It.‘

any

 

 

U. ..
a- I.“
$5

A!

v. .
5%.}! I:
, In “.289. I),
.11!
in”. an.

1a,. in?! I
s I (1.2
.l’) 1,3!” 50.7 1.!
I5”) huh-u r41 .1

at I...
. b bf
b . A.

a K.

.t
it!

,!

 

is
«d.
s!!!) it)!!! (a
{$.15 . an
917% filtuuua.’

r...
«‘09..

{\\ pt.

 

91A..."
I .3:
1.. .1. l1

 

.5

“$5.?9
$1.;
. . arsnfifl

..
In ‘ a . . 33.1. v‘
an. abavfimvffiarm; fin .
, . x 1

0’? v.
y... 5!", a
{3.14 {1“}:
$3.... slung»;
\ is: us

37:!
m7!»

1!. §.:.l

Lil .-

125! A 3.3.2...
is; a:
.93 35:1...

1.} .l

5‘. If“.
I It“
rail/‘11»...

.r...5u‘ls.a{t.. .

1 .’..1. :ill

Unit). 1!».
.33 9) re...

(.11..

_ Mats L'BHARY
m7 Michigan State
University

This is to certify that the
thesis entitled

CATALYTIC HYDROGENATION OF 2,5-PIPERAZINEDIONE TO
PIPERAZINE

presented by
Andrea Sue Molengrafl

has been accepted towards fulfillment .
of the requirements for the a

 

 

 

 

MS. degree in Chemistry
/ i Mi Major I'Drofes’sor’s Signature
28 P; 9m! 2006 3
Date :

MSU is an Affinnative Action/Equal Opportunity Institution

 

 

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

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2/05 p:/ClRC/DateDue.indd-p.1

 

CATALYTIC HYDROGENATION OF 2,5-PIPERAZINEDIONE TO PIPERAZINE

By

Andrea Sue Molengrafi

A THESIS
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
MASTER OF SCIENCE
Department of Chemistry

2006

ABSTRACT
CATALYTIC HYDROGENATION OF 2,5-PIPERAZINEDIONE TO PIPERAZIN E
By
Andrea Sue Molengraft

Catalytic hydrogenation of nearly all the carbonyl functional groups is well
established and widely practiced with one exception: the amide moiety. Stable,
ubiquitous in nature, and easily synthesized, amides’ value as chemical intermediates
would increase if a general, practical catalytic hydrogenation which preserved the C-N
bond were available. As an underutilized biomass fraction, proteinaceous substrates are a
particular interesting class of polyamides to explore; the resulting amine products could
have a wide variety of potential applications as pH adjusters, surfactants, detergent
additives, metal ion scavengers, or chiral building blocks for pharmaceutical synthesis.
Water, the native environment of peptidic materials, presents any hydrogenation must
compete with C-N cleavage via hydrolysis. We aimed to develop ‘green’ selective
heterogeneous aqueous-phase catalytic amide reduction. The cyclic anhydride of glycine,
2,5-piperazinedione, was used as a model peptidic amide substrate. The reaction was
optimized using a systematic experimental design, the Box-Behnken model. At optimal
conditions (1800 psi H2, 90°C, 3 g 5% Ru/C), 2,5-piperazinedione was reduced to 37%
piperazine at 24 hrs. Glycylglycine, glycine. and ethanolamine formed in 10, 1 and 2%
yield, respectively. Development of this ‘green’ process, and the formation of
polyamido and polyamino alcohol polymers from polypeptides opens the door to
chemical upgrading of new renewables-based feedstock fraction, the proteinaceous

component, for economical production of value-added products.

To Mom and Dad

iii

ACKNOWLEDGMENTS

I would like to express my gratitude to my major professor, Dr. James Jackson,
for his knowledge and support throughout my studies at Michigan State University.

