to date, offers
the most promising and flexible strategy for the development of renewable resource
refining. Such a partnership is illustrated in Scheme 1 for the net transformation of

glucose to propylene glycol, a commodity chemical with US production of ca. 5 x 108

 

 

 

kg/yr.
CHO
H——OH
H0—— H fermentation A OH H2: catalyst _ OH
H——OH ' Acoou ' /|VOH
Hj—OH Lactic acid Propylene
CH2OH glycol
D-glucose

Scheme 1. Biomass based organic acids in the production of commodity chemicals

An additional advantage of this “environmentally friendly” approach is the domestically
centered economic potential of converting simple sugars to high-valued and large volume

commodities. In the US in 2006, 9178 thousand tonnes of oil equivalent ethanol were

produced, up 24% from 2005 (7389).2 A tonne of oil equivalent is equivalent to the
amount of energy given off by burning one tonne of crude oil (~ 42GJ). The US.
Departments of Energy and Agriculture determined that 1.3 billion tons of US. biomass
feedstocks are potentially available for the production of biofuels. Thus, there is no
shortage of bio-based raw materials for chemicals production. The development of low-
cost, high efficiency hydrogenation routes to exploit this opportunity is a major goal of
the work described herein. In addition, a series of methylated amino acids was evaluated
in terms of H/D exchange to better understand the mechanism of hydrogenation and
possibly build on the opportunity for stereoretentive reduction to high-valued amino-

substituted alcohols.

1.3 Literature Review

1 .3.1 Hydro genation/Dehvdro gen_ation

Hydrogenation refers to a general chemical reaction that results in the addition of H2 to
either double or triple bonds. Hydrogenolysis (to be discussed later, section 1.3.7) refers
to a general chemical reaction that involves breaking bonds and the addition of hydrogen.
In the work described here, the hydrogenation reactions that are being considered are
mostly the reduction of X=Y double bonds. The exact ways in which the hydrogen
atoms are delivered are varied depending on how polar the substrate is, etc. The largest
and best studied class of hydrogenation includes hydrogenation of hydrocarbons. The
first paper described below discusses the effect of hydrogen pressure in a reaction with 1-
methylcyclohexene and following this is a brief discussion of the effect of hydrogen

pressure on the hydrogenation rate of organic acids (our research).

Hydrogen pressure is a critical parameter in catalytic hydrocarbon-refining processes,
controlling reaction rates and selectivity. Traditionally the rate of dehydrogenation and
fragmentation of hydrocarbons is slowed by the presence of hydrogen adsorbed on a
catalyst surface. Surprisingly, Yang et al. found that in the presence of excess hydrogen
at 0.3—0.7 psi (reported as 15-40 Torr) and at temperatures between 290-480 K,
dehydrogenation of l-methylcyclohexene on Pt(1 1 1) to form toluene dominates over

hydrogenation to form methylcyclohexane (Scheme 2).

\
...__.5

5 “d?
——> +
0
i—aiw—Cw

Scheme 2. Dehydrogenation of l-methylcyclohexene onto Pt(1 l 1) surface

 

The authors acknowledge that this dehydrogenation is somewhat counterintuitive and
suggest that perhaps the excess hydrogen might assist in displacing adsorbed
hydrocarbon fragments, which inhibit the chemisorption of the reactant. Hydrogenation
to 1-methylcyclohexane is temperature dependent reaching a maximum at 370 K,
whereas dehydrogenation to form toluene increases, starting around 350 K. In general, as

temperature increases, n-bonded alkenes, which are weakly bonded to the surface, easily

desorb, or they are converted into strongly di-c-bonded species.

Taking a SFG (sum frequency generation) spectrum of l-methylcyclohexene at 403 K in
0.03 psi (reported as 1.5 Torr) of substrate and 0.3 psi (reported as 15 Torr) of hydrogen
pressure and again at 300 K and below 9.7x10'll psi (reported as 5 x 10'9 Torr), Yang et
al. found no change in the SFG spectrum (two bands at 2845 and 2925 cm'I). SFG
spectroscopy is a second-order nonlinear optical process for analyzing surfaces and from
the output, structural information regarding the molecules on the surface can be
determined. The process involves an infrared laser beam combined with a visible laser
beam to generate a sum frequency output. Other surface analytical techniques exist, such
as thermal desorption spectroscopy (TDS), high-resolution electron energy loss

spectroscopy (HREELS), and reflection absorption infrared spectroscopy (RAIRS).

The results from Yang indicated that the species responsible for the spectrum was 0-

bonded to the surface (Scheme 3).

 

Scheme 3. Suggested molecular structure of methylcyclohexenyl species (C7Hlo)

They suggest that excess strongly chemisorbed hydrogen molecules react with weakly 1t-
bonded l-methylcyclohexene to form the hydrogenated methylcyclohexyl. C7H13 which

is o-bonded to the surface. The methylcyclohexyl intermediate can undergo a

disproportionation reaction with a weakly n-bonded l-methylcyclohexene to form
methylcyclohexane, and a dehydrogenated species C7H1 1, which can continue losing

hydrogen to form toluene, C7Hg.3

The effect of pressure has been studied in our research group on the hydrogenation of
organic acids (X=Y). Jere et al. the Jackson/Miller group determined the kinetics of
aqueous-phase hydrogenation of L-alanine to L-alaninol over 5 wt% Ru/C. Reaction
temperatures were between 353—398 K with 0.22-0.46 M feed concentration, 0-1.2 M
phosphoric acid concentrations and hydrogen pressure between 1.7-13.8 MPa (6.9 MPa =
1000 psi). Results indicated only a mild pressure dependence of the hydrogenation rate
above 1000 psi H2 partial pressure. Jere et al. found that at pressures above 1000 psi H2,
the rate was independent of the hydrogen concentration concluding that the surface was

saturated with hydrogen.4

Literature describing hydrogenation of organic acids to high-valued chemicals exists;
however, the fundamental chemistry involved at the catalyst surface has been little
discussed, especially for processes in aqueous phase. For producing chemicals that
replace traditional petroleum-derived species, hydrogenation or other reductions must
play a key role to reduce the overall oxidation state of the oxygenated feed material. The
ability to successfully and selectively hydrogenate organic acids will open pathways to
the use of biomass-derived materials in polymers, foods, and pharmaceuticals
applications. The development of low-cost, high-efficiency hydrogenation routes is the

goal of our research group.

Research conducted on hydrogenation of organic acids to alcohols will be discussed first,
followed by a discussion on the research conducted on hydrogenation of esters and

aldehydes.

Somewhat recently (1998), Antons et al. described the preparation of optically active
alcohols from optically active carboxylic acids using a ruthenium catalyst. For the
hydrogenation of lactic acid the enantiomeric excess of propylene glycol dropped with
increasing temperatures (from 383 K to 413 K).5’6 Recent patents describing
heterogeneous catalysis for the hydrogenation of carboxylic acids to alcohols provide
yields and selectivities, but do not go into the mechanistic aspects of the chemistry

involved.

Carnahan and co-workers examined hydrogenation of carboxylic acids using a ruthenium
catalyst. They used both ruthenium dioxide and ruthenium-on-carbon at temperatures of
about 423 K (reported as 150 oC); platinum and palladium under similar conditions were
not successful. They obtained an 80% yield of ethylene glycol from 20 wt% glycolic
acid (hydroxyacetic acid) after a reaction time of 1 hour using 10 g of 10% Ru on carbon
and extremely high pressures. The hydrogenation was performed at a temperature of
418-423 K (reported as 145-150 0C) and 9,500-10,400 psi (reported as 650-710 atm) in

water (Scheme 4).

10% Ru/C;

 

0 H2O
(“\‘m 420 K; (\0“
0” 9550-10,400 psi 0”

t = 1hr
yield = 80%

Scheme 4. Hydrogenation of glycolic acid

The yield was reduced to 68% ethylene glycol when using 3 g of 10% Ru on carbon at
the same reaction conditions. In an experiment with glycolic acid and 10% Ru on carbon
at 523-528 K (reported as 250-255 °C) and 7,900-10,300 psi (reported as 540-700 atm),
Carnahan et al. analyzed the gas in the head space using mass spectrometry and reported
the following distribution: 78.3 mole % hydrogen; 1.8 mole % methane; 0.8 mole %
ethane; 0.04 mole % carbon dioxide; 0.8 mole % oxygen; 0.09 mole % argon; 18.2 mole
% nitrogen (they mention that the last three constituents may have resulted from
contamination of the sample with air and with nitrogen used to purge the sample lines).
The yield of ethylene glycol from this reaction was 32% with side products methane and
ethane, lower than the 80% obtained at the conditions in scheme 4. The lower yield was
a result of reducing the catalyst amount from 10 to 3 g and increasing the reaction
temperature from 420 K to 523 K. The yield of 32% was obtained after 15 minutes

compared to the yield of 80% after a reaction time of 1 hour.7

Behr et al. reported homogenous bi-metallic catalyzed reduction of carboxylic acids to
alcohols with hydrogen. One drawback of traditional heterogeneous hydro genations
using metal catalysts is the use of high pressure, namely ~2900-4300 psi and expensive

high pressure reactors. Thus, homogeneous catalysis can often be advantageous in that it

can be performed at more moderate reaction temperatures and hydrogen pressures. Both
homogeneous and heterogeneous catalysis were investigated in the reduction of 2-
ethylheptanoic acid to 2-ethylheptanol. Catalysts included [Mo(CO)6], [Rh(acac)(CO)2],
[Ru(acac)3], Rh/C, Rh/Al2O3, and Ru/C (Scheme 5).

0 1 mol% catalyst;

 

dioxane

OH 9' OH
473 K; V\/\l:\
~2200 psi t = 2 hrs

Scheme 5. Hydrogenation of 2-ethylheptanoic acid

Yields of 10% (for the heterogeneous case of 1 mol% Ru/C) or less were obtained for all
catalysts, however, nearly quantitative yields were possible with homogeneous bimetallic
catalysts. An alcohol yield of 99% was found with Mo(CO)6, a group 6 early transition-
metal carbonyl, and Rh(acac)(CO)2, a group 8 late transition-metal complex, at 473 K
(reported as 200 °C), 2200 psi (reported as 150 bar) H2 pressure in dioxane for 4 hours.
They suggest that the acetylacetate anion, C5H7O' (acac) oxygen-donor ligands in general

show the highest activity for the reduction of carboxylic acids.

Behr and co-workers also looked at the reduction of 2-ethylheptanoic acid with
homogeneous/heterogeneous bimetallic catalysts. A yield of 69% alcohol was achieved
using the homogeneously dissolved Mo(CO)6 and heterogeneous Ru/C at 473 K (reported

as 200 °C), 2200 psi (reported as 150 bar) H2 pressure in dioxane for 2 hours.

Although the yields are not as high as in the case of the homogeneous bi-metallic

catalysis, there seems to be a collaborative effect. In the case of Mo(CO)6 or Ru/C alone

10

yields of only 2 and 10%, respectively were reported after 2 hours. It may be that the
main catalytic reaction is occurring on the Ru surface because in the case of reduction of
2-ethylheptan-5-olide to 2-ethyl-1,5-heptanediol (Scheme 6) with Mo(CO)6 and
[Rh(acac)(CO)2] reducing the amount of rhodium (reduction of ~66%) leads to a greater
loss of activity (from 92% to 19% yield), however, less molybdenum (reduction of

~66%) has less of an effect on the reaction (from 92% to 53% yield) (Scheme 6).

 

Mo(CO)6/
Rh(acac)(CO)2;
dioxane
> OH
0 o 473 K;
2200 psr OH t = 2 hrs

_ _ yield = up to 92%
Scheme 6. Hydrogenation of 2-ethylheptan-5-ollde

Behr also expanded the work to evaluate reduction of commercially available lactones
with the rhodium/molybdenum catalyst. They found that the reduction of lactones was
more difficult if the compound was branched in the a-position to the carboxylic acid and
suggested perhaps it is because the compound is more sterically hindered.8 A
disadvantage of this chemistry is that homogeneous bimetallic catalysts can be difficult to

prepare and often cannot be reused because they are not separated from the substrate, etc.

Mao and co-workers found that a magnesia-supported poly-y-aminopropylsiloxane-Ru
complex (MgO-NH2-Ru) was extremely effective in the aqueous catalytic hydrogenation
of several carboxylic acids (acetic, propanoic, lactic, and isobutyric acid). Optimal

reaction conditions included 513 K (reported as 240 °C), 725 psi (reported as 5.0 MPa)

11

H2 in 5 mL of water for 18 h. Upon examination of the hydrogenation of 0.1 mL (1.35
mmol) propanoic acid, results revealed that the amount of Ru in the MgO-NH2-Ru
complex, reaction temperature, and the amount of water significantly affected the yield of
desired alcohol. An increasing Ru content up to 0.12 mmol/g (from 0.048 mmol/g)
increased the yield of propanol formed. As the temperature increased from 433 K to 523
K (reported as 160 °C to 250 °C), the yield of propanol went from 0% to 95%.

Increasing the water amount from 1 to 5 mL increased the yield of propanol to 100%
(from 5%), however, more than 5 mL (up to 8 mL) reduced the yield of propanol (to
28%). In summary, the optimal conditions for 0.1 mL of propanoic acid in water is 513
K, 725 psi H2 pressure, 5 mL of water, and a Ru content of 0.12 mmol/g in the MgO-

NH2-Ru complex.9

F uchikami and co-workers discussed transition-metal catalyzed hydrogenation of
pentadecanoic acid using 1.0 mL of dimethoxyethane as a'solvent. Optimum reaction
conditions included 1.0 mmol substrate and 0.01 eq. 5% [Rh/A1203] and [Mo(CO)6] at
423 K (reported as 150 °C), 1470 psi (reported as 100 atm) H2 pressure for 16 h. The
reaction in the presence of only Mo(CO)6 resulted in no conversion to the desired alcohol
indicating that bimetallic catalysts consisting of Group 8 to 10 late transition-metal
compounds and Group 6 or 7 early transition-metal carbonyls were extremely effective in

the hydrogenation of organic acids. '0

Summarizing the hydrogenation of organic acid reactions, Antons et al. found that with a

ruthenium catalyst the enantiomeric excess of propylene glycol from hydrogenation of

12

optically active lactic dropped with increasing temperature (similar to results observed by
our research group, to be discussed later). Carnahan obtained yields of 80% ethylene
glycol from glycolic acid after 1 hour using Ru/C but at high H2 pressures (~10000 psi).
Research in this paper presents yields up to 50% for ethylene glycol from glycolic acid
using Ru/C at similar conditions except for lower H2 pressures (1000 psi) and 1/3 the
catalyst : substrate amount. Behr reported low yields (less than 10%) for hydrogenation
of 2-ethylheptanoic acid to the desired alcohol using 1 mol% Ru/C at conditions similar
to what our research used, but again Behr used higher pressures (~3000-43 00 psi)

compared with our research at 1000 psi H2.

Behr expanded his work to lactones and found they were more difficult to hydrogenate
than organic acids. Expanding the use of ruthenium to a magnesia-supported poly-y-
aminopropylsiloxane-Ru complex (MgO-NH2-Ru), Mao et al. found this catalyst very
effective in hydrogenating several carboxylic acids in water. Reaction temperatures were
about 100 °C higher than what we used in our research (to be discussed later). Fuchikami
et al. found that bimetallic catalysts (similar to the results from Behr) exhibited extremely
powerful reducing ability for the hydrogenation of organic acids focusing mostly on
[Rh/A1203] and [Mo(CO)6]. It may be suggested to expand the use of heterogeneous

Ru/C hydrogenations to include bimetallic catalysts.

Hydrogenation of esters and aldehydes has also been studied. Adkins et al. looked at
optically active esters, the effect of , free acids on the hydrogenation of esters, and also (1-

and B-oxygen substituted esters. Vaidya et al. looked at hydrogenation of n-

13

valeraldehyde and fit a power law model to the experimental data in order to generate a
rate expression for the reaction. We will first consider the hydrogenation of esters then

follow this discussion with Vaidya’s work with aldehydes.

Adkins and Bowden described the hydrogenation of optically active esters to alcohols
using copper-chromium oxide at a temperature of 523 K (reported as 250 oC) and
pressures of 2200-2900 psi (reported as 150-200 atm). A yield of 81% 1,2-propanediol

from butyl lactate (91% conversion) was obtained after 2 hours at these conditions.ll

Earlier (1932) studies by Adkins and coworkers also showed that hydrogenation of esters
over copper-chromium oxide at 523 K (reported as 250 °C), 2900-4400 psi (reported as
200-300 atm) H2 in ethanol is inhibited in the presence of free acids, presumably due to

competitive adsorption resulting in less desired product formation. 12,13

In 1947, Adkins and Pavlic studied alpha amino ester hydrogenation over Raney nickel in
a non-aqueous solvent, namely dry ethanol, using lower temperatures (25-75 °C) and
high pressures (150-200 atm or 2200-2900 psi). The reaction was carried out in a 96 5 M
stainless steel reactor with about 20 to 35 mL of solution. The hydrogenated products

were isolated by fractional distillation. Yields of desired alcohol were as high as 98% for

hydrogenation of a-ethylpiperdinoacetate. ‘4

Adkins and Billica reported effective hydrogenations of esters to alcohols at modest

temperatures (298-423 K reported as 25-150 °C), yet with the requirement of higher

14

catalyst loadings. They suggest using similar mass amounts or greater of catalyst
compared with the mass used of ester. In most cases, they used 10 g of the ester substrate
(roughly 0.085 moles for the case of ethyl lactate) and 15 g of catalyst. When using an
excess of Raney nickel the hydrogenation at 373 K (reported as 100 °C) and 5000 psi of
a-amino esters was quantitative and reasonably rapid (e. g. ethyl piperidinoacetate was
complete in 1.7 hrs). One limitation with Raney nickel is that hydrogenation of phenyl
substituted esters unavoidably results in benzene ring reduction to form the cyclohexyl

substituted amino alcohol.

Raney nickel was also used for the hydrogenation of or and B-oxygen substituted esters.
A quantitative yield of 1,2-propanediol from ethyl lactate in dry ethanol at 423 K (150
°C) and 5000 psi was obtained after 1.5 hrs. The research presented in this thesis
includes a-oxygen substituted organic acids and found an enhanced hydrogenation rate

with these substrates compared with the unsubstituted substrates (to be discussed later).

 

0 Raney Ni;
V/k Et0H<dry)
0“ 423 K; 5000 psi 0”
t= 1.5hrs
Scheme 7. Hydrogenation of ethyl lactate yield = quantitative

Adkins et al. report that for most of the substrates examined as good or better yields can
be obtained from the esters using copper chromium oxide as opposed to Raney nickel.
They suggest using copper chromium oxide unless hydrogenation is being performed at

temperatures below 100 °C, in which case Raney nickel is preferred. An added benefit of

15

running reactions at increased catalyst loadings and lower temperatures, is that undesired

hydrogenolysis of 1,3-glycols is avoided.ls

Vaidya et al. report kinetic data from liquid phase hydrogenation of n-valeraldehyde to n-
amyl alcohol over a 5% Ru/Al2O3 catalyst. Reaction conditions included temperatures
and pressures between 323-348 K (50-75 °C), 0.69-2.76 MPa (100-400 psi), and 2-
propanol as the solvent (Scheme 8). The Ru/Al2O3 catalyst (size 25-60 pm; diameter
46.6 um; BET surface area 598 ng'l; pore volume 0.17 cm3g‘l; pore diameter 10.1 nm)

could be used three times before any loss in activity.

 

5% Ru/A12O3;
2-propanol
Wok (10% w/w H2O)
' \/\/\
H 323-348 K; 0”
100-400 psi

Scheme 8. Hydrogenation of n-valeraldehyde

Reduction of the catalyst was carried out at 648 K (375 0C) under pure hydrogen. They
comment that, “our experience is that heterogeneous Ru catalyst works better in the
presence of water.” To support this claim they present a finding in which hydrogenation
of 0.3 kmol/m3 n-valeraldehyde was carried out in 2-propanol containing water (10%
w/w) at 323 K, and a hydrogen pressure of 400 psi (1.29 kg/m3 catalyst loading). At 10,
20, and 30 minute intervals samples were collected and analyzed. Results indicate the
conversion rate is nearly doubled in the presence of water at 10 minutes, but the

enhancement drops off over time with only about a 5% increase in conversion at 30

16

minutes (for the case with water added). Examining the mass transfer limitations, the gas
phase mass transfer resistance is neglected due to the gas phase being almost all pure
hydrogen (attributed to the low vapor pressure of the solvent under these conditions).
The speed of agitation controls the liquid phase mass transfer resistance and transfer of
dissolved hydrogen and valeraldehyde to the external surface of the catalyst particle.
They found an impeller speed of 20 rps was enough to overcome all mass transfer

resistances (gas-liquid, liquid-solid, and intraparticle diffusion).

There results also indicated that initial rates increased linearly with catalyst loading
(between 0.26 — 1.29 kg/m3). The initial rate of hydrogenation was also increased with
both an increased hydrogen pressure (between 100 - 400 psi reported as 0.69 — 2.76 MPa)

and increased initial valeraldehyde concentration (between 0.075 — 0.3 kmol/m3).

They fit a power law model to the experimental data and the resulting rate expression
was: r = 69.9 exp(-2587/T)[A]O'44[B]0'46 ([A] = conc. of hydrogen in bulk liquid phase,
M; [B] = conc. of n-valeraldehyde in liquid phase, M). Thus the order of the reaction
with respect to n-valeraldehyde and hydrogen is 0.46 and 0.44, respectively. They also
examined using Langmuir-Hinshelwood type models for both single and dual site
mechanisms as well as dissociative and non-dissociative adsorption of hydrogen. There

results supported a single site mechanism and dissociative adsorption of hydrogen.

The final model derived was represented as: r = {k3KAKB[A][B]} / {1 + (KA[A])”2 +

(KB[B])}3 (k3 = surface reaction rate constant, kmol/kg cat/min; KA and KB = adsorption

17

equilibrium constant for A and B, m3/kmol). The product term, [P], in the denominator
was omitted based on an experiment at 348 K in which the addition of n-amyl alcohol
(10% w/w of n-valeraldehyde) to the beginning reaction mixture had no effect on the rate
of reaction. The surface reaction step activation energy, 38.5 kJ/mol was calculated from
the temperature dependence of k3. Using the van’t Hoff equation and from the
temperature dependence of KA and K3, the heats of adsorption of hydrogen and
valeraldehyde were found to be 48.6 and 29.1 kJ/mol, respectively (no measure of the

uncertainties was provided).16

1.3.2 H/D Exchange

Deuteriurn-labeled compounds are used in many different facets of research, from
biological to environmental, and organic synthesis. In the catalytic hydrogenation of
double and triple bonds, ring-opening of epoxides, etc., the use of H/D exchange using

either D2O and/or D2 gas offers an efficient route to deuterium-labeled compounds.

Deuterium gas is quite expensive, so the development of a method for preparing D2 gas
in-situ could be valuable. Sajiki et al. used 1.0 mL (55 mmol) of D20 and 10 mg of 10%
Pd/C (0.0094 mmol of Pd metal) in the presence of ~2.0 L (~83 mmol) H2 gas at room
temperature to study the amount of D2 produced. Using 1H NMR, the signal intensity of
HOD was monitored and continued to increase over time. The reaction they monitored is

presented in Scheme 9.

18

2 D20 + H2 (balloon, ~2L) ———> 2 DHO + D2
(55 mmol) (~ 83 mmol)

Scheme 9. Reaction scheme for generation of D2 gas

After 48 hours, 87 mL of D2 gas was generated from 1.0 mL of D2O (TON = 829). In the
absence of 10% Pd/C or H2 gas (under Ar) results indicated a negligible increase in the

HOD signal over time.

Sajiki et a1. expanded this work and examined the amount of deuterium exchange into
organic substrates in a closed system, in the presence of a limited amount of H2 gas.
After 24 h of stirring, deuterium was incorporated via reduction of 0.5 mmol cinnamic
acid (C6H5CHCHCOOH) into both methylene positions of dihydrocinnamic (3-
phenylpropanoic) acid. The reaction consisted of 7.4 mg of 10% Pd/C in 5.0 mL or 3.0
mL (277 mmol or 166 mmol) of D2O in a 100 mL hydrogen-charged sealed flask at room
temperature. The actual internal volume of the hydrogen-charged sealed flask (100 mL)
was 160 mL. They also found that the efficiency of incorporation was greatly reduced in
reactions which used a larger flask. The mechanism of D2 gas generation is unclear.
However, Sajiki et al. propose oxidative insertion of Pd(0) activated by H2 or HD as a
ligand into D20 to give Pd(II) species followed by the key step, a ligand exchange
between H2-D2O, in the catalytic isotope exchange reaction. Subsequent reductive
elimination gives the corresponding Pd(0)-HD or D2 complex. At the end of the reaction

D2 gas is liberated.'7

19

Performing the reaction under an atmosphere of D2 and in dry EtOAc as an inert solvent
as opposed to H2 and D20, resulted in significantly less deuterium exchange, even at a
temperature of 180 °C. Thus, Sajiki, et a1. concluded that D20 is required for efficient

isotopic exchange.

Deuterium-labeled organic compounds can be prepared using two main strategies: (a) a
multistep synthetic method starting from deuterium-labeled synthons and (b) selective
catalytic replacement of carbon-bound hydrogen with deuterium. Both homogeneous and
heterogeneous catalytic conditions have been employed, with a wide variety of

- - I8 19 20 21 22 23
catalytically active metals (e.g. Ir , Rh , Co , Pt , Ru , and Mn ). However, because
they are easily separated from reaction mixtures by filtration, heterogeneous catalysts are
most convenient, where they work for these applications. The work described (later) in

this thesis discusses aqueous-phase heterogeneous catalytic HfD exchange.

Sajiki et al. using heterogeneous catalytic H/D exchange evaluated the effect of
temperature and also described the H/D exchange observed at carbons alpha to an acid,
phenyl ring, primary alcohol or ether. The work discussed in this thesis centers around
the H/D exchange at the alpha carbon of glycine and its derivatives. A maximum amount
of deuterium incorporation into 5-phenylvaleric acid occurred at a temperature of 160 °C.
Deuterium incorporation decreased with increasing distance from the benzene ring. Over
the time course of the reaction, rapid exchange was observed at the benzylic carbon (C1)
as well as at the beta carbon (C2) and also at position C3 with respect to the phenyl ring,

however, the carbon alpha to the acid (C4) and the aromatic carbons show only a gradual

20

increase of deuterium exchange (Scheme 10). At a temperature of 110 °C, there was no

measurable incorporation at the carbon alpha (C4) to the acid.
0 D D o

10% Pd/C, H2 A P, 002”

COH 7
Ph/W 2 D2O,24h,heat o D o D

Scheme 10. 10% Pd/C H/D exchange on 5-phenyl-l-valeric acid

 

Sajiki et al. found that the hydrogens on the alpha carbon adjacent to a primary alcohol or
ether as a substrate exchanged with deuterium to a lesser extent as compared with the
methylenes closer to the phenyl ring. Reactions with substrates which lacked an aromatic
ring, such as octanoic acid, led to almost no deuterium incorporation (at reaction
temperatures ranging from 110-160 °C). In addition, substrates with an oxygen between
the phenyl ring and the alkyl chain, such as 5-phenoxyvaleric acid underwent almost no
deuterium incorporation. Thus there is apparently a need for an aromatic ring to be
directly connected to the alkyl chain for H/D exchange using 10% Pd/C. The exact
mechanism of the exchange is not clear, but the authors proposed that the reaction goes
via Pd/C-catalyzed C-H bond activation combined with deuteration from D2O. The
results presented provide a deuterium gas-free, totally catalytic and postsynthetic

deuterium labeling method in D2O.24

Our research group, specifically Jere et al. report aqueous stereoretentive hydrogenation
of L-alanine to L-alaninol using a 5% Ru/C catalyst at 100 0C and 1000 psi in the
presence of added acid (Scheme 11). Alanine had become a compound of interest

because an enhanced rate of hydrogenation of a-substituted organic acids such as alanine

21

and lactic acid (to be discussed in section 1.4 summary of previous work) had been noted.
A slight excess, 0.29 M H3PO4 (pKa = 2.15) was added to protonate 0.22 M alanine (pKa

= 2.34) and afford a 91% yield of alaninol after 6 hrs.

5% Ru/C;
D20;
equimolar H3PO4
NH OH 373K; f0”
2 1000 psi D2 2 t=6hrs
yield = 91%

 

Scheme 1 l. Hydrogenation of L-alanine

To track the reaction’s stereochemistry, D2O was used as the solvent at 1000 psi D2 and
100 °C. Surprisingly, although rapid deuterium incorporation at the a-carbon was

observed for alanine and alaninol, little or no racemization was seen.

   

 

OH H
100°C

OH OH

 

cam-I3 @NH3

Scheme 12. Generalized amino acid hydrogenation sequence

22

Scheme 12 shows a generalized amino acid hydrogenation sequence which includes an
aldehyde and two potential enols as intermediates. Jere et al. suggest that the enhanced
rate of hydrogenation with lactic acid and protonated alanine is due to the electron
withdrawing a-OH and a-N+H3, enhancing the carbonyl group’s preference for sp3

hybridization.”

Looking again at enolization, Frey et al. using 2-acetylthiamin pyrophosphate (AcTPP, 1)
and 2-acety1-3,4-dimethylthiazolium (2) performed an H/D exchange kinetic analysis of
sequential reactions. Frey et al. found that an electrophilic thiazolium ring enhances the
ionization rate and that the acetyl methyl hydrogens of AcTPP, l and 2 ion exchange with

water at a pD of ~4.0 (Scheme 13).

"'3C OH O
OH CH3
8 s
\ \
= N\
R2 \ g\R1 R2 \ <9 R1
CH3 CH3
H2N /NIYCH3
\ N 3-
12 R1 = HZC R2 = CH2CH20P206
22 R1 = CH3 R2 = H

Scheme 13. Ion exchange with water of 2-acetylthiamin pyrophosphate and 2-acetyl-3,4-

dimethylthiazolium

23

The rate of ionization for AcTPP is on the order of 107 faster than acetone. The
following kinetics were employed to describe the exchange of the acetyl methyl
hydrogens (R-CH3):

R—CH3 —-) R—CDH2 —> R—CDzH

Defining a = [R-CH3] and b = [R-CDH2], at time t = 0, a = a0 = [R-CH3]0, and b = b0 = 0.
Describing consecutive (pseudo) first-order reactions the differential equations and

solutions are given by:

 

3$=kl*a=kl*ao*eXP(Tk1*t) a:aO’I‘CXP(—kl*t)
512:,“ *ao *exp(—ki *t)-k2 *b b = 1" *ao *lexP(—ki *1)“CXP(“ k2 *0]

(11 (k2 _k)

The rate constant k is defined for exchange of a single proton, thus the values of k1 and k2
are 3k and 2k, respectively. The substitution of these variables into the above equations

can be expressed as:
a ==a0 *exp(—3*k*t) and b =3*a0 *[exp(—2*k*t)—exp(—3*k*t)]

From these equations, the following ratios can be expressed.

_fl_
[3a+b)
3

Using lH NMR, normalized integral ratios were measured as a function of time to the

iR—CH3]
[[3*(R-CH3)+R—CDH2]:|
3

 

= exp(—- k * t), thus = exp(—k * t)

 

 

extent of 70-80% exchange of one proton. Frey et al. found that the rates of exchange
with water were the same (within error) for the hydrate and ketone forms and thus the
interconversion between them is fast compared with the exchange reactions. The rate of

exchange was not observed at pD S 2, thus the reaction is not acid catalyzed but base

24

catalyzed and presumably occurs via an enolization mechanism. Using the iodine
scavenging method, rate constants for the uncatalyzed and base catalyzed enolization
were determined (koE = (3.8 x 0.5) x 10'6 s" and (6.73 r 0.22) x 10'5 s" at pD 4.17 (46

mM Na oxalate)).

Frey et al. suggest that the exchange reaction occurs via the rate-limiting loss of a proton
to produce an enolate, which then can be rapidly deuterated in the presence of D20. The
rate of enolate formation from a carbonyl compound is related to the carbon acidity. In
the case of AcTPP, Frey et al. believe the largest contributor to the fast enolization is due
to through-space electrostatic effects of the positively charged thiazolium ring. Fast rates
of enolization have been reported for other cationic ketones, namely PhCOCH2N+C5H5
(kOHE = 1.8 x 105 M'ls'l). The rate enhancement in the ionization of carbon acids and the
enolization of carbonyl compounds is poorly understood, but Frey et al. suggest that rates
of base-catalyzed enolization are accelerated by the through—space electrostatic effect of a

nearby positive charge.26

Proteins have stereochemical integrity due to the weakly acidic nature of the a-protons
on the repeating amino-acid monomers. Richard et a1. calculated aqueous carbon
acidities for the zwitterions of glycine and its derivatives: H3N+CH2COO', pKa = 28.9 :1:
0.5; H3WCH2CO2Me, pKa = 21.0 :1: 1.0; Me3N+CH2CO2Me, pKa = 18.0 i 1.0;

Me3N+CH2CO2', pKa = 27.3 i 1.2 (Scheme 14).

25

0CD

O
G

0 CH3 0
(+3
pKa=28.9 H3N\/“\ pKa=27.3 H3C\rL
H3C/@
0

CH3 0

HC
pK,=21.o Hf} pKa=18.0 3\|

N
OCH3 H3C/® OCH3

Scheme 14. Calculated acidities for the zwitterions of glycine and its derivatives

The ionization state of the amino acid relates to the acidity of the or-proton. A decrease
in pH results in protonation of glycine and an increased reactivity towards enolization,
whereas enolization of the anionic form of glycine is far less favorable. The pH is
mathematically related to pKa by the following general equations. The acid dissociation
constant, Ka, is a quantitative measure which defines the equilibrium for a chemical
reaction. It can be written in terms of equilibrium concentrations, Ka = [H+] [A-]/ [HA].
Thus, [H+] = Ka[HA]/[A-] and due to the large orders of magnitude spanned by Ka
values, the log is of Ka is commonly used. The logarithmic measure of the acid
dissociation constant is written as, log[H+] = log Ka + log[HA] - log[A-]. Since pH = -
log[H+] the equation can be rewritten in the form of the Henderson-hasselbalch equation,

pH = pKa + log[A-]/ [HA].

In glycine and its derivatives, Richard et al.28’29 found that modifying the a-substituent
from N+H3 to NH2 increases the pKa values of the a-proton by 5 pKa units, and changing

the a-substituent from COO' to COOCH3 decreases the pKa of the oc-proton by 8 pKa

units. Malthouse and co-workers used the above values and estimated the pKa values and

26

tryptophan-synthase-catalyzed exchange rates of a-protons of alanine, serine,
aminobutyrate, norvaline, norleucine, and tryptophan. They assumed a linear relationship
and estimated that ApKa= -oer*, where 0t=l.6-2.2 and 0* represents the appropriate Taft
substituent constant. From the calculations, the pK,1 values of alanine, aminobutyrate,
norvaline, norleucine, and tryptophan are all raised by about 1 pKa unit, namely ~35,

relative to glycine and serine, ~34.27

Richard et al. performed kinetic studies in D2O at 25 °C using a substrate concentration
of 20 or 40 mM glycine with a constant ionic strength of 1.0, maintained with potassium
chloride. In the 1H NMR spectra, they monitored the replacement of the or-CH2 group’s
singlet resonance with the triplet due to a-CHD sites. This mono-deuterated product
signal exhibited an upfield shift of 0.014 ppm for glycine and its derivatives or 0.016
ppm for betaine and its derivatives (J HD = 2.5 Hz). To quantify the progress of deuterium
' exchange, values of R, reaction progress, were calculated: R = (Acriz/(ACH2+ACHD))-
Note: There is a mistake in this journal in that there is a two missing in front of the ACHD
term to represent the fact that for CH2 there are two hydrogens available for exchange
and for CHD there is only one hydrogen. ACH2 and ACHD are the integrated areas of the
singlet due to the oc-CH2 of the substrate and triplet due to the a-CHD of the

monodeuterated product.

They did not account for the oc-CD2 group, because it was their assumption that no
dideuterated glycine was formed. For the early reaction time samples, the ‘3 C satellite of

the signal due to the a-CH2 group was not resolved from the most downfield peak of the

27

triplet due to the or-CHD group. Due to this, the total integrated area of the triplet due to
the or-CHD group, ACHD was calculated by multiplying the integrated area of the most
upfield peak of the triplet by three. The area of the singlet due to the a-CH2 group, ACH2,
was calculated by subtracting the calculated value of ACI—ID from the total integrated area

of all signals due to the or-protons. Semi logarithmic plots (In R = -k0b5dt) of R vs time
were linear with negative slopes equal to kobsd (5"), which is the rate constant for

exchange of a single proton of the or-CH2 group. The fraction of deuterated glycine, fD,
was calculated: fD = 2ACHD/(ACH2 + 2ACHD). Their assumption was that there was
essentially no formation of dideuterated glycine, oc-CD2 during the hydrolysis of glycine
methyl ester in D2O. They suggested a mechanism whereby the zwitterionic glycine
enolizes to form the anionic intermediate which they believe is not stabilized via
protonation of the substrate carbonyl oxygen (pKa < 7.0). This intermediate quickly

forms the carboxylate compound and undergoes deuteration at the a-carbon (Scheme 15).

 

O
0 D20; 298K 3
1=1.o KCl 9
M < >, OAK/K I O
3 3
o / o e
e G) H D

Scheme 15. H/D exchange at the or-carbon on glycine

Calculating rate constants, Richard et al. found that simple amino acid zwitterions
(glycine and betaine) are weaker carbon acids than the simple oxygen ester ethyl acetate
(pKa = 25.6). In addition, there is about a 2-3 unit increase in acidifying effect of an or-
N+Me3 as compared to an or-N+H3 substituent on the carbon acidity, corresponding to the

difference in polar effects of these substituents. There is more of a polar effect with the

28

N+Me3 due to the formation of hydrogen bonds between the (x-N+H3 and the solvent,
which moves the positive charge away from the nitrogen and onto the solvent, increasing
the effective separation between the interacting cationic and anionic centers at the
enolate. There is greater stabilization of negative charge by electrostatic intramolecular
interactions with the (l-N+M€3, wherein the methyl groups reduce the effective dielectric

constant of the local medium.28

Richard et al. found that hydrolysis of N-protonated glycine methyl ester is 10-times
faster than deuterium exchange of the first a-hydrogen at a neutral pD (pD = 7.4). The
glycine obtained after the completion of the hydrolysis reaction (>10 half-times) is
enriched with 0.10 atom of deuterium (Scheme 16). N-protonated glycine methyl ester
has a second-order rate constant of km = 6.0 M'ls'l for DO' catalyzed. exchange of the
first a-proton, whereas the rate constant for H/D exchange of acetone is kDO = 0.3 M'ls'l

(20-fold difference).

 

O
H
@ OCH3 OCH3
D3N
H
In D20; \
25°C
H G H G
o 0 e O
D3N D3N

H

Scheme 16. Hydrolysis of N-protonated glycine methyl ester

29

The fast H/D exchange of N-protonated glycine methyl ester is presumably due to polar
stabilization by interactions between the or-NH3+ group and the developing enol’s
partially negative carbon next door. The oc-protons of N-protonated glycine methyl ester
(pKa “~" 21) show roughly the same acidity as the corresponding protons for the thiol ester

and mandelic acid.29

In general, the extent of H/D exchange for a given species increases with the basicity of
the exchange reagent. Cox et al., using an external ion source 7-T FTICR (Fourier
transform ion cyclotron resonance) mass spectrometer, found that charge-solvated
sodiated glycine oligomers (Gllea+ see Scheme 18 to GlysNa+) are of the lowest energy
(gas phase). Both experimental and computational methods were employed. An F TICR
is one of the highest performance mass spectrometers and measures the mass-to-charge
ratio of ions based on the cyclotron frequency of the ions in a fixed magnetic field with
excellent resolution and mass accuracy. Using a low level of theory calculations, 3-21G.
and comparing and confirming with higher theory calculations (Jensen used 6-3 10* see
JACS 1992, 114, 9533, Moison and Armentrout used MP2/6-31G* see J. Phys. Chem. A.
2002, 106, 10350, and Wyttenbach et al. used B3LYP/6-311++G** see JACS 1998, 120,
5098) in the literature, the most stable structure was the one in which the glycine was not
in the zwitterionic form but rather when the sodium cation was bound to both the

nitrogen and the carbonyl oxygen (Scheme 17).

30

8,2;- Ho

CS C2H5NO2Na+ ZW Czl‘lsl\102l\lf:l+

Scheme 17. Sodiated glycine in the charge-solvated (CS) and zwitterionic (ZW) forms

Experimentally, sodiated peptide ions were generated by matrix-assisted laser
desorption/ionization (MALDI) and samples were deposited onto a stainless steel probe
tip. The MALDI solutions consisted of a mix (623:2) of 1 M 2,5-dihydroxybenzoic acid
in ethanol, 0.03 M peptide in water/acetonitrile (3:7 v/v), and 1 M D-fructose in water. A
static pressure of the H/D exchange gas, ND3 (7 x 10'8 Torr) was maintained. In the gas
phase for sodiated pentaglycine they experimentally found the rate constants for

deuterium exchange with ND3 were 3.0 x 10'”), 2.5 x 10'”), and 1.5 x 10"0 cm3 molecule‘l

s'1 for the three observed hydrogen exchanges (nitrogen and oxygen hydrogens). The
exchange rate ratios were about 3:2: 1, thus three equivalent hydrogens are presumed to
be exchanging with ND3. They also suggest that the mechanism requires a carboxylic
acid hydrogen due to the fact that no exchange occurs with methyl esters of glycine

oligomers.30

Matsubara et al. examined deuterium exchange of cyclododecene in hydrothermal or
subcritical water (523 K reported as 250 °C, 5 MPa 3 725 psi), in the absence of
hydrogen. At these conditions the pKw of water is lower (ca. 11) and thus the water

ionizes more to form OH' and H3O+ than water at ambient conditions. They propose that

31

It

in these conditions the Pd(0) might oxidatively insert into O-H bonds of water to form the
Pd(II) species, H-Pd-OH. The key species is the palladium hydride that exchanges with
the solvent water’s hydrogens. Using cyclododecene (5.0 mmol), 10 wt% Pd on active
carbon (100 mg, 0.1 mmol Pd) and deuterium oxide (20 g) in a 30 mL Teflon vessel
under hydrothermal conditions for 12 h, cyclododecene was converted into the fully

deuterated product (E/Z = 72:28) (Scheme 18).

10 wt% Pd/C (2 mol%) L
D2O, 250 °C, 12h

I

E/Z=33:67

 

E/Z= 72:28

Scheme 18. H/D exchange on cyclododecene

The product was extracted with hexane and analyzed by GC, 1H NMR, 2H NMR, and
mass spectra. Matsubara et al. explain the deuteration of alkene containing compounds
via an interaction of the alkene and the palladium hydride. However, they also report full
deuteration of saturated hydrocarbons, and in this case suggest that direct C-H activation
may be the probable route to forming the fully deuterated alkanes. T err-butyl benzene
was subjected to the above conditions and no exchange was observed at the sterically
hindered ortho positions, however full deuteration occurred at the meta- and para-
positions. Nearly 45% of the C-H bonds on the tertiary butyl group were exchanged with
deuterium after 10 h (10 wt% Pd/C (2 mol%), D2O , 250 °C). Using the same conditions,

biphenyl was fully deuterated. Mechanistic studies have not yet been examined. It is

32

worth noting that hydrogen gas was needed for the 10% Pd/C-catalyzed deuterium

exchange reaction below 180 oC.31

1.3.3 Alkynylation of sp3 C-H Bonds Adjacent to a Nitrogen Atom

A challenging, yet attractive organic synthesis reaction is C-C bond formation via C-H
bond activation. The possibility of enantioselective catalytic C-C bond formation via this
method could enhance asymmetric synthesis both theoretically and practically: In order
to construct a chiral carbon center, a prochiral sp2 carbon center is generally necessary as
the precursor; however, Li et al. discuss the synthesis of enantioselective C-C bond

formation based on activation of a sp3 C-H bond.

Three papers by Li et al. will be discussed, starting with alkynylation of sp3 C-H bonds
adjacent to a nitrogen atom followed by enantioselective alkynylation of prochiral sp3 C-
H bonds adjacent to a nitrogen atom, and finally cross-dehydrogenative coupling between
sp3 C-H bonds and sp3 C-H bonds. In the first paper, Li et al. report a copper-catalyzed
(5 mol%) oxidative cross-coupling of amines with various alkynes via a direct sp3 C-H
bond alkynylation in the presence of tert-butyl hydroperoxide (2:1 :1 ratio, respectively)
at 100 °C for 3h with yields as high as 74%. The reaction did not take place in the
absence of the tert-butyl hydroperoxide. There was less desired product formed in
reactions with aliphatic amines, as opposed to aryl amines. Interesting to note is that
N,N-dimethylbenzylamine produced alkynylation of the methyl group rather than the

benzyl position.

33

Li et al. propose a tentative mechanism for the direct oxidative coupling suggesting the
formation of an imine-type intermediate (coordinated to copper) through activation of sp3
C-H adjacent to nitrogen. Copper also activated the terminal alkyne and coupling of the

two intermediates resulted in the desired product and regenerated the copper catalyst.32

Li et al. expanded on this work to include tetrahydroisoquinoline alkaloids which exist
widely in nature and are of biological and pharmacological interest. Specifically, they
discuss the development of a novel catalytic asymmetric l-alkynylation of
tetrahydroisoquinoline derivatives via activation of the sp3 C-H bonds of prochiral CH2,

yielding optically active C-l-substituted tetrahydroisoquinoline derivatives.

Li et al. found that the optimal conditions were with CuOTf (as opposed to CuBr seen in
their previous work), at 50 °C (as opposed to 80 °C) in THF with a chiral ligand, yielding
63% of the desired product. Asymmetric alkynylation also proceeds in water or without
solvent but the yields are lower. Aromatic substituted alkynes provided good yields and
enantiomeric excess in most cases, whereas aliphatic substituted alkynes resulted in fair
or low enantiomeric excesses. The factors which influence the enantioselectivities of the

reaction are still being investigated.33

Vicinal diamines are found in biologically active natural products and are used in
medicinal chemistry as well as in chiral ligands for asymmetric catalysis. Li et al.
previously discussed both alkynylation of sp3 C-H bonds adjacent to a nitrogen atom and

enantioselective alkynylation of prochiral sp3 C-H bonds adjacent to a nitrogen atom.

34

They expand on this work and discuss what they consider to be the first simple and
efficient cross-dehydrogenative-coupling reaction between two sp3 C-Hs. Due to the
importance of vicinal diamines they explore an alternative approach to these compounds.
Typically a nitro-Mannich (aza-Henry) reaction to afford the new C-C bond B-nitroamine
derivative might be employed. Following the formation of the B-nitroamine derivative,
reduction of the nitro group to give vicinal diamines can be obtained. Using
tetrahydroisoquinoline and nitromethane, Li et al. report a 90% yield of the C-C coupled
product at room temperature with 2 mol% of CuBr catalyst with 1.0-1.2 equivalents of
tert-butyl hydroperoxide, 0.1 mmol tetrahydroisoquinoline, and 1.0 mL of nitromethane.
The exact mechanism of this coupling is unknown. However, Li et al. propose three
possible intermediates: 1) an imine-type intermediate formed via coordination to copper
through H-abstraction of the sp3 C-H adjacent to nitrogen. 2) In addition to the previous
intermediate, the copper catalyst activates the nitroalkanes to form an oxygen coordinated
copper species which couples with the previous intermediate to form the product, while
regenerating the copper catalyst. Lastly, 3) an intermediate, in which tart-butyl
hydroperoxide is involved, which further is converted into the cross-coupling products

catalyzed by CuBr.34 The mechanism of this reaction is still under investigation.

1.3.4 Secondary Deuterium Isotope Effects for Enolization Reactions

Murray et al. report a— and B-deuterium isotope effects for the enolization reaction and
equilibria for acetone or acetophenone in deuterium oxide using IH NMR spectroscopy,
triton exchange kinetics, and ab initio calculations. For the C-H,D bond adjacent to the

carbonyl in the ionization of protonated acetone, an inverse secondary isotope effect of

35

0.97/D was calculated. This isotope effect is consistent with the decrease in
hyperconjugation upon ionization. Upon further loss of hyperconjugation in the out-of-
plane CH in the enolization reaction, a secondary a-deuterium kinetic isotope effect of
1.06 i 0.02 was calculated. Murray et al. described the kinetics according to the
following first order equations for a-CH3 singlet (P), triplet a-CH2D (Q), and quintet or-
CHD2 (R):

P = P0 =I‘exp(—7tl *t)

9‘1 =1: :1: _ :1: Mi— :1: _ =1:
Q-[m] P0 exp( )‘1 t)+|:(Kl-X2)+Q0:l exp( )‘2 t)

 

 

 

R=[( AlarkzakPo )]*exp(—kl*t)+[[( klakkzakPo ]+ kzrth ]*exp(—?t2*t)

rpm-(2.4. t,_r,)r(2.3_22) ((3%)
Apfltzirpo 12*Q0 *ex — *
+ii[(x1_)‘3)*(}‘2 ‘x3)i+ (A, *l3):i+Roi p( l3 t)

P0, Q0, and R0 correspond to the initial peak area, and M, A2, and A3 are the pseudo-first-

 

 

order rate constants for deuteroxide ion catalyzed exchange of protons. Murray et al.
display the 1H NMR spectra for acetophenone illustrating the incomplete resolution of the
multiplets for the a-CH2D and a-CHD2 signals (J: 2.2 Hz for both). The data were fit to
first-order consecutive process equations with rate constants 3kHHH [OD-] = 4.20 x 10“1
s", ZkHHD[OD-] = 2.52 x 10'4 s", and kHDD[OD-] = 1.25 x 10“ s". A secondary isotope

effect of kHHH/kHHD = 1.11 and kHHD/kHDD = 1.01 was calculated. (kLij where L is the

primary hydron, hydron meaning hydrogen cation, that is cleaved and i and j refer to the
remaining secondary hydron sites). The isotope effects determined by 1H NMR for
proton exchange of CHH and CHD are not significantly different from one another.

Murray et al. examined several methods of accounting for the lack of complete resolution

36

of the multiplets including adding 0.1 Hz of line broadening, Gaussian weighting of the

F IDs, and using peak heights instead of integrals.35

1.3.5 Catalytic Hydrogenation of a-Amino and a-Hydroxy Esters

Carboxylic acids and esters are among the least reactive functional groups in catalytic
hydrogenation reactions. They usually require higher temperatures (>200 °C) and higher
pressures, and thus can result in racemization of stereogenic centers. However,
Broadbent et al. demonstrated with rheniurn heptoxide reduced in situ as catalyst,
propylene glycol yields as high as 84% from ethyl lactate at 150 °C, 250 atm, and neat.36
Studer et al. report using 10% Rh/Pt oxide (N ishimura catalyst 45.9% Rh, 19.9% Pt) for
the mild reduction of various oc—amino and a-hydroxy esters in methanol at 100 bar (1450
psi) hydrogen pressure and 25 °C while maintaining the integrity of the a-stereogenic
center. Antons et al. report using Ru oxide to hydrogenate alanine at 100 °C and 200 bar

with moderate to good yields and some racemization (93-98.5% ee).37

In looking at the available catalyst systems for the hydrogenation of acids and esters,
many of them are not commercially available, catalyst loading can be high, and
racemization can be a problem at higher temperatures. Studer et al. found that the
Nishimura catalyst was the only catalyst effective (compared with Ru and Pt oxide which
showed weak activity; and Rh, Pd, Ir, and Re oxide as well as 2% Pt, 4% Rh/C were
completely inactive) in the reduction of or-substituted esters at room temperature, and
methanol and ethanol as solvents yielded the highest amount of desired product.

Monomethylation of the amino group resulted in ~17-25% reduction in yield of

37

hydrogenated product. In addition, the alanine derivatives, N(n-Pr)2 or an NH-BOC

group did not react at all.

Studer et al. also found that Nishimura catalyst can selectively reduce an activated ester
in the presence of an unactivated ester, as is the case with the dimethyl ester of glutamic
acid. The non-activated ester forms a lactam, while the or-functionalized carboxylic acid
derivative is hydrogenated to the corresponding alcohol. Studer et al. also found that or-
methoxy and B-hydroxy esters did not react over the Nishimura Catalyst at these
conditions. In preparative-scale experiments with alanine methyl ester and methyl lactate
(2 g substrate and 6 g substrate; 10:30 catalyst to substrate ratio) yields of greater than 80
and 90%, respectively were obtained and a distinct induction period of 4 minutes was
observed after hydrogen saturation. The ease of ester reduction decreased in the order of
a-NH2 > oc-OH, NH-alkyl > B-NH2 using these mild reaction conditions and the oxidic,
bimetallic Nishimura catalyst. The addition of acid lowered the hydrogenation rate and

deactivation of the catalyst was observed at higher temperatures.

Studer et al. propose that pre-hydrogenation of the catalyst in the absence of substrate
reduces the catalytic activity. They suggest a working hypothesis of a multi-functional
active center, both basic and multi-metallic sites in which the substrate is adsorbed, then
transformed in several steps, and the desired product is desorbed. Studer et al. think the
a-NH2, a-NHR, or or-OH functionalities serve to interact with the basic (oxidic) sites,
thus anchoring the activated ester near the metallic centers. This interaction would not be

possible with the OR or NR2 groups and less favored with B-functionalities.

38

A drawback of this research is that the Nishimura catalysts are very expensive and these
reactions require high catalyst loadings.38 In terms of direct relevance to green chemistry
goals, the use of esters as substrates and alcohols as solvents imply additional steps and

disposal requirements for any practical implementation of these processes.

1.3.6 Catalytic Properties of Ru/C

Ruiz and co-workers explored the interactions between Ru nanoparticles and carbon
support.39 Carbon materials for support in heterogeneous catalytic reactions such as
hydrogenations (and hydroprocessing of petroleum fractions) have several advantages.
They are chemically inert, stable (in the absence of oxygen), have high surface area and
optimum porosity. In addition to a large surface area, the carbon-support surface must
also be chemically accessible so that active sites (unsaturated valences) must be present.
Ruiz et al. discuss that the narrow microporosity of activated carbons can hardly
participate in the adsorption process of the metallic precursors from solution, so there is
no simple relationship between the surface area of the activated carbon and the final

metal dispersion of the carbon-supported catalysts.

They prepared Ru catalysts with three different carbon material supports to evaluate how
the physical and chemical properties of the support and the Ru precursor on the surface
affect the Ru nanoparticles generated after reduction using hydrogen. The three carbon
materials were 1) powdered carbon black (Seer = 135 m2/ g); 2) mesoporous high surface
area graphite (SBET = 297 mz/g); and 3) microporous carbon molecular sieves (SBET=1500

mz/g). All were prepared using the incipient wetness impregnation technique, using

39

metal precursor solutions with the proper precursor quantity in order to obtain a 2wt% Ru
in the final catalyst. The reduction of Ru is from Ru(III) to Ru0 via a continuous flow of
20 cm3/min of a H2/He gas mixture (10% H2) with 200-300 mg of prepared sample. The
reduction started at room temperature and continued to 800 K at 2 K /min. In the
presence of chlorine, a higher reduction (470 K as compared to 400 K and 440 K for RuS
and RuNH, respectively) temperature was used due to the chlorine chemisorbing on the
metal particles decreasing their reducibility. Furthermore, Ruiz and co-workers found
that residual chloride species reduce absorption capacity of the sample by poisoning the

surface of the Ru particles.

1.3.7 Hydrogenolysis of Glycerol to Propylene Glycol

About 1 kg of crude glycerol by-product is produced for every 9 kg of biodiesel
produced. Nearly 1 billion pounds of propylene glycol is generated annually in the
United States40 and retails for about $0.71 per pound.41 The commercial route of
preparing propylene glycol is the hydration of propylene oxide derived from propylene
by either the chlorohydrin process or the hydroperoxide process. Dasari et al. discuss the
selective hydrogenolysis of concentrated glycerol to propylene glycol at lower
temperatures and pressures using various catalysts. All catalysts were pre-reduced at 300

°C for four hours under a constant stream of hydrogen.42
Shanks et al. discuss the kinetics in the hydrogenolysis of glycerol to form either

propylene glycol or ethylene glycol. Looking first at the degradation rate of the products,

they found the rate was zero order (independent of initial concentration) for the

40

concentration levels of 2.5 — 10 wt% (0.3-1.5 M) in water. The catalyst is saturated with
the glycols even at very low concentrations. The calculated degradation rates for
ethylene glycol and propylene glycol were 40 and 50 mol/(kg of catalyst)-s, respectively.
Therefore, the degradation of products needs to be included in the model for glycerol

hydrogenolysis.

Shanks et al. also examined the competitive adsorption of ethylene glycol and propylene
glycol. Results indicated that even though propylene glycol degraded slightly faster than
ethylene glycol, it was less competitive for active sites. Previous work suggested that the
polyols adsorb to the ruthenium through oxygen, thus the non-oxygenated end of
propylene glycol may be responsible for a lower binding energy compared with ethylene

glycol.43 The following model (Scheme 19) was used for ethylene glycol (EG):

 

EG+S

 

EG-S (fast)

EG-S = EG'OS (slow)
EG'°S = EG’ + S (fast)
EG’ ——* X (pH dependent)

Scheme 19. Proposed model for adsorption of ethylene glycol

A similar model was proposed for propylene glycol, where S is the catalytic site, EG' is
the aldehyde formed via dehydrogenation in the first step, and X is the degradation
products. From Gibbs free energy data, the equilibrium concentration of glycol is

strongly favored over the formation of the aldehyde (second step) under these conditions.

41

Assuming Langmuir-Hinshelwood kinetics, and assuming competitive adsorption
between the glycols, the following rate equation (Scheme 20) was initially derived:

k'iG *(iG)
kEG *EG+kpG *PG+1)

 

"r'r‘G=(

Scheme 20. Langmuir-Hinshelwood kinetic expression for competitive degradation of

ethylene glycol or propylene glycol

where i = E for ethylene glycol and P for propylene glycol, k',-G is the degradation rate

constant, iG is the respective concentrations of the glycols, and km and kPG are the

adsoption constants. Experiments were conducted at two pH levels, 11.7 and 8.0. The
glycols had about twice the degradation at the higher pH, even though, the adsorption
constants did not vary significantly. Another observation was that as the pH increased,
the amount of product degradation increased, but presumably due to ethylene glycol
degradation. At the higher pH, a 5-fold increase in the glycerol reaction correlated to a 5
times higher production in propylene glycol, but less than two times higher production in
ethylene glycol. Including glycerol in the competitive adsorption, it was found that
ethylene glycol had an inhibitory effect, whereas propylene glycol did not. Fitting
experimental data to the proposed reaction rate model, a modified equation was obtained
which included a reaction order of 1.5 for the glycerol. The mechanistic reason for this
1.5 order is not clear, but may be explained by a reaction intermediate. The modified

equation (Scheme 21) is shown as:

42

_rr. _ k10 *(IG)—SiG *k’G *(Glj)
10 (k0 *G'l'kEG *EG+kPG*PG+1)

Scheme 21. Langmuir-Hinshelwood kinetic expression for hydrogenolysis of glycerol

They performed the reactions at 500 rpm, although data was also obtained at 1000 rpm.
No difference in reaction rates was found between the two mixing speeds, confirming

that the system is not limited by external diffusion.

1.3.8 Deuterium NMR

There are several differences between lH and 2H including spin, natural abundance,
frequency, shape, quadrupolar relaxation, isotope shift, and conventional coupling.
Unlike the familiar spin 1/2 proton (1H), deuterium (2H) with its extra neutron has a
nuclear spin of 1 and a natural abundance of 0.01 5% (proton 99.985%), and an NMR

frequency (at 7.05T) of 46.05 (as compared to the proton NMR frequency of 300).

When I=1, as in 2H, there are three stable nucleus orientations in the magnetic field. The
orientations are parallel, orthogonal, and antiparallel. The mechanism is distinguished
from dipole-dipole relaxation in two ways. First, it does not require a second nucleus in
motion; the quadrupolar nucleus experiences its own fluctuating field due to its motion in
the unsymmetrical electron cloud. Second, because the mechanism is extremely effective
when the quadrupolar moment of the nucleus is large, T; can become very short
(milliseconds or less, in contrast to multisecond T; values typically seen for 1H). When

the lifetime of the spin state, as measured by the relaxation time, is very short, the larger

43

uncertainty in energies implies a larger range of resonant frequencies, or a broader signal

in the NMR spectrum.

Deuterium has a very weak quadrupole moment, with the result that neighboring protons
exhibit normal couplings to 2H. Thus nitromethane with one deuterium (CH2DNO2)
shows a 1:1:1 triplet because the protons are influenced by the three spin states (+1, 0, - l)
of deuterium. Nitromethane with two deuteriums (CHD2NO2) shows a 1:2:3:2:1 quintet
from coupling to the various combinations of the three spin states (++; +0, 0+; +-, 00, -+;

-0, 0-; --).44

Lambert et al. summarized the general isotopic shift observed for hydrogen exchanged
with deuterium. As the number of bonds increases between the perturbing and resonating
nuclei, the magnitude of the isotOpic shift decreases. Typical shifts for hydrogen
perturbed by deuterium are on the order of 0.04 ppm, 0.015, and 0.01 for 1, 2, and 3
bonds between the participating nuclei.45 Isotopic proton shifts have been observed in
several pairs of compounds. Gutowsky reports an upfield proton shift for CH2D in Ph-
CH2D of 0.015 t 0.002x10'6 with respect to CH3 in Ph-CH3. Gutowsky suggests that
differences in vibrational amplitudes of D and H atoms are the primary causes of such
isotopic shifts. The difference between the electronic wave functions of the isotopic
compounds shift the protons in CHD upfield relative to CH2. Since C-D has a smaller
vibrational amplitude, it is closer to the nucleus and imparts an enhanced shielding

effect.46

44

A coupling interaction between a spin 1/2 and spin >'/2, as in the case of a deuterium, may
lead to broadening of the individual components of both multiplets assuming a relatively
slow rate of relaxation. If the rate of relaxationbecomes fast enough it may lead to a
broadening that results in a merging of broad signals. Pople discusses the theory behind
broadening of a component with spin 1/2 coupled to a nucleus with a spin of I = 1. His
results revealed that for slow quadrupole relaxation there is individual and unequal line
broadening and as the lines continue to broaden, the outer peaks move inward until they

merge with the center to become a single signal.47

1.4 Summary of Previous Work

Over the past several years, the J ackson/Miller group has probed the Ru/C-catalyzed
aqueous-phase reductions of several organic acids, focusing on lactic acid hydrogenation
to propylene glycol and amino acid hydrogenation to the corresponding amino alcohols.
Related aqueous hydrogenolysis reactions of glycerol and higher sugar alcohols to high-

valued polyols have also been studied, providing valuable additional insights.

In 2001 , Zhang et al. examined several factors including catalyst, effect of pressure and
temperature, feed concentration, and catalyst loading in the aqueous-phase hydrogenation
of lactic acid to propylene glycol.48 Reactions were carried out in a 300 ml stirred batch
reactor (Parr Instruments) with a typical reaction time of 5 hours. Zhang found that the
RWC catalyst achieved nearly 100% conversion of 0.55 M lactic acid at 423 K (reported
as 150 °C), 2100 psi (reported as 14.5 MPa) H2, and 1.0 g of catalyst within 5 hours. In

comparison at the same conditions, Ru/Al2O3 achieved ~60 % conversion, Ru/TiO2

45

achieved ~38 %, Raney Nickel achieved ~20 %, and CuCrOx, Pd/C, and Ni/Al2O3
achieved conversions all less than ~5%. Zhang also characterized the product mixture
from hydrogenation of lactic acid using Ru/C < 170 °C the major byproducts at < 170 °C
are light hydrocarbons (methane, ethane, and propane). For reactions at temperatures >
170 °C the byproducts are 1-propanol and ethanol, with trace amounts of 2-propanol. An
optimum reaction temperature of 150 0C was determined; at higher temperatures (170

°C) cracking and excessive byproduct formations become an issue.

Due to low solubility of hydrogen in water, higher pressures are required to increase the
hydrogen concentration in the liquid. Pressures between 500 psi and 2,000 psi (reported
as 3.5 MPa and 13.8 MPa) were evaluated. Higher pressures resulted in an increased
yield of propylene glycol. Zhang calculated the absolute reaction rate (moles lactic
acid/h/g catalyst) for three different feed concentrations (0.55, 1.15, and 3.6 M) and
found that the reaction rate increases at higher lactic acid concentrations, but leads to a
lower conversion. These results suggest that the reaction may not be first-order in lactic
acid and a Langmuir-Hinshelwood rate expression might better describe the reaction

kinetics.

Zhang compared three different catalyst loadings and found that with higher catalyst
loadings, conversion rates increased, yet selectivity to propylene glycol slightly decreases
due to the increased exposure of propylene glycol to the catalyst. Zhang also looked at
the addition of potassium and calcium salts on the effect of lactic acid hydrogenation.

Typically calcium lactate is the usual product of fermentation and Zhang found that it can

46

be easily converted to propylene glycol by adding a stoichiometric quantity of sulfuric
acid followed by hydrogenation of the free acid. Zhang et al’s achieved 90-95 % lactic
acid conversion to propylene glycol at optimal conditions (423 K and ~2100 psi).
Zhang et al. continued with this work and determined the kinetics of aqueous-phase
hydrogenation of lactic acid to propylene glycol (423 K, 2,000 psi, 1.5 g of catalyst, and

1.15 M lactic acid in water).49 Mass transfer was considered in a three-phase reaction.

For the hydrogenation reaction to occur several physical and chemical steps need to
occur: (1) transport of hydrogen from the gas phase to the liquid phase; (2) mass transfer
of both hydrogen and lactic acid from the liquid phase to the catalyst surface; (3)
diffusion of both hydrogen and lactic acid within the porous catalyst to the metal surface
sites; and (4) hydrogenation of the acid to the desired alcohol via a sequence of surface
chemical reaction steps. To calculate the reaction rate of lactic acid hydrogenation to
propylene glycol, a fourth order polynomial was fit to lactic acid conversion vs time data.
The equation was then differentiated and evaluated at various times during the reaction.
Using stirring rates between 400 and 1200 rpm, Zhang et al. evaluated the influence of
gas-liquid mass transfer on the rate. They found a weak dependence of reaction rate on
hydrogen concentration. And although at lower stirring speeds (400 rpm) there is a
reduced concentration, as the reaction proceeds, the reaction rate declines and mass
transfer has less effect and the result is no observable influence of stirring speed on the

reaction rate over the course of the conversion.

47

Zhang et al. used Langmuir-Hinshelwood kinetics and described the hydrogenation of
lactic acid by the following steps (Scheme 22), where lactic acid (LA) and hydrogen (H2)
adsorb on the catalyst surface (S) and react to form propylene glycol (PG). The exact
mechanism of the surface reactions are unknown.

(1) LA + S = LA-S (fast)

(2) H2 + S = H2°S (fast)
(3) H2°S + LA-S —> P1°S + S (slow)
(4) PI'S + H2°S = PG-S (fast)
(5) PCS = PG + S (fast)

Scheme 22. Langmuir-Hinshelwood model of hydrogenation of lactic acid

Zhang et al. found that the above model which suggests molecular adsorption of both
lactic acid and hydrogen (as opposed to dissociative adsorption), the rate-limiting surface
reaction step between lactic acid and hydrogen, and desorption of propylene glycol gave
all positive kinetic and adsorption constants with the correct temperature dependence.
The rate-limiting step was found to be irreversible. The effect of propylene glycol on the
reaction kinetics was evaluated by adding propylene glycol in the amount of 2.5 times
that of lactic acid feed concentration and the rate was unaffected. So, propylene glycol
was not used in the final Langmuir-Hinshelwood model. The final form of the kinetic
expression is given in Scheme 23.

kCLAPHZ
(1+KH2PH2 +KLAC1..4 )~

 

-R G“(kmol/kg of cat/s) =

Scheme 23. Langmuir-Hinshelwood kinetic expression of hydrogenation of lactic acid

48

Also, when Zhang et al. used only the initial rate data, the Langmuir-Hinshelwood model
gave similar regression coefficients thus the catalyst deactivation was not a major factor
at these experimental conditions. This work was later re-visited by Chen (2007,
discussed later). Chen found that using a multivariable regression on the initial rate data
resulted in error from an unrealized interdependence of the kinetic parameters. But,
when using the initial rate data from Zhang et al, an activation energy of 46 kJ mol was
calculated which agreed with later results. From the model, heats of adsorption of lactic
acid and hydrogen were estimated based on the adsorption constants (KLA and KH2) at the
two temperatures. Lactic acid had a AHLA = 47 kJ/mol and hydrogen had a AHH2 = 79
kJ/mol. They also looked at the relative surface concentrations of hydrogen and lactic
acid on the catalyst. They found the ratio of empty, hydrogen-occupied (KH2PH2), and
lactic acid-occupied (KLACLA) sites is 1 : 0.05 : 0.6 at ~2000 psi hydrogen pressure, 403
K and 1.13 M lactic acid. This supports the finding that increased hydrogen pressure
increases the reaction rate. There is significant lactic acid adsorbed but relatively little
hydrogen. In summary, Zhang developed a useful kinetic model which could be used a
tool for further investigation into the mechanism of hydrogenation of lactic acid to

propylene glycol.

Jere et a1. expanded on this work by including hydrogenation of the amino acid L-alanine
((5)-2—aminopr0panoic acid) to L-alaninol (S-(+)-2-amino-1-propanol).25 This simple
chiral amino acid was chosen to further probe the mechanism and stereochemistry of the
hydrogenation by using isotopic labeling and reaction rate data. It was found that the

addition of phosphoric acid (slight excess) is required to pronate alanine (pKa = 2.34)

49

enabling the carboxylic acid functionality to be undissociated. Jere et al. found that with
sufficient acid (0.29 M), hydrogenation of protonated L-alanine (0.22 M) to protonated
L-alaninol at 398 K (reported as 125 °C) and 1800 psi D2 in D2O occurred with more
than 95% selectivity and 99% enantiomeric excess (ee). At 448K (reported as 150 °C)
and 1000 psi hydrogen pressure, the product L-alaninol was found to racemize to D-
alaninol, however, at no point during the hydrogenation reaction was D-alanine detected.
At lower reaction temperatures, less racemization is observed with essentially no product
racemization detected at 100 °C. Also at 100 °C they found incorporation of deuterium
at C2 without loss of optical purity for both alanine and alaninol (greater than 90% D
incorporation at six hours reaction time). At higher temperatures (150 oC), H/D
exchange is observed at the C3 positions with racemization of alaninol (65% D

incorporation; 81% ee alaninol) (Scheme 24).

      

     

O O
100 °C 150 °C
—_'> H3C DH2C
5 ‘ OD
Hs NH3 Us N03
100 °C , 100 °C
D D
D o D 0 DH C
H3C 100 C H30 150 C 2
‘. OD .. OD

   

s‘ s‘
s‘ s‘
H D

ND3 ND3

Scheme 24. Deuterium incorporation in D2/D2O using Ru/C

Jere et al. also found that the catalyst binding sites were saturated with hydrogen above

1000 psi and that at pressures at or above this neither hydrogen nor substrate

50

concentrations limit the reaction rate. From their results, they postulate a general amino
acid hydrogenation sequence (Scheme 12). The scheme does not explicitly represent the

role of the catalyst.

It is expected that hydrogenation of amino acid is fast relative to dehydration of the 1,1-
diol and reduction of the aldehyde. The aldehyde has not been detected but it is
postulated that the hydrogenation reaction does go through this intermediate as opposed
to enol which is catalyst-independent and not seen in the conditions investigated (Scheme
12). The H/D exchange reaction presumably occurs separately from hydrogenation. One
reason for this is that in the H/D exchange reaction there is little or no loss of chirality at
C2, despite exchange. So, a free enol, spz hybridized at C2 cannot play a role in this
reaction. Also, in the absence of additional acid, hydrogenation does not occur but H/D

exchange at C2 occurs rapidly.

These results suggest that deuterium is incorporated via the catalyst directly removing
hydrogen from the arnine-bearing carbon yielding a surface-bound intermediate that
retains the original C2 configuration. Jere et al. continued with this work and developed
a Langmuir-Hinshelwood kinetic model for predicting alanine conversion over Ru/C
catalyst.50 Reaction conditions included alanine concentrations between 0.22 to 0.46 M,
phosphoric acid concentrations between 0 and 1.2 M, 80 °C to 125 °C (reported as 353 K
to 398 K), and hydrogen pressures from 250 psi to 2,000 psi (reported as 1.7 MPa to 13.8
MPa). Jere et al. made some new assumptions than what Zhang had previous proposed.

Jere modeled hydrogen as dissociatively adsorbing on the Ru metal surface and also

51

suggested that the hydrogen and the acids (protonated alanine and phosphoric acid)
absorb on different sites. The hypothesis that protonated alanine and phosphoric acid
compete for the same sites is based on experimental data. The reaction rate of
hydrogenation of protonated alanine starts to decrease upon excess phosphoric acid
addition, which can not be explained unless competitive adsorption is at work. The
assumption that hydrogen dissociatively adsorbs comes from recent work done by both
Vannice and Neurock.51’52 Jere et al. proposed the following series of elementary
reactions to model the hydrogenation of alanine (Scheme 25).
(1) A+ + S, = A+-S1 (fast)
(2) P + S1 = P-S. (fast)
(3) H2 + 2S2 = 2H-S2 (fast)
(4) 214-82 + A+-Sl —> A1+°S1 + 282 (slow)
(5) 2H-S2 + Af-Sl = Alol+-Si + 2S2 (fast)
(6) Alol+°S1 = Alol+ + S1 (fast)

Scheme 25. Langmuir-Hinshelwood model of hydrogenation of protonated alanine

Earlier research in the J ackson/Miller group looked at aqueous-phase catalytic
hydrogenolysis of glycerol to propylene glycol.53‘54’55‘56’25’57‘58 Glycerol is a by-product of
biodiesel formation, however, there is more of an industrial demand for propylene glycol
than glycerol. Propylene glycol is used as a solvent, coolant, moisturizer in medicines
and cosmetics, and chemical intermediate. Peereboom et al. sought to gain a deeper
understanding of glycerol hydrogenolysis by studying the activated carbon support used

in the catalytic reaction.59 Two carbon supports were used, a 0.8 mm extrudate activated

52

carbon referred to as ROX and a power activated carbon designated as 3310. The

properties of the carbon supports are presented in Table 1.

Table 1. Carbon characterization by N2 adsorption at 78 K

 

 

 

 

 

Carbon type 3310 ROX
BET surface area (mz/g) 716 834
Micropore area (mz/g) 375 586
Total pore volume (cm3/g) 0.654 0.536
Micropore volume (cm3/g) 0.173 0.272

 

 

 

 

 

Typically there are local concentrations of the substrate and product (glycerol and
propylene glycol) in the activated carbon micropores. These concentrations are
presumably different than those in the bulk solution phase. In an effort to gain a bigger
picture regarding the reaction kinetics it is necessary to try and understand the difference
in concentration between solution and pore. Typically only the concentration in the bulk
solution is being measured and used for the chemical kinetics. Peereboom et al.
performed carbon adsorption experiments using propylene glycol and glycerol (0.01 M to
2.0 M each) at room temperature (298 K reported as 25 °C) and modeled the results using
both F reundlich and Langmuir isotherms. Both models are relatively simple and describe
non-competitive adsorption. The F reundlich is based on a least-squares linear regression
of experimental data from ln(CAS) vs ln(CA), where C AS is the concentration of A
adsorbed onto the activated carbon and C A is the observed A concentration in solution in
equilibrium with carbon. The data is fit to the equation C A3 = KanA to give SIOpe, n, and

intercept ln(Kp). The Langmuir is based on a plot of (C A/C A5) vs C A and fit to the

53

equation CA3 = KACACTA/(HKACA). This gives a slope of l/CTA and intercept,

(l/KACTA), where CM is the maximum concentration of A in activated carbon absorbent.

In general for ROX, the Langmuir isotherm gave the best fit of the data and the indicated
a similar value for the maximum concentration of absorbed glycerol and propylene glycol
(CTA = 1.77 and 1.64 mol/kg, respectively). However, at concentrations below the
Langmuir maximum, the quantity of propylene glycol adsorbed is greater than that of
glycerol. This is deduced from the equilibrium constants for each, KA is 0.0163 M'l for
propylene glycol and 0.00475 M'1 for glycerol (3.5 times more for propylene glycol).
The enhanced adsorption for propylene glycol is due to the less hydrophilic nature of
propylene glycol favoring it presence in the carbon micropores rather than the bulk
solution. There is also a lower extent of solvation of propylene glycol via hydrogen

bonding.

Peereboom et al. also studied elevated temperature adsorption (313 to 433 K) with
substrate concentrations of 0.1 to 0.5 M using the assumption that all adsorption sites
have equal binding energies. The Van’t Hoff equation is described as ln(KA) =
-(AH/R)*(1/T)+ln(Ko) and the heat of adsorption (AH) is the slope calculated from a plot
of ln(KA) vs l/T. The heats of adsorption for glycerol and propylene glycol were
calculated to be about 15 kJ/mol each (3.6 kcal/mol); suggesting similar van der Waal’s

type forces and weak hydrogen bonding.

54

They also studied two-component adsorption and used the extended Langmuir model to
describe the adsorption of the two species into the activated carbon micropores. Aqueous
solutions of propylene glycol and glycerol with a total species concentration of 0.05 to
0.5 M were evaluated. From the data, the results show a clear preference for propylene
glycol adsorption, by about a factor of about 3.5 times. Using a total species
concentration of 0.5 M, the heat of adsorption was calculated and verified to be constant
at 353 K, 393 K, and 433 K (reported as 80 °C, 120 oC, and 160 °C). Using the measured
micropore volume at a temperature of 353 K for different bulk glycerol and propylene
glycol concentrations it was found that the glycerol solution concentration is not very
different than the pore concentration, however, the pore concentration of propylene
glycol is about five times larger than the corresponding glycerol solution concentration.
The presence of propylene glycol lowers the glycerol pore concentration significantly. In
conclusion, they found that the pore concentrations for both propylene glycol and
glycerol deviate from the bulk solution concentrations enough that they should be

accounted for in the kinetic modeling of catalytic glycerol hydrogenolysis.

Dalavoy et al. recognized the importance of the commodity chemical propylene glycol
and decided to explore electrocatalytic hydrogenation (ECH) as a mild alternative to the
traditional chemical catalytic method (involving higher temperatures 423 K, and
pressures 1200 psi H2).60 In this method, the acid is reduced by hydrogen formed from
electrolytic acid decomposition at the surface of a catalytic metal electrode. The
hydrogen is produced in situ on the catalyst surface and this presents some challenges.

Due to the low hydrogen gas solubility, there is a need for high pressures in order to

55

achieve adequate concentrations of the gas. Also mass transport and the splitting of

hydrogen atoms on the catalyst surface can be a challenge.

A series of ECH experiments using lactic acid and a RVC electrode suffused with 5%
Ru/C powder catalyst (same catalyst as used in the chemical catalytic hydrogenation
studies) were conducted. Using a two-compartment cell, the anode and cathode
compartment were separated by a glass frit preventing the diffusion of gaseous products.
The electrode was an Ag/AgCl and the counter electrode was Pt wire. Depending on the
electrolyte and the temperature, the potential of the electrolyte was -100 to -400 mV (vs
Ag/AgCl). The final concentration of lactic acid in solution was 11.1 mmol (0.1 wt%)
and the total volume was 75 mL. The final electrolyte concentration was about 0.01 M
and argon gas was bubbled through the solution to remove oxygen and other dissolved
gases. Water was circulated through a glass ATR cell (~20 mL in volume) to help

maintain a constant temperature.

At a temperature of 363 K (reported as 90 oC), yields of 5.7% and 9.8% lactaldehyde and
propylene glycol were achieved by ECH in 0.01 M H2SO4 after 9 h and current = 40 mA.
Increased yields of lactaldehyde were achieved using HCl instead of H2SO4. Again using
11.1 mmol of lactic acid but at 343 K (reported as 70 °C), 0.0] M HCl and with a current
= 100 mA a yield of 80.3% lactaldehyde was achieved. Only 2.9% of propylene glycol
was formed at these conditions. The G3 composite ab initio method (in Gaussian 98
code) was performed to determine the heats of formation calculated from atomization

enthalpies (Scheme 26).

56

OH OH OH
OH
OH H H -1420
—-2—H—> F: "2H ""> /|\/OH
OH
O OH O

lactic acid 1,1,2-propanetriol lactaldehyde 1,2-propanediol
(lactaldehyde hydrate)

Gas -l47.9 -151.5 -142.8 -161.0
Phase
Liq. -158.8 -163.5 -156.7 -177.3
phase

Scheme 26. Heats of formation (kcal/mol) for lactic acid hydrogenation

To confirm the formation of the intermediate aldehyde, glyceric acid, glycolic acid, and
alanine were subjected to the same conditions and were all converted to the intermediate
aldehyde (glyceraldehydes, glycolic aldehyde, and 2-amino propanol, respectively). The
confirmation of the aldehyde intermediate was an interesting discovery because it was a
proposed intermediate in the catalytic hydrogenation experiments and proposed reaction

mechanism but was not confirmed.

Dalavoy et al. looked at several reaction conditions to try and optimize the efficiency of
the reaction. Results indicated that with increasing temperature, yield of lactaldehyde
and propylene glycol increased. Using identical conditions (11.1 mmol lactic acid in 0.01
M H2SO4 for 9 h at 363 K) currents between 10 and 100 mA were examined. They found
that with increasing current, the propylene glycol yield increases and the lactaldehyde
yield decreases, while the total conversion remains nearly constant. A simple

interpretation of this based on scheme G is that propylene glycol is formed by two

57

discrete intermediates (lactaldehyde hydrate and lactaldehyde). At low currents, the
lactaldehyde hydrate is weakly bound and easily replaced by the lactic acid substrate.
The, replacement occurs faster than the dehydration and capture by surface hydro gens.
However, at higher currents the lactaldehyde would more quickly be converted to
propylene glycol due to the availability of surface hydrogens. In addition, the increased
amount of propylene glycol formed at higher currents may be due to the fact that there is
higher electrode surface energy at more negative potential, this results in a higher
mobility of H atoms. Three different electrolytes at concentrations of either 0.01 or 0.1
M were studied, HCl, HClO4, H2S04, using 11.1 mmol of lactic acid, 343 K and 100 mA
current. In general, the rate of formation for lactaldehyde was found to decrease in the
order HCl > HClO4 > H2SO4. The differences in ECH reaction rates are correlated to the

extent of surface poisoning by the anions of the electrolyte.

Chen et al. took the aqueous phase Ru/C catalytic hydrogenation reactions a step further
by looking at competitive reactions with lactic acid and propanoic acid to their respective
alcohol products, propylene glycol and 1-propanol.6' The reactions were carried out
using several feed concentrations (0.05 to 5 M), temperatures (343 to 423 K), and
hydrogen pressures (500 to 1500 psi reported as 3.4 to 10.3 MPa). However, for most
reactions a pressure of 1000 psi and temperature of 403 K were used for most
experiments. Catalyst loading was 0.5 g (dry basis) in 50 mL of aqueous solution.

Scheme 27 shows the general hydrogenation reaction, conditions, and products.

58

‘ f...

OH + 2H2
+ H O
. . 5% Ru/C 403 K /l\/0H 2
OH Lactrc Acrd ? ’ g <

1000 psi H2 1,2-Propanediol

o
'1' 2H2
\/U-\0H J KH°\/\ + H2O

Propanoic Acid l-Propanol

 

 

Scheme 27. Catalytic hydrogenation of lactic acid and propanoic acid

Chen et al. found that the reaction rate of lactic acid was nearly 4x faster than that of
propanoic acid and that the conversion rates of both increased dramatically with
temperature. These results support previous findings (discussed earlier) in our research
group. In addition the selectivity to propylene glycol from lactic acid ranged from 80 to
99%, while selectivity to l-propanol from propanoic acid was less than 60% for most
conditions. The lower selectivity to l-propanol is due to a faster degradation rate of l-
propanol and also the slow hydrogenation rate of propanoic acid. Also, it was found that
l-propanol adsorbs more strongly than propylene glycol on the catalyst Ru metal and
carbon support, creating locally high pore concentrations of l-propanol near the Ru sites

which contribute to its elevated degradation rate.

At four feed concentrations (0.1, 0.5, 2, and 5 M) and otherwise identical concentrations.
initial lactic acid and propanoic acid hydrogenation rates were calculated. A nonlinear
dependence of the initial rate on the initial concentration indicates that the reactions are

not first order and as previously found in our research group, suggests that a Langmuir-

59

Hinshelwood rate expression would best describe the kinetics. Activation energies were
also calculated from plot of the natural log of the initial rate vs UT. The average

activation energy for lactic acid was 52 kJ/mol and 68 kJ/mol for propanoic acid.

In competition experiments, it was found that the addition of a second acid has an
inhibitory effect on the conversion and this effect increases with concentration. However
at concentrations over 2 M the catalyst is saturated with acids and thus the inhibitory
effect becomes constant. The influence of the product alcohols were also and it was
found that the presence of either propylene glycol of l-propanol in the starting solution
had little effect on the lactic acid hydrogenation rate. However, l—propanol did have an
effect on the hydrogenation rate of propylene glycol. Together these results suggest that
the kinetic model should include these inhibitory effects in order to properly describe the

reaction.

Chen et al. performed a mass transfer analysis on the three-phase catalytic hydrogenation
reaction and used these data and experimental reaction data to develop a kinetic model.
Again using a Langmuir-Hinshelwood kinetic model and the same assumptions from Jere
(discussed previously) a set of elementary reactions were used to describe the lactic acid
and propanoic acid hydrogenation reactions. Again the rate limiting step is the
irreversible surface reaction of the adsorbed acid and the hydrogen to form a surface
bound intermediate. The resulting rate expression for single acid hydrogenation is

presented in Scheme 28.

60

k . C .
__ rAcid (kmol/kg of catalyst/s) = ACld AcrdPH 2 2
(1 + KAeidC Acid + KA1coho1CA1cohollll + KHz PH2 )

2
kAcid = ksAKAcidKH2 Ct] (Ct2)

Scheme 28. Rate expressions for single acid hydrogenation

Performing a least squares regression analysis on 35 experiments at 403 K the adsorption
constant for lactic acid was slightly more than propanoic acid (1.75 vs 1.36 m3/kmol) and
both significantly more than that of propylene glycol (0.063 m3/krnol). However, as
noted in the competition experiments, the adsorption constant of l-propanol is much
higher (2.41 m3/kmol) and it’s presence in the starting solution decreased the propanoic
acid hydrogenation rate. The rate constants for lactic acid and propanoic acid were

calculated to be 1.03 x 10'5 and 1.66 x 106, respectively.

The Langmuir-Hinshelwood model was extended to include adsorption of additional
species for the mixed acid/alcohol hydrogenation reactions and worked reasonably well.
Using the model, it was found for 0.5 M concentrations of lactic acid and propanoic acid
at 403 K resulted in site occupancies of 39.1% vacant, 34.3% lactic acid, and 26.6%
propanoic acid. However, with 0.5 M l-propanol added the fractions change to 26.6%
vacant, 23.3% lactic acid, 18.1% propanoic acid, and 32% l-propanol. Propylene glycol
was not included in the study due to its weak adsorption. Also, at the typical hydrogen
concentration of 1000 psi, the fraction of occupied sites is 60%. These results suggest

that at high concentrations (>2 M), the surface is saturated with reactants.

61

Zhang et al., who originally (2001, discussed above) studied aqueous phase Ru/C
catalytic hydrogenation of lactic acid to propylene glycol expanded this work to.
understand the effect of biogenic fermentation impurities on the hydrogenation reaction.62
The work done in our research group consisted of using reagent grade lactic acid, which
allowed us to study different catalysts and reaction conditions as well as measuring
intrinsic reaction kinetics. However, in order for propylene glycol production from lactic
acid to compete with existing petroleum-based routes it is necessary to consider the use
of lower cost, less pure lactic acid feedstock. The current petroleum-based routes
produce propylene glycol from propylene involving either hydroperoxide intermediate of
direct hydration of propylene oxide. Conversion of propylene to propylene glycol
requires only 0.6 kg propylene/kg propylene glycol compared with 1.25 kg lactic acid/kg
propylene glycol using lactic acid as the feedstock. So, in order to compete, the

renewable feedstock must be on the order of one-half that of propylene.

Currently, lactic acid is made from glucose via fermentation with Lactobacillus
microorganisms. The yields obtained are nearly quantitative and the fermentation media
has a complex mixture of nutrients, minerals, and base (Ca(OH)2) to neutralize the
forming acid. After fermentation lactate is formed and is acidulated with H2SO4 to form
lactic acid, which is then purified (~88% wt acid in solution). Characterizing the key
biogenic impurities present in the fermentation process and their effect on catalyst

performance in producing propylene glycol was the key objective in this work.

62

Three different lactic acid samples from along the purification train of a demonstration-
scale lactic acid production facility were studied. The three samples were afully refined
lactic acid (1 ), lactic acid that was withdrawn after the extraction/back extraction step
(light brown color) (2), and lactic acid that was not refined beyond acidulation and
filtration (dark brown color) (3). The catalyst used in this studied was 5 wt% Ru/C and
was prepared in-house. Experiments were conducted in a three-phase trickle bed reactor
using the following general conditions: 403 K (reported as 130 °C), 990 psi (reported as
6.8 MPa), 10 g catalyst bed mass, 1.0 mL/min feed rate, 1.35 M feed concentration, and

H2: LA molar feed ration of 3.5 : 1.

Results from the refined lactic acid (1) indicated that increased pressure (range studied:
200 to 1200 psi) improves both conversion and selectivity. In addition, a temperature of
373 K (reported as 100 °C) gave the best combination of activity and selectivity (range
studied: 353 to 423 K). A long-term catalyst evaluation was performed (conditions: 120
kg refined lactic acid/kg catalyst at 403 K and ~1000 psi H2) over 370 h of operation and
the conversion was found to be nearly constant at 90%. The liquid by-products identified
included ethanol, n-propanol, and i-propanol in decreasing amount and together
accounted for 3.5% of the lactic acid fed. The gas by-products identified included
methane, ethane, and a small amount of propane in decreasing amounts and together

accounted for ~12% of the lactic acid fed.

To study the conversion of the partially refined lactic acids (2 and 3), a steady-state lactic

acid conversion was established using the refined lactic acid (1) followed by switching

63

the feed to one of the partially refined lactic acids (2 or 3). For partially refined lactic
acid 2, the conversion drops rapidly from 55% (established steady-state conversion using
1) to 30% for the first 5 h and slowly continues to decrease thereafter. For partially
refined lactic acid 3, the conversion drops again from 55% to 10% in the first 5 h then
essentially levels off. Switching back to the refined lactic acid feed results in a partial
recovery and stabilization of the lactic acid conversion. In order to try and improve the
conversion using the partially refined lactic acid (2), both heat treating (273 K at 5 h
under helium) and bubbling ozone gas through the sample was tried, but there was no

improvement in conversion.

In an effort to model the effect of the impurities, the various impurities (salts, organic
acids, extraction solvents, proteins, etc) were added to the purified lactic acid solution
and the results interpreted. Three different salts (N aOH, Na2SOa, Na2HPO4) were added
to the refined lactic acid feed. When adding NaOH, conversion of lactic acid decreased
by an amount proportional to the amount of sodium lactate formed. These results are
supported by previous work in our research group (Zhang et al., 2001). However, when
Na2SO4 and Na2HPO4 were added there was little reduction in lactic acid conversion. So,
addition of salts (PO43' and SO42) did not appear to poison the Ru/C catalyst. Several
organic acids (glucose, sorbitol, succinic acid, or propanoic acid) and extraction solvents
(dodecane, l-octanol, or trioctylamine) were added to the lactic acid feed and the results

for each indicated no effect on the conversion of lactic acid.

64

To study the effect of protein and protein fragments, 0.12 wt% bovine albumin was added
to the lactic acid solution. Results indicated a steady decline in conversion before
essentially leveling off. The behavior observed partially matches the results from the
partially refined lactic acid (2 and 3), indicating that proteins are partially responsible for
the decreased activity for those feeds. Using nitrogen adsorption at 78 K the pore volume
of the protein deactivated catalyst was studied. Results indicated that the deactivation
was partly caused by a combination of pore plugging and sulfirr poisoning (sources:
cysteine and methionine). The main sources of the sulfide were studied and when adding
cysteine or methionine to the feed solution, the conversions declined nearly linearly over
time. It was found that both cysteine and methionine strongly adsorb on the Ru/C and it

results in poisoning of the catalyst.

Due to previous interest in the substrate alanine, it was chosen to be studied as a non-
sulfur-containing amino acid impurity. Results showed a significant drop in lactic acid
conversion with the addition of alanine. However, when the feed stream was replaced by
pure lactic acid, the conversion returned and remained at its original value. It was
postulated that alanine reversibly and more strongly adsorbs on the catalyst surface,
blocking the adsorption and reaction of lactic acid. These results are consistent with Jere
et al.’s results (discussed earlier). Based on all of the studies, it appears that partial
plugging of the pores and reversible amino acid (sulfur containing) adsorption are the
main contributors to the reduced catalyst efficiency. Initial design considerations have
been identified from this work and depending on purification requirements and cost, a

more efficient biorefinery could be pursued.