I am also grateful to my committee members for their prolonged support. I would -
like to acknowledge Dr. Dennis Miller for all his support and enthusiasm for this project,
especially for his guidance with the project, use of the equipment in his laboratory, and
help with data analyses and interpretation. Thank you to Drs. Baker and Walker for their
support and interest of this project and serving as committee members. I also want to
express my thanks to Drs. Simona Marincean and Mikhail Redko for their guidance in
the laboratory as well. I would also like to thank past and present colleagues in my

laboratory, for all their support in the lab.

iv

TABLE OF CONTENTS

 

 

 

 

 

 

 

LIST OF FIGURES vii
LIST OF TABLES - x
INTRODUCTION - - - 1
CHAPTER ONE LITERATURE REVIEW - _ _ - - 4
1.1 Carboxamide Group ...................................................................................... 4

1.2 Amide Reactivity .......................................................................................... 5

1.2.1 Classical amide reduction using complex metal hydrides .................... 7

1.2.1.1 LiA1H4 ............................................................................................. 8

1.2.1.2 BH3 ................................................................................................ 10

1.2.1.3 NaBH4 ............... 14

1.2.2 Catalytic Hydrogenation ..................................................................... 24

1.2.2.1 Mechanisms for Catalytic Hydrogenation .................................... 30

1.2.2.2 Bimetallic Catalyst Systems ......................................................... 32

1.3 Relevant Laboratory Studies ....................................................................... 36

1.3.1 Catalytic Hydrogenation of Amino Acids ........................................... 36

1.3.1.1 L-Alanine ...................................................................................... 36

1.3.1.2 Glycylglycine ................................................................................ 37

1.3. 1.2.1 Hydrolysis of Glycylglycine .................................................. 40

CHAPTER TWO ANALYTICAL METHOD DEVELOPMENT _ 42
2.1 Introduction .......... 42
2.2 Experimental Procedures ............................................................................................ 44
2.2.1 Materials .................................................................................................. 44

2.2.2 Methods ................................................................................................... 45

2.3 Results & Discussion .................................................................................................. 46
CHAPTER THREE - 50
CATALYTIC HYDROGENATION OF AMIDES TO AMINES -- 50
3.1 Introduction ................................................................................................................. 50
3.2 Experimental Procedures ......................................... A ................................................... 55
3.2.1 Materials .................................................................................................. 55

3.2.2 Methods ................................................................................................... 56

3.3 Results & Discussion .................................................................................................. 58
3.3.1 ,Box-Behnken Model ................................................................................ 58

3.3.1.1 Effect of Temperature .................................................................. 59

3.3.1.2 Effect of Pressure .......................................................................... 6O

3.3.1.3 Effect of Catalyst Loading ........................................................... 60

 

 

 

 

 

3.3.2 Beyond the Boundaries of Box-Behnken Model .................................... 73

3.3.3 Control Experiments ................................................................................ 79

3.3.4 Reaction pathways for 2,5-piperazinedione hydrogenation to piperazine
....................................................................................................................................... 85
CHAPTER FOUR 92
SUMMARY AND CONCLUSIONS 92
CHAPTER FIVE 93
FUTURE RESEARCH ...... 93
LIST OF REFERENCES _ 95