65

Most recently published, Pimparker et al. studied the hydrogenation of three amino acids
(alanine, serine, and valine) and their mixtures to their respective alcohols.63 The
hydrogenation conditions were at 403 K (reported as 130 0C), 1100 psi (reported as 7.0
MPa H2), with 0.12 to 0.5 M feed concentration and 5-20% excess H3P04. Results
indicated nearly comparable hydrogenation rates for alanine and serine (nearly complete
conversion after 3.5 h), with significantly lower reactivity for valine (~10% conversion
after 3.5 h). However, a selectivity of ~90% was achieved with both alanine and valine,
compared with 70% for serine. A second-order polynomial was fit to the concentration
vs time data and differentiation of the curves allowed rate data to be calculated. Initial
rate data was used to estimate an activation energy and as seen previously in our research
group, a nonlinear dependence on initial concentration was observed. The hydrogenation
reactions appear to not be first order with respect to the amino acid concentration. The
activation energies were calculated: alanine, Ea = 88.5 kJ/mol; serine, E5 = 83.1 kJ/mol;
and valine, Ev = 95.5 kJ/mol and were comparable to values reported earlier by our

research group (Jere et al. reported alanine, E,l = 86.4 kJ/mol).

Mixed amino acid studies were conducted and the results indicated that as serine or
valine is added to alanine, the alanine conversion decreases. The conversion continued to
decrease as the amount of competing amino acid is increased (similar results as Chen et
'al., discussed earlier). Using an equimolar amount of competing amino acid (either
serine or valine), the alanine conversion drops from nearly 100% to ~40% after 5 h of
reaction time. A kinetic model was again developed to describe the hydrogenation

reactions. Using the model developed by Jere et al. (discussed above) a two-site

66

Langmuir-Hinshelwood kinetic mechanism was proposed. Again, it was assumed that
the protonated amino acids adsorb on one type of surface catalytic site (8,) and the
hydrogen dissociatively adsorbs on a second type of site (S2). The optimized rate
constants at 403 K for serine, alanine, and valine were 0.018, 0.016, and 0.0015
kmol/(kgcat h), respectively. The adsorption coefficients were calculated to be 56, 24,
and 124 M‘1 for serine, alanine, and valine, respectively. The larger adsorption affinity
for valine (as evidenced also by its essentially zero-order behavior) is not completely
understood. Lastly, the Langmuir-Hinshelwood rate expression allows for insight into
the relative surface concentrations of hydrogen and amino acid on the catalyst. For a
single amino acid the relative concentrations (conditions: 0.22 M amino acid
concentration; 0.3 M H3PO4) of vacant, H3PO4 occupied, and amino acid-occupied
surface sites was 9%, 52%, and 39%. In experiments with two amino acids the site
distribution was 4% vacant, 35% H3POa, 15% alanine, and 46% serine. Thus, the amount
of active sites occupied by the original amino acid (alanine) decreased from 39% to 15%,

resulting in a decreased conversion rate.

67

Chapter 2. EXPERIMENTAL
2.1 Hydrogenation Studies
The catalytic studies were conducted in water, a “green” solvent, as opposed to the more
traditional gas-phase hydrogenation widely used in the petrochemical industry.
Advantages of developing aqueous-phase conversion reactions include the following:
a) most bio-based feedstocks are produced and readily handled in water solution;
b) nonvolatile organic acids can be converted without having to be esterified, as required
for vapor-phase conversion; c) energy costs associated with vaporizing and condensing
the process streams are reduced; d) feed purity is less critical; e) the low concentration of
non-volatile residuals from fermentation need not be removed in liquid-phase processing
as they must be in vapor phase conversion; and 1) conversion temperatures well below the

boiling point of the acid or its ester are possible.

2.1.1 Materials

Several three-carbon and two-carbon organic acid substrates were chosen in order to
evaluate the extent to which the alpha substituent affects the hydrogenation rate.
Substrates included the following substituents: -H, -CH3, -OH, -N+H3, -OCH3, -
N+H2CH3, -N+H(CH3)2, and -N+(CH3)3 (note: purchased as commercially available HCl
salt (CH3)3N+CH2COOH). Feeds were purchased from Aldrich Chemical Co. The acid
used in the reaction was ACS grade 85% phosphoric acid (J .T. Baker). Water used was
of HPLC grade purity (J .T. Baker, USA). Ultra high purity hydrogen (AGA Gas
99.999%) was used for all runs. One wt% dry 5% Ru/C, pre-reduced under H2 with 0.25

M substrate and 0.3 M H3PO4 in water was used. These quantities corresponds to a 50:1

68

substratezRu ratio (5% x 1 g /101.07 = ~0.5 mmol oftotal Ru and 100 mL x 0.25 M = 25
mmol of substrate). As detailed by Jere, the acid was needed for successful
hydrogenation of the amino acids due to the zwitterionic nature of amino acids and the
requirement of a protonated carboxylic acid group for hydrogenation. In order to be

consistent, the H3P04 was added to all experiments.

A 5% ruthenium on carbon powder catalyst obtained from PMC, Inc. (Severville, TN)
was used for all experiments. The particles had a N2 BET surface area of 716 m2/ g,
porosity of 0.6, a mean diameter of 150 microns, and a bulk density of 800 kg/m3. The
as-received ruthenium metal dispersion was 8.8% (measured by volumetric hydrogen
chemisorption in a Micromeritics ASAP 2010 instrument). This corresponds to a metal
surface area of 1.6 mZ/g catalyst. Upon receiving the catalyst it was 50.9 wt % slurry in

water. For all experiments and associated reaction rates a dry catalyst basis was used.

2.1.2 Batch Reactor System

A 300 mL continuously stirred steady-state batch reactor (Parr Instrument Co, USA)
equipped with a gas entraining impeller was used for all of the reactions. The
temperature was maintained to $1 °C using a temperature controller that would either
heat or cool the vessel in order to ensure the temperature set point. The reactor was
constructed of 316 stainless steel and rated to a maximum pressure of 3000 psi at 350 OC.
Analyses were performed during the course of the reaction via a liquid sampling system.
The experimental conditions include reaction temperatures between 70 °C and 150 °C

with the majority of the experiments conducted at 130 °C, 1000 psi H2, and six hours

69

reaction time. Hydrogen was added throughout the reaction, as needed, in order to
maintain the pressure. The sample composition was analyzed every 30-60 minutes using
high-pressure liquid chromatography, for which ~1-2 mL of solution was removed from

the reaction vessel.

2.1.3 Reaction Procedure

The reaction procedure consisted of pre-reducing 1 g dry weight (2.036 g wet) of Ru/C
catalyst under 250 psi hydrogen at 250 °C, overnight. The reactor vessel was purged
three times with nitrogen to flush out oxygen, then heated to the desired temperature.
The impellor was set to ~100 rpm to facilitate a consistent temperature throughout the

vessel. Following the reduction the reactor was cooled to room temperature and the

hydrogen vented.

A carrier vessel was attached to the reactor and a 0.25 M aqueous organic acid solution
was charged into the reactor using hydrogen. After charging the reactor it was again
purged three times with hydrogen. After the final purge, the heating and stirring (~1000
rpm) were initiated. When the reactor vessel reached the desired temperature, it was
pressurized with hydrogen and a sample was taken. This sample was known as time
zero, or the start of the reaction. Reaction temperatures were between 70 and 150 °C.
Hydrogen pressure was set to 1000 psi for all reactions. Reaction times usually were six
hours with samples being taken and analyzed every 30-60 minutes. At the end of the

experiment, the vessel was cooled and the pressure was released.

70

2.1.4 Organic Acid Analysis
2.1.4.1 HPLC Analysis

High-pressure liquid chromatography (HPLC) was used in the analysis of the
hydrogenation reactions. Two analytical methods for each substrate and product set
evaluated were required. A meticulous protocol of analytical standards and techniques
was used. Internal standards were used for all experiments and the calibration of each
compound was verified before and after all samples for a particular experiment. These
methods allowed the absolute error of each data point to be assessed. In addition, several
control experiments were designed to evaluate both the accuracy and repeatability of the

analysis method.

Method 1

A Bio-Rad Aminex HPX-87H column was used for all substrates, excluding the amino
acids. The mobile phase consisted of 5 mM H2SO4. A Waters 410 Differential
Refractometer and Hitachi UV Detector set to a wavelength of 210 were used for the
detection. To ensure a constant injection volume a LDC/Milton Roy Autosampler was
employed. The pump was a Spectra-Physics SP8810 Precision Isocratic Pump. The
sample loop was 10 mL and the flow rate was optimized to 0.6 mL/min. Before each
analysis a standard mix was run to evaluate the daily performance of the instrument with
respect to the calibration curve. If a measurement error of more than 2-5% was obtained,
new calibration standards were prepared and a new calibration was performed. The
calibration curves were 3-point curves for all compounds (0.02, 0.06, and 0.1 M). The

mobile phase consisted of 5 mM H2SO4.

71

2.1.4.2 pH Analysis
The pH of the liquid samples was measured and recorded using a bench top pH meter.

The meter was calibrated before each use using solutions of pH 4, 7, and 10.

2.2 H/D Exchange Studies

Valuable information related to retention of stereochemistry during hydrogenation, has
been ascertained by our group using H/D exchange studies. To build upon the
knowledge gathered from the hydrogenation of (Jr-substituted organic acids, an extensive
H/D exchange study was done with glycine and its N-methylated analogs and will be

discussed-in this thesis.

2.2.1 Materials

All substrates were reagent-grade, including glycine (crystalline, 99%), sarcosine (98%),
N,N-dimethylglycine (99%), betaine (98%), betaine hydrochloride (99%), ethanolamine
(99%), N-methylethanolamine (98%), N,N-dimethylethanolarnine (98%), choline
chloride (98%), and choline bicarbonate (80%) were obtained and used as purchased
from Aldrich Chemical Co. HPLC grade purity water was used for all aqueous phase
experiments (J .T. Baker, USA). The hydrogen used was ultra high purity (Aga Gas,
99.999%) and the deuterium was ultra high purity (Spectra Gases). For deuterium
exchange experiments, deuterium oxide (Sigma Aldrich, 99.9%) was used in place of

water.

72

2.2.2 Batch Reactor System

Paramagnetic broadening was observed in early 1H NMR studies, presumably caused by
solution contact with the stainless steel reactor and its parts. Thus, extensive effort was
put forth to minimize the amount of stainless steel exposed to the reaction mixture. The
300 mL stainless steel batch reactor (Parr Instrument Co.) was custom coated with a
Teflon based material by Unconventional Solutions (Milford, MI). The stainless steel dip
tube used for liquid sample removal was replaced with a glass dip tube. A Teflon tube
was placed over the stirring shaft. The stainless steel impellor was replaced with a
titanium impellor. The stainless steel cooling loop was removed from the system. This
eliminated the opportunity for both heating and cooling; therefore, in order to compensate
for the lack of dual-control, the rate of heating was slowed. A glass tube was placed over
the stainless steel thermocouple. The glass tube was filled with a small amount of water

to facilitate the heat transfer.

2.2.3 Reaction Procedure

The reaction procedure consisted of pre-reducing 1 g dry weight (2.036 g wet) of Ru/C
catalyst under 500 psi hydrogen at 150 °C, overnight (although the hydrogenation studies
above had used 250 °C, it was necessary to reduce the temperature from 250 °C to 150
°C in order to maintain the integrity of the Teflon lining). The lower temperature of 150
°C proved to be sufficient; however, a longer reduction time (overnight) was used (as
opposed to 4 h). Several experiments were conducted using either a catalyst reducing
temperature of 250 °C or 150 °C to evaluate whether there was an effect on the

hydrogenation reaction. For the substrate glycine (0.25 M) with 0.3 M H3PO4 at 130 °C

73

and 1000 psi H2, the hydrogenation rate was nearly identical for both reducing
conditions. The reactor vessel was purged, heated to the desired temperature, and stirred

using the same method as described previously.

The carrier vessel was attached to the reactor and the 0.25 M aqueous (D2O) solution was
charged into the reactor using hydrogen or deuterium. The charged reactor was then
again purged three times with hydrogen or deuterium. After the final purge, the heating
and stirring (~1000 rpm) were initiated. When the reactor vessel reached the desired
temperature, it was pressurized with hydrogen and a sample was taken. This sample was
defined as time zero, or the start of the reaction. Reactions were usually conducted at
temperatures between 50 and 100 °C. Hydrogen or deuterium pressure was set to 1000
psi. Reaction times usually were six hours with samples being taken and analyzed every
30-60 minutes. At the end of the experiment, the vessel was cooled and the pressure was

released.

2.2.4 Amino Acid Analysis
Both HPLC and lH NMR were used in the analysis of the I-I/D exchange studies. Results

from the HPLC allowed for quantitation of the various substrates and products for each
sample (samples collected every 30-60 min for six hours). Results from the 1H NMR
allowed for individual isotopomer ratios to be calculated for each sample. Using these
data, kinetic models were developed in order to model and understand the deuterium

exchange behavior.

74

2.2.4.1 HPLC Analysis

Method 2

A Hewlett Packard Series II 1090 liquid chromatograph instrument with a Waters 410
differential refractometer and Hitachi UV detector model L-4000H was used for the
amino acid analyses. The mobile phase was 0.01% acetonitrile in 0.05 M phosphoric

acid.

2.2.4.2 ‘H NMR

 

Liquid samples were analyzed using a Varian 500 MHz superconducting NMR-
spectrometer operating at 499.738 MHz interfaced with a Sun Microsytems Ultra 5
UNIX console. NMR tubes with an outer diameter of 5 mm were used (Wilmad-
Labglass, Buena, NJ). A glass insert (2 mm outer diameter; 1.25 mm inner diameter)
filled with 0.25 M acetic acid internal standard was used for all experiments. The
chemical shift of the internal standard was set to 1.9008 ppm. The integration area of the
internal standard was set to 30 for all experiments and based on calibration curves, the

concentrations of the observable amino acid isotopomers were calculated.

75

Chapter 3. CALCULATIONS
3.1 1H NMR Peak Fitting and Deconvolution
3.1.1 NMR Calibration Calculations
The slope of the area versus the concentration for all compounds was determined both
experimentally and mathematically. Experimentally, three-point calibration curves were
constructed for all amino acid and amino alcohol compounds. On average, the slope of
the calibration curves was 9. Mathematically, the ratio of outer NMR tube to inner insert
tube volume was calculated based on the dimensions of the tubes. Equation 1 below was

used for the calculation.

Equation 1. NMR tube ratio: outer-to-insert volume

z ((oo — wt)2 - (013(3))

2 , where
IDin

 

OD = outer diameter (mm) of outer NMR tube
wt = wall thickness (mm) of outer NMR tube
OD = outer diameter (mm) of NMR insert

ID,n = inner diameter (mm) of NMR insert

Using the low and high values specified by the manufacturer for the dimensions, an
average of 9.008 :t 0.298 was calculated for the volume ratio. The prepared
concentration of the internal standard, acetic acid, was 0.249 M. The peak integral for
the internal standard, acetic acid, was set to 30. Thus, the peak area per hydrogen for the

internal standard was 10.

76

3.1.2 Peak Fitting Ecmations
The data from the IH NMR was exported into Excel and using a Lorentzian function, the
peak integrals were determined by fitting the data and using deconvolution. The general

equation for a Lorentzian function is found in equation 2:

Equation 2. Lorentzian lineshape

= (g) * - 1 ), where g = half-width and x0 = unshifted peak center

7‘ ((x-xo)2 +g2

The integrated area of this function is unity.

 

Global values and local values were defined. For each H/D exchange reaction, after
Fourier transform and phasing, the NMR spectra were exported into Excel as x,y data
where y represented the peak height and x represented the peak position, ppm. The data
were then analyzed in parallel as a set. For each such set, the fitting procedure defined
quantities that were global (i.e. common to all spectra) and those that were local (i.e.
specific to a given spectrum). Global values included isotope shifts, coupling constants,
and multiplet proportions, while local values included chemical shift, baseline offset,
linewidth, and of course, the individual deconvoluted peaks' intensities. Least-squares
fitting to all datapoints was used to align and quantitate the Lorentzian functions. The
respective fractional areas were then sealed to molarities using the ratios of measured
total substrate peak integral to that of the internal standard. Modeled spectra and their
integrals were plotted and compared to the originals to ensure accuracy. In the case of a
triplet, a triplet scaling factor was defined. And in the case of a quintet, a quintet outer

and inner scaling factor was defined.

77

The equation used for fitting the data is below (3, 4, 5). It is broken into three parts
representing the singlet (CH2 or CH3), triplet (CHD or CH2D) and quintet (CHD2)
resonances. In the case of fitting the data which represented the methylene, only the first
two parts of the equation (3 and 4) were used. The portion of the equation used to fit a
quintet was omitted due to the fact that for the methylene the singlet represented CH2,
and the triplet represented CHD. However, in the case of the methyl, all three parts of the
equation were used (3,4,5). The singlet represented CH3, the triplet represented CH2D,

and the quintet represented CHD2.

Equation 3. Singlet peak fit

r (a0*[§] \

 

 

 

 

 

= baseline offset + 2
(X-Xo)2+g
\ J

Equation 4. Triplet peak fit

/
+ a3 * {—g—D

( 31!

s3 1 53
+ + +
((x—xo +.IS+HI))2 +g 2) ((x—xo +IS)2 +g2) ((x—xo +IS—HD)2 +g 2)

 

78

Equation 5. Quintet peak fit

(sell

 

 

 

 

 

 

 

SSout 2"‘S‘Sin
+ 2 2
U ~x0+2*IS+2*HD)2+(350ut*g)2) (x—x0+2*IS+HD) +(55in*g)
:1: + 3
((x—x0+2*IS)2+(g)2)
((x—xo+2*IS—HD)2+(35in*g)2) (x—XO+2*IS—2*HD)2+(SSOut*g)2

Samples were analyzed using 1H NMR every 30-60 minutes of the reaction. Using the
above models to fit the full time point sets of NMR spectra, the global values described
above were optimized for all time points. The local values were individually optimized
for a particular time point (i.e. NMR spectrum). The local linewidth was defined by the
global linewidth multiplied by the local broadening. The global linewidth for the singlet
was usually in the range of 0.0006 to 0.001 ppm. The global linewidth for the triplet was
usually slightly larger and in the range of 0.0009 to 0.002 ppm. The global isotope shift
was between 0013-0015 ppm. The global H-D coupling constant was 0.003 to 0.006.

The local values were dependent on the individual experiment.

Starting with a good estimate of the global and local values, the sum of the squares error
was minimized for each spectrum using Solver. Once the fits were optimized for
individual spectra, Solver was run on all spectra to obtain one set of values for each

experiment. The sum of the integrals was calculated for both the actual data and the

79

fitted data and compared. Using the amplitudes (a1, a3, and a5), the mol% of singlet,

triplet, and quintet was calculated.

3.1 .3 Peaflrtegration Calculations

MestReC is a powerful NMR software package used for analyzing high-resolution NMR
data (Mestrelab Research, Spain www.mestrec.com). The raw F ID data was exported to
Excel, then to calculate the individual concentration of isotopomers, the below equations

were solved.

Equation 6. Total H concentration for each spectrum at time t

[IS]* (areasfl )

area [SJ
#of H

1H1=

, where

 

9.008 *(

[IS] = concentration of the internal standard, acetic acid (0.24881)

area s,t = area of the sample at time t

area 13,1 = area of the internal standard, acetic acid, at time t (set to 30)

# of H = 3, for the 3 methyl hydrogens in acetic acid (internal standard)

The value 9.008 is the slope of the calibration curve for the sample versus the standard
(see NMR Calibration Calculations 3.1.1). The area fractions as [H] were determined
using the following equations (7,8,9). The variables a1, a3, and a5 represent the mol% of
the singlet, triplet, and quintet, respectively. For example, to calculate the singlet mol%,
al, for methylene the amplitude of the singlet is divided by 2; divided by the sum of the

amplitude of the singlet divided by 2 plus the amplitude of the triplet.

80

Equation 7. Singlet area fraction

al
= H =1: __
. [ ]t(a1+a3+a5)

Equation 8. Triplet area fraction

a3
= H * ____..__
[ ]t(a1+a3+a5)

Equation 9. Quintet area fraction

a5
= H * .______
[ 1‘ (al+a3+a5]

The concentration of isotopomers at each time point can then be calculated using the

equations below.

Equation 10. CH2 concentration

Equation 11. CHD concentration

[CHD] t: (area) CHD

Equation 12. CD2 concentration

[CD21t=(lCHzlt0+lCHDItol—(ICHzlt+lCHDl t)

The equations for the methyl are analogous to the above equations. Next, the peak areas

of each of the isotopomers can be derived using the following equation.

81

Equation 13. Peak area of individual isotopomer

[s] ”(Till ]*9.008
0

[IS]

 

 

(areaS )t = , where S = individual isotopomer

The conversion, yield and selectivity can be calculated from the concentrations of the

individual isotopomers. Equations 14, 15, and 16 were used.

Equation 14. Conversion

= [S] to- [S] t
13] t0

Equation 15. Yield

_ [Slt

’1—81—5

Equation 16. Selectivity

, where S = individual isotopomer and to = time zero

_ yield
conversion

3 .2 H-D Content Inventory

3.2.1 Percent H in Water

The molecular weight of H20 is 18.015 g/mol and the molecular weight of D20 is 20.028

g/mol, and at 25 °C, their respective densities are 0.9971 and 1.104 g/mL. Thus their

concentrations are 55.34 and 55.14 M, respectively. Their equivalent hydrogen or

deuterium concentrations are 110.68, [H]H20, and 111.28, [D]D2o. The internal standard,

acetic acid, in the insert had a bench concentration of 0.248 M (at 25 °C), thus the

equivalent hydrogen concentration was 0.746 M. The peak area of the internal standard

82

was set to 30. Using the ratio of the outer-to-insert volume of 9.008, the concentration in
the insert was calculated to be 0.083 M (not the true concentration, but would be
equivalent to this if it had been in the whole volume). During all calculations the

untruncated values were used until the end.

If the water was H2O (no D20 or HOD) then the water peak should integrate to 40,173.

Equation 17. Peak area of water in sample

[1H] H20

— * (area),S = 40,173

[15} in J

The acetic acid was prepared in deuterium oxide and the 1H NMR of acetic acid was
taken in the outer NMR tube (not the insert). If the water in the internal standard, acetic
acid, was composed of H20 (no D20 or HOD) then the water peak should integrate to
4,462.

Equation 18. Peak area of water in internal standard

[M

US] )*(area),5 = 4.462

When measured, the peak area of the water in the acetic acid was 52.52, thus the %H in

the deuterium oxide of the internal standard was taken as 52.52/4,462, thus 1.2%.

3.2.2 Percent H in Sample
The molarity of hydrogen and deuterium could be determined experimentally from the

NMR integrals and also mathematically by the following equations.

83

Mathematical:

The initial liquid volume in the reactor, VL, was 100 mL. The pressure, P, was 1000 psi
(68.046 atm). The temperature, T, was 100 °C (373.15K). The initial gas volume, Vg in
the reactor, was 200 mL. Using the below equations, an example calculation for

sarcosine will be carried out.

Equation 19. Moles of deuterium in D20 (1)

 

* V
moles of DL 2 [p L j , where mw = molecular weight
mw

The moles of DL are 5.55, thus in D2O there are 11.11 moles of D. In practice, we will

assume [H] and [D] are equal to ~111 (used 111.11).

Equation 20. Moles of hydrogen in H2 (g)

 

, P * Vg .
moles of Hg = 2 * R T , where R = 1deal gas constant
*

The moles of Hg are 0.438, thus in H2 there are 0.876 moles of H.

Equation 21. Substrate — NHx contribution

moles of HN = x * [substrate]* VL , where [substrate] is the prepared bench concentration

The typical bench substrate concentration is 0.25 M, thus for sarcosine the moles of HN

are 0.05 (x=2).

84

Equation 22. Substrate - methylene (CH2) and methyl contribution y*(CH3)

moles of HC = (2 + y * 3)* [substrate]* VL , where y = number of methyls

Again, using the typical bench substrate concentration of 0.25 M, the moles of Hg in
sarcosine are 0.125 (y=1). Using equation 19 to calculate the moles of deuterium and
equations (20, 21, and 22) for the sum of the moles of H in the different isotopomers, the

H-D inventory can be determined.

Equation 23. Time zero H-D inventory

 

ZmolesofH ]*100
Zmoles of H + Zmoles of D

o/°Ht0 =(

The sum of the moles of H at time zero is calculated from equations 20, 21, and 22 and
equal 1.05 for sarcosine. The sum of the moles of D at time zero was calculated from

equation 19 and equal 11.11. Thus, the %Hto in the sample (sarcosine) is 8.6%.

Equation 24. Molarity of H at time zero

 

2 moles of H j

H =H *
[ ]t0 [ ]H20 l2molesofH+ZmolesofD

Based on the above calculations, the molarity of H at time zero for sarcosine is 9.602

mol/L (again where [H] H20 = 111.11).

Equation 25. Molarity of D at time zero

[DI t0= [H] H20 — [HI t0

85

The molarity of D (for 0.25 M sarcosine in D2O) at time zero is calculated to be 101.51

mol/L.

Once the reactor has reached the reaction temperature, the vessel is pressurized and the
reaction is started. A sample (~2.5 mL, Vremoved, which is equal to 0.14 moles) is

removed and is referred to as the “time zero” sample. Then, the hydrogen and deuterium
composition of the remaining solution is calculated from the NMR analyses using the

below equations.

Equation 26. Moles of D (l) at time zero

 

 

moles of D = 2 * ( Vremoved * p :1: 2 moles of D
to mw (2 “10165 0f H)’ “10135 0f He + 2 moles of D

Equation 27. Moles of H (l) at time zero

 

 

V *
molesof H10 = 2*( removed p]*[(

(2 moles of H) — moles of HC
mw

2 moles of H)- moles of HC + Z moles of D

The number of moles of Hc are subtracted out from the total moles of H at time zero
because at time zero a negligible amount of methylene and methyl hydrogens are
exchanged, thus more deuterium is removed. However, after time zero a H-D
equilibrium is assumed. After the sample (~2.5 mL) is removed, the reactor is re-
pressurized to 1000 psi. The amount of hydrogen re-introduced to the reactor is ~100 psi.
Thus the moles of hydrogen added at that time can be calculated from the equation

below.

86

Equation 28. Moles of H (g) at time zero

P*Vg
R*T

 

moles of H gto = 2 *[ J, where R = ideal gas constant

From this point, it is an iterative process whereby the composition can be re-calculated

after each sample (~2.5 mL) is removed.

Equation 29. Moles of D (l) at time t

moles of D” = moles of DL — moles of Dto
Equation 30. Total moles of H (1) and H (g)

moles of H” = (Zmoles of H)— moles of Hto + moles of Hg,t0

Equation 31. %H at time t

moles of Ht]

 

°/H =
0 t1 molesothl +molesof D”

Equation 32. Molarity of H at time t
[HItl = lHlHZO * (%H)tl
Equation 33. Molarity of D at time t

lDlt1 = [HIH20 -[HItl

After this point, the process repeats. The time 2 sample is taken (~2.5 mL) and the new

composition of the solution is calculated. After the reaction is complete, the final sample
composition (%H, [H]tf, [D]tf) can be compared with the experimental values obtained

from the NMR integrals. The results from both methods were in agreement.

87

3.3 Kinetic Rate Data Determination

3 .3 .1 Concentration Calculations

The peak areas from the 1H NMR fitting are used to calculate the concentration of the
individual isotopomers for each sample time. The methylene carbon is referred to as “A”
and the methyl carbon is referred to as “C”. The concentrations at time zero were
determined using the internal standard, acetic acid, and the ratio of area to concentration.
Again, the slope was determined to be 9.008, the [IS] = 0.25 M, and the peak area was set
to 30. All concentrations after time zero were based on the initial concentration of the
individual isotopomer and ratio of the area at time, t, to the area at time zero. However,
the total concentration of all isotopomers was verified using HPLC to ensure the starting
material was not consumed before time zero (once the reactor has reached reaction

temperature and pressure) of the experiment. The below equations were used.

Equation 34. AH2 concentration at time zero

(area AH2//)
2
- . t0 :1: [[5]

(area 1.8/3')t0

9.008

 

 

IAHzltO =

Equation 35. AH2 concentration at time t

(area AH2/j
t

l H I l H l /2
A 2 t — A 2 t0*(areaAH2/)
t0

/2

 

88

Equation 36. AHD concentration at time t

* [area AH%) t
t0 (area AHD/2 ) t0

Equation 37. AD2 concentration at time t

[AHD], =1AH21

 

[AD2], =1AH21 ,0 —11AH21.+IAHD1,)
Equation 38. CH3 concentration at time zero

(areaCHg, )
3
- . ‘0 *[IS]

(area 1% )t0

9.008

 

 

[CH3]t0 =

Equation 39. CH3 concentration at time t

(area CH 3 /)
I

/
[CH3], =ICH31,0.( 3

area CH3/j
/ 3 t0

 

Equation 40. CH2D concentration at time t

area CH2D

I /21.
area CH3

( /3th

Equation 41. CHD2 concentration at time t

[€11let = [CH3lt0-

 

(area CHD2 )t

t0* areaCHy
3 t0

Equation 42. CD3 concentration at time t

 

[CHD2]t = [CH3]

[CD3], =[CH31t0-IICH3I ,+[CH2DI t+lCHD21 t)

89

Equation 43. Aqueous hydrogen concentration in H20 and HOD

 

average([AH2],[CH3])
average((area AH2/2),(al’€a CH3/3)) ,

[H]L =(area HL )*[

 

where [ average([AH2 ],[CH3])

is equivalent to a calibration based
average((area AH2/2),(area CH3/3))

on the internal standard (IS) acetic acid concentration to area and equals 9.008.

Equation 44. Aqueous deuterium concentration in D20
1000
D = --— — H
The sum of the hydrogen area can be calculated for each time point, and during the

course of the reaction the areas for C-H sites decrease as hydrogen is replaced with

deuterium.

Equation 45. Sum of the H area

2 Harea = (area AH2 ) + (area AHD) + (area CH3 ) + (area CH2D) + (area CHD2)

The aqueous hydrogen concentration data was fit to the following equation over time.
Equation 46. Aqueous hydrogen concentration equation of fit

[HlL = A * exp(- B * t)+ C * t + D , where A, B, C, and D are fitted variables and t = time

The sum of the squares error is minimized to determine the fitted variables.

90

3.3.2 Kinetic Model A: 10-variables

The time zero concentrations for the isotopomers of the methylene and methyl sites were
determined from the time zero areas for each. The kinetic model assumed that the
adsorption/desorption of H20 (or D20) is fast and that the ratio of H:D in the solution
represents the H:D on the surface. This assumption proved to be adequate for the
experiments with 1000 psi D2 and D20. However, after reviewing the results from the
experiments with 1000 psi H2 and D20 a modified kinetic model was used (to be
described later). Using the original kinetic model, for experiments with 1000 psi H2, the
initial hydrogen on the surface was assumed to be 111.11. The initial deuterium on the
surface was assumed to be zero. For experiments with 1000 psi D2, the initial deuterium
on the surface was set to the amount of deuterium in the liquid (usually less than 3%) and
the hydrogen was equal to the difference between 111.11 and the amount of deuterium.
The hydrogen and the deuterium in the liquid at any time, t, was calculated using the rate

equations in 47.

The experiment describing the hydrogen-deuterium exchange on the methylene carbon is
depicted by the following two reversible, consecutive first-order reactions. Again the
role of the catalyst, including the adsorption/desorption events are represented in the

aqueous hydrogen (HL) and consequently the aqueous deuterium (DL).

91

Equation 47. Rate equations for H/D exchange at the methylene position

km

AH2+DS (TIA—2 AHD+HS

sz
AHD + Ds <__—-> AD2 + HS
2A

Hence, doing a numerical step-by-step fit, where the time-step increment (At) was 0.5
min, the equations become:

Equation 48. CH2 concentration at time t

[AHZL = [AHZL—At _]k [A *[AH211-AI*[D]L,t.—At _ k _]A *[AHDlt—At *[H]L,I—Al )* At

Equation 49. CHD concentration at time t

lk-IA * [AHDIt—At * [HlL,t—At + sz * [AHDlt—At * [D]L,t-At)

* At
- (km * [AWL—At * [D]L,t-At 3‘ k—ZA * IADzlt_At * [H]L,t—At)

[AHDIt : [AHDlt—At '—

Equation 50. CD2 concentration at time t
[AD2]. await... — It-.. wins]... 41312.2, - t2. *[AHDI._..*1D1,,_,.1*At

The experiment describing the hydrogen-deuterium exchange on the methyl carbon is
depicted by the following three reversible, consecutive first-order reactions.

Equation 51. Rate equations for H/D exchange at the methyl position

I 5
CH3 '1' DS

k C
$—— CHzD + H8
‘IC

k2C

CH2D+DS (‘15—'— CHD2 +HS
'2C

k3C

CHD2 + Ds 61:— CD3 + HS
~3c

92

Hence, doing a numerical step-by-step fit the equations are defined as follows.

Equation 52. CH3 concentration at time t

[CH3]t = [CH3]t—At ‘ 1le *[CH31t—At * [D]L,t-At — k-IC * lCHlet—At * [HIL,t-At )* At

Equation 53. CH2D concentration

lk—1C * [CHlet—Ar * [Hint-At + kZC * [CHZDL—At * [DlL’t—At )] * A

= D —
[CHZDL [CH2 L—At i'lklc * lCHslt—At * [D]L,t—At + k~2C * [CHDllt—At * [mm-Al)

Equation 54. CHD2 concentration at time t

(k-ZC * [CHD2 lt-At * [H]L,[—At + k3C * [CHD2 lt-At * [D]L9t_At )] * A

_ = H _
[CHD2 L [C D2 L—At |:— (kzc * [CHlet—At * [D]L,t—At + k—3C * [CD3 lt—At * [H]L,t—At )

Equation 55. CD3 concentration at time t

[CD3 ]t = [CD3 lt—At - lk—3C * [CD3 lt—At * [HIL t-At _ k3C * [CHD2 lt—At * [D]L, t-At )* At

The 10-variables: km, k-1 A, k2A, k-2A, k1c, klc, k2c, k-2c, k3c, and k-3c were calculated by
minimizing the sum of the squares error between the experimental and the calculated
concentrations for all measured time points. To verify that this was a short enough time
increment for direct analysis of these multi-hour reactions we confirmed that the resulting
k values were not significantly affected by further 2-fold shortening or lengthening of the

time increment.

3.3.3 Kinetic Model B — 6 variables
The second kinetic model used to fit the data incorporated primary and secondary isotope

effects. All of the rate variables above (Model A) represent the same hydrogen trading

93

processes except for the isotopes involved. Thus it should be possible to simplify the
model. The equations from Model A were used in Model B, but here, the 10 rate
variables were re-defined as isotope-perturbed instances of the same events in order to

incorporate the possibility of a primary and or secondary isotope effect. The six rate
variables used to optimize the data fit were: kiA, pA, SA, kic, pc, and sC. The kinetic 2
isotope effect at the methylene carbon and methyl carbon(s) is represented by km and kic,

respectively. A primary kinetic isotope effect arises when the bond to the isotopic

substitution is broken at or before the rate-limiting step and for the methylene carbon and
methyl carbon(s) is represented by p A and pc, respectively. When hydrogen-deuterium
substitution is not involved in the bond breaking or bond making step, an isotope effect
(albeit smaller) may still be observed and is represented by SA and sc for the methylene

carbon and methyl carbon(s), respectively.

Equation 56. Forward and reverse rate constants for H/D exchange at the methylene

position
kIA = 2 * kiA
k.
k-IA =45-
PA
k.
sz =4}:
SA
k 2*kiA
-2A =————
(PA *SA)

94

Equation 57. Forward and reverse rate constants for H/D exchange at the methyl position

 

 

le=3*kiC
k.
2*k'C
k2C= SCI
k _ 2*kiC
'ZC (PC*SC)
k.
k =_£_
3C (SC*SC)
k 3*kiC

'3C :(PC *SC ”CI

The time step used in calculating the kinetic rate variables was 0.5 min.

Kinetic Model Modification

The kinetic model for isotope exchange was varied to explore the effects of various
assumptions with the intent of optimizing the fit to the experimental data. Experiments
were conducted with either 1000 psi H2 or 1000 psi D2 and for reactions at the same
temperature, the calculated rate constants should be the same. Using the kinetic model
discussed above, the H/D exchange for the experiments with 1000 psi H2 and D20
appeared to be going to fast for some of the early experimental points. This is partly due
to the fact that in the kinetic model, the hydrogen and deuterium available for exchange
with the substrate was assumed to be in equilibrium with the concentration of hydrogen

or deuterium in the liquid. But from the experiments with H2 and D20, it appears that

95

there is a delay in the substrates' H/D exchange reaction due to the time it takes to reach

equilibrium among liquid, gas, and surface hydrogens.

For experiments with 1000 psi D2, the aqueous hydrogen concentration is almost an order
of magnitude lower than the typical 15% or so that is observed with H2 (Figure 67). This
time evolution is important, but also reflects the gas/liquid exchange. To attempt to
address this, the hydrogen and deuterium available were defined by equations below.

The equations for calculating the hydrogen and deuterium on the surface, assuming the

amounts are in equilibrium with the liquid concentrations, at any time are below.

Equation 58. Expression for H on the surface at time t

Hat = [Hlear +(2*IAH2].-a +[AHDIt_At +3*[CH31._A. +2*[CH2DI..ai +lCHDzl.-At)
— (2 *[AH2]1+[AHD]t + 3 * [CH3]t + 2 - [CHZD]t + [CHDZM

- k H * llD]L,t—At * “st—At — [H]L,t-At * Ds,t—At)

Equation 59. Expression for D on the surface at time t

D“: 111.11—H53

The kinetics for the surface hydrogen incorporate the hydrogen-deuterium surface-liquid

exchange represented by kH. At any point in time, in the solution the sum of hydrogen
and deuterium concentrations equals the total concentration [Ct]. During the reaction, we

will assume the hydrogen on the surface and the deuterium concentration in solution is in

equilibrium with the deuterium on the surface and the hydrogen concentration in solution.

96

Once the equilibrium between the hydrogen and deuterium is reached, their
concentrations in the liquid and their respective amounts on the surface can be expressed

by the equilibrium equation below.

Equation 60. Equilibrium expression for hydrogen and deuterium in the liquid and on the
surface

[DIL + H5 ‘—— lle. + Ds

Let HS + DS = Ct, the total catalyst site density (Ct) for hydrogen adsorption in the

reactor. From this, the rate of deuterium on the surface versus time can be derived.

Equation 61. Rate of deuterium on the surface versus time

“135%” = k a.r([[)]L :1: H5 ”l-Hll. =I= D5), thus

= k*([D]L -(c. -D.)-D. *(lctl—[Dlr))

Assuming the total concentration in solution is in equilibrium with the surface, then the
total catalyst site density for hydrogen adsorption, Ct, equals the concentration of
hydrogen and deuterium in the solution, [Ct]. Factoring this out of the equation, the
effective rate constant can be described as k*Ct and is defined as kH in the rate equations.

D, is a variable and represents the fraction of total sites occupied by deuterium. This

variable can be used in all kinetic expressions recognizing that all the rate constants will

then also be effective rate constants with C, in them, kH=k*Ct.

97

Chapter 4. RESULTS AND DISCUSSION
4.1 Control Eigeriments
4.1 .1 Hydrogenation Studies
Early research in our group highlighted the need for both catalyst and hydrogen pressure
in the hydrogenation of organic acids.25’48’49‘50 Several control experiments have been
performed and results confirm that molecular hydrogen and catalyst are required for the
hydrogenation reaction. Using otherwise identical procedures, a reaction was performed
replacing H2 with helium. The catalyst was exposed to 1000 psi He pressure. No product
formation was observed for the reaction of 0.25 M refluxed lactic acid at these conditions

(H20, 130 °C, 1000 psi He, 1 g dry 5% Ru/C).

 

0.30
+ Lactic Acid

+1.2 Propylene Glycol

A

o.25~——- :g— -— v
o/T—

l
l
l
l
41
0
l

Concentratlon (moIlL)
O
a

l
l
l
l
l
l
l
l
l
l
l

g:
8
i
l
l
l
l
l
l
l
l
l
l

 

 

0.05-~——-—~—-»e- -- . f. - , ~ ~ ~ ._

 

 

 

0.00 FRY“. I I r . r I I l I .
0 50 100 150 200 250 300 350 400

Tlmo (mln)

Figure 1. Hydrogenation of 0.25 M refluxed lactic acid (0.3 M H3PO4, H20, 130 °C,

1000 psi He, 1 g dry 5% Ru/C)

98

The concentrations of lactic acid and product alcohol, 1,2-propylene glycol, are shown in
Figure 1. The purchased (Sigma-Aldrich) lactic acid was 85% (w/w) in H2O, so before

using, the water was removed by refluxing the lactic acid.

A reaction with 0.25 M refluxed lactic acid in the absence of 5% Ru/C catalyst was
performed and the results indicated no product formation after six hours (0.3 M H3PO4,
H20, 130 °C, 1000 psi H2, no catalyst). The concentration data for lactic acid and the

desired alcohol product, 1,2—propylene glycol are shown in Figure 2.

 

 

 

 

0.30 -~—-~—---r,
A A 9 ¢ e e 4
02516:-“ L- L L- a-
+Lactic Acid
-—I— 1.2 Propylene Glycol
5' 0.20 1%.. ~- —— ~A ~~ - ----- - --~
0
.5.
8
3015 — r - 74 i '
c i
e .
O l
e .
8 0.10 - - - ~— -— — , — t
0.05 -..._._- , , - ~~ -72 77~ , 7 ,, , ~ emu-g
i
0.00 I———I———-r—I . I . I r I . I . I i
0 50 100 1 50 200 250 300 350 400

Time (min)

Figure 2. Hydrogenation of 0.25 M refluxed lactic acid (0.3 M H3P04, H20, 130 0C,

1000 psi H2, no catalyst)

For all experiments discussed in this thesis, the reaction procedure involves pre-reducing
the 5% Ru/C catalyst under H2 pressure (150 °C overnight with ~300 psi H2). The

sample is then injected and the reactor is heated to the reaction temperature, 130 °C (one-

99

step introduction of the solution). One-step introduction involves adding the substrate
(0.25 M) at the point in which the reactor has reached the desired temperature. The next
step is to immediately pressurize the vessel with H2. The point when the H2 is introduced

is the point that the reaction is considered to have started (t = 0 min).

In some experiments a small amount of conversion was observed at time zero. So, a

' control experiment was designed to evaluate the effect of adding the substrate after the
reaction vessel was at temperature and pressure. This was done in an effort to avoid any
pre-reaction conversion that can occur during the heat up phase of the reactor. The
refluxed lactic acid was mixed with 25 mL of water and set to the side. The other 77 mL
(all H2O) of the 102 mL solution was injected into the reactor and allowed to heat to the
reaction temperature. Once at the desired temperature, the lactic acid solution was
injected (two-step introduction of the solution). Results indicated no significant
difference between this and introducing the sample at room temperature and allowing it
to heat up (in the absence of H2) in the reaction vessel to 130 °C (Figure 3). The bench
concentration for the two-step was 0.32 M (1.3 M in 25 mL) and slightly higher than the
traditional experiment in which the bench concentration is 0.25 M (one-step

introduction).

100

0.40 a —— _____-_.__. .. ..-.---_.-._ .-

 

 

—I-— Lactic Acid -—I~-~~ 1.2 Propylene Glycol
- A— - Two-Step Lactic Acid — -A~- - Two-Step 1.2 Propylene Glycol

.0
N
or

Concentration (molIL)
p o
—b N
0’! O

 

 

 

0 50 100 150 200 250 300 350 400
Time (min)

Figure 3. Hydrogenation of refluxed lactic acid (0.3 M H3PO4, H20, 130 °C, 1000 psi

H2, 1 g dry 5% Ru/C)

All hydrogenation experiments were carried out in aqueous solvent in the presence of 0.3
M H3PO4. A control experiment using 0.25 M glycolic acid (GA) in the absence of
additional acid was conducted to determine the effect of equimolar H3PO4 on the
hydrogenation rate. No significant difference in the conversion of glycolic acid and the
yield of ethylene glycol (EG) was observed over six hours (H20, 130 °C, 1000 psi H2, 1 g
dry 5% Ru/C). The concentrations of glycolic acid and product alcohol, ethylene glycol,

are presented in Figure 4.

101

025 W ______.-._. ....._-_._..- .. .........-... . .
—I— GA

4 - EG

 
  
 

 

 

~A~ GA with addtl acid
A EG wlih addtl acid
0.20 -. ~ , , _.._-_ -L "L,
3
=6
E 0.15
c
0
=1
5
c
8 0.10 - _ ,
,2/
lf’rd/l
,/"l/ in I
0.05 - ————— 7- / ’14- 7 47 if
// I .
I .
,-/
0.00 I r i a . . T . .
0 50 100 150 200 250 300 350 400
Time (min)

Figure 4. Hydrogenation of 0.25 M glycolic acid in the absence and presence of 0.3 M

H3PO4 (H20, 130 °C, 1000 psi H2, 1 g dry 5% Ru/C)

It is known from previous research in our group that the catalyst surface interacts with the
solvent, substrate, and presumably the acid.25’59‘63 Thus, the presence of additional acid
in the case of glycolic acid may not have a significant effect due to a high affinity of the
glycolic acid for the catalyst surface, but this may not be true for every substrate and all
conditions. The effect of H3PO4 on hydrogenation and H/D exchange will be discussed

further in section 4.4.

4.1.2 H/D Exchange Studies
Several control experiments were conducted in the absence of the 5% Ru/C catalyst. An
experiment with 0.25 M glycine in D20 at 130 °C and 1000 psi H2 pressure showed no

deuterium incorporation after five hours of reaction. The 1H NMR specifics for samples

102

taken at time zero and after five hours are shown below (Figure 5). Based on the 1H
NMR integrations using acetic acid as the internal standard, the concentration, at time
zero and after five hours, was essentially unchanged. A very small, but visible amount of

deuterium incorporation was observed, partly due to a small amount deprotonation at the

 

 

 

 

 

 

a-carbon.
9000 _-._-__ ._ -___-__. .---__- ., _ , .- 8000
—I—Time: 0 Hours + Time: 5 Hours
8000 .. -—- i # 4 — a , — - - A f e , 7000
7000 -_ 4 4 e — - - - a» , , ~ 6000
6000 ~»— — - - —- — - - -~ - e 5000
5000 .. — i i A A A” - — --------- e - . 4000
4000 a» — — — — —— --— l —— — —— -— .. 3000
3000 4» e i - A 7 ~— ' e -F V» -~ 2000
2000 --————- I 1 -~ - - 1000
J k .
0 V -1000
-1000 ----—----——-—— —- ---__-__-- -2000
3.42 3.40 3.38 3.36 3.34 3.32 3.30

Figure 5. 1H NMR of H/D exchange at the methylene position of 0.25 M glycine (D2O,

130 °C, 1000 psi H2, no catalyst)

An experiment with 0.25 M sarcosine in D2O at 130 °C and 1000 psi H2 pressure was
allowed to react for six hours in the absence of the 5% Ru/C catalyst. About thirty
percent of the methylene was partially deuterated during the reaction. The 1H NMR of
time zero and after six hours is shown in Figure 6. No observable deuterium
incorporation was obtained at the N-methyl position on sarcosine after six hours. The 1H

NMR of time zero and after six hours is shown in Figure 7.

103

 

2250 -- »~--——-~—~~-~--~-~-—~~- ----_____--_____-- i 2000
+Time: 0 Hours + Time: 6 Hours

2000 ~

1750 -

1500 3

1250 ~

1000 ~

750 *

 

500 ‘

 

 

250

 

 

3.47 3.45 3.43 3.41 3.39 3.37

Figure 6. 1H NMR of H/D exchange at the methylene position of 0.25 M sarcosine (D20,

130 °C, 1000 psi H2, no catalyst)

 

 

 

 

 

 

3500 _..-.. _..__..A_“_-.__-_- .. --------.-.-- - - W - - -..._ .. _ W.---.-- - ...----_n_,_ 3000
+Time: 0 Hours a» Time: 6 Hours
3000 -—- -—— -— — —-#—- -#---~— i ~#r ~ — -—-——#--- - ~ 2500
2500 «- ___,--- - # -# - ## ##n --- ,__ - 2000
2000 --—- __-,- ##w .# -## . __-_,, - # 1500
1500 ---—— --— —~- # -# —---— - ##- # 1000
1000 ~- — # #- # ,k-- #- ‘ ' -—-- ~ # - - ~ ~ 500
500 0
0 - -~ «- -- -500
-500 -1000
2.58 2.57 2.56 2.55 2.54 2.53 1 2.52 2.51 2.50

Figure 7. 1H NMR of H/D exchange at the methyl position of 0.25 M sarcosine (D2O,

130 °C, 1000 psi H2, no catalyst)

104

The data were fit to the kinetic models discussed in chapter 3 (Figure 8). The H20 and
HOD experimental (symbols) and fitted (line) data refers to the percent H in water and is
on the right y-axis (secondary y-axis). The experimental (symbol) and fitted (line)
concentration of CH2 (referred to as AH2), CHD (referred to as AHD), and CD2 (referred

to as AD2) are on the left y-axis (primary y-axis).

The rate constants for the hydrogens on the methylene carbon were calculated using the
IO-variable model. The forward rate constant for the first hydrogen exchanged with
deuterium was found to be k1A=9.8x10'6 L/mol/min, but the forward rate constant for
exchanging the second hydrogen with deuterium was found to be k2A=O (both reverse rate
constants were zero k- 1 A=O, k-2A=O). All rate constants have the units L/mol/min. There
is an appreciable amount of H/D exchange at the methylene position (Figure 8). Results
from the H/D exchange studies (to be discussed later) have shown that sarcosine is the
most reactive substrate (as compared to glycine and N,N-dimethylglycine). So, the H/D
exchange observed in this control experiment (no catalyst) may be partly due to some
residual catalyst being present in the reactor lines or reactor wall. However, the majority
of the H/D exchange is presumably due to the non-catalytic process of deprotonation of
the weakly acidic zwitterion. The compound is most stable in the zwitterionic form. But,
again at 130 °C there is a larger population of the higher energy form that can undergo

deprotonation.

105

 

 

 

 

 

030 -_.-__.._._.___._..-._-.___-._..._-.-...._...__--_.-_.... ...__..-..._.__-_._.-_.-__._-..___-_-_-.._W- ..,,......_..._.-......_...._-_.._...----.-..-.. “WWW-T 42
——AH2 ——AHD —ADZ
I AH2_exp A AHD_exp o A02_exp
—H20 and HOD 0 H20 and HOD_exp

0.25 —* ~— - - *—~ ,. — i 4- 3.5
g . + 2.8
.E. x

a
o c
g - + 2.1 g
0
g o.
5
0 0.10m - — ~ — — —— — - 7 ~~1.4
A
A /l/
A ./’”"’
0.05 « — / -. 0.7
//,.W
0.00 —/. +1 . . c . c . .s o
O 50 100 150 200 250 300 350 400
TImo(mln)

Figure 8. Fit of the kinetic rate data for 0.25 M sarcosine (D20, 130 °C, 1000 psi H2, no

catalyst)

A control experiment with 0.25 M choline bicarbonate in D20 at 130 °C and 1000 psi H2
was allowed to react for four hours in the absence of 5% Ru/C catalyst. Again, there was
no observable deuterium incorporation at either of the two methylenes (3.88 ppm next to
-OH and 3.32 ppm next to -N+(CH3)3) or on the nitrogen methyl groups (3.02 ppm) '
(Scheme 29). This is an expected result and differs from the case of the amino acids in

that there is not an activating carboxylate group to facilitate non-catalytic H/D exchange.

l/ 0
Ho/\/ \

HO 0'

Scheme 29. Structure of choline bicarbonate

106

The 1H NMR taken at time zero and after four hours for the more downfield methylene
position (3.88 ppm) corresponding to the methylene next to the -OH is shown below
(Figure 9). The 1H NMR of time zero and after four hours for the more upfield
methylene position (3.32 ppm) corresponding to the methylene next to the -N+(CH3)3 is

shown in Figure 10.

An experiment with choline bicarbonate was performed with 1 g dry 5% Ru/C (D20, 130
°C, 1000 psi H2). Hydrogen exchanged with deuterium at all three positions on choline
bicarbonate, however, the 1H NMR peak data was too convoluted to distinguish between
the CH2 and CHD for the two methylene positions and the CH3, CH2D, and CHD2 for the
three N-methyl groups. So, the overall integrals were compared. From the results, the
methylene hydrogens adjacent to the -OH group exchanged the most quickly reaching a
total conversion of 58% after six hours (sum of CH2 and CHD). The methylene
hydrogens adjacent to the -N+(CH3)3 and the N-methyl hydrogens exchanged at nearly an
identical rate reaching a total conversion of 35% each after six hours. These results
confirm that H/D exchange occurs faster at positions adjacent to an oxygen as opposed to
a nitrogen. Further evidence for this will be discussed in the H/D exchange of

ethanolamine.

107

 

—l—- Time: 0 Hours + Time: 4 Hours

150 #7 7

100 -» 7

 

 

 

 

 

 

 

3.96 3.94 3.92 3.90 3.88 3.86 3.84 3.82 3.80

Figure 9. 1H NMR of H/D exchange at the methylene position next to the —OH of 0.25 M

choline bicarbonate (D20, 130 °C, 1000 psi H2, no catalyst)

 

 

 

 

 

 

 

250 2.00
-I— Time: 0 Hours + Time: 4 Hours
200 -— 7 . . . » 150
150 .- 100
100 ~~ * 7" - _ 5o
50 w y 0
0 WWW L—» -50
-50 -100
3 40 3 37 3.34 3 31 3 28 3 25

Figure 10. 1H NMR of H/D exchange at the methylene position next to the —N+(CH3)3 of

0.25 M choline bicarbonate (D20, 130 °C, 1000 psi H2, no catalyst)

108

A control experiment with 0.25 M betaine ((CH3)3N+CH2COO') at 130 °C with 1000 psi
H2 in the absence of catalyst was compared with the identical experiment in the presence
of l g dry 5% Ru/C. Figure 11 depicts the 1H NMR for the methylene position at both
time zero and after five hours for the experiment conducted in the absence of catalyst.
The H/D exchange reaction occurs at a comparable rate in both the presence and absence

of catalyst.

Figure 12 shows the 1H NMR for the methylene position at both time zero and after five
hours for the experiment conducted in the presence of catalyst. In the experiment with
catalyst, after only three hours nearly 50% of the CH2 is partially or fully deuterated.
After six hours, 86% of the CH2 is partially or fully deuterated. In contrast, the CH3
peaks show no evidence of H/D exchange (in either the absence or presence of catalyst).
The 1H NMR of the nitrogen methyls after five hours of reaction time are presented in
Figure 13 (no catalyst). The integration area and peak shape is unchanged throughout the

course of the reaction.

The lH NMR at time zero and after five hours for the methyls on nitrogen for the
experiment with betaine in the presence of l g dry 5% Ru/C are shown in Figure 14.
Again, there is no HfD exchange on the nitrogen methyls even in the case of catalyst.

The peak shape is the same for time zero and after five hours.

109

 

 

 

 

 

 

1000 800
+Time: 0 Hours +Time: 5 Hours
800 ~ - 600
600 <~7 , 7 ~ 400
400 -» - 7 . 200
200 O
0 -200
3.80 3.78 3.66 3.64

 

Figure 1 l. 1H NMR of H/D exchange at the methylene position of 0.25 M betaine (D20,

130 °C, 1000 psi H2, no catalyst)

 

 

 

 

 

 

 

 

 

900 ._.__._____ 700
'— 120l ,. Time: 0 Hours .{- Time: 5 Hours
800 ~ .. .. 7 , . - . ~ » 600
700 . O ~ 500
600 - » 400
500 7777777~ 300
400 i~ , ~ 200
300 - ,W- 7 W — 100
200 0
100 - - -100
0 -200
3.77 . . . . . . 3.63 3.61

 

Figure 12. 1H NMR of H/D exchange at the methylene position of 0.25 M betaine (D20,

130 °C, 1000 psi H2, lg dry 5% Ru/C)

110

 

4000 ~~~~7 7~-77—---~-~-- 3500
+Time: 0 Hours ~I— Time: 5 Hours

 

 

 

 

 

3500 77#-- ---#7 # #7 # # # — #7- #7 #7 #7 # # #7 3000
3000 77— #- 7- ——--# #7 77 # 7# —- l ~# # # 77# # -# 7 2500
2500 77 #- 7777777 7 # 77# 7# I 7# #7 7# #7 # # 7 2000
2000 77 -—# # # # 7 7 ~ — 7 1500
1500 77-——#— --# -— #7 7# 7 # 77 7 1000
1000 77 #7 # # # # 500
500 o

0 ,_ - 7500

3.16 3.14 3.12 3.10 3.08 3.06 3.04 3.02 3.00

Figure 13. 1H NMR of H/D exchange at the methyl position of 0.25 M betaine (D20, 130

°C, 1000 psi H2, no catalyst)

 

 

 

 

 

+ Time: 0 Hours 7:7— Time: 5 Hours
300077-7—#—# 77 --—# -- -#77 -—-——##7 . ##7## 77 # 7 72500
I
I
2500 777- ## 7#7 #7 ##7-7 ' --7777— 77 ##7##7- 7 2000
200077 # # #7 # # : 777## 7 7 --7-77771500
u
I
1500 #- # ## ##7 #77# = — #### 7777 71000
'
1000 7 #-#~—-7 -— #7 # 7 . ——-- 500
s;
500 7 3~ o
o 7500
3.20 3.15 3.10 3.05 3.00 2.95 2.90

Figure 14. 1H NMR of I-I/D exchange at the methyl position of 0.25 M betaine (D20, 130

°C, 1000 psi H2, 1g dry 5% Ru/C)

111

Figure 15 shows the concentrations of CH2, CHD, CD2 and the 3 CH3’s over time. The
hydrogens on the methyls of the positively charged nitrogen do not exchange. Although
the peaks areas not the same height, the peak areas after—six hours are about 6% less than
the peak areas at time zero. This percentage loss is consistent across all experiments and
may represent initial adsorption of the substrate into the catalyst. The rate constants for
both experiments were calculated and using kinetic model A (Chapter 3.1.1). The
forward rate constants for the experiment with no catalyst were: k; A = 3.5x10'5 and sz =
2.6x10'4 L/mol/min. The forward rate constants for the experiment with 1 g dry 5% Ru/C
catalyst were: km = 4.1x10'5 and kg», = 1.6x10'2 L/mol/min. The rate of km is very
similar in both reactions suggesting that betaine is undergoing H/D exchange at the same
rate irrespective of whether catalyst is present. This suggests that betaine is undergoing a
non-catalytic mechanism for H/D exchange. This is supported by the work of Richard et
al. They evaluated the acidity of the or-proton of betaine at 25 °C in D20 and observed
nearly 90% of exchange of the first or-proton via deprotonation by the hydroxide ion and

formation of the intermediate enol.

112

+0112 —I—CHD +CH3
- t - - CH2_no cat - A ~ - CHD_no cat 7 A CDZ_no cat — t - - CH3_no cat ‘
0.30 ~- - — - —# —77 — #77 777— _. 77-. -. . -, 7

0.35

 

 

“7' l

0.25

.0

N

O
i

  
 

Concentration (moIIL)
O
a:

 

 

 

0.10 .._ — i — 2 7
0.0577 -— -- 7- ,777/5 " -- 7 , 77 -- g
' ‘Wfl7l" I ' T T 2
-_. ”“5 i
0.00 1 i
150 200 250 300 350
Tlmo(mln)

Figure 15. Concentrations of CH2, CHD, and CD2 of 0.25 M betaine with and without 1

g dry 5% Ru/C catalyst (D20, 130 °C, 1000 psi Hz)

4.2 Substrate Adsorption on the Catalyst/Carbon Support

4.2.1 Adsorption on Ru/C

An early observation in this research was the initial disappearance of substrate in the
reaction vessel before reaction conditions were achieved. The concentration of the
substrate was measured after preparing the solution on the bench (bench concentration).
Again, the concentration was measured after it was exposed to the catalyst (reactor
concentration). It was proposed that the substrate was being adsorbed (possibly
irreversibly) onto the carbon support. When doubling the catalyst loading using refluxed
lactic acid as the substrate, the initial concentration of substrate was reduced by more
than half (Figure 16). This provides further evidence that the reaction is kinetically

limited by catalyst sites and not by interphase mass transport. In addition, a control was

113

conducted with no H2 pressure at room temperature, knowing that the hydrogenation
reaction would not be occurring at these conditions, and an initial conversion of about
3.5% was observed, yet no products were detected. The amount “lost” or converted after
the bench concentration is exposed to the catalyst is fairly consistent across all

experiments.

0.30

 

I Bench Concentration

I Reador Concentration
0.28 -# # # 7 777 7 7 7-

 

S3

M

a)
i

Concentration (mol/L)
_o o
M M
m A
I l
i

0.207# . ,, ,,

     

 

 

   

0.18 ~

19 Dry Catalyst 29 Dry Catalyst 2 9 Dry Catalyst - Repeat

Figure 16. 0.25 M Refluxed lactic acid concentrations measured (bench and reactor)

It is not clear yet, whether all of the initial material adsorbed is desorbed and allowed to
react or whether some amount of the substrate is permanently lost on the carbon support.
In our research group, Peereboom et al. performed carbon adsorption experiments at
room temperature (25 °C) using propylene glycol and glycerol (0.01 M to 2.0 M each) at
room temperature and modeled the results using both F reundlich and Langmuir
isotherms. Peereboom et al. observed that substrates adsorb on the catalyst and that

propylene glycol has a greater preference for the carbon micropores over the bulk

114

solution than glycerol, due to its less hydrophilic nature. Using the Langmuir model,
they found that the bulk glycerol concentration is not very different from the pore
concentration. However, the pore concentration of propylene glycol is 5x larger than the
glycerol concentration.59 In general, it appeared that after time the molecules were getting
torn apart and did not desorb. In order to further follow up, a Ru-free sample of the
carbon support used in our Ru/C catalyst from PMC was obtained and adsorption studies

with different substrates were conducted.

4.2.2 Adsorption on Carbon Support

Six substrates were evaluated to determine the amount of substrate adsorbed onto the
carbon support. The carbon was 3310 lot #28850 from PMC. The percentage of water in
the carbon support was 8.81%. Three-point calibration curves were nm for each
substrate using the methods described in chapter 2. The six substrates studied were:
lactic acid, glycolic acid, propylene glycol, ethylene glycol, phosphoric acid, and n-
butanol. Normally, in the experiments discussed in this thesis the ratio of moles of
substrate to grams of dry catalyst has been 0.025 mol/g dry catalyst. Thus, in this work 5
mL of a concentration of 0.25 M substrate was combined with 0.05 g of dry catalyst

resulting in the same ratio as most experiments.

The results indicated that 7% of lactic acid adsorbed onto the carbon support, more than
any other substrate. Figure 17 presents all of the substrates (lactic acid = 7.1%, butanol =
5.7%, glycolic acid = 4.8%, propylene glycol = 3.4%, phosphoric acid = 1.0%, ethylene

glycol = 0.8%). The error bars represent the measurement error as determined by the

115

relative standard deviation from the calibration curve for each substrate. The amount of
lactic acid adsorbed on the carbon support is similar to the amount of lactic acid lost after
being exposed to the 5% Ru/C catalyst. It can thus be assumed that most of the substrate

lost is being adsorbed on the carbon support.

100% _
90% 7# # #7 #7 # #

80% 7 #7 #77 # # # # #7 #- #7 77 7 #
70% 7 ##7 # 7 #7 ##7 77# 77# 7# #77 # 77 #7 7 # i
60% 7- #7 7# # ## #7 #7 7 # #- 7# #7

50%. WW W W, W., --W W- . W. ___W_ -
40% _ ____ ____ _W_ W_ .W W _ _W_ __

30% 7# 77 # #77 #7 #7 #7

20% 7-—--#-— ## #77 7 -#-~-- —~ #77 # 7## 7##7

Percent Adsorbed on Carbon Support We)

 

10%

 

 

 

 

 

0% - i 7
Lactic Acid Butanol Glycolic Acid Propylene Phosphoric Ethylene
Glycol Acid Glycol

Figure 17. Substrate adsorption on the carbon support

4.3 Effect of HCI

4.3.1 Hydrogenation Studies

A precedent for chloride inhibition has been observed when using a ruthenium catalyst.
In our research group, Dalavoy et a1. performed electrocatalytic hydrogenation
experiments using 11.1 mM lactic acid and a RVC electrode suffused with 5% Ru/C
powder catalyst. The electrolytes examined were H2SO4, HCl, and HCIO4. At 90 °C in
0.01 M H2804 a yield of 9.8% propylene glycol was achieved after 9 hours and at the

same conditions but with 0.01 M HCl instead, only 2.9% of propylene glycol was

116

achieved. The major product was lactaldehyde. Increasing the electrolyte to 0.1 M HCI
yielded more propylene glycol after 48 hours (6.6%) than 0.1 M H2SO4 (2.7% yield of

propylene glycol) (11.1 mM lactic acid, 100 mA, and 70 °C).

We continued this work and performed a pair of experiments with 0.25 M betaine
hydrochloride ((CH3)3N+CH2COOH-Cl') in the presence of 0.3 M H3PO4 at both 100 °C
and 130 °C, 1000 psi H2, and 1 g dry 5% Ru/C. The results were compared to identical
experiments using 0.25 M betaine ((CH3)3N+CH2COO' (chloride-free» at 100 °C, 130 °C,
and 150 oC. None of the desired product, choline chloride, was formed from betaine
hydrochloride hydrogenation. The effect of a slight excess of HCl may be to inhibit the
hydrogenation reaction. Figure 18 shows the hydrogenation reaction of betaine

hydrochloride.

Results for chloride—free 0.25 M betaine ((CH3)3N+CH2COO') at the same conditions
(130 °C) indicated 41% conversion and formation of 7% product, choline, after six hours.
When the reaction was run at 150 °C the reaction was faster and yielded a conversion of
74% and 17% choline yield. As the temperature was increased, more conversion of the
substrate was observed. At a reaction temperature of 100 oC there is minimal conversion
and no product formation. The results for all three temperatures (100 0C, 130 °C, and

150 °C) are presented in Figure 19.

117

 

 

 

 

—I— Betaine'HCl 130C --I-- Choline Chloride 1300
k 7 Betaine‘HCl 1000 A Chioline Chloride 1000
0.25 I 7 # 7 #
7t A A A
A '
£0.20 7 7# 7 7 # # # # # #
e
.E-
S
= 0.15 7 7# # # 7 7
.2.
B
8
fl
8 0.10 - 7 # 7 # # #
0.05 7 # 7 # # # # # 7 7
0.00 7 = i = i = r : . ---------- [I ----------- . -l
0 50 1 00 1 50 200 250 300 350 400
Time (min)

Figure 18. Hydrogenation of 0.25 M betaine hydrochloride (0.3 M H3PO4, H20, 100 °C
and 130 °C, 1000 psi H2, 1 g dry 5% Ru/C)

100%

 

I 100C

121130C B1500

w#'w-

 

 

 

80%-7

70% 17

60% 7

50% 4

40%7 7

30% -

20%- 7

10%- -

 

 

 

Choline Yield

Betaine Conversion

Figure 19. Hydrogenation of 0.25 M betaine after 6 hours (0.3 M H3PO4, H20, 100 °C,

130 °C, and 150 °C, 1000 psi H2, 1 g dry 5% Ru/C)

118

When comparing the experiments using betaine hydrochloride and betaine at the same
conditions, it appears that the presence of the chloride is inhibiting the hydrogenation
reaction presumably due to the competitive interaction of the chloride with the catalyst

surface.

There is an unknown peak that grew in at roughly the same rate as the choline. The
unknown peak is most likely trimethyl amine. Due to the formation of the unknown peak
in the experiment with 0.25 M betaine ((CH3)3N+CH2COO'), an additional experiment
was run in the absence of H3PO4 (1 50 °C, 1000 psi H2, and 1 g dry 5% Ru/C) to evaluate
whether the unknown peak was still formed. Results indicated no choline formation and
no unknown peak formation after five hours. Part of the reason the unknown peak was
not observed may be that if it is trimethylamine this good leaving group would be less
stabilized in the absence of acid. No choline was formed because neutral betaine is in the
zwitterionic form and hydrogenation of the carbonyl of the carboxylic acid will only

occur if the carboxylic acid is protonated.

An experiment with 0.25 M glycine was run at 130 °C, 1000 psi H2 in the presence of 0.3
M H3PO4 with 1 g dry 5% Ru/C in both the absence and presence of 0.3 M HCl.
Addition of the HCl in proportion to the substrate was strongly inhibitory, cutting glycine
conversion from 96% to 58% and ethanolamine yield from 80% to 25% after six hours.
Figure 20 illustrates the results. In this experiment the amount of additional acid (0.3 M
H3PO4 and 0.3 M HCl) is twice the amount (0.3 M H3PO4) used in our experiments with

0.25 M substrate and 1 g dry 5% Ru/C. Jere et al. examined the effect of excess H3PO4

119

and found a ~20% reduced alanine conversion (0.22 M alanine, 100 °C, 1000 psi H2, and
l g dry 5% Ru/C) with 0.6 M H3PO4 (as compared to 0.3 M H3PO4).50 When further
increasing the amount of H3PO4, the conversion continued to decrease slightly (730%

reduced conversion with 1.2 M H3PO4 as compared to 0.3 M H3PO4).

 

 

+ G'Ycine Conversion I Ethanolamine Yield
+ Glycine Conversion - with H01 0 7 Ethanolamine Yield 7 with H01
100% 1—— 7 -—-7 777—77 —#7 #7 77—. _-. - W. -W- _W_ _

80% ~

60% ~

Conversion or Yleld (‘16)

40%

20%77 - 7,

 

 

 

0 50 100 1 50 200 250 300 350 400
Time (min)

Figure 20. Hydrogenation of 0.25 M glycine in the absence and presence of 0.3 M HCl

(0.3 M H3PO4, H20, 130 °C, 1000 psi H2, 1 g dry 5% Ru/C)

An experiment was run with 0.125 M glycine at 130 °C with 1000 psi H2 and slight
excess equivalents of both H3PO4 and NaCl (0.15 M each) and 0.5 g dry 5% Ru/C to
determine if NaCl had the same effect on the hydrogenation reaction. The same volume
of solution was used (100 mL) so the substrate concentration was more dilute (not
allowing a direct comparison). It appears that the HCI exerted a larger effect than NaCl
but a direct comparison using the same substrate concentrations and catalyst amounts

should be done. Figure 21 displays the conversion of glycine in the hydrogenation

120

reaction with no chloride (0.25 M glycine, 1 g dry 5% Ru/C, and 0.3 M H3PO4), with 0.3
M HCI (0.25 M glycine, 1 g dry 5% Ru/C, and 0.3 M H3PO4), and 0.125 M NaCl (0.125
M glycine, 0.5 g dry 5% Ru/C, and 0.15 M H3PO4). In the experiment with HCl,

additional inhibition may be occurring due to the fact that HCI is a very strong acid (pKa

~ -4) and may be changing the sample media enough to slow the hydrogenation reaction.

—I— 0.25 M Glycine (0.3 M H3PO4. 1 g dry 5% RulC)
7&7 0.125 M Glycine with 0.15M NaCl (0.15 M H3PO4, 0.5 g dry 5% RulC)
7.7 0.25 M Glycine with 0.3 M HCI (0.3 M H3PO4. 1 g dry 5% RulC)

100%

 

80%1~

60% +

Conversion (7.)

40% 1

20% ~'

 

 

 

0%

 

0 50 100 1 50 200 250 300 350 400
11me (min)

Figure 21. Hydrogenation of glycine alone and in the presence of NaCl or HCI (H3PO4,

H20, 130 °C, 1000 psi H2, 1 g dry 5% Ru/C)

In the hydrogenation experiments with amino acids, the role of the H3PO4 is to protonate
the carboxylate, due to the preference of the amino acid to be in the zwitterionic form.
The result is H2PO4'. In the experiment with both 0.3 M H3PO4 and 0.3 M HCl the HCI
which is the stronger acid (pKa ~-4 as compared to ~2.15) is protonating the glycine
carboxylate. Thus, H3PO4 is still in solution (as opposed to H2PO4'). The idea that

perhaps it is the H3PO4 that is causing inhibition could be considered. However, in our

121

hydrogenation experiments with glycolic acid (discussed in section 4.4) we see no effect

with the addition of an equimolar amount of H3PO4.

From the experiments with betaine and glycine, it is evident that the presence of chloride
slows or completely inhibits (as in the case of betaine hydrochloride) the hydrogenation
reaction. Chloride inhibition is also observed in the H/D exchange studies with betaine

hydrochloride (which will be discussed next).

4.3.2 H/D Exchange Studies

As found in the hydrogenation experiments (Figure 18), betaine hydrochloride
((CH3)3N+CH2COOH-Cl') was also unreactive toward H/D exchange. There was no
measurable amount of exchange of the methylene or methyl hydrogens with deuterium.
An experiment with 0.25 M betaine hydrochloride in D20 was run at 100 °C with 1000
psi H2 and 1 g dry 5% Ru/C. The results were compared with an experiment using 0.25
M betaine ((CH3)3N+CH2COO') at 130 °C. The experiments with betaine hydrochloride

showed no deuterium incorporation on the methylene or methyl carbon sites (Figure 22).

This common inhibition by chloride could suggest that the hydrogenation and the H/D
exchange catalyzing sites may be the same. But, we know from the experiment with
betaine that the hydrogens on the methylene carbon undergo exchange with deuterium in
the absence of catalyst due to the fixed cation promoting solution deprotonation (Figure

l 1, 130 °C). So, if the chloride was deactivating the catalyst it should only affect the

122

hydrogenation reaction. The reason for the chloride in betaine chloride inhibiting the

H/D exchange reaction is still unknown.

 

 

5000 T #7777~7~-77- 77777 4500

+ Time: 0 Hours + Time: 6 Hours
4500 +

   
 
 
 
 
 
 
 
 
 
 
 

-- _ 777 77 3500
4000 7
3500 7

#77 7 77 77 777 7 72500
3000 7
2500 7- 77 7777 ##7 777 7 77 777 7 7777 7 771500
2000 7

r 500
1500 77

1000 1*
L ~ 7500

500 <-

0....1

4.2 4.1 4.0 3.9 3.8 3.7 3.6 3.5 3.4 3.3 3.2 3.1

 

 

 

Figure 22. lH NMR of H/D exchange at the methyl (3.2 ppm) and methylene position

(4.1 ppm) on 0.25 M betaine hydrochloride (D20, 100 °C, 1000 psi H2, 1 g dry 5% Ru/C)

To try and push the H/D exchange reaction with betaine hydrochloride, an additional
experiment was conducted at 130 °C and ran for 24 hours. During the course of the
reaction, degradation of the betaine hydrochloride occurred yet still no observable H/D

exchange.

In the experiment with 0.25 M betaine at 130 °C with 1000 psi H2 and l g dry 5% Ru/C
in D20 exchange of hydrogen with deuterium is occurring at the methylene position
albeit relatively slow. Kinetic rate data were determined and using the kinetic model

described in Chapter 3, the fit of the methylene data can be observed in Figure 23. The

123

H20 and HOD experimental (symbols) and fitted (line) data refers to the percent H in
water and is on the right y-axis (secondary y-axis). The experimental (symbol) and fitted
(line) concentration of CH2 (referred to as AH2), CHD (referred to as AHD), and CD2

(referred to as AD2) are on the left y-axis (primary y-axis).

0‘30 ,-W____-W..WW_-_ . -...W --- .._.. WW. -. . .. -,- W. -...-.
--AH2

W--- _- .W, ..__ _. .. . ..., 18

AHD — A02

 

 

 

 

 

I AH2_exp A AHD_exp e ADZ_exp
—— H20 and HOD 0 H20 and HOD_exp
0.25 ‘77 ~--#77 7 --# #7 ##7## # #7- ## ~- #7 #77 7 77 15
.. 0.20 # 12
i
v I
8 E
=3 0.15 77 9 o
g s
. m
U
c
8
0.10 i’ - W -. 6
0.05 77 i 3
0.00 1 T r r r T r 0
0 so 100 150 200 250 300 350 400

Time (min)

Figure 23. Fit of the kinetic rate data for 0.25 M betaine (D20, 130 °C, 1000 psi H2, 1 g

dry 5% Ru/C)

The rate constants for H/D exchange on the methylene carbon are: k. A = 4.1x10'5, k-l A =
0, k2A = 1.6x10'2, k-2A = 2.2x10'l L/mol/min; and the sum of the squares error (SSE) =
1.9x10'3 . At these low conversions, the back reaction rate constants are not likely to have
much significance. A more detailed discussion considering the significance of the back
reaction rate constants (km, k2A, k.1c, k-2c, and k.3c) will be discussed in 4.7 (as well as

H/D exchange of betaine ((CH3)3N+CH2COO')).

124

4.4 Effect of H3PO§

4.4.1 Hydrogenation Studies

In the case of the amino acids, a slight excess amount of H3PO4 was required to protonate
the carboxylic acid. Glycine and the N-methylated glycines are zwitterionic in nature and
thus the carboxylic acid functionality is dissociated unless the additional acid is added.
The H3PO4 is only required for the hydrogenation of the amino acids, but was added to
all hydrogenation experiments in order to be consistent. To determine the effect of the

H3PO4 on substrates that were not zwitterionic in nature, glycolic acid was studied.

A pair of experiments with 0.25 M glycolic acid were conducted at 130 °C with 1000 psi

H2 in both the absence and presence of H3PO4 (0.3 M).

 

100% , ' “‘1
+ Conversuon

+ Yield
[Conversion (with 0.3M H3PO4)
.31 ie'dM-QQQMEE’QQ _

90% ~—

 

80°/o -

70% -____-____ ___ g —

60% *

 

50% ~

40% -~

Conversion or Yield

30% ~

20% ~-

 

10% --

 

0%

 

0 50 100 150 200 250 300 350 400
Tlmo (min)

Figure 24. Hydrogenation of 0.25 M glycolic acid in the absence and. presence of 0.3 M

H3PO4 (H20, 130 °C, 1000 psi H2, 1 g dry 5% Ru/C)

125

The addition of H3PO4 did not significantly affect the hydrogenation of glycolic acid to
ethylene glycol. The conversion and yield for the pair of experiments were essentially

the same (Figure 24).

A pair of experiments with 0.25 M glycine were conducted at 130 °C with 1000 psi H2 in
both the absence (Figure 25) and presence (Figure 26) of H3PO4. In the presence of
H3PO4, a conversion of 83% and a yield of 71% ethanolamine were obtained after four
hours of reaction time. However, in the absence of H3PO4 much lower conversion and
yield values, both only 15% were obtained (Figure 27), presumably due to the amino acid

being in the zwitterionic form a majority of the time.

  

 

 

 

 

 

Glycine 4" 240 min
180 min
120 min
Ethanolamine 60 m!“
30 mm
l / 0 min
i
\2.‘ "‘N I
-. .- . \ , , V. . -. _.
4, . .. 7 ._ . ... ..\.‘.. . J" ... 2;? . .A J: , .. . .. .1. {‘~.1»;~-~-- ;'*;
”_...—.--_._..,\g p , -. WI ,,,,,,, 4 \\;- —~ — a _ _ _
2 3 4 5 6 minutes

Figure 25. Hydrogenation of 0.25 M glycine (no H3PO4, H20, 130 °C, 1000 psi H2, 1 g

dry 5% Ru/C)

The H3PO4 is thought to protonate the carboxylic acid, allowing for the hydrogenation of
glycine to form ethanolamine. In order to further determine whether or not the acid is

competing with glycine for surface sites, H/D exchange studies can be done to see if the

126

presence of the additional acid slows or inhibits the H/D exchange. This is assuming that
the hydrogenation and H/D exchange reactions are occurring on the same sites. The
phosphoric acid is not necessary for H/D exchange and proved to be inhibitory (to be

discussed next), but due to the need for the amino group to be free and electron donating.

Glycine

Ethanolamine

240 min
180 min
120 min
60 min
30 min
0 min

 

2 3 4 5 minutes

Figure 26. Hydrogenation of 0.25 M glycine (0.3 M H3PO4, H20, 130 °C, 1000 psi H2, 1

g dry 5% Ru/C)

In the experiment with H3PO4 present, glycine is essentially fully converted after three
hours of reaction (as opposed to almost no reaction in the case of no H3PO4 added). The
retention times in the HPLC analyses of glycine and ethanolamine are slightly shifted due
to the H3PO4. It is assumed that the H3PO4 is adsorbing on the same sites as the glycine.
So, competitive adsorption of the H3PO4 and glycine should be considered. Peereboom
et al. did find that compounds that are more hydrophilic tend to stay more in the bulk

solution and are less competitive for the catalyst due to the hydrogen bonding ability with

127

the solvent water.59 Competitive adsorption studies should be compared with the
knowledge that H3PO4 was not significantly adsorbed on the Ru/C (Figure 17) and did
not inhibit the hydrogenation of glycolic acid (Figure 24) to further determine whether

the adsorption of H3PO4 on the catalyst should be included in the kinetic model.

1000/O -..._ _ _.- - . .. ..-..-__._.. ........-,_-.-_ __.--.-._ .. _ "._.-M..- _. . __ _...- .. ._.-._ _.- _... ..-_... , _ ._..- ...- . .- - .
+ Conversion (with 0.3M H3PO4)

90% w +Yield (with 0.3M H3PO4)
+ Conversion /.
+Yield

70% . —

 

 

80% ~

 

60% ~

50% 4

40% «~

Conversion or Yleld

30% ‘-

20% r

10% «

 

 

 

0%

 

0 30 60 120 180 240
Time (mln)

Figure 27. Hydrogenation of 0.25 M glycine in the absence and presence of 0.3 M

H3PO4 (H20, 130 °C, 1000 psi H2, 1 g dry 5% Ru/C)

Another approach to evaluating the potential effect of H3PO4 on the hydrogenation
reaction would be to perform an experiment with equimolar amounts of glycine, lactic
acid, and a slight excess of H3PO4. There were several competition experiments
performed by members in our research group.49’61’63 Results indicated that with
equimolar amounts of alanine and lactic acid, the hydrogenation of alanine occurs but
lactic acid is initially unreactive. One explanation for this may be that alanine binds

stronger than lactic acid on the surface or has a greater affinity for the surface. But, a

128

second explanation may be that in the experiment with lactic acid and alanine, the lactic
acid may serve to protonate the zwitterionic alanine resulting in deprotonation of the
lactic acid thereby preventing the reduction. In the suggested experiment above, the
H3PO4 would serve to protonate the alanine and allow a more direct comparison of

alanine and lactic acid as well as evaluate the effect of H3PO4.

4.4.2 H/D Exchange Studies

Several experiments were conducted in the presence of H3PO4 to try and determine if the
additional acid caused an effect on the H/D exchange reaction. One goal was to try and
establish a link between the hydrogenation and H/D exchange studies. A pair of
experiments with glycine and sarcosine were run at 50 °C to try and slow or eliminate the
competing hydrogenation reaction and isolate the effect of H3PO4 on H/D exchange.
Glycine (0.25 M) with 0.3 M H3PO4 at 50 °C with 1000 psi D2 in D20 was allowed to
react for six hours. Results indicated a negligible amount of H/D exchange. Figure 28
shows the 1H NMR at time zero and also at six hours. The yield of CHD and
subsequently CD2 was zero. We can compare these results to the identical experiment in

which no H3PO4 was added (Figure 29).

129

 

1250 -fi-_- - ~ 1000

+Time: 0 Hours -—I—— Time: 6 Hours
1000 .2- , ,. 7 ,, t. 7‘ ~ 750
750 «~ ~ 500
~ 250

500 ~~

250

 

 

 

0 .. -250
3.75 3.70 3.65 3.60 3.55 3.50

 

Figure 28. 1H NMR of H/D exchange at the methylene position of 0.25 M glycine (0.3 M

H3PO4; D20; 50 °C; 1000 psi D2, 1 g dry 5% Ru/C)

4000 «~—-——-———~~ - e Wm ~ ~————— 3500

 

+Time: OHours ---l-Time: 6Hours

3500 ~~ ~ 3000

3000 ~ -~ 2500
2500 ~— ._ 2000
2000 ~ ,- 1500
1500 - ~1000
1000 _ 500

500

 

 

 

o - - -500
3.38 3.36 3.34 3.32 3.3 3.23 3.26 3.24

Figure 29. 1H NMR of H/D exchange at the methylene position of 0.25 M glycine (D20;

50 °C; 1000 psi D2, 1 g dry 5% Ru/C)

130

The concentration data for the experiment with H3PO4 showed essentially no H/D
exchange at the methylene carbon on glycine (Figure 30). The concentration of CD2 is
not measured, but is the difference between the summed amount of AH2 and AHD at time
zero minus the summed amount of CH2 and AHD at time t. Due, to some of the substrate
being adsorbed on the surface ("lost") it may cause the calculated CD2 to be slightly
higher. However, in this experiment the amount "lost" was equal to less than 3% (it is
reflected as the CD2). The experimental percent H in water and the fitted line are also

presented in Figure 30.

 

 

 

 

 

I AH2_exp A AHD_exp O A02_exp
—H20 and HOD 0 H20 and HOD_exp
i I I I
0.25 — 77 7 7 77 7 —- I ———-7-- I . l - -- «7 2.5
g 0.20 — l» 2.0
o
E
‘2' 2
a . - _ o
g 0.15 15 §
= I;
§
8 0.10 - ° ° r A * 1-0
’ o
0.05 -— _- 7 77 7 7 ~- 7 77777 + 0.5
‘ A A A 0
000 e e . é . e T c c . A 00
0 50 100 150 200 250 300 350 400

Tlmo (mln)

Figure 30. Concentration data for 0.25 M glycine (0.3 M H3PO4, D20, 50 °C, 1000 psi

D2, 1 g dry 5% Ru/C)

The catalytic H/D exchange of the amino acids is promoted by having an electron-rich C-
H bond available for insertion by the catalyst. When the amino group is positively

charged (as in the case with H3PO4 added), the OH insertion by the catalyst is not as

131

likely to occur. This is also compounded by the fact that the experiment was run at such
a low temperature (50 °C). A higher temperature reaction with sarcosine will be
discussed later. The reason betaine (with a quartenary nitrogen) undergoes H/D
exchange is because it is occurring non-catalytically through deprotonation of the or-

carbon in solution.

The same experiment but with 0.25 M sarcosine was performed (0.3 M H3PO4, D20, 50
°C, 1000 psi D2). Results again, as in the case of glycine, indicated a negligible amount
of H/D exchange with H3PO4 added. Figure 31 shows the 1H NMR at time zero and also
at six hours. The yield of CHD and subsequently of CD; was zero. There was a small
amount of conversion of the CH2 peak area (2.4%), but is presumably due to degradation
of the substrate over six hours and experimental error. When calculating the amount
"lost" from the difference between the time zero and at six hours for the methyl

concentration, the percent lost was zero.

We can compare this reaction to the experiment (same conditions) in which no H3PO4
was added (Figure 32). The concentration data for the experiment with H3PO4 showed
essentially no H/D exchange at the methylene carbon on sarcosine (Figure 33). The

experimental percent H in water and the fitted line are also presented in Figure 33.

132

1000 -r__-_~__-_.- _ -...--.-..- -_ ... _..-. .---... _ .. -_.-. . ..-. - -....-.. -..... - .. ... ..--.--- .........-----..... .. _- ...- - _....-..._---_-. _ ._ -.. . _- _ _ . ._.-.k ... .... - .-.- . .. _ 900
+ Time: 0 Hours 7'7 Time: 6 Hours

77—7- 77 7747 800

 

900 77-

800~ --

700 '1

 

600
500 i 7 -
400 .

300 ~

 

 

200

 

 

100

 

3.76 3.74 3.72 3.70 3.68 3.66 3.64 3.62 3.60 3.58

Figure 31. 1H NMR of H/D exchange at the methylene position of 0.25 M sarcosine (0.3

M H3PO4, D20, 50 °C, 1000 pSl D2, 1 g dry 5% RU/C)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2200 1800
"' _ +Time: OHours 47 Time: 6Hours
2000 H - 7+ - 7 1600
1800 J T" 77 777 1400
1600 17' i — 77777 —1200
1400 -- I * 7—7777 "r 1000
[3.45 3.43 3.41 3.39 I 0
1200 -_—_.-._.-.--__ .__— _ — _ 0 ._ _ _ -_____ __ 800
0 '
1000 777—77—— - 777——— - x 7777 7 777777 - 7 600
0
0
800 4k 7—7- -~ —--~---— -7 ”i' -i 7777 7 Wfi 2. 400
0
600 - 200
400 0
200 ‘ ~ -200
400

 

o -
3.52 3.50 3.48 3.46 3.44 3.42 3.40 3.38 3.36 3.34

Figure 32. lH NMR of H/D exchange at the methylene position of 0.25 M sarcosine

(D20, 50 °C, 1000 psi D2, 1 g dry 5% Ru/C)

133

It is important to note that the amount "lost" after six hours with respect to time zero is
small. This was confirmed in all experiments in which the desired reaction was not
occurring. The catalyst and carbon adsorption studies indicated that up to 5-7% can be
"lost" once the substrate is exposed to the catalyst. But, presumably this occurs
immediately (at time zero) and does not increase with time. For the kinetic studies as
well as conversion, yield, and selectivity the time zero concentration was used. It was
assumed that the initial amount of substrate adsorbed on the carbon support does not

eventually react but is degraded.

As discussed above, the ability for the amino acids to undergo H/D exchange is affected
by protonation of the amino group. Thus, it would be of value to do a detailed pH study
to measure the extent of protonation with increasing concentrations of H3PO4. Slight
changes in pH could drastically change the H/D exchange reactivity. In addition, it could
be suggested to explore the effect of added base on the H/D exchange reaction to evaluate

whether it would increase reactivity.

We can also look at the rate of exchange on the methyl carbon (Figure 34). Again, there

is essentially no I-I/D exchange occurring at the methyl position at these reaction

conditions.

134

 

 

 

 

 

 

I AH2_exp A AHD_exp 9 AD2_exp
--H20 and H00 0 H20 and HOD_exp
0.25 I7 .7 7. 7-7 77.7 7- -7 7777- 777 7 I 7 I 7 7 2.5
t
g 0.20 7 7 77 7 7 7 7 77 77 2.0
E
Ev J:
E
30157 .- 77771—7 77.7 77 77- — - +150
g d
5
o 0.10 7 77 77 7 7 77 7 77 7 +1.0
0.05 77 — — 77 7 7 7 —7 7 77 77 0.5
000 {H r . r g I 3 T _e r z r z 00
0 50 100 150 200 250 300 350 400
Time (min)

Figure 33. Concentration data for 0.25 M sarcosine (0.3 M H3PO4, D20, 50 °C, 1000 psi

D2, 1 g dry 5% Ru/C)

1 600

 

+Time: 0 Hours 7&— Time: 6 Hours

1400 7777—7 777-

1200 777 -7 7-

 

"rioe '\ x

1000 777-77-7777 - 77

 

 

800 77

 

 

600 7

400 7777--

 

200

 

o .-m_ ..
2.68 2.64 2.60

 

2.52

- 1400

7 1200

1000

r 800

600

7 7 400

77 200

 

77 7 -200
2.48

Figure 34. 1H NMR of H/D exchange at the methyl position of 0.25 M sarcosine (0.3 M

H3PO4, D20, 50 °C, 1000 psi D2, 1 g dry 5% Ru/C)

135

Sarcosine is more reactive than glycine in the H/D exchange reactions, so a higher
temperature, 70 °C, was also evaluated. It was desired to have some H/D exchange
occurring so that the rate constants could be compared (absence and presence of H3PO4).
In addition, higher temperatures promote dissociation and allow for more of the higher

energy form (non-zwitterionic) to be present.

Figure 35 shows the lH NMR for the methylene carbon at time zero and after six hours
for 0.25 M sarcosine with 0.3 M H3PO4 at 70 °C with 1000 psi D2 in D20. A small
amount of H/D exchange is occurring. H/D exchange is also observed at the methyl
position. Figure 36 shows the 1H NMR for the methyl carbon of sarcosine at time zero

and after six hours.

The kinetics were determined using both kinetic models and the reaction data fit is shown
below in Figure 37 and 38 (methylene and methyl). Again, by definition the rate constant
klA should be twice that of km. The model defines the rate constants on a per hydrogen
basis and in the case of converting CH2 to CHD there are two available hydrogens for
exchange whereas in the conversion of CHD to CD2 there is only one. The same theory
is applied to the methyls in that the rate of km should be 1.33 times that of kzc and 1.67
times that of k3c. This assumes there are no isotope effects (primary or secondary). This
assumption as well as a detailed discussion on the rate constants will be included in

section 4.7.