vi

LIST OF FIGURES

Figure 1.1 Resonance structures of amide ......................................................................... 4
Figure 1.2 Mechanism of amide reduction with a metal (M) hydride ............................... 5
Figure 1.3 Possible hydrogenolysis sites ........................................................................... 6
Figure 1.4 Hydrogenolysis of an amide forming an amine and water ............................... 6
Figure 1.5 Hydrogenolysis of the carbon-nitrogen bond of amides .................................. 6
Figure 1.6 Hydrolysis of an amide to a carboxylic acid and an amine under aqueous
conditions ................................................................................................................ 7
Figure 1.7 Secondary product formation from amide aminolysis or amine alkylation ..... 7
Figure 1.8 Reduction of triisopropylamide and benzamide to a primary amine via a
nitrile intermediate .................................................................................................. 9
Figure 1.9 Proposed mechanism of amide reduction to a nitrile with LiAlH4 .................. 9
Figure 1.10 Reduction of N,N-diethylpivalamide and N,N-diisopropylbenzamide to
their corresponding amines with BH3/T HF (0-25°C) ........................................... 10
Figure 1.11 Amine yields from amide reduction with BH3fT HF (0-25°C) ..................... 11
Figure 1.12 Relative rate of reduction of a substituted N,N-dimethylamides with
BH3/T HF (25°C) ................................................................................................... 1 1
Figure 1.13 Rate of reduction of N substituted N,N-dialkylbutyramides with BH3/T HF at
25°C ....................................................................................................................... 12
Figure 1.14 Rates of reduction of substituted N,N-dimethylbenzamides with BH3/T HF at
25°C ....................................................................................................................... 12
Figure 1.15 Representative selective amide reduction in the presence of an ester
(BH3/T HF (0-25°C)) ............................................................................................. 13
Figure 1.16 Reduction of methyl hippurate with BH3/T HF (0- 25°C) .............................. 14
Figure 1.17 Proposed intermediate of the corrdinate covalent bond of an ester oxygen
with boron ............................................................................................................. 14
Figure 1.18 Amine yield from the reduction of amides with NaBH3(OCOR) in THF or
dioxane ......................................................................... ._. ....................................... 16
Figure 1.19 Thioamide reduction to an amine with Et3O+BF4 /CH2C12 and NaBHJMeOH
............................................................................................................................... 16
Figure 1.20 Amine yields from the reduction of its thioamide with Et3O+BF4 /CH2C12
and NaBH4/MeOH ................................................................................................ 17
Figure 1.21 Reduction of an amide to an amine via direct formation of an imino ether
fluoborate intermediate ......................................................................................... 18
Figure 1.22 Mechanism of reduction of imino ether fluoborate intermediate to an amine
with BH4' ............................................................................................................... 18
Figure 1.23 Reaction sequence of an amide or lactam reduction to an amine via an imino
derivative ............................................................................................................... 19
Figure 1.24 Selective reduction of amides over olefin, nitrile and esters groups with
l)POC13/glyme and 2)NaBH4/EtOH ..................................................................... 19
Figure 1.25 Amine yields from their corresponding amide reduced With POC13/glyme
and NaBH4/EtOH ....................................................................... V ........................... 20
Figure 1.26 Amide reduction to an amine via a Vilsmeier intermediate ........................ 21
Figure 1.27 Lactam reduction to an amine with NaA1H2(OCH2CH20CH3)2 in benzene
(100°C) .................................................................................................................. 22

vii

Figure 1.28 Amides reduced with NaBH4/C0C12 in dioxane or methanol (100°C) ......... 22

Figure 1.29 Proposed mechanism of amide reduction with NaBH4/g1yme ..................... 23
Figure 1.30 Amides reduced using Na metal/n-propanol ................................................ 24
Figure 1.31 First reported hydrogenation of an amide by Mailhe ................................... 24
Figure 1.32 Hydrogenation of benzamide over Ni (1500 psi H2, 200°C) ........................ 25
Figure 1.33 Amides reduced over 2CuO-Cr203/dioxane (1470-4400 psi H2, 175-250°C)
............................................................................................................................... 26
Figure 1.34 Ester reduction over 2CuO-Cr2O3/dioxane (1470-4400 psi H2, ZOO-260°C) in -
the presence of an amide ....................................................................................... 28
Figure 1.35 Lactam hydrogenation over 2CuO-Cr2O3/dioxane (1470-4400 psi H2, 200-
260°C) ................................................................................................................... 29
Figure 1.36 Hydrogenation of quinolones over 2CuO-Cr203/dioxane (1470-4400 psi H2,
ZOO-260°C) ............................................................................................................ 29
Figure 1.37 Transalkylation of an amine with alcohol over 2CuO-Cr2O3/H2 .................. 30
Figure 1.38 Formation of ether and ammonia via an amino alcohol over 2CuO-Cr2OJH2
............................................................................................................................... 30
Figure 1.39 Side reaction forming a secondary amine via an imine intermediate over
2CDO‘CI‘203/I‘IZ ..................................................................................................... 31
Figure 1.40 Secondary amine formation via nitrile over 2CuO-Cr2O3/H2 ....................... 31
Figure 1.41 Nitn'le hydrogenation over Ni (R)/NH3 (125°C) to an amine ....................... 32
Figure 1.42 Transalkylation of tertiary amides with alcohols to form amines over copper-
chromium barium oxide (I)/dioxane (4500-12000 psi H2, 250°C) ........................ 32
Figure 1.43 Proposed catalytic cycle of an amide reacting with dialkylsilane ................ 35
Figure 1.44 Reaction pathway of L-alanine to L-alaninol hydrogenation ..................... 36
Figure 1.45 Catalytic hydrogenation of glycylglycine with H3PO4/H2O (1000 psi H2
100°C, 1 g 5% Ru/C) ............................................................................................ 37
Figure 1.46 Catalytic hydrogenation (1000 psi H2, 100°C, 5% Ru/C)ofglycy1g1ycine
(ZnHPO4/H2O) and glycinamide (Ca(H2PO4)2/H2O) ........................................... 38
Figure 1.47 Catalytic hydrogenation of 2,5-piperazinedione with H3PO4/H2O (1000 psi
H2, 100°C, 1 g 5% Ru/C) ...................................................................................... 39
Figure 1.48 Equilibrium between 2,5-piperazinedione and glycylglycine at 25°C ......... 40
Figure 2.1 Amino acids and corresponding amines of interest ........................................ 43
Figure 2. 2 Sample chromatogram of amides, corresponding amines, and aminoalcohols
.............................................................................................................................. 47
Figure 2. 3 Sample Chromatogram of amino acid separation .......................................... 49
Figure 3.1 Catalytic hydrogenation of glycylglycine and 2 ,S-piperazinedioine with
H3PO4/H2O (1000psi H2, 100°C, 1 g 5% Ru/C) ................................................... 51
Figure 3.2 Possible reaction pathways of 2,5-piperazinedione ............................ ' ........... 52
Figure 3.3 Box-Behnken Model ..................................................................................... 54
Figure 3.4 Regression analysis of predicted %yie1d versus experimental % yield ........ 59
Figure 3.5 The effect of temperature (90, 110 and 130°C), on yield (A) and conversion
(B) of piperazine from 2,5-piperazinedione at 1500 psi H2, 3 g 5% Ru/C ........... 63
Figure 3.6 The effeCt of pressure (500, 1500, and 1800 psi H2) on % yield of piperazine
from 2,5-piperazinedione at 110°C, 3g 5% Ru/C ................................................. 65
Figure 3.7 Product yields from the conversion of 2,5-piperazinedione at (A) 1000 (B)
1500 and (C) 1800 psi H2, 90°C, 3g 5% Ru/C ...................................................... 67 .