136

 

1800 7
+Time: OHours I Time: 6 Hours

1600777 777 7 77 7 77 7

 

1400 77

1200 77 777 77 777

 

 

 

10007777 77 7

 

1
800 77 777 77 77 7 777 -7 ...IL.‘_ ._A

 

600 +- 7 7

"T 1600
7 71400

7 ~ 1200

7 ~ 7-71000

77 — 800

7 600

7 400

 

 

400 — ,

200

 

 

0

3.68 3.60

3.66 3.64 3.62

Figure 35. 1H NMR of H/D exchange at the methylene position of 0.25 M sarcosine (0.3

M H3PO4, D20, 70 °C, 1000 psi D2, 1 g dry 5% Ru/C)

 

 

3500 - —-—-7~
+Time: 0 Hours 7I-Time: 6 Hours

 

300077- 7 7—--——7-- 7 7-7 7 77

 

2500-7 7 7

 

 

3.58

2000 7 77 - 7 7 - 77

1500 77 7 77 7

1000777 7777 77 7-7...-- --___H

500

 

2.52

o . .... mun-._...." . .
2.60 2.58 2.56 2.54

Figure 36. 1H NMR of H/D exchange at the methyl position of 0.25 M sarcosine (0.3 M

H3PO4, D20, 70 °C, 1000 psi D2, 1 g dry 5% Ru/C)

137

2.50

777 200

 

-200
3.56

P--. 7.777 3000

7 7 2500

—- —- 72000

 

77 1500

7 7* 1000

77777 - 7500

 

7 -500
2.48

 

—— AH2 —— AHD — AD2
I AH2_exp A AHD_exp o ADZ_exp
—H20 and HOD 0 H20 and HOD_exp
0.25 7. -—~~— — 7 - -— 77' --—-—— - v —* 7 —- - -— 2.0

9

N

o
[
i
b

 

 

 

 

3 , , .
3
E
v :r:
S ‘a
3015 7 77 7 —- f \r *1-2 8
ir~7_\ .
é 3
c
8 0.10 — —- — - - /1:’“/_,3_..77—1 .. 08
M” ‘
0,05 7 77 7/ 7 7 7 7 7 7. 7777 7 7 7 777 0,4
“/A/
M/
0.00 v I w ’ v r v . r 0.0
0 50 100 150 200 250 300 350 400

Time (min)

Figure 37. Fit of the kinetic rate data for 0.25 M sarcosine (0.3 M H3PO4, D20, 70 °C,

1000 psi D2, 1 g dry 5% Ru/C)

 

 

  

 

 

 

 

 

0'30 .. _-....- -a- --_--.....-.-.--.__.._-...........-_ 3.0
~-—-- CH3 W CHZD -— CHDZ m- CD3
I CH3_exp A CHZD_exp O CHDZ_exp O CD3_exp
—H20 and H00 0 H20 and HOD_exp
0.25 <~———i- -- ~ _ > - , «- 2.5
g 0.20 77 2.0
8.
V I
C o-I
0 c
E 015 * ~— 1 5 g
‘E E
8
8
o 0.107 77 7-- 7 — 7—7— 7 77 77 771.0
0.05 . + 0.5
4
0.00 7 . 0.0
0 50 100 150 200 250 300 350 400

Tlmo (min)

Figure 38. Fit of the kinetic rate data for 0.25 M sarcosine (0.3 M H3PO4, D20, 70 °C,

1000 psi D2, 1 g dry 5% Ru/C)

138

The rates for H/D exchange on the methylene carbon are: k1 A = 0.21x10‘4, k-1 A =
9.3x104, km = 0.18x10‘4, and k-ZA = 1.9x10'3 L/mol/min. The reaction rates for H/D
exchange on the methyl carbon are klc = 0.18x10'4, k-1c = 1.1x10'3, kzc = 0.30x10'4, k-zc
= 1.2x10'2 k3c = 0.31x10'4, and k-3c = 0 L/mol/min. We can compare the rate constant
data for the experiment at 70 °C with 1000 psi D2 in D20 in both the absence and

presence of 0.3 M H3PO4 (Figure 39).

Again at these low conversions, the back reaction rate constants do not have much
significance. In addition, these experiments were conducted with 1000 psi D2 and D20,
so the available H in the water (which is in equilibrium with the surface) is only ~2%.
Using a purely statistical approach and assuming the reaction has reached equilibrium the
percent of CH2 would be (0.02 x 0.02) 0.04%, CHD would be (0.98 x 0.02 + 0.02 x 0.98)
3.92%, and CD2 would be (0.98 x 0.98) 96.04%. The large amount of deuterium

available promotes the forward reactions to occur.

The rate constant results show that the rate of H/D exchange is significantly suppressed in
the presence of H3PO4 (Figure 39) Again, this is presumably due to the amino group
being protonated and not in the free form. The ratios for the forward rate constants for
both the methylene and methyl carbon are not equal to the expected ratios (as discussed
previously), indicating that isotope effects may be playing a role in the H/D exchange
reaction. Kinetic model B (chapter 3) incorporates the possibility of primary and

secondary isotope effects and will be explored in section 4.7.

139

 

2.5E-04 7 ..._- ”MN“- “a-..” -_W,

 

 

Iwithout 0.3 M H3PO4
lwith 0.3MH3PO4

2.05704 7 7 77 ---. __ _ ._ A i
’5‘
E
3
5 1.55704 «77 i ,
.3.
5
a
5 1.05-04 7
0
3
I
n:

5.05705 77

o.oe+oo 7

 

 

   

k3_C

Figure 39. Rate constants for 0.25 M sarcosine in the absence and presence of 0.3 M

H3PO4 (D20, 70 °C, 1000 psi D2, 1 g dry 5% Ru/C)

An experiment was conducted with 0.25 M betaine ((CH3)3N+CH2COO') and 0.3 M
H3PO4 at 100 °C with 1000 psi H2 in D20. Results indicated that even with the H3PO4
present, H/D exchange is still occurring at the methylene carbon. The 1H NMR of the
methylene position on betaine is compared at time zero and at six hours of reaction time
(Figure 40). The peak shape of the CHD (triplet) from the 1H NMR for betaine was not
as resolved as compared to most of the other experiments. In Figure 42 the inset shows
the CHD peak after six hours. At six hours of reaction, 100% of the CH2 had been

converted.

140

 

 

 

300 7777777 7 . 250

—+7 Time: 6 Hours

250 1 ~ 200

 

 

 

200 3 - 150

150 1 ~ 100
100 ~ ~ 50

50‘

 

 

0 = 7 :‘r . j -50

4.04 4.00 3.96 3.92 3.88 3.84 3.80

Figure 40. 1H NMR of H/D exchange at the methylene position of 0.25 M betaine (0.3 M

H3PO4, D20, 100 °C, 1000 psi H2, 1 g dry 5% Ru/C)

1500 _... _.... _-....- _-... “7“.-. -_..-- .7. . . .. 7.. ----.-m. _-.--~---....__ -.- _--.---“ .-...___.._..._.._. .--....--_.___._ .__..__ l" 1250

1250 —~ ~ 1000

 

1000 ~ ~ 750

750 7 - L 500

500 1~ ~ 250

 

 

 

 

 

 

250 o
o . 77 77 77 7250
3.22 3.20 3.18 3.16 3.14 3.12 3.10 3.08

Figure 41. 'H NMR of H/D exchange at the methyl position of 0.25 M betaine (0.3 M

H3PO4, D20, 100 °C, 1000 psi H2, 1 g dry 5% Ru/C)

141

This is not surprising considering that betaine exhibits H/D exchange at the methylene
position even in the absence of catalyst. Thus, if the H3PO4 is competing for the catalyst
it still should not inhibit the H/D exchange in betaine, as the exchange does not involve
the catalyst. Consistent with this notion, H/D exchange was not observed at the N—methyl
position of betaine. The 1H NMR peak of the three methyls decreased in size over time

due to the conversion of betaine to choline, but no H/D exchange occurred (Figure 41).

In this experiment the presence of choline was also detected over time. From HPLC
analyses, a 38% yield of choline was obtained after 6 hours of reaction time. A
conversion of 100% of the methylene carbon was measured after 6 hours, due to both
hydrogenation and H/D exchange. Using the N-methyls as a reference for the amount of
hydrogenation that occurred, roughly 55% of the N—methyl 1H NMR peak was converted

to choline. The HPLC analysis matched this, reporting 49% conversion after 6 hours.

For this experiment, the percent H in water was much variable. So, the [HOD] fit
parameters were modified to better represent the actual data. A 4th order polynomial was
used to fit the aqueous hydrogen concentration. For all other experiments, equation 46
(Section 3.2.2) was used. The concentration data versus time and the fitted rate equations
are presented in Figure 42. Betaine is undergoing uncatalyzed H/D exchange that does

not appear to be inhibited by the H3PO4.

142

0.30

 

0.25 ~

~—-— AH2
I AH2_exp

_:'_H20 “(1-590

 

Concentration (mol/L)
.0
a:

.0
_L
o

0.05

 

0.00

77 AHD
A AHD_exp
__°_._t'20 amLHQD—eXP

 

 

 

 

 

Figure 42. Fit of the kinetic rate data for 0.25 M betaine (0.3 M H3PO4, D20, 100 °C,

1000 psi H2, 1 g dry 5% Ru/C)

The experiment with 0.25 M betaine in the absence of H3PO4 at 130 °C with 1000 psi H2

50 100

150 200 250

Time (mln)

and D20 will be discussed further in section 4.7.

4.5 Solvent Effects on the Hydrogenation of Organic Acids

A key aspect involved in understanding the mechanisms controlling the catalytic
hydrogenation of organic acids is the effect the solvent has on the hydrogenation. In
catalytic hydrogenation of acids, it has been determined that catalyst and hydrogen are
essential in order to convert the starting material. In addition it appears that the acids

need to be in their protonated form. It is less clear what role, if any, the solvent plays in

the overall catalytic hydrogenation scheme.

143

300

—36

~~ 30

7- 24

7713

7— 12

400

Percent H

There are several advantages of using aqueous-phase conversion routes including:

1) most bio-based feedstocks are produced and handled in water solutions and 2) feed
purity is less critical. However, the question of whether water is involved in the
hydrogenation mechanism has not been answered. It has also been suggested that a more
"organic" environment might favor conversion of the more polar amino and hydroxyl
acids to their less polar alcohol reduction products. To examine this issue, an experiment
was conducted in which no water was present. Glycolic acid (0.25 M) in dry
tetrahydrofuran (THF) at 130 °C with 1000 psi H2 revealed a suppressed hydrogenation
reaction rate. The yield of ethylene glycol formed after six hours was about 14%
compared with nearly 50% with water as the solvent. Though THF appeared to hinder
the reaction some ethylene glycol was formed, indicating that the reaction did still occur.
Additional experiments were conducted using ratios of 80/20 (v/v) water/THF and 20/ 80
(v/v) water/T HF . From this series, it is clear that as the amount of organic solvent is

increased the reaction is further suppressed (Figure 43).
To promote hydrogen solubility in water, high pressures are used (1000 psi H2). In the

case of THF as the solvent, the solubility of hydrogen was examined to determine if less

hydrogen was present in these reactions.

144

1000b 2.2.2-222 .-......__-... -..... -.. ”2-- . .. .... _-..“..-2... 2.... “-2... -... -_ _...- -..-2.-.“ . .. .. . -_......__ __ .. . .. _... . ,, ,_ , _ . . .. .. 2 ._ . . .. .. .,
+ Conversion of glycolic acid ~+» Yield of ethylene glycol

90% i

 

80%1----—-~77 777 7 7 7777 7- 7 77 7 77

094.777 777 7 7 7 7h 7 if a;

7
60%\—— 77 — 77777 7 77 --77 7
5

0% f: \777 -7- 7, fl "7’ 7 2-.-, ,,_ 7 7 fl
\\ \

\\
30% ‘ "** “W 7 7:1- -2 4f; ,__ _. J; -- '5'“"f“"""‘~\~-77777._:

20% 2. ,7 ,___ "7, , ”A7, - , . - W _ T. 7... 72,,

10%77 77 77 7 -- 7 7 77::

 

0%

 

0 10 20 30 40 50 60 70 80 90 100
“/o THF In Water

Figure 43. Hydrogenation of 0.25 M glycolic acid after 6 hours (THF/H2O, 130 °C, 1000

psi H2, 1 g dry 5% Ru/C)

In contemplating the role of the solvent, the various reactions in which it can participate
need to be considered. From previous work in our group it was known that catalytic H/D
exchange is possible between solvents and both the organic acids and the alcohol
products. To understand whether hydrogenation and H/D exchange are occurring at the
same sites on the surface THF was chosen as a solvent. An experiment with 0.25 M
THF at 130 °C with 1000 psi H2 in D20 and l g dry 5% Ru/C was run and results
indicated that THF readily undergoes H/D exchange at these conditions. The H/D
exchange was observed at both C1 (83% a-CH2 conversion) and C2 (42% B-CH2
conversion) carbons. The lH NMRs for the methylene sites were difficult to convolute
once the exchange started due to the overlap of the CH2 and CHD peaks. So, the total

amount of conversion was based on the disappearance of both CH2 and CHD. If H/D

145

exchange and hydrogenation were occurring on distinct sites then the rate of
hydrogenation should not be slowed as much in the presence of a solvent that
predominantly undergoes H/D exchange. As mentioned above, the rate was significantly
slowed as more THF was added, suggesting the sites may not be distinct and that there is

competition at the surface.

Due to the a C-H activation present with the THF solvent, a solvent was chosen that was
expected to be unreactive to H/D exchange or hydrogenation, namely tert-butanol. The
results from 0.25 M glycolic acid hydrogenation in tert-butanol suggested that a
competitive reaction is occurring involving the solvent on the catalyst surface, thereby
possibly tying up the available sites for acid hydrogenation or perhaps simply excluding
the water needed by the reaction. tert-Butanol in the presence of acid (0.25 M glycolic
acid) at 130 °C with 1000 psi H2 may also undergo selective dehydration to
isobutylene.“65 lsobutene could undergo hydrogenation on the surface of the catalyst.
No significant yield of ethylene glycol was detected. H/D exchange seemed unlikely in
this context, but an experiment with neat tert-butanol and D2 and D20 was run to evaluate
the amount of deuterium incorporation at the methyl groups of tert-butanol. After four
hours, little H/D exchange was observed using lH NMR, eventhough there was some loss

(~25%) of the CH3 as assessed by the 1H NMR (discussed further in section 4.7.2).

From previous research in our group and the work presented in this thesis, we know that
alcohols can undergo C-H activation. But, we also know that, in general, they have less

of an affinity for the Ru/C catalyst as compared to the organic acid substrates. So,

146

ethanol and ethylene glycol were both explored as solvents for the hydrogenation reaction
of glycolic acid. Hydrogenation experiments with glycolic acid (0.25 M) in ethanol were
run in both the absence and presence of H3PO4. The presence of H3PO4 promoted
esterification (9%) at t=0 min, defined as the point where the reaction temperature
reaches 130 °C and the H2 pressure (1000 psi) is introduced. Figure 44 shows the HPLC
chromatogram (mV versus time (min)) for glycolic acid in ethanol (with 0.3 M H3PO4)
for the pre-reaction solution, time zero, and at 6 hours. In the absence of acid, the ester is
not present until after 5 hours of reaction (Figure 45). However, there is some conversion
of glycolic acid that is occurring over time. This may be due to the glycolic acid
adsorbing into the catalyst. In either case, nearly the same amount of product, 8% of
ethylene glycol, was formed. The conversion and yield over time is presented in Figure
46. This was an interesting result in that it also begs the question of whether the ester

formed is able to undergo hydrogenation itself, albeit at a slower rate.

147

ii
\ Glycolic

\ acid

. , Ethyl ti 8 2“

I I. l colate Ethylene t' rs.
I, g y glycol Pre—reactlon

 

 

 

 

Figure 44. Hydrogenation of 0.25 M glycolic acid (0.3 M H3PO4, CH3CH20H, 130 °C,

1000 psi H2, 1 g dry 5% Ru/C)

 

 

 

 

 

 

Glycolic
acid
t= 6 hrs
t= 0 hrs
Ethyl Ethylene Pre-reaction
glycolate glycol
ii \i\ ._-_ -7 ----- , 777.--} . /‘\ _
-,__- i \\ * ' r ~-:--~~~~---
_...—._.J/ \777777—
5/ ¥n V

 

 

 

 

 

 

 

 

Figure 45. Hydrogenation of 0.25 M glycolic acid (CH3CH20H, 130 °C, 1000 psi H2, 1 g

dry 5% Ru/C)

The literature suggests that esters undergo hydrogenation more rapidly than the

corresponding acids with certain catalysts (Raney Ni, Copper/chromium oxide, rhenium

148

black). However, in the case of 100% ethanol there is little desired product formation.
This result is not unexpected; early studies by Adkins (Raney Ni and Copper/chromium
oxide) showed that esters hydrogenate poorly in the presence of free acids. Thus, the
presence of both the organic acid and ester leads to competitive adsorption resulting in
less desired product formation. If the ester formed does not undergo hydrogenation, less
hydrogenation product is expected in the presence of acid due to more ester formed and
depletion of free acid substrate available. However, these results indicate the same
amount of ethylene glycol yield which suggests that some amount ester hydrogenation
may be occurring (albeit small). In addition, in the presence of ethanol (and H3PO4)

esterification is occurring faster than the hydrogenation of glycolic acid.

 

 

 

 

 

 

 

 

 

100% ~7
+ Conversion of Glycolic Acid (with 0.3M H3PO4)
90% _ I ~ Yield of Ethylene Glycol (with 0.3M H3PO4)
r-A— Conversion of Glycolic Acid
80% +__ + Yield of Ethylene Glycol
70% _.7 ___, ".7 —' - ‘ .... ' ' 'u7i734" " -:::’ 477:;_:_4 *7’ ""
M,,7~k——————O“*"’ 4*
60% .. 7- . _./.V:__ . _ -7 777-7 .. . ,. 777-777
500A, _,_,_ _/__ _ ._ __ . _ __ ___..__._ _ _..-...7_ _ .. __.__ . . _ . . .. . ___ __.___
400/0 "‘ / ““ _ ,7 ’ ‘ 77,77, ’ " " ‘ r 77__..‘ "— ‘‘‘‘
30% .. 7 7 77 "—7777 7A — 7 7
20% WP — ~~7 — 77 7" 77 7 - 77 7 7 77
10% 7 _,...7A 77- 7 7
0 50 100 150 200 250 300 350 400
Time (min)

Figure 46. Hydrogenation of 0.25 M glycolic acid in the absence and presence of 0.3 M

H3PO4 (CH3CH20H, 130 °C, 1000 psi H2, 1 g dry 5% Ru/C)

149

An experiment with 0.25 M ethyl lactate in water was exposed to the hydrogenation
reaction conditions (130 °C; 1000 psi H2). The ethyl lactate quickly hydrolyzed to form
lactic acid. The lactic acid peak is overlapped by the ethyl lactate peak (Figure 47). The
product propylene glycol was not observed until lactic acid was formed. However, the

yield of propylene glycol at 6 hours was comparable to the yield using lactic acid as the

substrate, nearly 72%.

   

.1": 1’2-
i l Propylene
lllglycol
l
1 “‘1 l.
l i l Ethanol
Lactic Et h l i l. in
. '7’ l "~ /
ac1d 1 ityt {It i, \\ /\~.
r. acae 1' ix
r/ x I; I ‘1 \\ a“.
‘-- [1" .. “M w ‘___' 7...-'7'7“77:7r714.17-2317: 1;": /7' 1‘7" m;.T.i.;;'.;;_;7‘,:;;.\;;.;._;7;i _-///\\ l¥uvi7lJiub t = 6 hrs

"' .( t: 1 hrs

77 Pre-reaction

 

 

 

T?"

Figure 47. Hydrogenation of 0.25 M ethyl lactate (H20, 130 °C, 1000 psi H2, 1 g dry 5%

Ru/C)

This experiment suggests that the hydrolysis of ethyl lactate is faster than its

hydrogenation, and that lactic acid hydrogenation is occurring with little hindrance from

the ester or the product alcohol.

In the case of glycolic acid in ethanol, the hydrogenation is suppressed due to the solvent

(solvent/catalyst interactions) and it appears that hydrogenation of ethyl glycolate is slow.

150

The reduced amount of hydrogenation observed with the ester be due to the ester
functionality or may be more of a result of the overwhelming presence of the alcohol as
the solvent. In the case of ethyl lactate in water, ethanol is formed as the ethyl lactate is
hydrogenated. However, the amount of alcohol present is small compared with the
solvent water and thus in this case the alcohol is not inhibiting the hydrogenation

reaction.

An experiment that would help verify whether ester hydrogenation is possible is ethyl
lactate in ethanol. However, as mentioned earlier, the solvent ethanol significantly
suppresses the hydrogenation reaction and would presumably suppress it for an ester in
the same way as it does for an acid. Our group has shown, using isotopic labeling
experiments, that both glycerol and sorbitol exhibit rapid deuterium incorporation at C1
without reaction, suggesting reversible dehydrogenation on the metal surface as the first
step in hydrogenolysis. This previous work sets a precedent for ethanol to undergo the
same mechanism. Indeed, early qualitative studies by Kovacs et al. found deuterium
incorporation at the 1-position of 1-propanol and 1,3-propanediol sites are very similar to
the 1-position of ethanol.53 The dehydrogenation/hydrogenation of ethanol is a possible

reaction competing with the glycolic acid hydrogenation for catalyst sites.

Additional experiments using 0.25 M glycolic acid at 130 °C with 1000 psi H2 in 80/20
(v/v) water/ethanol and 20/80 (v/v) water/ethanol mixtures were conducted and revealed
that the yield of desired glycol continued to decrease as the organic solvent proportion

was increased. However, net acid conversion was relatively the same (Figure 48) due to

151

increasing esterification and less reduction product formation as the solvent’s ethanol
fraction increased. The ethanol is known to undergo C-H activation and may be
undergoing H/D exchange on the catalyst surface sites thereby reducing the amount of

glycolic acid hydrogenation.

 

 

 

 

 

 

 

 

 

 

 

1000/0 “F” __-__....... ...-_...-..._..--- ._-- ...... . .. ..
7l— Conversion of glycolic acid 79— Yield of ethylene glycol
90% 777 7 - 7 7 77 7 7
800,6 ..7- . _77 - -7 7 .. 7777.. 77-7- 77 7 77 777 7.7 7
70% 77 7 - .77-- 7 7777
600/0 -\\ _ 7 7 7
\\
50% K \ 7 7 — — é
_ 7 I
40% ...17-77\'“——~“~77777:77 ._.-7-7-7’“
\
\ i
30% +7 777 77“ 7.7 7 7 777 7 7
20% ~ 7 77 - — — 7 ~ T-\~ "‘ “r 7- 7— 77 77 -—7 -—7 777:1
‘ r ‘7 7-
10% — 7-— - ——-7-—7--- —-—“—‘-=-77-. 4‘
0% 1 i
0 20 40 60 80 100

% Ethanol In Water

Figure 48. Hydrogenation of 0.25 M glycolic acid after 6 hours (CH3CH20H/H20, 130

°C, 1000 psi H2, 1 g dry 5% Ru/C)

Previous work in the J ackson/Miller group revealed that lactic acid had a higher surface
affinity than propylene glycol.66 These results were deduced from H/D exchange studies
in which the rate of exchange and chirality loss in propylene glycol decreased as lactic
acid was introduced to the system. Using this logic, ethylene glycol was chosen as a
solvent. If glycolic acid has a higher affinity for the catalyst than ethylene glycol then
some hydrogenation product might be expected. Studies of 0.25 M glycolic acid in

ethylene glycol (130 °C; 1000 psi H2) revealed that the dominant reaction occurring was

152

the esterification of the acid. The ester peak overlapped on the HPLC chromatogram
with the glycolic acid peak and it was not possible to quantify the conversion of glycolic
acid to the ester. Thus, the response factor was not determined for this suspected ester.
Qualitatively, the amount of glycolic acid converted (based on manual integrations) and
the amount of ester formed indicates that not much, if any hydrogenation product was

formed. However, a small amount of ethanol was formed.

In the presence of hydridic hydrogens on the Ru surface, the mechanism to form ethanol
as a side product from the degradation of ethylene glycol may involve a hydrogen from
the surface displacing a water molecule.67 To further examine the possible degradation of
ethylene glycol a control experiment in which ethylene glycol alone is subjected to the
reaction conditions should be run. However, we have seen ethanol formed after six hours
in experiments with glycolic acid that resulted in high conversion (62%) yielding

ethylene glycol (49%).

Up to this point, it seems apparent that the solvent may be occupying the surface and
undergoing either H/D exchange or hydrogenation/dehydrogenation thereby
competitively preventing the hydrogenation of the organic acid. If an organic acid is
chosen as the solvent, then one would hypothesize that the product alcohol from the
solvent would be formed. Glycolic acid has a much higher rate of reaction than acetic
acid, 49% yield compared with 5% yield after six hours (H20, 130 °C; 1000 psi H2, 1 g
dry 5% Ru/C). This may be partly due to the difference in adsorption affinity as well as

the activation due to the or-substituent.

153

From the present work, results with 0.25 M glycolic acid at 130 °C with 1000 psi H2 in
acetic acid as the solvent indicate that neither ethylene glycol nor ethanol is produced.
The acetic acid forms esters with the glycolic acid’s —OH side chain, and there is 0%
yield of either alcohol products. If the solvent, acetic acid, were saturating the surface
then some ethanol would be expected. However, the pKa of acetic acid is about 4.75 and
it may be that the acid is reacting with the surface and causing deactivation of the catalyst

due to the amount of acetic acid present in the system.

Though less reactive than lactic or glycolic acid, propanoic acid is more reactive and
slightly less acidic (pKa = 4.85) than acetic acid. After six hours of reaction, nearly 13%
of propanol is formed from 0.25 M propanoic acid (H20, 130 °C; 1000 psi H2, 1 g dry
5% Ru/C). 0.25 M glycolic acid at 130 °C with 1000 psi H2 in propanoic acid revealed
about 3% propanol and detectable ester. Ethylene glycol elutes at the same time as the
solvent propanoic acid and therefore could not be quantified, although 70% of the
glycolic acid was converted after six hours. Some of the conversion may have yielded
ethylene glycol and the ester of propanoic acid. Again, because there was hardly any
propanoic acid hydrogenation product formed it begs the question of what reaction is

taking place to prevent the acid hydrogenation.

Competitive adsorption is a key component in trying to understand these phenomena.

Due to the fairly unreactive nature of the unsubstituted organic acids, it may be that the

competitive reactions or esterification and self dimerization are enough to suppress the

154

hydrogenation of the propanoic acid. In order to further explore this, an experiment in
which there is an equimolar amount of propanoic acid and glycolic acid in water was
conducted. The hydrogenation of glycolic acid may be slightly suppressed if there is
competitive adsorption with an equimolar amount of propanoic acid, but production of
the alcohol products would be expected. In this case, the esterification and self
dimerization are not expected to be dominant. Competition experiments will be

discussed in section 4.8.

To summarize the solvent studies, it is evident that if the solvent is able to undergo
reactions on the catalyst surface, these reactions compete with the organic acid
hydrogenation. Solvent H/D exchange and hydrogenation/dehydrogenation, and catalyst
deactivation (whether temporary or permanent) can almost completely stop the organic
acid hydrogenation reaction. Thus, the water is playing an important role in the overall
catalytic hydrogenation of organic acids in that it is not competing with the reaction. The
water may be playing an additional role of assisting H2 cleavage into hydrogen atoms on
the surface. The potential role of water in the overall hydrogenation mechanism using
5% Ru/C will continue to be explored via detailed H/D exchange reactions and discussed
later. Again, there are several advantages of aqueous phase processing. The nonvolatile
organic acids can be converted without having to be esterified, as would be required for

vapor-phase conversion. And energy costs associated with vaporizing and condensing

process streams is reduced.

155

4.6 Organic Acid Hydrogenation Studies

4.6.1 Effect of the a-Substituent on the Hydrogenation Reaction

One of the key findings from the previous research done in the Jackson/Miller research
group was that simple carboxylic acids such as propanoic acid and acetic acid are
substantially more difficult to reduce than lactic acid and L-alanine. Thus, the presence
of the heteroatomic side chains modify reactivity. Specifically, protonated alanine and
lactic acid undergo aqueous-phase hydrogenation over a carbon-supported ruthenium
catalyst ca. 10 and 3-5 times as fast as their unsubstituted analogue, propanoic acid. The
present work explores the substituent effects on organic acids' reactivity and selectivity
toward hydrogenation by comparing reduction rates over a wide range of electron-

withdrawing, hydrogen-bonding, or sterically bulky vicinal substituents.

Several three—carbon and two-carbon organic acid substrates were chosen in order to
evaluate the extent to which the oc-substituent affects the hydrogenation rate. Substrates
included the following a-substituents: -H, -CH3, -OH, -N+H3, -OCH3, -N+H2CH3, -
N+H(CH3)2, -N+(CH3)3CI', -N+(CH3)3 .7 The experiments were conducted at 130 °C with
1000 psi H2 using 1 gram of pre—reduced 5 wt% dry Ru/C with 0.25 M substrate and 0.3
M H3PO4 in water. Amino acids represent an attractive group of compounds for liquid-
phase hydrogenation, both because of the diversity of substituents available and because
of the recognized importance of optically active amino alcohol products as intermediates

in pharmaceuticals.

156

Amino acids typically form dipolar ions in aqueous solution at their isoelectric point (pH
~5-7); the actual form is thus determined by the pH of the solution. The reason the
H3PO4 was required (as discussed previously) for the successful hydrogenation of the
amino acids was to protonate the zwitterionic carboxylic acid, thus enabling

hydrogenation. In order to be consistent, the H3PO4 was added to all experiments.

Three carbon organic acid substrates:

0 o
0 .
OH
propanoic acid isobutyric acid pivalic acid
0 0 0
OH OH HO OH
OH NH2 OH

lactic acid alanine glyceric acid

Two carbon organic acid substrates:

0 o o o
O\/“\ HO\/“\ HZN \i
iOH / OH OH OH

acetic acid methoxyacetic acid glycolic acid glycine

O ’ O I O
H \ +\/U\
/”\/lkOH /”\)]\OH /” O-

N-methylglycine N,N-dimethylglycine N,N,N-trimethylglycine (betaine)

157

Glycine and the N-methylated glycine series were protonated due to the presence of
H3PO4. If additional acid is not present, they are in zwitterionic form. In addition, the

hydrogenation of both betaine with and without the chloride counterion was compared.

A set of optimum reactor conditions was selected for this study based on previous work
done by our research group.25‘48'49‘50 The results are presented in the graphs (Figures 49
and 50) and table (Table 2) below. Table 2 presents the overall data after six hours of

reaction time and details the conversion based on the initial concentration; selectivity, in
units of mole product per mole reactant converted, and yield, in units of mole of product

per mole of reactant fed. Carbon balances were calculated to account for most products

(>90%).

Equation 62. Conversion of substrate in the hydrogenation reaction

 

_[ reactant moles consumed J
reactant moles fed

Equation 63. Yield of desired product in the hydrogenation reaction

 

_ [ product moles formed J
reactant moles fed

Equation 64. Selectivity of desired product in the hydrogenation reaction

 

_{ product moles formed ]
reactant moles consumed

158

 

 

 

 

100%
. Alanme
90% 7 7 - 7 777—7 777 7
o
80% 7 —- 7—7—7 77777 7-~—7- - ~77
’ Lactic Acid
70% . 7-7- 7 7 -- —7777 : Glyceric Acid
. I
A 60% 777—7- - - -7 7 7 -- 7- 7777—7377 -— o
O
a: 0
IE 50/0 4L*—— _____7 if 7 _. ”if " ' a i 7
>.
40% 7777 — --—-7—---7777—— —---————77— 77777 7 7 7 7 — 7 —
I
o
30% -7—— — 77777
O n u
20% 2 777—l -— 7777777 777 77 7 7 7777777 — - Isobutyrlc Acrd
o . .
‘ A Propanorc Acrd
o . . .
10% ,, ._. - 7 7 7 ‘6 . .777— § 87.77 Plvallc Acrd
A
0 x
4 9
00/0 ‘ 5 g T r I r r .
O 50 IOO ISO 200 250 300 350 400
Time (min)

Figure 49. Hydrogenation of C3 organic acids to alcohols (0.3 M H3PO4, H20, 130 °C,

1000 psi H2, 1 g dry 5% Ru/C)

The various a-substituents were chosen in order to evaluate the effect of hydrogen
bonding, electron withdrawing ability, and heteroatom coordination to the surface. In
acid, the OL-substituent if it is an amino group is presumably protonated. This allows for
two types of hydrogen bonding. The hydrogen can interact with the hydridic hydrogens
on the surface thereby increasing the reactivity by promoting interaction between the
substrate and the catalyst surface. The other type of hydrogen bonding is intramolecular
hydrogen bonding with the free electrons on the carbonyl oxygen, stabilizing the partial

positive charge on the carbonyl carbon. In the case of the unsubstituted alkanoic acids

159

(acetic, propanoic, isobutyric, and pivalic acid) and the trimethylated amino species
((CH3)3N+CH2COOH'C1’) the hydrogen bonding ability is eliminated and it was observed
that they are very unreactive. In addition a comparison between glycolic acid and
methoxyacetic acid indicates a reduced rate of hydrogenation in the case of the methoxy
(it-substituent. Thus, hydrogen bonding may be playing a role in accelerating the

hydrogenation reaction.

 

 

 

 

 

100%
90%7777— 7 77 -7
80% - 7 7 77 7 7777. —+—— — Glycine
700/0 _,-,,__._____v,m_ W, 77#, a , ,. fi,~m ,9 __d W
A 60% - 7— . 7 7 7-777777-7 777777-
e:
E 50% ‘* , ‘ * i _ A * i ‘ Glycolic Acid
.2 O
>-' o ' -
40% ‘ .____g_*___ __ _ o___ _‘ Methortyacetic Acrd
Sarcosme
o I
A
30% 7 -- -- 77 7777 - 7 77 - 7
O O l .
N,N-dmg
20% ———--— — “—6 7} — 7 7 7*
x .
° = A Propanoic Acrd
X . .
10% 7— ——— — 7—7— 77 -- X777 7 —— 7— A 77 7 —- Acetic Acrd
9 f . 8 0 Betaine (w/o Cl-)
0% i a I I a 1 I; 01 I? To Betaine (w Cl-)
0 50 100 150 250 300 350 400
Time (min)

Figure 50. Hydrogenation of mostly C2 organic acids to alcohols (0.3 M H3PO4, H20,

130 °C, 1000 psi H2, 1 g dry 5% Ru/C)

It was observed that (CH3)3N+CH2COOH°C1' was not very reactive (<2% yield of choline

chloride) and the extent to which the chloride counter-ion was playing a role was not

160

fully understood. So, betaine ((CH3)3N+CH2COO') was also run under the same
conditions (0.3 M H3PO4, H20, 130 °C, 1000 psi H2, 1 g dry 5% Ru/C). A conversion of
41% was obtained, but only 7% yield of the choline product of carboxylate reduction was
achieved. Thus, the presence of the chloride counter-ion appears to be inhibiting the

reaction. No other products were detected in the HPLC.

Table 2. Hydrogenation of C3 and C2 organic acids to alcohols after six hours (0.3 M

H3PO4, H20, 130 °C, 1000 psi H2, 1 g dry 5% Ru/C)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

(Substrate)
a-substituent Conversion (%) Yield (%) Selectivity to Alcohol
(c3) NH3 94 86 91
(c2) NH3 96 80 83
(C3) OH 93 79 85
(c3) 61,13 (0H)2 76 68 100
(c2) OH 56 47 82
*(c3) OCH3 48 47 98
(C2) OCH3 53 35 67
(c3) N+H2(CH3) 37 35 96
(c3) N+H(CH3)2 4o 20 51
(c3) (CH3)2 46 12 27
(c3) CH3 39 16 40
(c2) CH3 / (c3) H 33 13 39
(C3) N+(CH3)3 41 7 17
(C2) H 19 7 37
(c3) N+(CH3)3Cl' o 3 0

 

 

 

 

 

 

*Data from Norbert Varga’s thesis (Michigan State University)

Looking at the low selectivities to the desired alcohol for the unsubstituted alkanoic
acids, it is assumed that some of the by-products that are formed include methane, ethane,

and propane.”49 Zhang et al. found that at temperatures below 170 °C these were the

161

typical by-products and at temperatures above 170 °C, degradation to l-propanol,

ethanol, and 2-propanol were observed.

4.6.2 Amino Acid Substrates

Additional experiments at temperatures of 100 °C and 150 °C were conducted with
glycine and the N—methylated glycine series. The results for conversion, yield and
selectivity are presented in Table 3. In general, the selectivity decreased at higher

temperatures indicating increased degradation.

Table 3. Hydrogenation of glycine, sarcosine, N,N-dmg, and betaine at 100, 130, and

150 °C after six hours (0.3 M H3PO4, H20, X °C, 1000 psi H2, 1g dry 5% Ru/C)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

100 °C Conversion (%) Yield (%) Selectivity
Glycine 31.4 28.2 89.8
Sarcosine 10.2 11.7 100
N,N-dimethylglycine 13.8 11.2 81.1
Betaine 4.1 0 0

130 °C Conversion (%) Yield (%) Selectivity
Glycine 89.8 82.8 92.2
Sarcosine 40.5 30.4 75.1
N, N-dimetMglycine 29.4 18.6 63.2
Betaine 40.6 6.8 16.7

150 °C Conversion (%) Yield (%) Selectivity
Glycine 98.5 79.9 81.1
Sarcosine 71.6 50.7 70.8
N,N-dimetrflqucine 57.1 41.1 72.0
Betaine 73.7 17.4 23.6

 

 

The initial rates were calculated from plots of concentration (mol/L) versus time (min)
and are reported in units of kmol/kg of cat/min (Table 4). The hydrogenation rate of

reaction is reduced by a factor of ~3 as each methyl group is added. The activation

162

energy was calculated for each substrate and was 41, 56, and 50 kJ/mol for glycine,

sarcosine, and N,N-dimethyglycine.

Table 4. Rates of reaction for hydrogenation of glycine, sarcosine, and N,N-dmg (0.3 M

H3PO4, H20, 100, 130, and 150 °C, 1000 psi H2, 1g dry 5% Ru/C)

 

-Ri (kmol/kg of cat/s)*

 

 

 

 

 

 

 

100°C 130°C 150°C
Glycine 6.2 16.7 30.9
N-methylglycine 1.6 6.9 13.3
N,N-dimethylglycine 0.6 3.9 10.8
Betaine 0.0 0.0 0.0

 

*Rates are multiplied by 10"7

It is not clear whether the heteroatom at the oc-position is adsorbing on the surface. There
are a few different mechanisms to consider. It may be that the a-carbon adsorbs, or if the
acidic proton on the heteroatom is able to dissociate, then a pi bond can be formed and
interact with the surface. The ability to form this pi bond is only present in the hydroxy,
amino, methylamino and di-methylamino substrates. In traditional C-H activation, it is
the lone pair on the N that is active, so varying the amount of acid might lead to an
understanding of how the OH activation is affected. We know that for hydrogenation to
occur, acid must be present to ensure the carboxylic acid is protonated. Looking at the
amine species, the acid amount can be varied to slow or prevent hydrogenation. Jere et
al. examined the effect of phosphoric acid concentration on conversion of 0.22 M alanine
(100 °c and 1000 psi H2).50 The conversion was ~90% with 0.29 M H3PO4, ~65% with
0.14 M H3PO4, and 45% with 0.07 M H3PO4. H/D exchange will be exploited and
discussed in section 4.7 to examine how the reactivity of each of the amine species is

affected.

163

The methoxy substituent evidently benefits the reaction over the similar sized methyl
substituent. From the data, it appears that the hydrogenation reaction is not as sensitive
to sterics, with respect to a methyl group. In fact, the methylated substrate (isobutyric
acid) appears to be slightly more reactive than the simple carboxylic acid substrates
(propanoic acid, acetic acid). But, pivalic acid is about the same in reactivity as
propanoic and acetic acid suggesting that at some point the addition of another methyl (as
compared to isobutyric acid) may start to slow the reaction. In theory, the relatively
bulky methoxy or methyl group may adversely influence the adsorption of the substrate
on the catalyst surface as free rotation around the sigma bond between the C1 and C2
carbon is possible. The presence of a bulky substituent close to the sp2 carbon of the
carboxyl functionality could hinder the delivery of the hydride. However, this steric

effect does not appear for the simple alkanoic acids.

The methylated amine species have provided more information on the sensitivity of the
hydrogenation reaction to steric modifications at the a—amino position. Looking at the
trend for —N+H3, -N+H2CH3, -N+H(CH3)2 the reactivity falls off by a factor of three as
each hydrogen is replaced by a methyl. However, these replacements also represent
changes in the amino groups’ H-bonding abilities. To determine whether the addition of
a methyl(s) is hindering the substrate interaction with the catalyst surface, H/D exchange
studies can be done on this series. If less exchange is observed as methyl groups are

added, it may support the decrease trend in reactivity, partly due to a steric effect.

164

In the hydrogenation of organic acids, electronics may be more important than sterics due
to OH activation. To understand the relative rates of OH activation as well as gaining
more mechanistic understanding of the reaction occurring at C2 in the organic acids, a
series of substrates has been studied; Direct C-H activation has been observed with an
ether (THF) and direct C-H activation or indirect (via dehydrogenation) C-H activation in
terminal alcohols (ethanol, glycols) is commonplace with the heterogeneous Ru catalyst
systems studied here. For future work, it is suggested that (S)—2-hydroxybutanol, (S)-2-
methoxybutanol, and (S)-2-aminobutanol be evaluated because they are secondary
alcohols (similar to the organic acid substrates). It would be interesting to measure the
rate of deuterium incorporation at C2 as well as observe whether deuterium is
incorporated at either C1 or C3 in the hydroxyl and amino substrates, indicating the enol
and enamine intermediates. The rate of deuterium incorporation at C2 can be compared
to the amount of deuterium incorporated at C2 of the corresponding acid. If deuterium is

incorporated in (S)-2-methoxybutanol it may occur via direct or-C-H activation.

Another factor that is inevitably affecting the hydrogenation reaction is the adsorption
affinity of each of the substrates. Competitive adsorption studies with the four aniino
species (N+H3, N+H2CH3, N+H(CH3)2, and N+(CH3)3)C1' should also be carried out to
evaluate their relative adsorption affinity. If the trend follows their reactivity trend then a
deeper understanding of the decreased reactivity can be realized. If they have similar
adsorption affinities; however, it would suggest that the sterics or electronics are the
controlling factors. Understanding the product distributions in the amino species and the

reaction kinetics will help fill in details regarding their relative reactivities.

165

4.7 H/D Exchange Studies

4.7.1 Amino Acid Substrates

4.7.1.1 H/D Exchange at the Methylene Position

The catalytic hydrogenation of the methylated glycine amino acids to amino alcohols has
been discussed. In general, amino alcohols can serve as building blocks for numerous
products in pharmaceutical and fine chemical applications. Research conducted in the
Jackson/Miller group by Jere et al. and Pimparkar et al. has shown the stereoretentive
reduction of alanine, serine, and valine.50,63 In an effort to build upon this research and
probe the reactivity differences in hydrogenation of the methylated glycine series discuss

above, H/D exchange studies were conducted on this series.

The amino acid substrates that were studied included glycine, N-methylglycine
(sarcosine), N,N-dimethylglycine, betaine, and glycylglycine. Again, the rate equations
(Chapter 3) can be used to calculate rate constants k1 A, km, km, and km. The reactions
described consider the back reactions (as opposed to the work by Frey et al.26 and
Richard et al.28’29). Although, the back reactions were not expected to have much
significance, especially in the reactions with 1000 psi D2 and D20. This will be

discussed more as the individual experiments are considered.

An early experiment with 0.25 M glycine at 130 °C with 1000 psi H2 and D20 for two

hours showed that the H/D exchange reaction occurred very quickly at this temperature.
A CH2 conversion of 99% was found and 18% and 80% (calculated) yield of CHD and

CD2, respectively. The 1H NMR spectra taken of time zero and after two hours of

166

reaction time are compared in Figure 51. Conversion of CH2 is already at 71% at time

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

ZCI'O.
110 ---“- n— w—-~~——--—--—--——v-——————~—~—-- ———~——~—---——-——-~~ 90
+Time: 0 Hours —+—- Time: 2 Hours

100 ~- — , ,, _._____.___ _, #-- v __V,, — 80
90 ~~ .,,,, 7— "—9 «W 7* A 7777777 . . ~~ r 70
80 ~—~ ~— — 777w, , ~ — WWW, - _--VVLW. . ~ 60
70 -.,___-_ 7* 777w , - - ,, , ~ 50
60 w - i—— - . r 40
50 - ———~~ ~77 — . ,Aifl ~ - , —~ ———~~We~ 30
40 . 20
30 ~ 10

20 0
10 ~——~—————————— ———————7—- —- -— - — m» -10
0 = = - ~ , -20

3.33 3.32 3.31 3.30 3.29 3.28 3.27 3.26 3.25

Figure 51. 1H NMR of H/D exchange at the methylene position of 0.25 M glycine (D20,

130 °C, 1000 psi H2, 1 g dry 5% Ru/C)

H/D Exchange Studies at 100 °C with 1000 psi Hg

In order to better understand and model the reaction rate kinetics, the reaction was
subsequently run at lower temperatures (100, 90, 70 and 50 0C). An experiment with
0.25 M glycine at 100 °C with 1000 psi H2 in D20 was allowed to react for five hours,
the singlet 1H NMR peak representing the methylene (CH2) position continued to
decrease in size over time, the triplet peak representing CHD grew in reaching a
maximum peak height at 60 minutes then started to decrease as the hydrogen at that
position was further exchanged with deuterium. The stacked 1H NMR spectra of time 0,

1, 2, 3, 4, and 5 hours of reaction time is compared in Figure 52.

167

 

 

 

Reaction " ” “ A

time My

 

 

 

 

 

Figure 52. Stacked lH NMR of H/D exchange at the methylene position on 0.25 M

glycine (D20, 100 °C, 1000 psi H2, 1 g dry 5% Ru/C)

Figure 53 shows the experimental and fitted concentrations of CH2, CHD, and CD2 over
time as determined by the equations in Chapter 3. The rate constants for H/D exchange
on the methylene carbon are: klA = 2.3x10‘4, km = 1.2x10'4, kZA = 1.2x10'4, k-2A =
3.2x10'4 L/mol/min and the sum of the squares error (SSE) = 3.9x10'5. The rate constants
for glycine, sarcosine, and N,N-dimethylglycine at these reaction conditions (D20, 100
°C, 1000 psi H2, 1 g dry 5% Ru/C) will be summarized after each experiment is
discussed. The sum of the squares error (SSE) is minimized for AH2 and AHD; thus it

does not include AD2 because AD2 is not directly measured. The plotted points as shown

168

also represent residuals, differences between the known initial glycine concentration and

the observable CH2 and CHD amount.

 

 

 

 

 

 

0.25 _..--.---- ..--.-.._..-__---.- -- _-_. ...-._- _ _.... _...--_--._...----._._-_F 20
l VAH2 —AHD —AD2
I AH2_exp A AHD_exp o AD2_exp
—H20 and HOD 0 H20 and HOD_exp
0.20 « L, . -----___ - - -7 9—— #4 4» 16
-. . a
3
=
g 015 -. 12
E’ ' :
° ‘5
a
E 010 g 8 IL
:
o
O
0.05 s» .. 4
0.00 V I T r T 7 0
0 50 100 150 200 250 300 350

Time (mln)

Figure 53. Fit of the kinetic rate data for 0.25 M glycine (D20, 100 °C, 1000 psi H2, 1 g

dry 5% Ru/C)

For this experiment, glycine reaches equilibrium at about 4 hours, thus the statistical
equilibrium end points can be calculated and compared to the experimental data. The
statistical concentrations of CH2, CHD, and CD2 based on the actual percent hydrogen in
water at five hours (14.5%) and the concentration at time zero (0.23 M) are 0.0048,
0.057, and 0.168 M, respectively. The actual concentrations of CH2, CHD, and CD2 at
five hours of reaction time are 0.0054, 0.075, and 0.166 M, respectively. Thus, the

values are in good agreement.

169

The same experiment using sarcosine (N -methylglycine) was performed at 100 °C with
1000 psi H2 and D20 for six hours. The 1H NMR spectra of time zero and after six hours
of reaction are compared in Figure 54. Figure 55 shows the experimental and fitted
concentrations of CH2, CHD, and CD2 over time as determined by the equations in
Chapter 3. The kinetics were determined using the models described in Chapter 3. Using
kinetic rate model A (Chapter 3.2.1) the rates for H/D exchange on the methylene carbon
are: km = 1.6x104, km = 0, k2A = 1.6x10'4, k-2A = 2.6x10'4 L/mol/min and the sum of the

squares error (SSE) = 4.5x10'3.

Again, we can compare the statistical equilibrium end points and the experimental data.
The statistical concentrations of CH2, CHD, and CD2 based on the actual percent

hydrogen in water at six hours (15.0%) and the concentration at time zero (0.24 M) are
0.0054, 0.062, and 0.176 M, respectively. The actual concentrations of CH2, CHD, and

CD2 at six hours of reaction time are 0.004, 0.054, and 0.199 M, respectively.

170

 

 

 

 

 

 

 

 

3500 3000
+Time: 0 Hours 7I7 Time: 6 Hours
3000 7 0 7 7 7 2500
1
1
2500 77 2 77 77 777 77 2000
1
4
2000 <— 7 77 7 1500
q I
1500 77 7 777 7 4 1000
1000 77 7 7 77 7 500
500 0
0 —500
3.47 3.45 3.43 3.35 3 33

Figure 54. IH NMR of H/D exchange at the methylene position of 0.25 M sarcosine

 

(D20, 100 °C, 1000 psi H2, 1 g dry 5% Ru/C)

 

 

 

 

0.30 18
7AH2 7AHD 7AD2
I AH2_exp A AHD_exp o AD2_exp
—H20 and HOD 0 H20 and HOD_exp

0.25 i 7777777 7 77 7 77 7 15
g 0.20—7 -7 777 7 .77 r 7 « 12
o o
E
v 3:
C a
3 015- 7 7 7 7 7 7 77 7779 c
s - 7 g
= a.
3
S
o 0.10 —7 77 7 7 7 7 77 7 77 7 77 77 6

\74
0.05 i 77 77 777 77 7—777—‘ 7 3
A I \\.K
0.00 7 1 . I , Y Q r I 0
0 50 100 150 200 250 300 350 400
Time (min)

Figure 55. Fit of the kinetic rate data for 0.25 M sarcosine (D20, 100 0C, 1000 psi H2, 1

g dry 5% Ru/C)

171

From the data, it appears the rate constants are too large or perhaps the fitted line should
be starting at some time, t, after time zero. In experiments with 1000 psi H2 and D20
there may be two reasons for the apparent need for a time offset or something equivalent.
The first reason is the slight delay in equilibrium between the liquid concentration and the
amount of available hydrogen or deuterium on the catalyst surface. And in the beginning
of the reaction, the amount of deuterium available on the surface is less so hydrogen
exchange at this point may be occurring but with the hydrogen on the surface. The
second reason may be a time delay due to the catalyst ramping up in activity. This will

be discussed further.

The same experiment using N,N-dimethylglycine (N ,N-dmg) was performed at 100 °C
with 1000 psi H2 and D20 for six hours. The 1H NMR of time zero and after six hours of
reaction time is compared in Figure 56. Figure 57 shows the experimental and fitted
concentrations of CH2, CHD, and CD2 over time as determined by the equations in
Chapter 3. Using kinetic rate model A (Chapter 3.2.1) the rates for H/D exchange on the
methylene carbon are: km = 7.1x10'5, km = 0, k2A = 6.4x10'5, k-2A = 1.6x10'4 L/mol/min
and the sum of the squares error (SSE) = 2.5x10'3. Again, we can compare the statistical
equilibrium end points and the experimental data. However, for the experiment with
N,N-dimethylglycine the equilibrium is not reached at six hours. The statistical
concentrations of CH2, CHD, and CD2 based on the actual percent hydrogen in water at
six hours (15.3%) and the concentration at time zero (0.24 M) are 0.0057, 0.063, and
0.175 M, respectively. The actual concentrations of CH2, CHD, and CD2 at six hours of

reaction time are 0.018, 0.084, and 0.142 M, respectively.

172

1400 1

 

1200 1

1000 1

 

800 1

600 7

400 7

200

 

3.59

 

 

 

|3.53 3.52 3.51 3.50 3.49 3.48

———__—_—‘

3.55

3.57

+Time: 0 Hours 7| Time: 6 Hours

1
------ A
v vvvv'- A

1
v -A‘
vav'

 

 

[ 1200
. 1000
w» 800
7 600
400

7 200

 

-200

 

3.45

Figure 56. 1H NMR of H/D exchange at the methylene position of 0.25 M N,N-dmg

(D20, 100 °C, 1000 psi H2, 1 g dry 5% Ru/C)

 

 

  

 

 

  
 

 

   
 

 

 

0.30 -_.__....._._---._-___-.. ..-.-. __ .- _---.-. -.. . -____ _...-..._.-.______-_-___ ..--.... 18
7AH2 7AHD 7A02
I AH2_exp A AHD_exp O A02_exp
—H20 and H00 0 H20 and HOD_exp
0.25 77777 7 7- 77 7—77 -—7777 777 7 7“ 15
g 0.20 7 77777 W~7~ 12
O
f 3:
o E
0.15 7 77 9
E i
8
8
0 0.10 7 ~ 6
0.05 .. 3
0.00 r , 0
0 50 100 150 200 250 300 350 400
11me(mln)

Figure 57. Fit of the kinetic rate data for 0.25 M N,N-dmg (D20, 100 0C, 1000 psi H2, 1

g dry 5% Ru/C)

173

It is important to note, that we need to be skeptical of the rate constants for the back
reactions (km and k-2A, deuterium to hydrogen) because these rate constants depend so
heavily on small variations in the “main event” of replacing the deuterium with hydrogen
and its rate variation with time and how well we are modeling the aqueous [H]
concentration. Unlike the hydrogen to deuterium exchanges, there is no initial rate
behavior we can easily compare to the fitted constants, because deuterium to hydrogen
only happens after substantial deuterium incorporation has occurred, and even then, it’s a
minor process because of the relatively low hydrogen concentration in the water. In
addition, our water hydrogen and deuterium concentration versus time function is an
empirical fit to fairly spread out data points and if it were off a bit, that could throw off
the back reaction rate constants significantly. In order to partially address this, a more
constrained model using essentially one k value for each site and then modified via

isotope effect numbers was also explored.

A kinetic isotope effect is observed when the reaction rate is affected as one atom in a
compound is replaced with its isotope. For the case of hydrogen and deuterium, the K”;
is expressed as KIE=kH/kD. All of the H/D exchanges are assumed to be occurring

through the same process, but now instead of four independent variables, the system is

redefined and has three independent variables, kiA, pA, and SA. As discussed in Chapter

3, k1 A is equal to 2*kiA indicating that there are two hydrogens on the methylene carbon
and subsequently two chances for H/D exchange. The equations for how the k, A, etc rate

constants could be effectively calculated from the kiA, pA, and SA values are in Chapter 3

(equations 56 and 57). It was assumed pA and SA are independent.

174

Looking at the same three experiments glycine, sarcosine, and N,N-dimethylglycine at
100 °C with 1000 psi H2 and D20 the chemical equations which incorporate the kinetic
isotope effects (6-variable kinetic model) were solved. For the forward reactions, the rate
constants using both models are similar. For the back reactions the rate constants are

somewhat different. Again we need to be skeptical of their meaning. When constraining

the model to use only one rate constant, ki and incorporating primary and secondary

isotope effects, the rate constant for glycine is kiA = 1.2x10“1 L/mol/min and the isotope
effects are pA = 0.76 and SA = 0.98. For sarcosine kiA = 0.82x104 L/mol/min, pA = 1.25,
and SA = 0.51. For N,N-dimethylglycine kiA = 0.36x10'4 L/mol/min, pA = 2.12, and SA =

0.69.

The rate constants, ki, for glycine, sarcosine, and N,N-dimethylglycine are about half of

k1 A and k2A that were calculated using the lO-variable model. This makes sense based

on the fact that k1 A = 2*kiA and the klA as computed from both the 10 and 6-variable
models are nearly the same. In addition, the fact that the kiA values are similar to k2A for

sarcosine and N,N-dimethylglycine makes sense because k2A = kiA/SA and SA for those
two substrates is close to 0.5. For glycine, the secondary isotope effect is essentially one,

so kiA is essentially the same as k2A and half the value of k1 A. It is important to note that
it is the value of p A that is most sensitive to the back processes, so the SA values should

be more reliable than the pA values. For each N-methyl group added, the rate drops off

by about one-third.

175

To further verify whether isotope effects are playing a role in the IUD exchange reaction

at 100 0C with 1000 psi H2 and D20, the primary and secondary isotope effects for the

experiment with 0.25 M sarcosine were fixed to 1. The rate constants km and kic were
then solved. The newly solved rate constants km and kic were equal to 1.1x10“4 and

0.51x10‘4, respectively. These were compared to the original km and kiC which were

equal to 0.82x10'4 L/mol/min and 0.40x104 L/mol/min, respectively. The quality ofthe
fits indicated that when excluding the potential isotope effects, the rate constants are too

large with respect to the experimental data.

Due to the knowledge of a delay in the available deuterium on the surface in the reactions
with 1000 psi H2 and D20 a modified kinetic model was developed. In addition, the
same set of experiments with glycine, sarcosine, and N,N-dimethylglycine were
performed with 1000 psi D2 (and otherwise identical conditions). The rate constants for
these pairs of experiments should be the same. The comparison of these rate constants as
well as the modified kinetic model will be discussed following the experimental results

with 1000 psi D2 (100 °C).

H/D Exchange Studies at 100 °C with 1000 psi D2

As mentioned earlier, it is assmned that the equilibrium between the hydrogen/deuterium
on the surface and the deuterium/hydrogen in the liquid at a temperature of 100 °C is fast
with respect to the experimental data points. It is expected that the equilibrium is reached
within the first hour. In the experiments with 1000 psi D2 and D20 there is only ~3% or

less H in the solvent water and the gas is D2, so the equilibrium is essentially reached at

176

time zero. In the experiments with 1000 psi H2 and D20 there is about ~15% H in the
solvent water and the apparent delay in equilibrium needs to be addressed. The delay in

equilibrium and its potential affect on the H/D reaction will be explored in this section.

The same three experiments were performed at 100 °C with 1000 psi D2 and D20. For
the experiment with 0.25 M glycine, the 1H NMR spectra of time zero and after six hours
of reaction time are compared in Figure 58. The singlet 1H NMR peak representing the
methylene (CH2) position decreased continuously while the CHD triplet grew in over
time, reaching a maximum peak height at sixty minutes of reaction time, and then
decreasing as hydrogen at that position was further exchanged with deuterium (Figure
59). Figure 59 shows the experimental and fitted concentrations of CH2, CHD, and CD2
over time as determined by the equations in Chapter 3. The forward rate constants for the
reaction with glycine at 1000 psi D2 are fairly similar to those with 1000 psi H2. The
rates for H/D exchange on the methylene carbon are: 1(1A = 1.7x104, k-1A = 0, k2A =

1.2x10‘4, k_2A = 7.0x10'5 L/mol/min and the sum of the squares error (SSE) = 7.9x10'5.

177

 

5000 4000

 

 

 

_ "" "" "" '_ — — "'25" + Time: 0 Hours 7!- Time: 6 Hours
4500 .J ‘ 3500
4000 77' 77 7 7 3000
3500 71 7 2500
3000 "I 3.30 .7 . id. 2000
I— —— —— —— —— — _ — —
2500 77 - 1500
2000 77 7 1000
1500 77 77 7 7 7 7 500
1000 0
500 7- 7 77 7 7 77 7 -500
f"""""'_'--"'1
o __________________ 1 4000

 

 

 

3.34 3.33 3.32 3.31 3.30 3.29 3.28 3.27 3.26 3.25

Figure 58. lH NMR of H/D exchange at the methylene position of 0.25 M glycine (D20,

100 °C, 1000 psi D2, 1 g dry 5% Ru/C)

 

0.30 3.6
AH2 7AHD —AD2
I AH2_exp A AHD_exp O A02_exp
—H20 and H00 0 H20 and HOD_exp
0.25 7 7 77 77 3.0
I

.0
N
o

 

Concentration (mol/L)
O
a?
3
Percent H

 

 

 

 

0.10 - 17 1.2

0.05 — «- 0.6

0.00 0.0
o 400

Tlmo (min)

Figure 59. Fit of the kinetic rate data for 0.25 M glycine (D20, 100 °C, 1000 psi D2, 1 g

dry 5% Ru/C)

178

As in the experiments with 1000 psi H2 and D20, we can compare the statistical
equilibrium end points and the experimental data. The statistical concentrations of CH2,
CHD, and CD2 based on the actual percent hydrogen in water at six hours (2.2%) and the
concentration at time zero (0.24 M) are 1.1x10'4, 0.0101, and 0.227 M, respectively. The
actual concentrations of CH2, CHD, and CD2 at six hours of reaction time are 4.7x104,
0.012, and 0.24 M, respectively. For this experiment, it appears that glycine has not quite

yet reached the equilibrium at six hours.

The concentration of CHD at time zero was compared in the reactions with 1000 psi H2

and with 1000 psi D2. For both experimental conditions, the presence of AHD was noted.
This suggests that even in the case with 1000 psi H2, there is some deuterium available on
the surface for exchange. In general, the experiments that were conducted in 1000 psi D2

and D20 resulted in larger concentrations of CHD at time zero.

The same experiment with 0.25 M sarcosine was performed at 100 °C with 1000 psi D2
and D20 for six hours. The 1H NMR of time zero and after six hours of reaction time is
compared in Figure 60. Figure 61 shows the experimental and fitted concentrations of
CH2, CHD, and CD2 over time as determined by the equations in Chapter 3. The reaction
of sarcosine started out quickly and the peak concentration of CHD occurred at about 60
minutes of reaction time but then rapidly fell off as the CD2 grew in. The H/D exchange
reaction rate constatns on the methylene carbon are: k1A = 2.8x10'4, k1 A = 6.9x104, k2 A =

2.9x10'4, k-2A = 8.3x10'4 L/mol/min and the sum of the squares error (SSE) = 5.7x104.

179

 

4500 3500

 

 

 

 

      
 

 

 

 

 

 

'— _ _ — _ _ _ "£01 +Time: 0Hours 7I-7T1me: 6Hours
4000 ..I 7 716 L 777 7 - 3000
7 7 7 12 I
3500 4 7’ 7 7 ' 8 777 77777 7- 2500
1 ’ ‘4 '
. r 1 0
3000 77 7 777 7 7 7 777 2000
I 3.40 3.39 3.38 3.37 3.36
2500 .-_-- _ — — .'._ f. :-: -, Li --- -.. - . 1500
200077 7 7 7 7 77 7 777777777 1000
1500 -7 7 7 77 500
AAA
1000 1 r , 0
500 «77 7 777 77 7 7 7 7 - 7500
l"""""""_-—__'_'-_l
0 _____________________ -1000
3.43 3.42 3.41 3.40 3.39 3.38 3.37 3.36 3.35

Figure 60. lH NMR of H/D exchange at the methylene position of 0.25 M sarcosine

(D20, 100 °C, 1000 psi D2, 1 g dry 5% Ru/C)

 

0.30 3 0
7 AH2 7 AHD —AD2
I AH2_exp A AHD_exp o AD2_exp
-—H20 and HOD 0 H20 and HOD_exp
0.25 77 77 777 7 77 77 77 2.5

 

 

 

 

 

 

 

g 77 2.0
O
E
V I
C a
O C
E 0.157 7 o 77 7 7 7 7 77 7 7 777 7 7 715 3
O
C n.
8
S
o 0.1077 77 77 777 7 77 77 77 7 7 7 7 771.0
0.054 7 7‘77 7 7 7 7 7 7 7 7 7 7 7 7 7 77- 0.5
A
NEW—...
0.00 r r r i r ' -— 00
0 50 100 150 200 250 300 350 400

Time (min)

Figure 61. Fit of the kinetic rate data for 0.25 M sarcosine (D20, 100 °C, 1000 psi D2, 1

g dry 5% Ru/C)

180

The experiment with sarcosine occurred very quickly, reaching an equilibrium at or
before 4 hours. The statistical equilibrium end points were again compared to the
experimental data. The statistical concentrations of CH2, CHD, and CD2 based on the
actual percent hydrogen in water at six hours (1.95%) and the concentration at time zero
(0.23 M) are 7.3x10'5, 0.0081, and 0.226 M, respectively. The actual concentrations of
CH2, CHD, and CD2 at six hours of reaction time are 1.2x104, 0.009, and 0.225 M,

respectively.

The same experiment using 0.25 M N,N-dimethylglycine was performed at 100 °C with
1000 psi D2 and D20 for six hours. The 1H NMR of time zero and after six hours of
reaction time is compared in Figure 62. Figure 63 shows the experimental and fitted
concentrations of CH2, CHD, and CD2 over time as determined by the equations in
Chapter 3. The reaction of N,N-dimethylglycine again started out quickly and the peak
concentration of CHD occurred at about 60 minutes of reaction time but then rapidly fell
off. The rate constants for H/D exchange on the methylene carbon are: k; A = 1.4x10'4,

kl A = 0, k2A = 1.1x10'4, k-2A = 2.2x10'4 L/mol/min and the sum of the squares error (SSE)
= 6.3x10‘4. The experiment with N,N-dimethylglycine was slower and did not reach
equilibrium after six hours. The statistical equilibrium end points were again compared
to the experimental data. The statistical concentrations of CH2, CHD, and CD2 based on
the actual percent hydrogen in water at six hours (2.7%) and the concentration at time
zero (0.20 M) are 1.5x10'4, 0.0105, and 0.188 M, respectively. The actual concentrations
of CH2, CHD, and CD2 at six hours of reaction time are 5.3x10'4, 0.0114, and 0.187 M,

respectively.

181

1200 1000
+Time: 0 Hours 7'7 Tlme: 6 Hours

 

   

 

 

 

0
3.49 3.48 3.47 3.46 3.45 3.44 3.43 3.42 3.41

 

Figure 62. ]H NMR of H/D exchange at the methylene position of 0.25 M N,N—dmg

(D20, 100 °C, 1000 psi D2, 1 g dry 5% Ru/C)

 

 

 

 

0.20 4
—AH2 —AHD —ADZ
I AH2_exp A AHD_exp o ADZ_exp .
—-H20 and HOD 0 H20 and HOD_exp
I
0.15 7 7 7 3
1 o
3
6 .
E o ‘ I
8 a
E 010 7 7 7 7 7 7 7 7 2 §
.- O
5 n.
8 A
o I
o
0.05 - . 7 7 7 7 7 7 7 7 7 77 1
1
A
\\
I I ‘
0.00 . 7 . 7 j 7 Y ._. 0
0 50 100 150 200 250 300 350 400
Tlme(m|n)

Figure 63. Fit of the kinetic rate data for 0.25 M N,N-dmg (D20, 100 °C, 1000 psi D2, 1

g dry 5% Ru/C)

182

Figure 64 compares the rate constants for glycine, sarcosine, and N,N-dimethylglycine on
the methylene carbon at 100 °C, 1000 psi D2, and D20 (reaction time = 6 hrs). For the
first hydrogen exchanged, sarcosine reacts the most quickly and for glycine and N,N—
dimethylglycine the rate falls off by roughly one-half. After the first hydrogen is
replaced with deuterium the rate of exchanging the second hydrogen with deuterium is
essentially the same for each compound as compared with the first hydrogen that is
exchanged (kl A compared to k2A).

3.5E-04

 

I Glycine I Sarcosine I NNdmg

3.0E-04 7 7 7 . i7

2.5E-04 7

2.0E-04 7

1 .5E-04 7 77

Rate constant (leollmin)

1.0E-04 77 7

5.0E-05 7

 

 

0.0E+00 7

 

Figure 64. Rate constants for H/D exchange at the methylene carbon (D20, 100 °C, 1000

psi D2, 1 g dry 5% Ru/C)

The sensitivity of the kinetic model to the back reactions is greatly reduced in the
experiments conducted with 1000 psi D2 and D20 at 100 °C. To verify this, the kinetic
model was modified and the back reaction rate constants were set to zero. So, the

number of fitted variable was decreased from 10 to 5 (Model A). The fits were compared

183

and for the experiment with D2, setting the back reaction rate constants to zero had

essentially no effect.

Using the more constrained model (6-variable) the rate constants k. A, k-1A, k2A, and k-2 A
were compared to the computed rate constants obtained from using the 10-variable
kinetic model. And, for the forward reactions the rate constants k1 A and k2 A are
essentially identical using either model (as previously mentioned). For the experiments
with 1000 psi D2 (D20, 100 °C, 1 g dry 5% Ru/C) the calculated secondary isotope

effects are comparable and less than one for glycine, sarcosine, and N,N-dimethylglycine
(Figure 65). For glycine kiA = 0.84x10“1 L/mol/min, pA = 4.4, and sA = 0.72. For
sarcosine ki A = 1.3x10‘4 L/mol/min, pA = 0.76, and sA = 0.46. For N,N-dimethylglycine
kiA = 0.71x104 L/mol/min, pA = 1.1, and SA = 0.65. The rate constants, ki, for glycine,
sarcosine, and N,N-dimethylglycine are roughly half of klA using the 10-variable model.
Again this makes sense because k, A is defined to be half of k1 A. And the k] A from the 10-

and 6-variable models are nearly the same. In addition, the secondary isotope effects, 3A,

are between 0.5 and 0.7, thus the calculated k2 A from the 6-variable model are all about

twice the value of k, A. This is again similar to the values computed for k. A.

' 184

 

 

 

 

  

 

 

 

 

 

 

 

 

1.5E-O4 7 7 7 5.0
DKi i
17315-04 Oprimary isotope effect , 4'5
. A secondary isotope effect
1.2E—04 «7 4.0
g 3.5
E
‘E’ 9.0E705 7 8.4E-05 ' 3.0 ,-
2 ‘ E:
E 1 7.1E-05 .. 25 g
g 6.0E-05 — p . 2,0 3‘
2 ’3
g 1.5
3.0E-05 . V .. , 7 1,0
7 ‘ -. O7 I
. A . I, 0 5
0.0E+00 V 0.0

 

Glycine Sarcosine NN—dmg

Figure 65. Kinetic rate constant and primary and secondary isotope effects for glycine.

sarcosine, and N,N-dmg (D20, 100 °C, 1000 psi D2. 1 g dry 5% Ru/C)

It is interesting to note that in the case of 1000 psi D2 and D20 the rate constants for
glycine and N,N-dimethylglycine are comparable and both about one-third to one-half

that of sarcosine.

Kinetic Model Modification — Linking experimental data 1 1000 psi H2 and D3)

The kinetic model for isotope exchange was varied to explore the effects of various

 

assumptions with the intent of optimizing the fit to the experimental data. Experiments at
100 °C were discussed for glycine. sarcosine, and N,N-dimethylglycine with either 1000
psi H2 or D2 and D20 as the solvent. The rate constants calculated for these pairs of
experiments should, in theory. be the same. Using the kinetic model discussed above. the

H/D exchange rate constants for the experiments with 1000 psi H2 and D20 appeared to

185

 

be too large. This is partly due to the fact that in the kinetic model the hydrogen or
deuterium available for exchange with the substrate was assumed to be in equilibrium
with the concentration of hydrogen or deuterium in the liquid. But from the experiments
with H2 and D20, it appears that there is a delay in the substrates' H/D exchange reaction

due to the time it takes to reach equilibrium among liquid, gas, and surface hydrogens.

Note that for the experiments with 1000 psi D2, the aqueous hydrogen concentration is

almost an order of magnitude lower than the typical 15% or so that is observed with H2
(Figure 67). This time evolution is important, but also reflects the gas/liquid exchange.
The rate of exchange for the second hydrogen is essentially the same in both conditions
(H2 or D2) which suggests an equilibrium is reached and at that point the reaction is not

limited by the presence of available deuterium.

 

0 H20 and HOD_1000psi H2
9 H20 and HOD_1000psi DZ '

16 -~—

 

14«

12 «-

 

 

10-

Percent H

 

 

 

—-—-'—‘f 4" “v- _A» . -L

 

0 50 100 150 200 250 300 350 400
Time (min)

Figure 66. Percent H in D20 with either 1000 psi H2 or D2

186

In terms of the delay of the deuterium available on the catalyst surface, this may be a
result of not having deuterium available to get on the surface until there is enough HD
formed. Scheme 30 shows the proposed mechanism for exchanging H and D between the
various isotopomers of H2, H20, and Ru-bound H atoms. Previous work in our research
group has shown that a significant amount of HD is generated before D2 is formed.
According to this mechanism, it is not until HD is formed (in the experiments with H2

and D20) that the catalyst surface is deuterated so that the substrate can start undergoing

H/D exchange.
,0 o D 1:
H D x’ \ / Dx‘ /
H O I ‘0
R. + H ——> ——» l
RU \RU
H—H
H-D
,H D H D 3i:
H
D H \0/ I ‘O/
, / «— ‘__
Ru ”—0 H‘\ l
RU ‘Ru

Scheme 30. Proposed mechanism for Ru/C catalyzed HD generation from H2 and D20
Note that exchange between H2 and D20 takes place without isotopic exchange on the Ru

surface.

Various kinetic model modifications were employed to address this background
equilibrium process. To start, the corresponding H2, D20 and D2, D20 experimental

datasheets were linked and a common set of rate constants was defined. The hydrogen

and deuterium available were defined by HS and DS (Chapter 3) as opposed to using the

187

liquid hydrogen and liquid deuterium concentrations. It would be ideal to assume that the

total concentration in solution is in equilibrium with the surface, but since this can not be
confirmed, it has to be treated simply in terms of rates. DS is a variable and represents the

fraction of total sites occupied by deuterium. For the experiments conducted with 1000
psi H2, the initial H on the surface was assumed to be 111.11. This is because at time
zero it is assumed that the amount of hydrogen on the surface is more similar to the gas
(which in this case is H2) than the aqueous D20. For the experiments conducted with
1000 psi D2, the initial H on the surface was assumed to be equal to the amount of H

measured in the liquid (usually less than 3%).

In addition, a catalyst tum-on function was included, as defined by:

1 -(tdep__coeft)*exp(tdep__exp*t),
where tdep_coeff is constrained to range 0-1 and can describe the fact that part of the
catalyst is fully active and there is a fraction that is not. This variable will ramp up to full
activity on a timescale described by the tdep_exp variable, an effective rate constant for

catalyst recovery.

Numerous fitting attempts were performed with the catalyst tum-on function included.
The results for both the lO-variable model and the 6-variable model (both now with kH as
defined in Chapter 3) indicated that with tdep_coeff and tdep_exp there was little to no
improvement of the fits. This finding is understandable in that much of the mathematical

flexibility the turn-on function offers is quite similar to that conferred by the kH variable.

188

The main focus was on using the 10- or 6-variable model with the addition of kH (as
defined by the equilibrium between the available hydrogen and deuterium on the surface
and in the liquid). It is important to mention that we do not know the exact amount of
available hydrogen and deuterium on the surface as it is not directly measured; only its
ratio is truly meaningful. Table 5 presents the rate constants for glycine at 100 0C with
1000 psi H2 or 1000 psi D2 and the linked H2 and D2 results for the 10- or 6-variable
model plus kH. It also includes the rate constants derived from the kinetic model
presented earlier in this thesis (based on the aqueous hydrogen and deuterium

concentrations).

The rate constants for the case of D2 and D20 are not affected much by the kinetic model
modification, due to the fact that this system essentially starts at equilibrium, so there is
not a delay in building up available deuterium on the surface (as in the H2, D20 case).
And kH was calculated to be zero with this model, again indicating that for the
experiments with D2, the assumption of using the aqueous H and D concentrations in the

kinetic model was adequate.

For the experiment with 1000 psi H2 and D20, the experimental (symbols) and fitted
(line) concentrations of CH2 (referred to as AH2), CHD (referred to as AHD), and CD2
(referred to as AD2) are presented in Figure 67. The fitted lines are a result of the 6-
variable model plus k“. This figure also shows the fitted (line) concentrations that

resulted from linking the experiment conducted with H2 with that with D2 and solving for

189

one common set of rate constants. Figure 68 presents the analogous results from the

experiment with 1000 psi D2 and D20. Again, there is not much difference in the fits.

Table 5. Rate constants for glycine (100 °C, 1000 psi H2 or D2) using the original kinetic

model, k10 + k“, k6 + kH

 

 

 

 

 

 

 

 

 

 

 

Glycine Glycine Glycine
Rate Constant 100 °C 100 °C 100 °C
(leollmin) H2 02 Linked H2 and D,

' k10 k10+kH k10 k10+kH k10+kH

k1_A 2.3E-04 2.7E-03 1.7E-04 1.7E-04 1.9E-04

k-1_A 1.2E-04 1.8E-03 0.0 0.0 1.3E-05

k2_A 1.2E-04 1.4E-03 1.2E-04 1.2E-04 1.2E-04

k-2_A 3.2E-04 4.0E-03 7.0E-05 1.2E-04 3.1E-04

kH(10) 8.0E-05 5.6E-05 1.0E-02

k6 k6 + kH . *k6 k6 + kH k6 + kH

ki_A 1.15E-04 2.20E-03 8.45E-05 8.49E-05 9.88E-05
p_A 0.76 0.69 4.39 2.36 0.80
s_A 0.98 2.10 0.72 0.71 0.81

kH(6) 8.20E-05 0.00E+00 8.22E-03

 

The apparent "blip" at the start of the curve fitting is a result from the initial H being

defined as all hydrogen (111.11), so that there is essentially no forward rate occurring,

k1 A but k.1A can take CHD back to CH2 (this is the same in the methyl case for sarcosine

and N,N-dimethylglycine for k2 A and k-3A). The rate constants and primary and

secondary isotope effects were calculated and presented in Figure 69.

190

 

 

 

 

 

 

 

 

 

 

- CH2-_exp A CHD_exp o c0235”
——CH2_H2 —CHD_H2 —c02_H2
0 25 7 7— -~ CH2_Iinked H2 and 02 CHD_|inked H2 and 02 ‘CDZ_linked H2 and 02
g 0.20 7 7 77 7 7 7
O
E _
g . _ -7 7'1?“ —#
E 0.15 7 7777— ,. 7”“ — — 777 —
g ,,
3 0.10 — 7 ‘7 7 /‘““~ 7 7h 7 —77
\\ a“ _ a. _ g
‘\
0.05 7 — 7 7 >7 —7 7 — 7 —— 77 — 7
A" // 7 at.
0.00 . T . . ' . ‘. 7
o 50 100 150 200 250 300 350 400
Time (min)

Figure 67. Fit of the kinetic rate data for 0.25 M glycine (D20, 100 °C, 1000 psi H2 or

linked H2 and D2, 1 g dry 5 % Ru/C)

0.30 -

 

 

 

 

 

 

i - CH2_exp A CHD_exp o CDZ__exp
—-—CH2_02 ——CHD_DZ —coz_02
0 25 ~ CH2_Iinked H2 and D2 CHD_Iinked H2 and DZ 77 CDz_linked H2 and D2
' l
§0.2077-7—~777—- —— — — 7 ,7 ' — 7 7 --7 7
o .2: //’,/
g '7‘ //'/
3 /'
30.1577 77 77 / 7 7 77 77 7 7 7 7 7 77
8 ‘ /
c . '”"‘—"‘"\\ .
80407—7577 77 7 7 7 7 7 7 7 7 7 7 7
/ ‘ i
005-" /7—\ 7777 7777777777 *
/ \
4 7‘ ”T" e _ -.h
0.00 7 7 . 7 7* , u
o 50 100 150 200 250 300 350 400
Time (mln)

Figure 68. Fit of the kinetic rate data for 0.25 M glycine (D20, 100 °C, 1000 psi D2 or

linked H2 and D2, 1 g dry 5 % Ru/C)

191

 

2.5E—03 ._- 2.5

 

 

. E1 KiA
0 primary isotope effect
A seconda isoto effect
2.0E-03 — ry pe 2.0
E
E
g 1.5E—03 — 1.5 A
n
3 i‘
3. I
'= L“,
g y
3 1.05-03 7 1.0 “
ti o
“5 A
5.0E-04 7 - 0.5
0.0E+00 7 - 0.0
02 Linked H2 and DZ

 

Figure 69. Kinetic rate constant and primary and secondary isotope effects for glycine

with 1000 psi H2 or 1000 psi D2 or H2 and D2

Table 6 presents the rate constants for sarcosine at 100 °C with 1000 psi H2 or 1000 psi
D2 and the linked H2 and D2 results for the 10- or 6-variable model plus kH. It also
includes the rate constants derived from the kinetic model presented in this discussion

prior (based on the liquid hydrogen and deuterium concentrations).
Figures 70 and 71 are analogous to 67 and 68 (glycine) and show the experimental and

fitted concentration data for the methylene carbon. The rate constants and primary and

secondary isotope effects were calculated and presented in Figure 72.

192

Table 6. Rate constants for sarcosine (100 °C, 1000 psi H2 or D2) using the original

kinetic model, k10 + kH, k6 + k”

 

Sarcosine Sarcosine Sarcosine
Rate Constant 100 °C 100 °C Linked

(leollmin) H2 02 Linked H2 and 02

       

k1_A 1.6E-04 4.6E-O4 2.8E 04 3.0E 04 2.7E—O4
k-1_A 0.0 1.3E-O4 7.0E-04 9.4E-O4 2.4E-05
k2_A 1.6E-04 9.5E-04 2.9E-O4 3.2E-04 2.9E-04
k-2_A 2.6E-O4 1.6E-03 8.3E-O4 7.8E-O4 5.3E-O4
k1_C 1.1E-04 2.1E-04 1.6E-04 1.6E-04 1.6E-04
k-1_C 0.0 0.0 0.0 0.0 0.0
k2_C 1.4E—04 3.5E-04 1.9E-04 1.9E-04 2.0E-04
k-2_C 2.5E-06 4.0E-04 0.0 0.0 1.2E-O4
k3_C 8.3E-05 2.2E-O4 1.4E-04 1.5E-O4 1.3E-04
k-3_C 2.0E-04 8.5E-04 8.5E-04 8.1 E-04 4.7E-O4
kH(10) 7.1E-05 5.2E-05 1.2E-O4

 

 

 

 

 

 

 

 

2.9E-04 1.5E-04
_A 1.25 1.34 0.76 0.80 1.08
s_A 0.51 0.16 0.46 0.44 0.49
ki_C 4.0E-05 9.4E-05 5.5E-05 5.6E-05 5.8E-05
_c 0.71 0.75 0.45 0.49 0.82
s_C 0.47 0.40 0.59 0.59 0.63
mg 6.3_E-05 505-05 1.3E-04

 

Comparing kiA using the linked experimental data and that obtained in the case of 1000
psi D2 and D20, the values are not much different. This is to be expected because the rate
constants that are derived from the experiments with D2 and D20 are the " gold standard".
Figures 70 and 71 compare the fitted concentration versus time data using the k6 + kH rate
constants as well as the linked H2 and D2 k6 + kH rate constants. The fits are extremely
close for both the individual data fit and the linked data fit. The results from the H/D

exchange on the methyl carbon of sarcosine will be discussed in section 4.7.1.2.

193

 

 

 

 

 

 

I AH2_exp A AHD_exp o AD2_exp
-—AH2_H2 AHD_HZ ADZ_H2
0 25 AH2_Iinked H2 and DZ AHD_Iinked H2 and DZ ADZ_linked H2 and D2
. l' "7’ ”” ' ’77””'
g 020 . a - .2./,. , artif—7Q a
a /”
/
é //
S o
'5 0.15 ~7 *5 , i ' ’ ’
2
‘0
E
d)
0
g _
o 0.10 7777 / \ 7 7 777 7
/ F x
/ \\ ‘\\.
I, \ ‘‘‘‘‘‘‘ 7““ 771i _ _.
0.05 - 1" i , "r - A , ,
0.00 7 7 . , ‘ . T , "
0 50 100 150 200 250 300 350 400
Time (min)

Figure 70. Fit of the kinetic rate data for 0.25 M sarcosine (D20, 100 °C, 1000 psi H2 or

linked H2 and D2, 1 g dry 5 % Ru/C)

 

 

 

 

 

 

0.30 7
I AH2_exp A AHD_exp o AD2_exp j
-—-AHZ_DZ ---AHD_D2 —AD2_DZ Q
025 7 :fl-12_IlnkedH2 anngZ 77 7—AHD:linkefl2 and DZ 7: ADZTlinkediHZ and D2
//
/
£0.20. 7 77 ’77 7 7 7 7 7
o
E
8
fiO.15-~77 i 777 77 7 7 7 7
.‘5
E
8
I:
80.1077; 777 7 77 7 7 77 7 7
/' \.
‘\
(105! .\ wt 7 “#7, r.- i a a 7 a,
\. A
\\7._
8“ \ 7 77.77 . _ *1
0.00 . r . - . - , = . =
0 50 100 150 200 250 300 350 400
Tlme(min)

Figure 71. Fit of the kinetic rate data for 0.25 M sarcosine (D20, 100 °C, 1000 psi D2 or

linked H2 and D2, 1 g dry 5 % Ru/C)

194

3.5E-04 1.6

 

El KiA
3 0E 04 0 primary isotope effect -. 14
' ' A secondary isotope effect
1.2
E 2.5E-04 r
E
3 1.0 A
5 205-04 7 g
E - 0.3 i
g 155-04 7 g
0 1- 0.6
3
fl
9‘ 1.05-04 - o
77 .4
5.0E-05 ‘ -e 02
0.0E+00 . > 0.0

 

 

 

H2 DZ Linked H2 and D2

Figure 72. Kinetic rate constant and primary and secondary isotope effects for sarcosine

with 1000 psi H2 or 1000 psi D2 or H2 and D2
The analogous data is presented below for N,N-dimethylglycine (Table 7, Figures 73-75).

The calculated fits for glycine, sarcosine, and N,N-dimethylglycine using the k5 + k“
kinetic model and linking the experimental data from the experiments with 1000 psi H2
and D2 were qualitatively very good and similar to the fits obtained when not linking the
experimental data. This indicates that incorporating kH into the kinetic model and using
the initial hydrogen and deuterium on the surface (inferred) is the best approach for
experiments conducted with 1000 psi H2 and D20. The original kinetic model that does

not include kH is adequate for experiments with D2 and D20.

195

Table 7. Rate constants for N,N-dimethylglycine (100 °C, 1000 psi H2 or D2) using the

original kinetic model, k10 + k“, k6 + kH

 

     
    
  

 

      

 

 

 

 

 

 

 

 

     

  

 

 

 

 

N,N-dimethylglycine N,N-dimethylglycine N,N-dmg
Rate Constant 100 °C 100 °C Linked
(L/mollmin)
‘ 33:11:10.1: -'fli3:.."-I:|EH23'E-‘i.€3.72:538.179 .- . ..ti..-. 3. 1 E.’ .- ' _
k1_A 7.1E—05 1.1E-04 1. E-04
k-1_A 0.0 1.1E-04 0.0
k2_A 6.4E-05 8.6E-05 1.1E-04
k-Z_A 1 .6E-04 3.15-04 2.2E-04
k1_C 1.0E-04 1.4E-04 1.9E-04
k-1_C 0.0 0.0 0.0
k2_C 1 .4E-04 1.8E-04 3.2E-04 .
k-2_C 1.8E-04 3.4E-04 0.0 . 0.0 4.8E-04
k3_C 1.5E-04 1.8E-04 2.8E-04 2.95-04 3.3E-04
k-3_C 5.8E-04 7.5E-04 8.8E-04 8.7E-04 1.3E-03
kH(10) 1.5E-04 3.0E-05 7.9E-05
2,, .. """" [(8 7 ‘ as + $7?
k1_A 3.6E-05 7.1E-05 7.2E-05 7. E-05
p_A 2.12 1.14 1.13 0.52
s_A 0.69 . 0.65 0.64 0.60
ki_C 3.5E-05 4.6E-05 6.9E-05 7.0E—05 6.6E-05
_C 0.83 0.81 0.87 0.93 0.68
s_C 0.52 0.58 0.47 0.46 0.48
kH(6) 1.7E-04 3.2E-05 8.2E-05

 

 

 

 

 

 

 

 

The rate constants for glycine, sarcosine, and N,N-dimethylglycine at 100 °C (k6 + k”,
linked H2 and D2) are kiA = 0.99x10'5, 1.5x104, and 0.74x10'4 L/mol/min, respectively.

The trend in reactivity (sarcosine > glycine > N,N-dimethylglycine) does not match the
reactivity observed in the hydrogenation of these same substrates (glycine > sarcosine >
N,N-dimethylglycine). This may be due to the fact that for H/D exchange an electron-
rich C-H bond promotes the reaction. Whereas, in the hydrogenation reaction, it is
presumed that steric effects are playing more of a role. It is observed that for each N-

methyl group added, the rate of the hydrogenation reaction falls by a factor of 3.

196

0.30

 

 

 

 

 

 

 

 

 

 

- I AH2_exp" . .. .. . A . AHD_epr .. -.. .. ... _ _. ;-.. AD2_exp - .. - ,_ - .. .. . .. _ ,
AH2_HZ AHD_HZ ADZ_HZ
0 25 ~ AH2_Iinked H2 and DZ ~ AHD_Iinked H2 and DZ ADZ_linked H2 and DZ
_\
g 0.20 7- -- —7 — 77— —7— — 77
O \
g .
g t
'3 0.15 T ’7 ' F;— ‘ i —i T T “ _
é \ 2...,
.
g \ ..77 ~ 7 ~r__2‘ _/
3 0.10 7 —7 , 7 77-7 , , .
I
A
o
,// Q /// \-
0.05 7— -- —- —7 /F 7 // v7-:‘~ i, A
' /
/. / o / ' .
/ /” ‘
7/ i
0.00 I V Y T T 1
0 50 100 150 200 250 300 350 400
Time (min)

Figure 73. F it of the kinetic rate data for 0.25 M N,N—dmg (D20, 100 °C, 1000 psi H2 or

linked H2 and D2, 1 g dry 5 % Ru/C)

 

 

0.30

 
 

 

 

 

 

 

I AH2_exp A AHD_exp o ADZ_exp
— AH2_DZ AH D_DZ ADZ_DZ
0 2, ___:Afiyinked H2 and- 9.2; 211977.912 and. 92 :_:_603—J"_“$Ed 32am 02
£02077 7 22 - _ _ i ..... - *
o
E
8
E 0.15 — —7
B
8
c
8 0.10 7
{a
0.05 7 ’ 7‘
1
0.00 .
0 50 1 00 1 50 200 250 300 350
11me(mln)

 

 

 

 

400

Figure 74. Fit of the kinetic rate data for 0.25 M N,N—dmg (D20, 100 0C, 1000 psi D2 or

linked H2 and D2, 1 g dry 5 % Ru/C)

197

 

1.0E-04 ___ wwww we 1_4
E! KiA
0 primary isotope effect

Aseconda isoto effect
8.0E-05 ~ W pe 0

 

   

6.0E-05 *

1
r
.o
m

 

KIE (kHIkD)

Rate constant (leollmln)
4r

> .~ } if; _
O
O)

4.0E-05 ‘ “*3; .
.ri: : ,. ' 1.
. . ‘ tr , ¢ ‘ "'_,{ .
,7 , , a "72“.: ‘7 “:55, ‘2 ‘1‘. , ,A ,
.m'r‘i; 2“. Tfizdh.‘ ‘_ Fifi-g" , 22::‘F‘L‘3 rift-3' . . .“fltr‘ fir
E'Ti'fiyifiifi-‘Jhr .. ' ‘ x. 3&3?“ 2 “153's 0 4
, , , , A . V _r,_ .:.,‘J;. 'b
"7 ‘h ‘ ‘ 7' .‘ I t r t‘ , _' ‘532
2 0E'05 ‘ ‘i i ' ' ‘1 '. ; . .t .t 411-2.: -
‘3‘ i 7 -' f 7.9.5 ,2‘»;'~‘.-, 1;; g: '
i "X"? ‘ “l "
+ 0.2

 

 

 

 

 

 

 

 

 

 

 

 

0.0E+00 ,...‘...’_ I 71:32:53.". "$331: ' I “_..ALZ‘L‘ 00
H2 DZ Linked H2 and 02

Figure 75. Kinetic rate constant and primary and secondary isotope effects for N,N-dmg

with 1000 psi H2 or 1000 psi D2 or H2 and D2

H/D Exchange Studies at 70 °C with 1000 psi D2

The reactions at 100 °C were occurring quickly reaching a maximum CHD concentration
at sixty minutes. Therefore, the reaction temperature was reduced to 70 °C in order to
obtain additional reaction rate data. The same three experiments were individually
performed at 70 °C with 1000 psi D2 and D20. For the experiment with glycine (0.25 M)
the 1H NMR of time 0.5 hours and 6 hours is compared in Figure 76. The singlet 1H
NMR peak representing the methylene (CH2) position continued to decrease in size over
time, while the CHD grew in, reaching a maximum peak height between 60 and 180
minutes of reaction‘time, and then started to decrease as the hydrogen at that position was

further exchanged with deuterium (Figure 77).