viii

Figure 3.8 The effect of catalyst loading (A) 0.3 g, (B) l g, (C) 3 g 5% Ru/C on all

product yields at 1500 psi H2, 130°C .................................................................... 69
Figure 3.9 The effect increased catalyst loading on product yields on from (A) 3 g 5%
Ru/C to (B) 6 g 5% Ru/C at 1500 psi H2, 110°C .................................................. 71
Figure 3.10 Representative chromatogram series of 2 ,5-piperazinedione hydrogenation
to piperazine at 1800 psi H2, 90°C, 3 g 5% Ru/C ................................................. 76
Figure 3.11 Effect of temperature on % yield at 1500 psi H2, 3 g 5% Ru/C .................. 77
Figure 3.12 Effect of temperature on % yield at 1800 psi H2, 3 g 5% Ru/C .................. 78
Figure 3.13 Product formation from control experiments ............................................... 80
Figure 3.14 Change of pH over time from the conversion of 2,5-piperazinedione to
piperazine at 1800 psi, 90°C, 3g 5% Ru/C ............................................................ 81
Figure 3.15 The effect of catalyst loading on initial % conversion of 2,5-piperazinedione
at 1500 psi H2, 130°C, with 0.3, 1, and 3 g 5% Ru/C ........................................... 83
Figure 3.16 The effect of catalyst loading on initial % conversion of 2,5-piperazinedione
(0.1 and 0.05M) at 1500 psi H2, 110°C, with 3 and 6 g 5% Ru/C ........................ 83
Figure 3.17 Modeling of proposed mechanism for amide hydrogenation in gas phase
and in solution ....................................................................................................... 89
Figure 3.18 Predicted reaction pathways for hydrogenation of 2,5-piperazinedione ...... 90
Figure 3.19 Predicted reaction pathways for hydrolysis of 2,5-piperazinedione ............ 91

ix

LIST OF TABLES

Table 3.1 Percent yield of all products for the reduction of 2,5-piperazinedione at 6 hrs
- ............................................................................................................................... 62
Table 3.2 Percent yield of all products for the reduction of 2,5-piperazinedione at 24 hrs.
............................................................................................................................... 75

INTRODUCTION .

The amide functionality in organic chemistry is one of the more intriguing
functional groups to work with, and must first be understood in order to discuss its
reductions. This functional group’s polarity, stability, and planar geometry are usually
explained in terms of resonance delocalization of pi electrons among the amide’s N, C,
and O atoms. When not decorated with hydrophobic alkyl groups, amides can hydrogen
bond as both donors and acceptors, enabling their diverse roles in biological structures
and reactions. As for their reactivity, like most polar moieties, amides are both
nucleophilic (at O or N) and electrophilic (at the carbonyl C).

In this study we illustrate a systematic experimental design to facilitate the
catalytic hydrogenation of amides. Even though the primary methods of amide reduction
are through the classic use of hydride reagents in organic solvents, later studies proved by
catalytic hydrogenation was also effective. Traditional amide reductions using
nucleophilic hydride donor reagents such as LiAlI-I4 in organic solvents exploit the latter
reactivity, ultimately reducing the carbonyl center to a methylene via imine or iminium
intermediate species. But another widely used amide reductant, the strong Lewis acid
BH3, exploits the amide’s reactivity as a nucleophile. Though effective, these
stoichiometric reactions have poor atom economy, form undesired byproducts, solvent
and work-up waste. Even the few known catalytic reductions, mainly transition metal—
based, involve reactive compounds that require handling in aprotic organic solvents, a
setting that is not ideal for the amide’s intrinsic hydrophilic nature.

This report describes work toward “green” reduction of amides. The “holy grail,”

direct aqueous-phase catalytic hydrogenation, would maximize atom economy by
avoiding reagent and protecting group byproducts and by minimizing reaction steps and
organic solvent streams. Such a process would offer new opportunities for agriculture via
formation of value-added chemicals from underutilized crop byproducts. Specifically,
afier separation from the carbohydrate and oil fractions, a crop’s peptide and protein
fractions could serve as starting materials for substituted polyamido and polyamino
alcohol polymers. Such compounds have a wide range of potential applications as pH
adjusters, surfactants, detergent additives, metal ion scavengers, gas scrubbers, support
materials, or chiral building blocks and auxiliaries for pharmaceutical synthesis.

Despite the challenges of working under aqueous conditions with amides, we
believe a systematic, mechanistic study will extend our understanding of the carbonyl
reduction and minimize the use of organic solvents. The cyclic dipeptide 2,5-
piperazinedione was chosen as a model subStrate. The “green” heterogeneous aqueous-
phase catalytic hydrogenation was optimized via a parametric study at modest
temperatures (90-130°C) and pressures (500-1500 psi H2) using a 5% Ru/C catalyst
originally found to be optimal for carboxylic acid reduction. To evaluate reaction results,
a direct analytical HPLC method was first developed to quantify all possible reaction
products (amino alcohols, amines, amino acids, alcohols and amides).

This thesis is organized into five chapters. The following chapter provides an
overview of the literature for this study. The subsequent chapters each cover a specific
objective and were prepared as manuscripts for publication preparation. Chapter 2 is

devoted to the first objective, i.e., analytical method development and Chapter 3

describes the parametric study for amide hydrogenation. Overall conclusions and

suggestions for future research are also provided in Chapters 4 and 5.

CHAPTER ONE

LITERATURE REVIEW

1.] Carboxamide Group

The carboxamide group, simply termed amide, is one of the more complex _
functionalities in organic chemistry. Amides are relatively weak bases with pKa values
~10-15 units lower than the corresponding amines. Protonation occurs at the carbonyl
oxygen, because (a) the oxygen is more negatively charged than the nitrogen, and (h) pi
delocalization is retained in the O-protonated form, whereas N-protonation forces pi
localization on the carbon oxygen double bond. The delocalization of the pi electrons
along the O-C-N chain moderates both the carbonyl reactivity and nucleophilic properties

of the amide group (Figure 1.1).

O 09 06
RJL NR'2 R/KSR'z R/l®\ NR';

Figure 1.1 Resonance structures of amide

Like most polar moieties, amides are both nucleophilic (at O or N) and
electrophilic (at the carbonyl C).l Nucleophilic addition to the carbonyl carbon requires
activation of the carbonyl either by protonation of the carbonyl oxygen or O-coordination
with an electron-deficient species, such as an organometallic reagent.

This report pays particular attention to the selective reduction of the carbonyl of

the amide, in the presence of other reducible functional groups.