198

 

 

 

 

    
 

 

 

 

 

2500 _________ 1750
_ 600 +Time: 0.5 Hours -—I—Time: 6 Hours
2250 J 7 7 »- 1500
2000 «7| 7 1250
1750 71 7 7 1000
‘1500 4 3.30 3.29 3.28 3.27 3.26 , 77W t 750
1250 " , 77 7 _ 500
1000 77 7 7 77 77 77777 250
750 o
500 777 """"""" 717 777777 - -250
I
I
250 -7 7 1 7 - -500
I
I
o . -750
3.35 3.33 3.31 3.25 3.23

Figure 76. 1H NMR of H/D exchange at the methylene position of 0.25 M glycine (D20,

70 °C, 1000 psi D2, 1 g dry 5% Ru/C)

 

 

 

 

 

 

 

0.30 4.8 _
—AH2 77AHD 7—A02
I AH2_exp A AHD_exp o ADZ_exp
—H20 d H D 0 H20 dHOD
0.25 3 ,, an 0 an 7 _expr , 0 4
I
Q 0.20 - 7 7 7 7 77 3.2
3
E z
5 E
= 0.15 < 7 7 7 7 7 77 7 77777 77 2.4 w
8 A M
C A _.A.
8 0.10 7 7 7 7'7 77 7 ‘7“1 7 7- 16
I
o
0.05-77 7 47 7 77 7 7 7 7 7 7 770.8
. I
0.00 c v t t . , , , 0
0 50 100 150 200 250 300 350 400
Tlme (min)

Figure 77. Fit of the kinetic rate data for 0.25 M glycine (D20, 70 °C, 1000 psi D2, 1 g

dry 5% Ru/C)

199

Figure 77 shows the experimental and fitted concentrations of CH2, CHD, and CD2 over
time as determined by the equations in Chapter 3. The rate constants for H/D exchange
at the methylene carbon are: km = 1.5x10'4, km = 2.8x10'3, k2A = 4.1x10'5, k-2A = 2.5x10'

3 L/mol/min and the sum of the squares error (SSE) = 2.9x104.

The same experiment using N—methylglycine (sarcosine) was performed at 70 °C with
1000 psi D2 and D20. The 1H NMR of time zero and after four hours of reaction time is
compared in Figure 78. Figure 79 shows the experimental and fitted concentrations of
CH2, CHD, and CD2 over time as determined by the equations in Chapter 3. A maximum
concentration for CHD was realized in less than sixty minutes of reaction time. The rate
constants for H/D exchange on the methylene carbon are: km = 1.9x10'4, k- 1 A = 0, k2A =

1.4x10'4, k-2A = 1.2x10'3 L/mol/min and the sum of the squares error (SSE) = 6.4x104.

The same experiment using N,N-dimethylglycine was performed at 70 °C with 1000 psi
D2 and D20. The 1H NMR of time zero and after four hours of reaction time is compared
in Figure 80. Figure 8i shows the experimental and fitted concentrations of CH2, CHD,
and CD2 over time as determined by the equations in Chapter 3. A maximum
concentration for CHD was realized after about 180 minutes of reaction time and was
much slower than either glycine or sarcosine at 70 °C. The rate constants for H/D
exchange on the methylene carbon are: km = 1.3x10'5, km = O, k2A = 2.2x10'5, k-2A = 0

L/mol/min and the sum of the squares error (SSE) = 1.2x10'3.

200

 

1500 77--___ 7 - ___ 1250

 

   

 

:— ——————— "" +Time: 0 Hours 'Tlme 74 Hours
30
1 |
1250 "I - 7 -- 7*» 20 l ”_-... 77777 77 r 1000
I l 77“- r“ 10 I ,
100074 . to ‘ ‘ i750

 

 

|3.43 3.42 3.41 3.4 3.39 3.38 I "

750 77 - 77 0777- 7 7 500

500 7

 

 

250

 

 

3.47 3.46 3.45 3.44 3.43 3.42 3.41 3.40 3.39 3.38 3.37

Figure 78. lH NMR of H/D exchange at the methylene position of 0.25 M sarcosine

(D20, 70 °C, 1000 psi D2, 1 g dry 5% Rqu)

   
   
  
  
  

 

 

 

 

 

 

 

 

0.30 777-7- 777-7777 7777 3.0
7AH2 7AHD 7ADZ
I AH2_exp A AHD_exp o ADZ_exp
—H20 and HOD 0 H20 and HOD_exp
0.25 77 7 777- - - 7 - 7 777—77 7 77 7 2.5
51 0.20 7 2.0
2
.. z:
5 a
'- 7— 77 0
g o 15 1-5 2
8
8
S
o 7 7 7 7 7 771.0
7 A — 7 77 0.5
A
N
v 7 l 1 00
150 200 250 300

Time (min)

Figure 79. Fit of the kinetic rate data for 0.25 M sarcosine (D20, 70 °C, 1000 psi D2, 1 g

dry 5% Ru/C)

201

 

3000 777 777-77 7777 7777 7- 77 2500
+Time: 0Hours --+- Time: 4 Hours i

2500 ~ 7 2000

2000 4* 7~ 1500

1500 ~ 7 1000

1000 ~* ~ 500

500

 

 

 

 

 

0 -~ 7 7 - » -5oo
3.58 3.56 3.54 3.52 3.50 3.48

Figure 80. 1H NMR of H/D exchange at the methylene position of 0.25 N,N-dmg (D20,

70 °C, 1000 psi D2, 1 g dry 5% Ru/C)

 

  
  
 
 

 

 

 

7AH2 7AHD —ADZ
I AH2_exp A AHD_exp e ADZ_exp
-9- H20 and HOD 0 H20 and HOD_exp
0.25 7 7 ~77 7 7 - 7 ~ 1.5
g 0.20 7 7 1.2
2
V I
C a
o c
E 015 — — 7 7 7 77 0.9 3
h
o
C n.
8
E
o 0.10 7 7 7 7 7 7 rt 0.6
A
0.05 7— 77 -- 7 7 7 7 M777: 7 77 7* 0.3
K/ e
4/ __fl///
0.00 3 r v 7 .7 , O r I o
0 50 100 150 200 250 300

Time (min)

Figure 81. Fit of the kinetic rate data for 0.25M N,N-dmg (D20, 70 0C, 1000 psi D2, 1 g

dry 5% Ru/C)

202

Figure 82 compares the rate constants for glycine, sarcosine, and N,N-dimethylglycine on
the methylene carbon at 70 °C, 1000 psi D2, and D20 (reaction time = 6 hrs). For the
first hydrogen exchanged, the rate constant, km, for sarcosine is the largest, but is
comparable to glycine. The km rate for N,N-dimethylglycine is significantly less (more
than a factor of ten) than the rate for either glycine or sarcosine. After the first hydrogen
is replaced with deuterium the rate constants representing exchange of the second
hydrogen with deuterium are less for both glycine and sarcosine but are comparable for

N,N-dimethylglycine (k; A compared to k2A).

 

2.5E-04

I Glycine I Sarcosine NNdmg

2.0E-04 7

1.5E-04 3

1.0E-O4 - 7 7

Rate constant (Umollmln)

5.0E-05 7

 

 

0.0E+00 7

 

Figure 82. Rate constants for H/D exchange at the methylene carbon (D20, 70 °C, 1000

psi D2, 1 g dry 5% Ru/C)

A comparison of rate constants for the temperatures 100 °C and 70 °C with 1000 psi D2
for glycine, sarcosine, and N,N-dimethylglycine will be presented at the end of this

section.

203

Using the more constrained model (6-variable) the rate constants k1 A, k-1A, k2A, and k-2A
were compared to the computed rate constants obtained from using the lO-variable
kinetic model. And, for the forward reactions the rate constants k] A and k2A are
essentially identical using either model (as mentioned previously). The rate constant and

primary and secondary isotope effects were calculated for the experiments with 1000 psi
I)2 (D20, 70 °C, 1 g dry 5% Ru/C) and for glycine km = 0.74x104, pA = 0.03, and sA =
1.5. For sarcosine kiA = 9.9x10'5 L/mol/min, pA % 0.24, and sA = 0.72. For N,N-
dimethylglycine kiA = 5.9x10'6 L/mol/min, pA = 1.0, and sA = 0.40 (Figure 83). The rate
constants, ki, for glycine, sarcosine, and N,N-dimethylglycine are roughly half of k1 A
using the lO-variable model. Again this makes sense because kiA is defined to be half of

k1 A. The secondary isotope effects, SA, are between 0.4 and 1.5.

204

 

1.2E—04 1.8

EIKi
Oprimary isotope effect __ 1 6
_ Aseconda isoto effect '
1.0504 _ A 9.9E 05 W P6
E
g 8.0E-05 -
E A
a E
E 6.0E-05 - 5
5 E
g ._
o X
o
3 4.0505 -
E
2.0E-05 7
0.0E+00 7

 

 

 

Glycine Sarcosine

Figure 83. Kinetic rate constant and primary and secondary isotope effects for glycine,

sarcosine, and N,N-dmg (D20, 70 °C, 1000 psi D2, 1 g dry 5% Ru/C)

The rate constants for glycine, sarcosine, and N,N-dimethylglycine can be compared for
the reactions at 100 °C and 70 °C with 1000 psi D2. Additional experiments were
conducted at 50 °C and 90 °C, but not for all three substrates. So, the individual

substrates will be compared at each temperature at the end of this section.

The comparison data of the rate constants is in Figure 84. All of the rate constants are

lower at 70 °C, which was expected.

205

3.5E-04 7

 

 

 

 

IGlycine 1OOC_DZ DGlycine 7OC_DZ
lSarcosine 1OOC_DZ E Sarcosme 7OC_D2
3.0E—04 — mNNdmg1ooc_02 DNNdmg 7OC_D2 W i 7— g
E 2.55-04 7
E
a
o
5 2.0504 «
2'
S
u
g 1.5E-04 ,
o 999
. 999 999
999 999
E 999 999
1 ()5 04 . :3: :3:
' ' 999 999
999 , 999
999 999
999 999
3:: 3::
5.0E—05 i 999 999
999 999
999 999
999 999
999 999
999 999 .
0.0E+00 7

 

 

 

 

Figure 84. Rate constants for H/D exchange at the methylene carbon at 100 °C or 70 °C

(D20, X °C, 1000 psi D2, 1 g dry 5% Ru/C)

H/D Exchange Studies at 50 °C with 1000 psi D_z

The reaction temperature was lowered to 50 °C and the same experiments were
individually performed using glycine and sarcosine with 1000 psi D2 and D20. The
substrate N,N-dimethylglycine was not evaluated at this temperature because of the low
reactivity observed at 70 °C. The singlet lH NMR peak representing the methylene
(CH2) position continued to decrease in size over time, while the CHD grew in albeit
slowly. This reaction was significantly slower than at 70 °C and 100 °C. Figure 85
shows the experimental and fitted concentrations of CH2, CHD, and CD2 over time as
determined by the equations in Chapter 3. The rate constants for H/D exchange on the
methylene carbon are: klA = 5.5x10'5, 1cm = 8.1x10'3, k2A = 3.9x10'5, k-2A = 5.6x10'3

L/mol/min and the sum of the squares error (SSE) = 6.0x10'4.

206

 

 

 

 

 

 

—— AH2 —— AHD — ADZ
l AH2_exp A AH D_exp 9 ADZ_exp
--H H D o H H
0,25 . -flggand 0 20 and 003w ._ 25
5‘ 0.20 « 2.0
o
E
V :I:
8 9a
3 0.15 i 1 5 §
In
5 &
u
8
o 0.10~~-—7v 7- 1.0
0.05 7- — 7» 0.5
A
0.00 l I T r Y Y 0.0
0 50 100 150 200 250 300 350 400

Time (mln)

Figure 85. Fit of the kinetic rate data for 0.25 M glycine (D20, 50 °C, 1000 psi D2, 1 g

dry 5% Ru/C)

An experiment with sarcosine at the same conditions (D20, 50 °C, 1000 psi D2, 1 g dry
5% Ru/C) was performed and results again showed a slower rate of H/D exchange.
Figure 86 shows the experimental and fitted concentrations of CH2, CHD, and CD2 over
time as determined by the equations in Chapter 3. The rate constants for H/D exchange
on the methylene carbon are: km = 5.8x10'5, k-1A = 6.4x10'4, k2A = 4.3x10'5, k-2A =

3.4x10'3 L/mol/min and the sum of the squares error (SSE) = 4.9x104.

207

0.25 »-- -

 

 

 

 

 

 

 

 

 

 

 

..._.__._~__.s-.~-.___--_.....-c__....- .W- _. --- _ 20
—AH2 ———AHD ——A02 i
I AH2_exp A AHD_exp 9 ADZ_exp
—H20 and HOD 0 H20 and HOD_exp
0.20 7— — ————- —— — ., J7 __,__ 7 a 1.6
3
=
g 0.15 .- . .— ~- 1.2 I
g A A g
f— ' 1 0.
§ 0.10 4 - l» 0.8
c
o
O \
9
0.05 7 ——~— J —— 7— 0.4
’ \-
. 9
0,00 I I I I I I I 0_0
0 50 100 150 200 250 300 350 400
Time (mln)

Figure 86. Fit of the kinetic rate data for 0.25 M sarcosine (D20, 50 °C, 1000 psi D2, 1 g

dry 5% Ru/C)

The kinetic rate constant for the H/D exchange on the methylene carbon will be

compared to the rate constant for the H/D exchange on the methyl carbon at all three

reaction temperatures in Figure 111.

In summary, we can compare the rate constants for glycine, sarcosine, and N,N-
dimethylglycine at 100 °C, 70 °C and 50 °C with 1000 psi D2. The rate constants at 100
°C were calculated from the linked H2 and D2 experimental data (and the kinetic model
includes kH). Using this set of data, activation energies were calculated. The BA for the
methylene carbon of glycine and sarcosine was 45 kJ/mol and 36 kJ/mol, respectively.

The activation energy for N,N-dimethylglycine was calculated based on two

208

experimental points only, and was equal to 88 kJ/mol. Figure 87 presents km for

glycine, sarcosine, and N,N—dimethylglycine at 100, 70, and 50 °C.

 

1 .8E-04
I Glyci ne__kiA_C H2

Sarcosine_kiA_CH2

 

1 .5E-04

1.25-04 ~

9.0E—05 “-7—

6.0E-05 7r“ ,

Rate constant. klA (Umollmln)

3.0E-05 7— '

 

0.0E+00 -

.5...

 

I N,N-dmg_kiA_CH2

L-_,_ -. J—d

 

 

 

50 °C

Figure 87. Kinetic rate constant for glycine, sarcosine, and N,N-dmg (D20, X °C, 1000

psi D2, 1 g dry 5% Ru/C)

Again, it was observed that the rate constant for sarcosine is the largest and it falls off by

about one-third for glycine and one-half for N,N-dimethylglycine at 100 °C. The same

general trend is observed at lower temperatures.

The H/D exchange reaction is presumably promoted by having an electron-rich C-H

bond. If the pKa's are examined for the three substrates (Table 8), the carboxylic acid of

sarcosine is the most acidic and may result in sarcosine favoring the zwitterionic form

more than glycine favoring the zwitterionic form. For H/D exchange to occur, the amino

must be free, so this may suggest a slower rate for sarcosine than that of glycine. But,

209

sarcosine has a more electron-donating amino group as compared to that of glycine which

would promote the H/D exchange reaction.

Table 8. pKa values for glycine, sarcosine, and N,N-dimethylglycine (25 °C)

 

 

 

Substrate 'NHX(CH3)Y(+) -COOH
Glycine (x=3, y=0) 9.6 2.34
Sarcosine (x=2, y=1) 10.01 2.23
N,N-dimethylglycine (x=1, y=2) 9.8 2.08

 

 

 

 

Mass Transport Considerations

It was noticed that for the H/D reactions at 100 °C, the exchange is occurring quickly.
So, a brief discussion on mass transport limitations is included. From our hydrogenation
experiments it was determined that the overall rate of reaction was proportional to the
catalyst loading (based on experiments with half and double the catalyst amount) and
thus there are no external mass transport limitations in the system. This can be assumed
to be true for the H/D exchange reactions as well. For our system with H2 and D20, the
reaction rate should be relatively insensitive to catalyst loading because the flux of H2
from the gas to liquid would define the reaction rate. But, it is still possible to have
internal mass transport limitations (even if the rate is proportional to the catalyst loading).
This is because the catalyst will have the same rate per unit mass even if there are

concentration gradients within the catalyst particles.

lntemal mass transfer limitations occur when the Thiele modulus for the reaction is

greater than 0.3. For an n-th order reaction, the Thiele modulus, (D, is equal to

210

L(k*Co(n+l)/2/Dc)0'5, where L = catalyst dimension (W3 for a sphere), k = rate constant,

Co = bulk fluid concentration, and D6 = effective diffusivity of reactant in the catalyst.
The H/D reactions are essentially second-order but will be treated as first-order in the
D2/D2O case due to the large molar excess of deuterium. So, the equation is simplified to
(D = L(k/De)0'5. The effectiveness factor for intraparticle resistance is the ratio of the
actual rate to the true kinetic rate. The rate constant in the equation for (I) is the true or
intrinsic rate constant, which cannot be determined experimentally until we are sure there
are no mass transfer limitations. It can be determined using the observable modulus. The
effective diffusivity of glycine can be found from the Wilke-Chang equation. For the

experiment with glycine at 70 °C in water, the diffusivity was equal to 3.5x10'5 cmZ/s.

The effective diffusivity in the catalyst pore takes into account the pore fraction, so
Defflgly = epsz"'Dg|y = 1.2x10'9 mz/s (based on eps = 0.59). The initial bulk concentration
is 0.25 M = 0.25 kmol/m3 but will vary throughout the reaction. To get the observable

modulus, the rate of reaction Rg = k*gly must be expressed in units mol gly/cm3

catalyst/s (rate per unit catalyst volume). Using the data for the experiment with glycine
with 1000 psi D2 and 70 °C, the slope from the concentration (mol/L) versus time was
multiplied by the volume of the reacting phase and dived by the volume of catalyst in the
reactor (volume = mass catalyst/density of catalyst). For this experiment the rate equaled

5.3x10'3 kmol/m3*cat/s.

211

The best way to assess the presence of mass transport is to use the observable modulus

(11*(D2 = Rg*L2/Co/De) and if this (11* (D2) is greater than about 0.3, then mass transport

limitations in the catalyst particles during exchange need to be considered. For the
experiment with glycine the observable modulus is equal to (5.3x10'3
kmol/m3/s)*(2.5x10'5 m2)/(0.25 kmol/m'i')/(1.2x10'9 mz/s) = 0.01. So, the conclusion is
that there does not appear to be serious mass transfer consideration in the H/D exchange

experiments. This is primarily because we are using such a small catalyst particle.

An additional experiment with 0.25 M glycine was performed at 90 °C with 1000 psi D2
and D20. The singlet lH NMR peak representing the methylene (CH2) position
continued to decrease in size over time, while the CHD grew in, reaching a maximum
peak height between 30 and 60 minutes of reaction time. Figure 88 shows the
experimental and fitted concentrations of CH2, CHD, and CD2 over time as determined
by the equations in Chapter 3. The rate constants for H/D exchange on the methylene
carbon are: k1A = 1.9x10'4, in“. = 1.0x10'5, k2A = 1.3x10'4, k2,. = 4.8x10‘4 L/mol/min and

the sum of the squares error (SSE) = 3.2x10'4.

212

      

     

 

 

 

 

025 _.______-__.___._ _ .-. -_-" _ __.__. _ _. -___ _-..._....-__...-..._ .. p 5
—AH2 -—-*AHD ——ADZ
I AH2_exp A AHD_exp 9 ADZ_exp
—H20 and HOD 0 H20 and HOD_exp ’
0.20 ~ ~ - -, ~ ~ ,-, r 4
Q
E 0.15 « — — .. 3 I
g e E
E v 9 . §
E 010 ‘ g
e
O 4 \ A
\
\
0.05 ~ . -—- \\ _,__ ~ , _J .- 1
I \\
A
0.00 r r T _F‘M. . 0
0 50 100 150 200 250 300 350

Time (min)

Figure 88. Fit of the kinetic rate data for 0.25 M glycine (D20, 90 °C, 1000 psi D2, 1 g

dry 5% Ru/C)

An experiment with 0.25 M betaine ((CH3)3N+CH2COO') at 130 °C was conducted to
evaluate the extent of H/D exchange at the methylene position (Figure 12). Results at
130 °C with 1000 psi H2 for six hours revealed 86% conversion of CH2 to both CHD and
CD2. Figure 89 shows the experimental and fitted concentrations of CH2, CHD, and CD2
over time as determined by the equations in Chapter 3. The rate constants for H/D
exchange on the methylene carbon are: k; A = 4.1x10'5, k-.A = 0, k2A = 1.6x10'2, k-2A =

2.2x10'l L/mol/min and the sum of the squares error (SSE) = 1.9x10'3.

213

 

0.30 ~-—----——--- —«

        

 

 

 

--- . -_._ .. -.____..._. ....___.___ . __ . -.__-__.______._ --...-Hm _-_ 18
—- AH2 —— AHD — A02
I AH2_exp A AHD_exp 9 ADZ_exp
—H20 and HOD 0 H20 and HOD_exp
0.25 . 777 , .___ — d 15
.2: 0.20 12
.. :r:
3 e
.l O
B 0 15 9 2
o
E :L
8 0.10 , J~ 6
0.05 T 4.- 3
0.00 . r . . . . . e o
0 50 100 150 200 250 300 350 400

Time (min)

Figure 89. Fit of the kinetic rate data for 0.25 M betaine (D20, 130 °C, 1000 psi H2, 1 g

dry 5% Ru/C)

Isotopic exchange at the two methylene positions on glycylglycine was evaluated using
0.25 M glycylglycine at 100 °C with 1000 psi H2 and D20. Position A, the more

downfield 1H NMR signal, exchanged much more rapidly than position B (Scheme 31).

B o
3 63 ppm H
H2N/\”/ A OH
O 3.67 ppm

Scheme 31. lH NMR positions for the methylene hydrogens on glycyl glycine

From the lH NMR results, it appeared as if both hydrogens at the methylene position

exchanged at the same time, rather than step-wise. In most of the other isotopic exchange

214

experiments, the NMR results indicate the disappearance of the singlet and the
appearance of the triplet over time. However, in this case the singlet (3.67 ppm)
disappears quickly and without detectable triplet formation. The methylene position
corresponding to the singlet at 3.63 ppm does appear to form CHD, as illustrated by the

triplet formation.

A .
i Reaction

time
2L -

 

 

 

 

Figure 90. 1H NMR of H/D exchange at the methylene positions of 0.25 glycylglycine

(D20, 100 °C, 1000 psi H2, 1 g dry 5% Ru/C

Again, the amount of triplet formed is quite small. Presumably much of the reason for
the decrease in both 3.67 and 3.63 ppm methylenes is due to the formation of glycine.
The results every hour over a six hour period are displayed in Figure 90. Hydrolyzed
glycine is formed over time (3.36 ppm) and undergoes H/D exchange as it grows in

(Figure 91).

215

i Reaction
time n

 

 

 

 

 

 

 

Figure 91. 1H NMR of H/D exchange at the methylene position of hydrolyzed glycine

from 0.25 M glyclglycine (D20, 100 °C, 1000 psi H2, 1 g dry 5% Ru/C)

4.7.1.2 H/D Exchange at the N-Methvl Position
In this section H/D exchange at the N-methyl positions will be examined. At the end of
this section a comparison will be made between the rate constants at the methylene and

methyl carbons.
Again, the following chemical reactions (Chapter 3) can be used to calculate rate

constants klc, k-1c, k2c, k2c, k3c, and k-3c. As mentioned in Chapter 3, the equilibrium

between the hydrogens on the catalyst surface and the hydrogen concentration in the

216

liquid is not reached at time zero for the experiments with 1000 psi H2 and D20, so the

hydrogen concentration in the liquid is used.

The experiment describing the hydrogen-deuterium exchange on the methyl carbon is
depicted by three reversible, consecutive first-order reactions (Chapter 3). The amino
acid substrates which were studied included sarcosine (N -methylglycine), N,N-
dimethylglycine, and betaine. An experiment with 0.25 M sarcosine at 100 °C with 1000
psi H2 and D20 was run for six hours. The singlet 1H NMR peak representing the methyl
(CH3) position continued to decrease in size over time, while the triplet peak representing
CH2D grew in, reaching a maximum peak height at 120 minutes of reaction. A quintet
peak representing CHD2 grew in, reaching a maximum peak height at 180 minutes of
reaction time. The 1H NMR of time zero and after six hours of reaction time is compared
in Figure 92. Figure 93 shows the experimental and fitted concentrations of CH3, CH2D,
CHD2 and CD3 over time as determined by the equations in Chapter 3. The rate
constants for H/D exchange on the methyl carbon are: krc = 1.1x104, k-1c = 0, k2c =

1.45.10“, k-2c = 2.5x10'6, k3c = 8.3x10'5, and k.3C = 2.0x10'4 L/mol/min.

An experiment with 0.25 M N,N-dimethylglycine at 100 °C with 1000 psi H2 and D20
was run for six hours. The 1H NMR of time zero and after six hours of reaction time is
compared in Figure 94. Figure 95 shows the experimental and fitted concentrations of
CH3, CH2D, CHD2 and CD3 over time as determined by the equations in Chapter 3. The
rate constants for H/D exchange on the methyl carbon are: klc = 1.0x10'4, lclc = 0, k2c =

1.4x10'4, k-2c = 1.8x10‘4, log = 1.5x10‘4, and k-3c = 5.8x10'4 L/mol/min.

217

 

 

 

 

 

 
 
 
  
     
   

 

 

 

 

 

 

 

4000 f— _ _ _ _ _ _— __ __.__ 3000
25 +T'Ime: 0 Hours 7G77TIme: 6 Hours
7 2500
7 7 2000
12.54 2.53 2.52 2.51 2.50 2.49 "‘ “___ ___“ * 150°
2000 7—————— - 7 - 77 7 7~— - —— ~ 1000
1500 77 7r 500
1000 0
50077—777 7w—7—7 — ———— ——-——- ——--~-5oo
: ------------------- -|
0 i ___ —-———-——L-— _________________ ’ b -1000
2.58 2.56 2.54 2.52 2.50 2.48

Figure 92. 1H NMR of H/D exchange at the methyl position of 0.25 M sarcosine (D20,

100 °C, 1000 psi H2, 1 g dry 5% Ru/C)

 

 

 

 

0.30 18
— CH3 —— CH2D — CHDZ
—CD3 I CH3_exp A CH20_exp
9 CHDZ_exp e C03_exp —H20 and HOD

0,25 . e H20 and HOD_exp 1 15
2 0.20 77—7- —- 7 —— 7777 —— —7777 77 - 7 7 77 12
o
5
= as
o c
E 015 7 —— 7 7 7 — — 7 — . 7 7 9 §
§ ’ a
8 0.10 - e — ° .. 6

‘ \i
0.05 7 7 777 7I-7 ~ - 7 - - 7 7 ~— 77 77 3
O
I N
0.00 - e r r r l . 4 . I o
0 50 100 150 200 250 300 350 400
Tlme(m|n)

Figure 93. Fit of the kinetic rate data for 0.25 M sarcosine (D20, 100 °C, 1000 psi H2, 1

g dry 5% Ru/C)

218

4000 3500

 

 

   
   

 

 

l" — _ _ _ "‘ " “'5'“ +11me: 0Hours +Time: 6Hours
3500 J " 23 l 7 7 3000
1_.
3000 7| 7 77 2500
250° ii 2.73 2.71 2.69 W ‘“ 2000
2000 7777 7 1500
1500 77 7 77 7 7 1000
1000 77 7 — 500
500 . o
I ------------- 'I
o l ' -5oo

 

 

 

2.82 2.80 2.78 2.76 2.74 2.72 2.70 2.68 2.66 2.64

Figure 94. 1H NMR of H/D exchange at the methyl position of 0.25 M N,N—dmg (D20,

100 °C, 1000 psi H2, 1 g dry 5% Ru/C)

 

 

 

 

0.30 18
777- CH3 777— CHZD — CHDZ
—coa l CH3_exp A CH2D_exp
9 CHD2__exp e CDa_exp --H20 and HOD
025 l 0 H20 and HOD_exp —“ 15
e
3 e
= 0.20 -- 77 7 7 77 7 7 7 7 7 7712
E
.. I
8 ‘e
=1 0.15 7 7 7 7 7 77 9 3
E o 3
: n.
E
. - ,0 .-
8 0.10 6
O
0.057 7A7 7 7777 7 777 7773
l I
9
I
0.00 . . . . E . - o
0 50 100 150 200 250 300 350 400
Time (min)

Figure 95. Pit of the kinetic rate data for 0.25 M N,N—dmg (D20, 100 °C, 1000 psi H2, 1

g dry 5% Ru/C)

219

The same pair of experiments were performed at 100 °C but with 1000 psi D2. An
experiment with 0.25 M sarcosine was run for six hours. The 1H NMR of time zero and
after six hours of reaction time is compared in Figure 96. Figure 97 shows the
experimental and fitted concentrations of CH3, CH2D, CHD2 and CD3 over time as
determined by the equations in Chapter 3. The rate constants for H/D exchange on the
methyl carbon are: kic = 1.6x10‘4, km 7 0, k2c = 1.9x10'4, k-2c = o, k3C = 1.5x10'4, and R.
3c = 8.5x104 L/mol/min. The H/D exchange on the N-methyl of sarcosine was slower
than the methylene (in general, for all conditions) and did not fiJlly reach equilibrium
after six hours. The statistical equilibrium end points were again compared to the
experimental data (for the experiment with 1000 psi D2). The statistical concentrations of
CH3, CH2D, CHD2 and CD3 based on the actual percent hydrogen in water at six hours
(1.76%) and the concentration at time zero (0.22 M) are 1.2x10'6, 2.0x10'4, 0.0113 and
0.210 M, respectively. The actual concentrations of CH2, CHD, and CD2 at six hours of

reaction time are 2.1x104, 1.0x10'3, 0.027 and 0.22 M, respectively.

220

6000

 

 

 

 

 

 

 

 

 

777-777777 777----7-77 777—7777r 5000
+Time: 0 Hours
+11me: 6 Hours
5000 7 7 7 4000
4000 77 7 7 77 7 3000
3000 7 7 2000
2000 7 -~ - 7 1000
1000 0
0 I __________ f: f _______ . -1000
2.55 2.54 2.53 2.52 2.51 2.50 2.49 2.48 2.47

Figure 96. 1H NMR of H/D exchange at the methyl position of 0.25 M sarcosine (D20,

100 °C,'1000 psi D2, 1 g dry 5% Ru/C)

 

 

 

0.30 777777 777777-777-77-77- 7 3.0
777-CH3 777CHZD —CH02
—CDB I CH3_exp A CHZD_exp
I CHDZ_exp e CDS_exp —H20 and HOD
0-25 e H20 and HOD_exp ‘“ 2-5
E 0.20 7 --77 77 ' —- 7 77— 77 7777 7 77 —7 77 2.0
e
5 x
5 .. 27:
E «h i 2
e
e.
8 7 77 1.0
77 77 0.5

 

 

 

“fig-"__n 00
400

 

Time (min)

Figure 97. Fit of the kinetic rate data for 0.25 M sarcosine (D20, 100 °C, 1000 psi D2, 1

g dry 5% Ru/C)

221

An experiment with 0.25 M N,N-dimethylglycine was run for six hours. The 1H NMR of
time zero and after six hours of reaction time is compared in Figure 98. Figure 99 shows
the'experimental and fitted concentrations of CH3, CH2D, CHD2 and CD3 over time as
determined by the equations in Chapter 3. The rate constants for H/D exchange on the
methyl carbon are: kic = 1.9x10‘4, kic = 0, k2c = 3.2x104, k-2c_ = O, k3c = 2.8x104, and

k-3c = 8.8x10'4 L/mol/min.

The H/D exchange on the N-methyl of N,N-dimethylglycine appeared to just approach
equilibrium after six hours. The statistical equilibrium end points were again compared
to the experimental data. The statistical concentrations of CH3, CH2D, CHD2 and CD3
based on the actual percent hydrogen in water at six hours (2.7%) and the concentration
at time zero (0.15 M) are 2.9x10'6, 3.1x104, 0.011 and 0.134 M, respectively. The actual
concentrations of CH2, CHD, and CD2 at six hours of reaction time are 6.7x10'6, 1.3x104,

0.0041 and 0.196 M, respectively.

222

 

 

3500 -~— 7 ._.T 3000

 

 

 

 

 

 

+Time: 0Hours I'—
+Timez 4 Hours
3000 77 7—7 777777- - 7 I 7 2500
2500 777—7 7 77 - -——7 -- I r 2000
1
I268 2.67 2.66 2.65 2.64 I
200077- 77 77‘ 7777-— L __ _ _ _ _ _ _ _ __ 71500

1500 7*

 

1000 1

 

 

500

 

 

2.71 2.70 2.69 2.68 2.67 2.66 2.65 2.64 2.63

Figure 98. 1H NMR of H/D exchange at the methyl position of 0.25 M N,N-dmg (D20,

100 °C, 1000 psi D2, 1 g dry 5% Rqu)

 

 

 

 

 

 

 

 

777-CH3 —-—CHZD —-CHDZ i
——CDB I CH3_exp A CHZD_exp
9 CHDZ_exp e CD3_exp —H20 and HOD
0.20 g 0 H20 and HOD_exp e g 4.0
2.7
a
~€0.15777—77—777777 7 77 7 7 7777777— 7 77 3'0:
e
= e.
8 0.10 7 -———————— 7 7 77 777 —- 77—7 7777777 77 77 -7—— 77 — —— — 7777 2.0
C
o
o
0.0577 77 - 777 -—- 77 77 -77 7777 7 7 7 7 7 771.0
4 9
9
. A
1 v
0.00 . . L 7W. . I 00
O 50 100 150 200 250 300 350 400
Tlme(m|n)

Figure 99. Fit of the kinetic rate data for 0.25 M N,N-dmg (D20, 100 °C, 1000 psi D2, 1

g dry 5% Ru/C)

223

Again, as in the case of the methylene carbon, the kinetic model can be redefined using
three independent variables, kic, pC, and SC instead of six. As discussed in Chapter 3, kic
is equal to 3*kic indicating that there are three hydrogens on the methyl carbon and
subsequently three opportunities for H/D exchange. The equations for how the k 1c, etc
rate constants could be effectively calculated from the kic, pc, and SC values are discussed

in Chapter 3. It was assumed pc and sc are independent.

For the reactions at 100 °C with 1000 psi H2 one k (ki) value for each site that is modified

via an isotope effect was calculated. The results show that sarcosine and N,N-

dimethylglycine the primary isotope effects are similar and about 0.7-0.8. For sarcosine
kiC = 0.40x10'4 L/mol/min, pc = 0.71, and sc = 0.47. For N,N-dimethylglycine kic =
0.35x10‘4 L/mol/min, pC = 0.83, and SC = 0.52. The secondary isotope effects for

sarcosine and N,N-dimethlyglycine are similar and equal to about 0.5. The trend for the
rate constants, kg, for sarcosine and N,N-dimethylglycine match well with be using either
the 10 or 6-variable model. The ki values for both sarcosine are N,N-dimethylglycine are

about 1/3rd that of km, respectively. Again this makes sense based on the fact that kic is

defined as being equal to 3*kiC-

For the reactions at 100 °C with 1000 psi D2 one k (k,-) value for each site that is modified

via an isotope effect was calculated. For sarcosine kg = 0.55x10'4 L/mol/min, pc = 0.45,
and sc = 0.59. For N,N-dimethylglycine kic = 0.69x10'4 L/mol/min, pc = 0.87, and sc =

0.47. The secondary isOtope effects for sarcosine and N,N-dimethlyglycine are similar

224

and equal to about 0.5. Again, the ki values for both sarcosine are N,N-dimethylglycine

are about 1/3rd that of kic, respectively.

As discussed in the IUD exchange on the methylene carbon section, the data from the
experiments with 1000 psi H2 and D2 can be linked in order to define a common set of
rate constants. Figure 100 presents the comparison of the kinetic fits using the k6 + kH
model with the data from the sarcosine experiment with 1000 psi H2 and also from the
linked data set (1000 psi H2 and D2). Figure 101 is the analogous plot showing the
comparison of the kinetic fits with the data from the experiment with 1000 psi D2 and

also from the linked data set.

225

0.30

Concentration (moIIL)

0.05

0.20 777- -

I CH3_exp

9 CHDZ_exp
-—CH3_H2
—CHDZ_H2

7 CH3_linked H2 and 02
77 7 CHDZ_Iinked H2 and 02

a CH20_exp
e CD3_exp
—CHZD_H2
-—CD3_H2
CHZD_Iinked H2 and 02
7 7C03_linked H2 and 02

 

 

 

 

 

 

 

0.00

 

 

 

 

 

 

Time (min)

Figure 100. Fit of the kinetic rate data for 0.25 M sarcosine (D20, 100 °C, 1000 psi H2 or

linked H2 and D2, 1 g dry 5 % Ru/C)

0.30

 

0.25 7—

   

9
N
o
J

 

 

 

I CH3_exp
9 CHDZ_exp
CH3_DZ
CHDZ_02
7 CH3_linked H2 and DZ
----- 7 CHDZ_Iinked H2 and DZ

 

Concentration (mol/L)
O
a

 

  

a CH20_exp l
e CD3_exp i
CHZD_DZ

coa_02

~ 7 CHZD_linked H2 and 02

 

 

 

..—7 -'“"

~~~~
.th.’

o
‘1
’-
,

 

 

 

200
Time (min)

250

300 400

Figure 101. Fit of the kinetic rate data for 0.25 M sarcosine (D20, 100 °C, 1000 psi D2 or

linked H2 and D2, 1 g dry 5 % Ru/C)

226

The fits from the linked experimental data (1000 psi H2 and D2) is comparable to those
from the individual experiments. For the experiment with 1000 psi H2, the fit is slightly
faster using the linked data. This result again gives credence to the common rate
constants. Figure 102 shows the kinetic rate constant and primary and secondary isotope

effects for the H/D exchange on the methyl carbon of sarcosine. The rate constants for

the H/D exchange on the methyl are slower than those for the methylene (kiA = 1.5x10“1
versus kiC = 0.58x10'4 L/mol/min). The primary and secondary isotope effects are fairly

close pA = 1.08 versus pc = 0.82 and SA = 0.49 versus sc = 0.63. Again, there is not

much significance to the primary isotope effect because the back reaction is not occurring

under these conditions.

Figures 103 through 105 present the results for N,N-dimethylglycine. The rate constants

for H/D exchange on the methyl are slower than those for the methylene (km = 7.4x10'5
versus kg = 6.6x10'S L/mol/min). The primary and secondary isotope effects were
compared for the methylene and methyl carbon. As with sarcosine, they are very close
pA = 0.52 versus pC = 0.68 and SA = 0.60 versus SC = 0.48. The secondary isotope effects
for both sarcosine and N,N-dimethylglycine at the methylene and methyl carbon are

about 0.6 (note: glycine SA = 0.81).

227

 

 

3.515704 16
EKiC
oprimary isotope effect ._ 1 4
3'0E'04 l Asecondary isotope effect
7* 1.2
E 2.5E-04 7
E
3 77 1.0
5 2.05704 - ’9‘
g . 9 71 0.3 g
8 155-04 7 '5‘":
o
O
%
0‘ 1.05-04 7
5.0E-05 '
0.0E+00 7

 

 

H2 DZ Linked H2 and DZ

Figure 102. Kinetic rate constant and primary and secondary isotope effects for sarcosine

with 1000 psi H2 or 1000 psi D2 or H2 and D2

 

 

 

 

 

0.30
I CH3_exp A CH2D_exp
9 CHDZ_exp o CDa_exp
——CH3_H2 CHZD_H2
0-25 I —CHD2_H2 —CD3_H2 *
\ - CH3_linked H2 and DZ . CHZD_Iinked H2 and DZ
\\ 77 7 CHDZ_Iinked H2 and D2 - CD3_linked H2 and 02
g 0.207 7 7 ‘777 7 7777 777 7777 777 7- 7 7 7 7 77 7
o 7.
E \
s
320157—77 1-7 7 7 7 7 7 7 7 7 7 7 7-
S ,9
C x
0 \‘ ’ _____
0 \
s 72 . - 7
00107777 777 \7 7 77:— 77 7 _fl 7
/’ I. I — K >“ it”- I
/ / F \77\___
0.05-7777 7 A I 7 -7 7 7 77.7.; 77 77 7 7 7 7
// I”, i FFFFFFF xx
’ . 7" F \ \\\\\\ - F“‘7777 7 __ ...A
0.00 . . . . I: . .
0 50 100 150 200 250 300 350 400
Tlme(m|n)

Figure 103. Fit of the kinetic rate data for 0.25 M N,N-dmg (D20, 100 °C, 1000 psi H2

or linked H2 and D2, 1 g dry 5 % Ru/C)

228

 

0.30

 

 

 

 

 

 

I CH3_exp A CHZD_exp
o CHDZ_exp O CDS_exp
0 25 ‘_ CH3_02 CHZD_DZ
' —CHDZ_DZ —CDS_DZ
7 7 CH3_llnked H2 and DZ 77 7 CHZD_linked H2 and 02
. 7 7 CHDZ_linked H2 and Dz 77 CDa_linked H2 and DZ
5! 0.207 7 7 7777 77 777 77 77 7 77 7 . 7
o
g - u
s
E 0.15i77 7 77 7 . 7 7 77 7 7 7 7
E *’ '
8 7’
C /
80.107 7 7 7 77 7 7 7 777 7 77 7 7
./
/
0.05 77 7 7 i 7777\\ ‘ 7 7 77 7 77 77 77 7 7 7
4 77\_ “*7
,/ '7. \\“797\
1 x “7 ¥ ~ ,
0.00 r , - . - 7 _ , - .7. i
0 50 1 00 1 50 200 250 300 350 400
Tlme(m|n)

Figure 104. Fit of the kinetic rate data for 0.25 M N,N-dmg (D20, 100 °C, 1000 psi D2

or linked H2 and D2, 1 g dry 5 % Ru/C)

 

   

1.0E-04 14
El KIC
0 primary isotope effect 1 2
A secondary isotope effect '
8.0E-05 4
7.0E-05 -t
,5 6.6E-05 1'0
E
3
g 6.0E-05 7 A
J -- 0.8 g
2' :I:
g 5.
g 7— 0.6 E
8 4.0E-O5 7
2
N
I! 7 0.4
2.0E-05 7
- 0.2
0.0E+00 7 - 0.0

 

 

H2

DZ Linked H2 and D2

Figure 105. Kinetic rate constant and primary and secondary isotope effects for N,N-

dmg with 1000 psi H2 or 1000 psi D2 or H2 and D2

229

A third pair of experiments was conducted at 70 °C with 1000 psi D2 and D20. An
experiment with 0.25 M sarcosine was run for four hours. Figure 106 shows the
experimental and fitted concentrations of CH3, CH2D, CHD2 and CD3 over time as
determined by the equations in Chapter 3. The rate constants for H/D exchange on the
methyl carbon are: km = 7.4x10'5, k-1c = O, k2c = 9.4x10'5, k-2c = 2.1x10'3, k3c = 1.7x10’4,

and 1m =6.Ox10'3 L/mol/min.

 

 

  
 
   

 

 

 

 

 

 

0.30 7777-- 3.0
7CH3 7CHZD 7CH02
1 7003 I CH3_exp A CH20_exp
0.25 _ o CHDZ_exp o CDS_exp —H20 and H00 _2_ 2.5
0 H20 and HOD_exp
’57 0.20 J7 2.0
E
v :I:
5 E
1 A 7 O
E O 5 15 g
= o.
8
S
0.05 7 77 0.5
0.00 r r w r 0.0
0 50 100 150 200 250 300

Time (min)

Figure 106. Fit of the kinetic rate data for 0.25 M sarcosine (D20, 70 °C, 1000 psi D2, 1

g dry 5% Ru/C)

An experiment with 0.25 M N,N-dimethylglycine was run for four hours. Figure 107
shows the experimental and fitted concentrations of CH3, CH2D, CHD2 and CD3 over
time as determined by the equations in Chapter 3. The rate constants for H/D exchange
on the methyl carbon are: klc = 3.0x10'5, k-lc = 0, k2c = 1.1x10'4, k-2c = 1.1x10'2, k3c =

5.9x10'5, and k-3c = O L/mol/min.

230

0.2 5 -....- WW .. __ . _ . ._-- «WWW WW .... .. ...___- ___-..__.__--_....._.. _. ._ .. . __._~W___.-____.-.__.._. .. . .. T 2_ 5
—CH3 ——CHZD -—CHDZ

7C03 I CH3_exp A CHZD_exp
o CHD2_exp e CDB_exp -—H20 and HOD
0 H20 and HOD_exp

0.20

 

 

 

i
_s
0

Concentration (molIL)
.0
a:
/
3.
Percent H

 

 

 

 

0.05 7 7 77 77 0.5
0.00 1 r 1 . . 0.0
o 50 100 150 200 250 300

Time (min)

Figure 107. Fit of the kinetic rate data for 0.25 M N,N-dimethylglycine (D20, 70 °C,

1000 psi D2, 1 g dry 5% Ru/C)

For the reactions at 70 °C in 1000 psi D2 we calculated one k (ki) value for each site that

is modified via an isotope effect (Figure 108). The results show that for sarcosine and

N,N-dimethylglycine the secondary isotope effects are both about 0.5 (similar to the

experiments discussed above). For sarcosine kg = 0.24x10“1 L/mol/min, pC = 1.5, and sc
= 0.56. For N,N-dimethylglycine kic = 8.7x10'6 L/mol/min, pC = 1.0, and SC = 0.38.
Again, the ki values for sarcosine and N,N-dimethylglycine are about one-third of their

respective klcs.

231

3.0E-05 2.0

 

EIKi
0 primary isotope effect
Aseconda isoto effect
2.5E-05 - ”I ”6
1.6
.‘s‘
g 2.0E—05 -
3
E 1.2 A
a 2
E 1.5E-05 7 g
g “2‘
u -7 0.8
§ 1.0E-05 7
I!
77 0.4
5.0E-06 7
0.0E+00 7 - 0.0

 

 

 

Sarcosine NN-dmg

Figure 108. Kinetic rate constants and primary and secondary isotope effects for

sarcosine and N,N-dimethylglycine; 70 °C, 1000 psi D2

We can compare the reactions in 1000 psi D2 at both 70 °C and 100 °C for the H/D
exchange on the methylene carbon and N-methyls of sarcosine and N,N-dimethylglycine
(Figure 109). Again for sarcosine the rate of exchange for the methylene hydrogens is
slightly larger than for the methyl hydrogens. Whereas in the case of N,N-
dimethylglycine, the methyl hydrogens exchange at a slightly faster rate than the
methylene hydrogens. Again, this is partly due to the fact that with N,N—dimethylglycine
there are twice as many N-methyl hydrogens available for H/D exchange. As seen
previously, the rate constants calculated in the reaction at 70 °C are slower than the

experiment with otherwise identical conditions at 100 °C.

232

1.6E-04

I kiA CH2_70
1.45704 I kiA CH2_100
u kiC CH3_70
1.2E-04
? akic CH3_1OO
:5
-° 1.05-04
E
a
‘E 8.0E-05
a
n
I:
8 6 0E 05
8 ,
N
0:
4.05705
2.05705
o.oe+oo

 

Sarcosine NN-dmg

Figure 109. Rate constants for H/D exchange at 100 °C or 70 °C (D20, X °C. 1000 psi

D2, 1 g dry 5% Ru/C)

An additional experiment was done with sarcosine at 50 °C with 1000 psi D2 and D20.
Figure 110 shows the experimental and fitted concentrations of CH3, CH2D, CHD2 and
CD3 over time as determined by the equations in Chapter 3. The rate constants for H/D
exchange on the methyl carbon are: klc = 1.6x10'5, k-1c = 3.lx10'5, k2c = 1.6x10'5, k-2c =

1.6x10'6, k3C = 3.2x10'5, and k_3c = 2.6x10'3 L/mol/min.

233

0.30

 

    

 

71.8

-- CH3 --- CHZD — CHDZ
7003 I CH3_exp A CHZD_exp
025 . o CHDZ_exp e CD3_exp —H20 and HOD __ 1 5
' 0 H20 and HOD_exp '
g 0.20 . — .77 1.2
v 0 I
c a
o c
E 0.15 i 7 ~77 7 7 09 §
:: \ 8
§ \-
3 0.10 -77 —7 7—7- 7 7 77— 7 7 777 7 - 0.6
0.05 7 -7 - 7 7 7 0.3

 

 

 

 

W
0_00 ‘ r r r f r 1 r 0

“me (mln)

 

Figure 110. Fit of the kinetic rate data for 0.25 M sarcosine (D20, 50 °C, 1000 psi D2, 1

g dry 5% Ru/C)

We can compare the rate constants for the H/D exchange at both the methylene carbon

and the N—methyl carbons for the experiments with sarcosine at 100 °C, 70°C, 50 °C with
1000 psi D2. Figure 111 compares km and he The activation energy was calculated for

this set of experiments and determined to be EA = 67 kJ/mol (note the EA for the H/D

exchange on the methylene carbon was calculated to be 36 kJ/mol).

234

 

1.8E-04

I kiA CH2 I kiC CH3

1.5E-04 7

1.2E-O4 7

9.0E-05 7

 

6.0E-05 7

Rate constant (Umoiimin)

3.0E-05 7

 

 

 

0.0E+00 7

 

100 70

Reaction Temperature (’C)

Figure 111. Rate constants for H/D exchange of sarcosine (D20, 50, 70, or 100 °C, 1000

psi H2 or D2, 1 g dry 5% Ru/C)

In summary, the kinetic rate constants from the 6-variable model can be compared for the

three substrates at the reaction temperatures of 50, 70, and 100 °C (Table 9).

An experiment with 0.25 M betaine ((CH3)3N+CH2COO') at 130 °C was conducted to
evaluate the extent of H/D exchange at the N-methyl positions. Results at 130 °C, 1000
psi H2 for six hours reveal no H/D exchange at the N-methyl positions (Figure 112). The
integration value for the N-methyls differed by about 6% comparing time zero and after

six hours. This is within the expected value for adsorption onto the Ru catalyst.

235

Table 9. Summary of the rate kinetic rate constants and primary and secondary isotope

effects for glycine, sarcosine, and N,N-dimethylglycine

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Glycine Sarcosine N,N-dmg
100 °C 100 °C 100 °C
Rate constant Dz D2 Dz
" '0 kiA 0.99 1.48 0.74 , '
DA 4.39 0.76 1.14
sA 0.72 0.46 0.65
kiC V 0.58 0.66
90 0.45 0.87
SC 0.59 0.47
aycine Sarcosine N,N-dmg
70 °C 70 °C 70 °C
Rate constant Dz Dz Dz
pA 0.03 0.24 1.00
SA 1.54 0.72 0.40
pC 1.50 1.00
5C 0.56 0.38
aycine Sarcosine
50 °C 50 °C
Rate constant Dz Dz
" kiA 0.09 0.25
pA 0.10 0.35
5A 0.42 1.00
kiC 0.05
PC 1.00
sc 0.70

 

 

All rate constants are multiplied by 104

236

 

3500

--.- Time: 0 Hours +Time: 5 Hours

30007777 -~ —-

2500 7777—777 7W- WWW,”

2000 777-

 

500 7

 

O
3.20 3.15 3.10

Figure 112. 1H NMR of H/D exchange at the methyl position of 0.25 M betaine (D20,

3.05

130 °C, 1000 psi H2, 1 g dry 5% Ru/C)

4.7.2 Other Substrates

Several additional substrates were examined. There were specific reasons for evaluating
each of the substrates. Ethanolamine was chosen in order to determine whether the
hydrogens on the methylene next to the oxygen or the methylene next to the nitrogen
would exchange faster and it was hypothesized that the hydrogens on the methylene next
to the oxygen would exchange faster. Using 0.25 M ethanolamine at 100 °C with 1000
psi H2 and D20 and l g dry 5% Ru/C for four hours, the hydrogens on the methylene

adjacent to the oxygen exchanged more quickly than the hydrogens adjacent to the

nitrogen (Scheme 32).

 

237

 

 

3000

77 2500

~ 2000

7 1500

7 1000

77 500

-500

2.90

l 2.6 ppm

HO\/\
NH

2
3.4 ppm T

Scheme 32. 1H NMR positions for the methylene hydrogens on ethanolamine

It was not possible to deconvolute the CH2 from the CHD due to the fact that the splitting
was overlapping. Thus, the concentrations are expressed as CH2 plus CHD. The 1H
NMR of both the methylene next to the oxygen (Figure 113) and the methylene next to

the nitrogen (Figure 114) are shown below.
After two hours of reaction time, the deuterium exchange at both positions reached

equilibrium and was essentially identical (Figure 115). The concentrations expressed in

Figure 115 are the sum of CH2 and CHD.

238

 

2250 2000
2000 7- 7 1750
1750 7 1 500

- 1250

 

1250 71— 77 - 1000

7- 750

7 500

- 250

 

 

 

 

 

Figure 1 13. 1H NMR of H/D exchange at the methylene position next to the 70H of 0.25

M ethanolamine (D20, 100 °C, 1000 psi H2, 1 g dry 5% Ru/C)

 

2250 2000
7 +Time: 0 Hours +Time: 4 Hours

   

 

 

 

 

 

 

20007 7 7 7 77 7 7 7 7 77 7 7 771750
17507 . H 7 777777 7 7 771500
1500 7 7 W .- 7 77777 7 7 — 1250
1250 77— 7 1000
1000 7 i 750
750 - 7 7 500
500 7 7 7 L250

250 o
0 7250
2.61 2.59 2.47

Figure 114. 1H NMR of H/D exchange at the methylene position next to the —NH2 of

0.25 M ethanolamine (D20, 100 °C, 1000 psi H2, 1 g dry 5% Ru/C)

239

0.25

 

 

 

 

+ CH2-OH 7'7 CH2-NH2 :
l
I
0.20 77 -— -- l
S.
g 0.15 —
S
a .
S ;
§ 0.10 77 7
c i
o l
o a
0.05 77 -7— i
0.00 i i + 4 + l

pre-rxn 0 60 120 180 240

Tlme(m|n)

Figure 115. Concentration data for 0.25 M ethanolamine (D20, 100 °C, 1000 psi H2, 1 g

dry 5% Ru/C)

An experiment with 0.25 M glycolic acid at 100 °C with 1000 psi H2 and D20 was run
for 6 hours. This experiment was conducted in order to evaluate the extent of H/D
exchange at the methylene carbon. Results indicate that very little H/D exchange .
occurred at the methylene carbon and no measurable amount of CHD was observed. The
concentration after six hours of reaction time was 0.23 M. However, the peak did
become broader over the course of the reaction. Figure 116 compares one hour and after
six hours of reaction. Glycolic acid will undergo hydrogenation under similar conditions,
however, at 100 °C the hydrogenation rate is slowed significantly and no hydrogenation

products were detected.

240

 

150 _...- _-..-.. -_.- was".-. .---W-mv.-.-._--w M, _-...-.“ _-.-.-._-.-.--,. 125
+ Time: 1 Hour + Time: 6 Hours

1257777 777 7 —— 7— 777—7— - A7 7777 i 777 7100
100 4777 7 75
~ 50

75*

50 r— ,p 25

 

25

 

 

 

4.24 4.20 4.16 4.12 4.08 4.04 4.00 3.96

Figure 116. 1H NMR of H/D exchange at the methylene position of 0.25 M glycolic acid

(D20, 100 °C, 1000 psi H2, 1 g dry 5% Ru/C)

An experiment with 0.25 M tert-butanol at 100 °C for with 1000 psi H2 and D20 for four
hours was run to evaluate the extent of H/D exchange at the methyl carbons, although
little was expected. Results indicate that little H/D exchange occurred at the methyl
carbons and a nominal amount of CH2D and CHD2 was observed. The conversion after
four hours of reaction time was 26%. However, the peak did become broader over the
course of the reaction. Figure 117 compares one hour and after four hours of reaction. In
the solvent study work tert-butanol was examined as a solvent and it was hypothesized
that H/D exchange was not occurring on the methyl carbons due to an inability for OH
activation and this experiment confirmed this. However, there is still some solvent
degradation and presumably the interaction with the solvent and the catalyst is preventing

the substrate (in the hydrogenation experiments) reduction to the alcohol product.

241

 

 

 

 

 

 

5500 ~- 77777777777777 5000
-O-Time: 0 Hours -..-.. Time: 4 Hours
5000 «777—- —---- - 77 — --7 77 —7—7—7- 7 77 — 7 7 4500
4500 «7 — —7 7-77 7 77 — i - 77 7 — 7 7 4000
4000 +7 7— 7777 7777 7 77 77 7 7 77 7 3500
3500 17 -77 7 777 77 7— 7 ~— 7—7 7 7 7 7 7 777 i 3000
3000 + 7 - -- 7 7 77 7—- 7 7 7 —- 7~ 7 7— 7 77 7 2500
2500 7777 77 7 7 7 77 7 77—7 777 7 77 - 7- 7 2000
0
2000 «~— 7 7 7 - — 7 777—- — 77 7 ——— 7777 7— 71500
n
1500 ~7 ~77 - 7 77 7 77 77—7 ’. —7 7 77 7 —77—7- 7 71000
1000 77—7— - 777—- 77— 7777- 7777— 7 $2 77 77777 77 7—7 77—7 7 500
i
0 7- -500
1.06 1.05 1.04 1.03 1.02

Figure 117. 1H NMR of H/D exchange at the methyl positions of 0.25 M (err-butanol

(D20, 100 °C, 1000 psi H2, 1 g dry 5% Ru/C)

4.8 Cometition Experiments

4.8.1 Hydrogenation Studies

Betaine hydrochloride did not undergo either hydrogenation or H/D exchange and it was
not clear if the substrate was interacting with the surface and whether the chloride ion
was poisoning or deactivating the surface. To evaluate the hypothesis that betaine
hydrochloride was not interacting with the surface (possibly due to sterics) a competition
experiment with glycine was conducted. Presumably, if betaine hydrochloride is not
competing for the surface then the hydrogenation of glycine should be unaffected.
However, if the chloride ion is reducing the efficiency of the catalyst, then the glycine

hydrogenation rate would be affected.

242

Two competition experiments with betaine and glycine were performed. The first was
hydrogenation of 0.25 M betaine hydrochloride, (CH3)3N+CH2COOH°C1' and 0.25 M
glycine in the presence of 0.5 M H3PO4 at 130 °C with 1 g dry 5% Ru/C catalyst. The
ratio of both the substrates and the additional acid to the catalyst was roughly double with
respect to the baseline reaction of glycine alone. The results of the reaction compared

with the baseline reaction are presented in Figure 118.

 

 

 

 

 

 

 

100% ".--“..-
- 7 I - - Glycine alone conversion

90% ._. --0--Ethanolamine alone yield .. ______ i
—I— Glycine conversion _ . I‘ ' '

80% ,_ +Ethanolamine yield -' . 4 ........ 9
—I- Betaine conversion g, .- "
_._ . . _.7 f

70% d» g”, Choline yield, __7 _ _7 ,_ _ —,—.-.7"..,.7' _ .

._I '

60%77 ~77 - 77 — - - 7 — —.

500/017 77 - 7 7 77 7 -— III-77,7- 7 7 77 77

40% 77 ~ "

30% 77 77

20% «7 -—

10% ‘77

0% 3‘-
0 3O 60 120 180 240 300 360
Time (min)

Figure 118. Hydrogenation of 0.25 M betaine hydrochloride and 0.25 M glycine (0.5 M

H3PO4, H20, 130 °C, 1000 psi H2, 1 g dry 5% Ru/C)

The conversion of glycine is significantly suppressed by the presence of betaine
hydrochloride, however, the amount of substrate and additional acid present is double
with respect to the amount of catalyst typically used. So, the increased substrate-to-

catalyst ratio may have contributed partly to the reduced rate. In addition, chloride was

243

present as a complex with betaine. To examine the effect of the counter-ion chloride, an

experiment without it was conducted.

The second experiment with glycine and betaine (chloride-free) was performed using
0.125 M of each and 0.3 M H3PO4. The results of 0.25 M glycine alone versus 0.125 M
glycine and 0.125 M betaine are presented in Figure 119 (130 °C, 1000 psi H2, 1 g dry
5% Ru/C). After six hours of reaction time, the conversion of glycine alone is 90%
compared with 71% when in the presence of an equimolar amount of betaine. The yield
of ethanolamine is 83% compared with 59% when in the presence of an equimolar
amount of betaine. Thus, it appears that the betaine (chloride-free) is competing for
surface reaction sites. This is consistent with the early findings that betaine will undergo
hydrogenation at 100, 130, and 150 °C. So, it was expected that betaine would be

reacting on the catalyst surface, thus a potential competitor to glycine.

244

100% 7-

‘ - ~ Glycine alone conversion
‘ 0-7 Ethanolamine alone yield .. ....... 17—1
“mi—- Glycine conversion .
\O— Ethanolamine yield

90%

 

80%

 

70%
60%
50°/°

40 t>4,

30 “lb

 

ROD/O fl

1 00A) w

 

 

0% , r
30 60 120 180 240 300 360
Time (mln)

Figure 119. Hydrogenation of 0.25 M glycine alone and 0.125 M glycine and 0.125 M

betaine (0.3 M H3PO4, H20, 130 °C, 1000 psi H2, 1 g dry 5% Ru/C)

The results of betaine alone and betaine in the presence of an equimolar amount of
glycine are presented below (Figure 120). Betaine is much less reactive than glycine so
presence of glycine appears to have less of an effect on the betaine. The experiment with
betaine hydrochloride resulted in a negligible amount of betaine hydrochloride
conversion and the glycine conversion was reduced from 90% to 21%. Presumably this
is mostly due to the presence of chloride and also partly to the substrate- and additional

acid-to-catalyst ratio.

245

100% 7 ---"-

‘ ‘ h - - Betaine alone conversion
90% ‘ ‘ --- Choline alone yield
“~3— Betaine conversion
#— Choline yield

 

 

80%

 

70% 7 -7 —-

 

 

 

 

0 3O 60 120 180 240 300 360
11mo (mln)

Figure 120. Hydrogenation of 0.25 M betaine alone and 0.125 M glycine and 0.125 M

betaine (0.3 M H3PO4, H20, 130 °C, 1000 psi H2, 1 g dry 5% Ru/C)

4.8.2 H/D Exchange Studies

A competition experiment was run with 0.25 M glycine and 0.25 M sarcosine at 100 °C
with 1000 psi H2 and D20. These results were compared to the results from the H/D
exchange studies using the individual substrates. The experimental and fitted
concentrations of CH2, CHD, and CD2 over time as determined by the equations in
Chapter 3 are presented in Figures 121 (with respect to glycine alone) and 122 (with
respect to sarcosine alone). When looking at the rate of H/D exchange for glycine at the
methylene position, the rate is slightly suppressed in the competition experiment

compared with the experiment of glycine alone.

246

Figure 121. Fit of the kinetic rate data for 0.25 M glycine alone and 0.25 M glycine and

 

 

 

 

 

! 7— Gly comp AH2 —— Gly comp AHD — Gly comp AD2
‘ I Gly comp AH2_exp A Gly comp AHD_exp 0 Gly comp ADZ_exp
-—Gly alone AH2 —Gly alone AHD —G|y alone AD2
0-20 ‘ D Gly alone AH2_exp A Gly alone AHD_exp 0 Gly alone ADZ_exp '
g
o
E 0.15 7 7 7 77 7 7
v o
g Q
fill
E
g 0.10 l 7 ‘ 7
8 o
0.05 i 7 7 7 7 7
i
0.00 . , . . . f 7
0 50 100 150 200 250 300 350 400
Time (mln)

0.25 M sarcosine (D20, 100 °C, 1000 psi H2, 1 g dry 5% Ru/C)

Concentratlon (mol/L)

Figure 122. Fit of the kinetic rate data for 0.25 M sarcosine alone and 0.25 M glycine

 

 

 

 

0.30
77— Sarc comp AH2 ~77 Sara comp AHD — Sarc comp A02
I Sarc comp AH2_exp A Sarc comp AHD_exp o Sarc comp ADZ_exp
0.25 i— —Sarc alone AH2 --Sarc alone AHD —Sarc alone A02
[3 Sarc alone AH2_exp A Sarc alone AHD_exp o Sarc alone AD2_exp
0.20 4 77 7 7 7 7 7 7 7 7 7 7 7 7 7 7
0.157 7 777 777777 77
A o
0.10 77 7 7 7 7 e 7 7 7 7
o
0.05-7 7 7 7 77 77777777
A
0.00 . l . . , , .
0 50 100 150 200 250 300 350 400
Tlme (mln)

and 0.25 M sarcosine (D20, 100 °C, 1000 psi H2, 1 g dry 5% Ru/C)

247

The conversion after 4 hours of reaction time is 97% in the experiment of glycine alone
and 95% in the competition experiment. The yields of CHD and CD2 after 4 hours of
reaction time are 30% and 68% in the experiment of glycine alone and 35% and 59% in
the competition experiment. However, the rate of H/D exchange for sarcosine at the
methylene position appears to be much slower than in the experiment of sarcosine alone.
The conversion after 4 hours of reaction time is 98% in the experiment of sarcosine alone
and 94% in the competition experiment. The yields of CHD and CD2 after 4 hours of
reaction time are 21% and 75% in the experiment of sarcosine alone and 37% and 48% in

the competition experiment.

The rate constants for glycine alone, sarcosine alone, and in the competition experiment
are presented in Figure 123. The kinetics were determined using the models described in
Chapter 3. The rate constants for H/D exchange on the methylene carbon of glycine in
the competition experiment (and glycine alone) are: k1 A = 1.6x10'4 L/mol/min; (2.3x10'4),
k1,, = 4.8x10"5 L/mol/min; (1 .2x10'4), k2A = 8.3x10'5 L/mol/min; (1.2x10'4), k-2A =
1.6x10'4 L/mol/min; (3.2x10'4). The rate constants are about 30% larger in the case of

glycine alone for the exchange of both the first and second hydrogen.

248

2.5E-04

 

   
   

   

 
   
 

I Glycine alone
BGlycine and sarcosine competition _glycine
fl Sarcosine alone
2 0504 _ , Elisaircosine and glycine competflion_§arcosine ,
E
E .
'o'
E 1.5E-04 7 3g ;
:l ,. I.»
v a
8 {gig '_’
= 3
i3 {0 i
§ 1.05-04 «7 7
o » . xiii".
:00; “I: Q.
E 3: ' ,1
no ~.
"’ 7] o:
.. ..
5.05-05 7 g " 9%
3° :
z: 2:
$07,. '9
1.7 . it
0.0E+00 7 :3:

 

 

 

k1_A k2_A k1_C k2_C k3_C
Figure 123. Rate constants for PUB exchange of 0.25 M glycine alone, 0.25 M sarcosine

alone, 0.25 M glycine in the competition experiment, and 0.25 M sarcosine in the

competition experiment (D20, 100 °C, 1000 psi H2, 1 g dry 5% Ru/C)

The rates for H/D exchange on the methylene carbon of sarcosine in the competition
experiment (and sarcosine alone) are: klA = 1.5x10“1 L/mol/min; (1.6x104), k.1A =
7.8x10'5 L/mol/min; (1 .4x10'5), k2,, = 8.4x10'5 L/mol/min; (1.5x10'4), k2A = 2.3x10'4
L/mol/min; (2.5x104). The rate of exchange for the first hydrogen for sarcosine alone or
in the presence of glycine is the same. Although the rate of the second hydrogen being
replaced with deuterium is roughly almost twice as fast in the case of sarcosine alone as

compared to sarcosine plus glycine.

Again, using kinetic rate model A the rates for H/D exchange on the N-methyl carbon of

sarcosine in the competition experiment (and sarcosine alone) are: klc = 6.5x10'5

249

L/mol/min; (1.1x10'4), 1m = 4.9x10'6 L/mol/min; (0), kzc = 8.0x10'5 L/mol/min; (1 .4x10'
4), 1m = 3.2x10“ L/mol/min; (9.7x10'7), k3c = 6.6x10'5 L/mol/min; (8.3x10'5), k-3c =
1.1x10'3 L/mol/min; (2.0x10'4). These results again illustrate a slower rate of deuterium
exchange on the methyl hydrogens of sarcosine in the presence of an equimolar amount
of glycine. The rate of H/D exchange for both the first and second hydrogen are nearly
2x faster in the case of sarcosine alone. The rate of exchange for the third hydrogen is
comparable for both the competition experiment and sarcosine alone but still quicker for

sarcosine alone.

250

Chapter 5. SUMMARY
From the studies discussed it was confirmed that for hydrogenation of organic acids, the
presence of hydrogen and catalyst (5% Ru/C) is required. Addition of equimolar H3PO4
resulted in a negligible effect on the hydrogenation reaction of glycolic acid, suggesting
that phosphates, at least in protonated form, do not bind strongly to the catalyst. It was
assumed that the additional acid would not affect the hydrogenation of the other
substrates evaluated. However, from the competition experiments discussed in this paper
as well as from the work by Chen, Jere, Peereboom, and Pimparkar et al. we know that
the organic acid substrates have different binding affinities for the catalytic Ru particles
and for the carbon support. Earlier amino acid reduction models included terms for
H3PO4 competing with the same surface sites as the organic substrate. However, from
the carbon adsorption studies it was found that H3PO4 has less of an adsorption affinity
for the catalyst than the set of organic acids and alcohols examined, indicating that it is
not in direct competition with the substrate for the actual hydrogenation sites on the

catalyst.

For the H/D exchange control studies it was observed that non-catalytic H/D exchange
does occur in the methylene positions of glycine, sarcosine and betaine. For sarcosine,
about 30% of the methylene was monodeuterated after six hours at 130 °C. This result
can be compared to a similar experiment run with catalyst (5% Ru/C) at the slightly lower
temperature of 100 °C (both in 1000 psi H2), where the methylene is ca. 30%
monodeuterated after just one hour. Comparing rate constants, k1 A and k2A for the two

experiments, the control experiment had a klA = 9.8x10'6 and k2A = 0 L/mol/min (i.e. CH;

251

and CHD peaks accounted for all starting substrate), whereas the catalyzed but 30 °C
cooler reaction was ca. 15 times faster, k1 A = 1.6x10'4 and k2A = 1.5x10‘4 L/mol/min.
For betaine the measured concentrations of CH2 and CHD are nearly identical in the
experiment with or without catalyst (over 6 hours), showing little or no evidence of
catalyzed exchange. H/D exchange in the absence of catalyst was not observed with

glycine or choline bicarbonate (130 °C; 1000 psi H2; D20).

Results from adsorption studies using 5% Ru/C illustrated that adsorption onto the
catalyst is occurring and when the amount of catalyst is doubled, the amount of substrate
lost to adsorption is also doubled. However, doubling the catalyst amount doubles the
conversion rate for organic acid hydrogenations. Taken together, these facts indicate that
the reaction is limited by the intrinsic kinetics and not by interphase mass transport.
Peereboom et al. also confirmed and explored the adsorption of substrates on the catalyst
surface and in the micropores. Results indicated that the less hydrophilic compounds
have more of an affinity for the pores rather than the bulk solution and also the
concentration of a substrate in the bulk solution could be very different from its

concentration in the pores.

Comparing six substrates and their relative adsorption onto the carbon support, it was
found that lactic acid (7% adsorbed) was absorbed more than any other substrate. The
other substrates examined were butanol (5.7% adsorbed), glycolic acid (4.8% adsorbed),
propylene glycol (3.4% adsorbed), phosphoric acid (1.0% adsorbed), and ethylene glycol

(0.8% adsorbed). Experiments evaluating the effect of additional acid, H3PO4 have

252

shown no effect on the hydrogenation rate and this is partly due to the fact that it may
have less of an adsorption affinity for the catalyst and is more hydrophilic than for
instance, butanol. It may thus prefer the bulk solution over the micropores. The amount
adsorbed on the carbon support was similar to the amount of substrate typically "lost"
after being exposed to the catalyst (5% Ru/C) at time zero of the experiment.

The hydrogenation of betaine hydrochloride (CH3)3N+CH2COOH°C1' was compared to
the hydrogenation of betaine (CH3)3N+CH2COO' (both experiments conducted with
equimolar H3PO4). Results indicated that none of the desired alcohol product was formed
using betaine hydrochloride. An experiment with 0.25 M glycine (130 °C, 0.3 M H3PO4,
1000 psi H2) in both the presence and absence of 0.25 M HCl showed a conversion drop
(83% to 34%) of more than half for the experiment with HCI. The effect of an equivalent
amount of sodium chloride was also evaluated. The hydrogenation rate dropped from
83% to 72% indicating that again the chloride ion may be partially deactivating the
catalyst, affecting the substrate, or competing with the surface. However, in this
experiment half the amount of glycine, NaCl, and additional acid, H3PO4 was added
compared with the experiment with HCl (for the same amount of catalyst). So, the

reduction in the amount of total acids may have reduced the effect seen by the NaCl.

Additional insights arise from deuterium exchange studies comparing betaine
hydrochloride (CH3)3N+CH2COOH°CI' and betaine (CH3)3N+CH2COO' substrates. In the
former case, no deuterium was incorporated at either the methylene or methyl carbons.
From the control experiments in which no catalyst was added, it was found that the same

amount of deuterium incorporation was occurring in betaine in both in the presence and

253

absence of catalyst. This suggests that the lack of H/D exchange seen in the case of
betaine hydrochloride is not due to a partially deactivated catalyst or competition by HCl
for catalyst surface sites. The chloride ion in the solution appears to somehow inhibit

methylene C-H activation in the bulk solution.

It was already noted that the presence of an equimolar amount of additional acid, H3PO4,
does not reduce the hydrogenation rate of glycolic acid. The need for the additional acid
is to ensure that the carboxylic acid functionality in glycine and the N-methylated
glycines is protonated. These amino acids are zwitterionic in nature and thus the
carboxylic acid functionality is dissociated unless the additional acid is present. For
glycine, a yield of 15% is obtained in the absence of catalyst, compared with 83%
conversion with additional acid. To determine if the additional acid slows the H/D
exchange reaction, betaine was initially chosen due to its slower competing
hydrogenation reaction. Results showed that both with and without the additional acid
present the amount of H/D exchange on the methylene carbons were essentially the same.
However, betaine undergoes non-catalytic H/D exchange. The k; A values were 0.41x104
and 0.43x10'4 for the case of betaine (130 °C) and betaine with 0.3 M H3PO4 (100 °C),
respectively. Presumably the rates should be different for the different temperatures, but
at 100 °C with acid, some of the betaine is hydrogenated to form choline. After six hours
a 38% yield of choline was detected by HPLC; thus, included in the H/D rates (methylene
carbon) are both the exchange and also the conversion to choline. The yield of choline in

this experiment is quite a bit higher than in the same conditions but at 130 °C. The

254

reason for this is that at higher temperatures more betaine degradation (presumably to

trimethylamine and gases) is occurring.

Glycine and sarcosine were also run at 50 and 70 °C in the absence and presence of 0.3
M H3PO4 to evaluate the extent of HfD exchange. The rate of exchange was significantly
decreased in the presence of H3PO4; presumably protonation of the amine makes the
neighboring C-H bond less electron-rich and therefore less activated for C-H attack by

Ru.

Various solvents were examined to explore their effects and the role of the water solvent
in the hydrogenation reactions. The hydrogenation rate of glycolic acid was significantly
depressed when using either dry THF or tert-butanol. Studies with THF indicate that it
readily undergoes H/D exchange at both C1 (83% H replacement) and C2 (42% H
replacement) positions over six hours. The fact that the H/D exchange reaction of THF
slows glycolic acid hydrogenation supports the claim that the hydrogenation sites and
H/D exchange sites are not distinct. Next, tert-butanol was tried based on the hypothesis
that it would not undergo exchange or interact with the surface. Indeed, no exchange was
observed in this solvent with or without added substrate, but neither was glycolic acid
hydrogenated. Some loss of tert—butanol was seen, a finding likely due to dehydration of

the solvent to isobutene and subsequent hydrogenation to isobutane.

The solvent ethanol was used for the experiments with glycolic acid both with and

without additional acid, H3PO4. As expected, the acid promoted a significant amount of

255

esterification, but even in the case with no additional acid, the ester ethyl glycolate was
still formed. In both experiments, only 8% of ethylene glycol was formed compared with
47% in the solvent water under otherwise identical conditions. The literature suggests
that esters undergo hydrogenation with certain catalysts (Raney Ni, copper/chromium
oxide, rhenium black) but that there is competition for the surface between the acids and
the esters, reducing the amount of ester hydrogenation. But, in this case, slow or no ester
hydrogenation occurs over 5% Ru/C . To examine this further, ethyl lactate was
subjected to the usual aqueous hydrogenation conditions. It quickly hydrolyzed to lactic
acid (100% at one hour). The product propylene glycol, was not observed until after
lactic acid formed, again supporting the claim that the ester reacts slowly or not at all.
This experiment’s propylene glycol yield was nearly the same as with lactic acid itself,
suggesting that in this case the lactic acid hydrogenation suffers little hindrance from the

ester or dilute ethanol.

The results from the organic acid hydrogenation studies indicate that simple carboxylic
acids such as propanoic acid and acetic acid are substantially more difficult to reduce
than lactic acid and alanine. The various substrates were chosen in order to evaluate the
effect of hydrogen bonding, electron withdrawing ability, and heteroatom coordination to
the surface. Hydrogen bonding was determined to be a key player in accelerating the
hydrogenation reaction. Reactions with the (Jr-substituent -N+H3, -OH, -OCH3, -
N+H2(CH3), and -N+H(CH3)2 were far more reactive than either the unsubstituted acetic
acid or an a-substituent of -CH3 or -N+(CH3)3. Sterics were considered in the N-

methylated glycine series (-N+H3, -N+H2(CH3), -N+H(CH3)2, and -N+(CH3)3) in that the

256

hydrogenation rate falls off by about 1/3 as each N-methyl group is added. However,
there is a trend for hydrogenation in that the reactivity falls off by nearly half as a methyl
is added. Also, when comparing the groups, -OH to a more bulky -OCH3 or -CH3, the
substrates with the or-OH group (glycolic acid and lactic'acid) are far more reactive. This
may be partly due to a less bulky group but also due to electronics (-OH is more electron-
withdrawing than —OCH3) as compared to -OCH3 or 7CH3. Considering the role of
electronics, it was found that the substrate with oc-OCH3 is much more reactive than on-

CH3 (yield of desired product is 35% compared with 13%).

In an effort to build upon the above research and probe the reactivity differences in
hydrogenation of the methylated glycine series, H/D exchange studies were performed.
Reactions run in D20 with either H2 or D2 and catalyst were effective at exchanging
deuterium into the CH2 and CH3 sites of amino acids. The Ru catalyst also exchanges H
(from H2) for D (from D20) to form HOD and HD, independent of substrate. Analysis by
internal standard-calibrated 1H NMR allowed complete isotopomer-specific inventory of
the H concentration in substrates and water. Simple sequential replacements of H by D
were observed in the amino acids. Two kinetic models were developed to determine the
rate constants for PUB exchange at both the methylene and methyl positions. Both

included possible time delays due gas/liquid isotopic equilibration (kH) and/or a catalyst
induction (ktum_on) period; ultimately the latter was discarded as redundant. The first

model treated each isotope replacement step as a process unrelated to the others. Thus,
between forward and reverse replacements in CH2 and CH3 sites and k“, eleven rate

constants were freely varied in the fitting. The second model used a single rate constant

257

for each site, further modified by primary and secondary isotope effect values, resulting

in six values + kH for this fitting. The latter, more economical and interpretive model

gave fits nearly as good as those from the eleven-variable one.

The H/D exchange took place at lower temperatures (SO-100 °C) than hydrogenation

(IOO-150 °C) and therefore it was possible to study H/D exchange while minimizing the

amount of competing hydrogenation reaction. Mapping out the optimal temperatures for

both reactions allowed H/D exchange to be studied. Glycine, sarcosine, and N,N-

dimethylglycine were run at 100 °C with 1000 psi H2 or D2 and the datasets from the

pairs of experiments were linked in order to define a common set of rate constants.

Experiments were also run at lower temperatures (70 and 50 °C). The rate constants for

H/D exchange at the methylene and methyl carbons are smnmarized in Table 10.

Table 10. Summary of rate constants for glycine, sarcosine, and N,N-dimethylglycine

 

 

 

 

 

 

 

 

 

 

Rate Glycine Sarcosine N,N-dmg Glycine Sarcosine N,N-dmg Glycine Sarcosine
constant 100 °C 100 °C 100 °C 70 °C 70 °C 70 °C 50 °C 50 °C
(L/mol/min) Dz Dz Dz Dz Dz Dz Dz Dz
ki_CHz 0.99 1.48 0.74 0.74 0.99 0.06 0.09 0.25
ki_CHa 0.58 0.66 0.24 0.09 0.05

 

 

 

All rate constants are multiplied by 104

Somewhat surprisingly, H/D exchange rates decreased in the following order: sarcosine

(CH3NH2+CH2CO2') > glycine (CH2 site) > N,N-dimethylglycine. This trend in reactivity

does not match that observed in the hydrogenation of these same substrates (glycine >

sarcosine > N,N-dimethyglycine. This may be due to the fact that for H/D exchange an

258

electron-rich C-H bond promotes the reaction, whereas in the hydrogenation reaction, it is

presumed that steric effects play more of a role.

The I-I/D exchanges on the N-methyl(s) of both sarcosine and N,N-dimethylglycine were
slower than on the methylene carbon (for all conditions) and in general did not fully
reach equilibrium after six hours (Table 10) at 100 °C. In addition it was found that the
H/D exchange occurs step-wise at both the methylene and methyl carbons (CH2 to CHD
to CD2) and (CH3 to CH2D to CHD2 to CD3). Betaine was found to undergo non-
catalytic H/D exchange at the methylene carbon. Glycine and sarcosine exhibited a small
amount of H/D exchange in the absence of catalyst but only at higher temperatures

(130 °C). With its fixed N+ next door, the acidity of betaine’s a-CH2 is enhanced,

increasing the rate of deprotonation, enol formation, and the resulting H/D exchange.

In summary, a detailed set of hydrogenation and H/D exchange experiments using a 5%
ruthenium on carbon catalyst were conducted for several two- and three-carbon organic
acid substrates and a kinetic model was developed for the H/D exchange. Simple
alkanoic acids (acetic, propanoic, isobutyric) and betaine, quatemized glycine, are not
very reactive. For the other methylated glycines, the rate of hydrogenation falls off by a
factor of ca. three with each additional methyl, and reduction requires acid, as found
earlier by Jere. Use of water as a reaction medium appears optimal. The H/D exchange
studies indicated a different trend in reactivity for the methylated glycines (sarcosine >
glycine > N,N-dimethylglycine) as compared to hydrogenation. The rate of H/D

exchange appears to be promoted by an electron-rich C-H bond. The addition of H3PO4

259

significantly slows the H/D exchange while enabling the reduction, suggesting that the

free amine is needed for the hydrogen replacement reaction.

Future projects could include additional probes of the roles of steric effects, pH more
precisely defined, additional catalyst site inhibitors, and stereoelectronic aspects of
control. For instance, to explore whether the electrons on the amine are coordinating with

the catalyst surface, an experiment with a t-butyl-locked piperidine (chair conformer)

with various N-alkyl groups could be run to determine the preference for Hax versus Heq.

The results would help to map out the elements controlling C-H activation alpha to amine

sites.

260

Chapter 6. REFERENCES

 

l Braun, B.; Charney, R.; Clarens, A; F arrugia, J .; Kitchens, C.; Lisowski, C.; Naistat, D.;
O’Neil, A. J. Chem. Ed. J. Chem. Ed. 2006, 83, 1126

2 British Petroleum, BP Statistical Review of World Energy, 2007

3 Yang, M; Somorjai, G.A. J. Phys. Chem. B. 2004, 108, 4405

4 Jere, F. Ph.D. Dissertation, Michigan State University, 2003

5 Antons, S. In US Patent 5, 731,479; Bayer Aktiengesellschaft: USA, 1998
6 Antons, S. In US Patent; Bayer Aktiengesellschaft: USA, 2002

7 Carnahan, J .E.; Ford, T.A.; Gresham, W.F.; Grigsby, W.E.; Hager, G.F. J. Am. Chem.
Soc. 1955, 77, 3766

8 Behr, A; Brehme, v. Adv. Synth. Catal. 2002, 344, 525

9 Mao, 13w; Cai, zz; Huang, M.Y.; Jiang, Y.Y. Polym. Adv. Technol. 2003, 14, 278
'0 He, D.I—I.; Wakasa, N.; Fuchikami, T. Tet. Lett. 1995, 36, 1059

“ Bowden, H; Adkins, H. J. Am. Chem. Soc. 1934,56, 689

‘2 Adkins, H; Folkers, K. J. Am. Chem. Soc. 1932, 54, 1145—1154

'3 Adkins, H; Cramer, H.I. J. Am. Chem. Soc. 1930, 52, 4349

‘4 Adkins, H.; Pavlic, A.A. J. Am. Chem. Soc. 1947, 69, 3039

'5 Adkins, H; Billica, H.R. J Am. Chem. Soc. 1948, 70, 3121

'6 Vaidya, P.D.; Mahajani, v.v. Chem. Eng. Sci. 2005,60, 1881

'7 Sajiki, H.; Kurita, T.; Esaki, H; Aoki, F.; Maegawa, T; Hirota, K. Org. Lett. 2004, 6,
3521

'8 (a) Faller, J.W.; Smart, C.J. Organometallics 1989, 8, 602. (b) Heys, R. J. Chem. Soc.,
Chem. Commun. 1992, 680. (c) Heys, J .R.; Shu, A.Y.L.; Senderoff, S.G.; Phillips,
N.M. J. Labelled Compd Radiopharm. 1993, 33, 431. (d) Lukey, C.A.; Long, M.A.;
Gamett, J .L. Aust. JChem. 1995, 48, 79. (e) Hesk, D.; Das, P.R.; Evans, B. J. Labelled
Compd Radiopharm. 1995, 36, 497. (1) Chen, W.; Games, K.T.; Levinson, H.H.;
Saunders, D.; Senderoff, S.G.; Shu, A.Y.L.; Villani, A.J.; Heys, JR. J. Labelled Compd

261

 

Radiopharm. 1997, 39, 291. (g) Klei, S.R.; Golden, J .T.; Tilley, T.D.; Bergman, R.G.
J.Am. Chem. Soc. 2002, 124, 2092. (h) Skaddan, M.B.; Yung, C.M.; Bergman, R.G.
Org. Lett. 2004, 6, ll

‘9 (a) Hesk, D.; Jones, J.R.; Lockley, W.J.S. J. Labelled Compd. Radiopharm. 1990, 28,
1427. (b) Hesk, D.; Jones, J.R.; Lockley, W.J.S. J. Pharm. Sci. 1991, 80, 887. (c)
Lenges, C.P.; White, P.S.; Brookhart, M. J. Am. Chem. Soc. 1999, 121, 4385

20 (a) Weil, T.A.; Friedman, S.; Wender, I. J. Org. Chem. 1974, 39, 48. (b) Lenges, C.P.;
Brookhart, M.; Grant, B.E. J. Organomet. Chem. 1997, 528, 199

2‘ (a) Garnett, J.L.; Hodges, R.J. J. Am. Chem. 1967, 89, 4546. (b) Gol’dshleger, N.F.;
Tyabin, M.B.; Shilov, A.E.; Shteinman, A. A. Zh. Fiz. Khim. 1969, 43, 1222. (c)
Buncel, E.; Clement, O. J. Chem. Soc., Perkin Trans. 2 1995, 1333

22 (a) Ogasawara, M.; Saburi, M. Organometallics 1994, 13, 1911. (b) Collman, J .P.;
Fish, H.T.; Wagenknecht, P.S.; Tyvoll, D.A.; Chng, L.-L.; Eberspacher, T.A.;
Brauman, J .I.; Bacon, J .W.; Pignolet, L.H. Inorg. Chem. 1996, 35, 6746

23 Perthuisot, C; Fan, M; Jones, W.D. Organometallics 1992, 11, 3622

24 Sajiki, H.; Aoki, F.; Esaki, H.; Maegawa, T.; Hirota, K. Org. Lett., 2004, 6, 1485

25 Jere, F .T.; Miller, D.J.; Jackson, J .E. Org. Lett. 2003, 5, 527

26 Haikides, C.J.; Frey, P.A.; Tobin, 1.3. J. Am. Chem. Soc. 1993, 115, 3332

27 Nibeilliu, M.B.; Malthouse, P.G. Biochem. J. 2004, 381, 847

28 Rios, A.; Amyes, T.L.; Richard, J .P. J. Am. Chem. Soc. 2000, 122, 9373

29 Rios, A; Richard, J.P. J. Am. Chem. Soc. 1997, 119, 8375

3" Cox, HA; Julian, R.R.; Lee, s.w; Beauchamp, J.L. J. Am. Chem. Soc. 2004, 126,
6485

31 Matsubara, S. Yokota, Y.; Oshima, K. Chem. Lett., 2004, 33, 294
32 Li, 2.; Li, C.J. J. Am. Chem. Soc. 2004, 126, 11810
33 Li, z; Li, C.J. Org. Lett. 2004, 6, 4997

34 Li, z; Li, C.J. J. Am. Chem. Soc. 2005, 127, 3672

262

 

 

 

35 Alston 11, W.C.; Haley, K.; Kanski, R; Murray, C.J.; Pranata, J. J. Am. Chem. Soc.
1996,118,6562

36 Broadbent, H.S., Campbell, G.C., Bartley, W.J., Johnson, J .H. J Org. Chem. 1959, 24,
1847-1854

37 Antons, s; Beitzke, B. Chem. Abstr. 1996, 125, 114175

38 Studer, M.; Burkhardt, s; Blaser, H.U. Adv. Synth. Catal. 2001, 343, 802

39 Ruiz, A. Guerrero; Alarcon, M. Cerro; Valiente, A. Maroto; Ramos, I. Rodriguez
Carbon 2005, 43, 2711

40 Chemical Marketing Reporter, 1091-0110, 260 (11) (September 24, 2001) 30
4‘ Chemical Marketing Reporter, 1092-0110, 265 (17) (April 26, 2004) 20 (4)

42 Dasari, M.A.; Kiatsimkul, P.P.; Sutterlin, W.R.; Suppes, G.J. Appl. Catal. A: Gen.
2005,281,225

43 Muller, P.; Rimmelin, P.; Hindermans, J .P.; Kieffer, R.; Kiennemann, A.
Transformation of Sugar into Glycols on a 5% Ru/C Catalyst. Stud. Surf Sci. Catal.
1991, 59, 237

44 Lambert, J .B., Mazzola, E.P. Nuclear Magnetic Resonance Spectroscopy: An
Introduction to Principles, Applications, and Experimental Methods

45 Lambert, J.B., Greifenstein, L.G. J. Am. Chem. Soc. 1974, 96, 5120

46 Gutowsky, HS. J. Chem. Phys. 1959, 31, 1683f

47 Pople, J.A. M01. Phys. 1963, 7, 201

48 Zhang, Z., Jackson, J .E., Miller, D.J. Appl. Catal. A 2001, 89-98

49 Zhang, Z., Jackson, J .E., Miller, D]. Ind. Eng. Chem. Res. 2002, 41, 691-696

5" Jere, F.T., Jackson, J.E., Miller, D.J Ind. Eng. Chem. Res. 2004, 43, 3297-3303

5‘ Singh, U.K., Vannice, M.A. Appl. Catal. A 2001, 213, 1-24

52 Pallassana, v; Neurock, M. J. Catal. 2002, 209, 289-305

53 Kovacs, D.G., Jackson, J.E., Miller, D.J. Abstr Pap Am Chem. Soc. 2001, 221 U177-8

5“ Lahr, D.G., Shanks, B.H. J. Catal. 2005, 232, 386-394

263

 

55 Lahr, D.G., Shanks, B.H. Ind. Eng. Chem. Res. 2003,42, 5467-5472

56 Kusunoki, Y., Miyazawa, T., Kunimori, K., Tomishige K Catal. Commun. 2005, 6,
645-649

57 Chaminand, J ., Djakovitch, L., Gallezot, P., Marion, P., Pinel, C., Rosier, C. Green
Chem. 2004, 6, 359-361

5“ Perosa, A., Tundo, P. Ind. Eng. Chem. Res. 2005, 44, 8535-8537

59 Peereboom, L., Koenigsknecht, 8., Hunter, M. Jackson, J .E., Miller, D.J. Carbon 2007,
45, 579-586

60 Dalavoy, T.S., Jackson, J .E., Swain, G.M., Miller, D.J., Li, J ., Lipkowski, J. J. Catal.
2007, 246, 15-28

6' Chen, Y., Miller, D.J., Jackson, J .E. Ind. Eng. Chem. Res. 2007, 46, 3334-3340

62 Zhang, Z., Jackson, J .E., Miller, D]. Bioresource Technology 2008, 5873-5880

63 Pimparkar, K.P., Miller, D.J., Jackson, J .E. Ind. Eng. Chem. Res. 2008, 47, 7648-7653
6“ Knifton, J.F., Sanderson, JR, and Stockton, ME. Catal. Lett. 2001, 73, 55

65 Abella, L.C., Gaspillo, P.D., Itoh, H., and Goto, S.J. Chem. Eng. Jpn. 1999,32, 742

66 WE. Smith. Tetrahedron 2002, 58, 2091

264