1.2 Amide Reactivity

The overall reactivity of amides depends on the substitution of the amides
themselves and activation of the carbonyl carbon. Metal hydrides are classically used for
the reduction of amides to amines. Figure 1.2 shows the scheme proposed by Brown and
Heim summarizing the reaction paths of amide reduction with a metal hydride. They
propose that a tetrahedral intermediate is initially formed, which can further react by
either carbon-oxygen (GO) or carbon—nitrogen (C-N) bond cleavage. Nucleophilic
attack by the hydride reagent on the CO bond can form the amine directly, or attack on
the ON bond with subsequent hydrolysis can form the alcohol. Cleavage of the
tetrahedral intermediate’s C-N bond can also form an amine and an aldehyde which can
be further reduced to the alcohol."2'3 Sterically demanding substituents on N tend to favor

C-N bond rupture products.

hydrolysis 0
NHR'

 

 

i w ° \
+
“”2 /H}\NR'2 C-N FIAOH + NHR'z
We W
M=AIH3,LiAIH3 MH RAOM + MNR'2
C-O MH
cleavage

 

Figure 1.2 Mechanism of amide reduction with a metal (M) hydride

Catalytic hydrogenation, another possible method for amide reduction, is the
primary focus for this project. Adkins and co-workers describe the possible reactions that
can take place during catalytic hydrogenation of amides.4 First, there are three possible

linkages in which hydrogenolysis can occur on an amide (Figure1.3).

Figure 1.3 Possible hydrogenolysis sites

In the presence of hydrogen, complete cleavage of the C-0 (c) double bond
results in the formation of the corresponding amine @rimary, secondary or tertiary) and
water. Further cleavage of the C-N bond (b) of the amine product, would result in

another amine and water (Figure 1.4).
0 H2 H2 , '
R/ILNR ———> R./\NR2 + H20 —-> RCH3 + NHR2 + H20
2

Figure 1.4 Hydrogenolysis of an amide forming an amine and water

When ethanol was used as a solvent during amide hydrogenation, cleavage of the
amide C-N (b), resulted in the formation of an alcohol and an ammonia (Figure 1.5)."'6
The alcohol formed in this process is believed to be the product of hydrogenation of an

intermediate ester.

H OH

Figure 1.5 Hydrogenolysis of the carbon-nitrogen bond of amides

It is known that amides react with several nucleophilic species; for example,

amines, alcohols, hydroxylamines, carboxylic acids etc. Most relevant to this project, is

6

amide hydrolysis to its parent carboxylic acid and amine fragments under aqueous

conditions (Figure 1.6).7

0 o
)LNRZ + H20 _. /“\OH + ~an

Figure 1.6 Hydrolysis of an amide to a carboxylic acid and an amine under aqueous
conditions

In addition to hydrolysis, other possible secondary reactions, such as aminolysis

of the amide or amine alkylation by an alcohol, should be noted (Figure 1.7).

o o
/u\ + NHR'2 __. )L 1 ann
R NHR R' NR2

H20

 

R'AOH + NHR2 > NHR'R +

Figure 1.7 Secondary product formation from amide aminolysis or amine alkylation

1.2.1 Classical amide reduction using complex metal hydrides

The classical approach for the reduction of an amide to its corresponding amine is
the use of various hydride reagents and solvents. Examples include lithium}8 sodium,
and potassium, and aluminum hydrides, tetrahydrido aluminates and diboranefw'10 Well
established reductions of amides have been reported using NaBHJTFA,”
NaBHngyme,” NaBH4/CH3SO2OH in DMSO,” and POC13/NaBH4,” However, the
choice of reducing agent depends on the functional groups present. Since amides are the

least electrophilic of the carbonyl functionalities, it is not surprising that nucleophilic

hydride reagents are generally not selective for amides, preferring to reduce better
electrophiles first. When used alone, NaBH4 reduces aldehydes, ketones, acid chlorides,
nitriles, nitro or halogenated organic molecules and amides."12 LiAlI-I4 is considered a
more active hydride donor and is often used over NaBI-I4, but it also reduces ketones,
esters, and even carboxylic acids in preference to amides.12 Consequently, electrophilic
borane complexes are more often used than nucleophilic hydrides. These reagents
exhibit amide selectivity, and are easy to use, but C=C double bonds are also susceptible

to reduction.2' '4

Mixed systems, such as sodium borohydride—cobalt chloride have been
used to successfully reduce nitriles, amides and nitro compounds to their corresponding

primary amines in both hydroxylic and non-hydroxylic solvents.15

1.2.1.1 LiAlH,

Kuehne and Shannon reported that the reduction of amides and lactams with
LiAlH4 proceeds with much difficulty, due to the formation of insoluble complexes.‘4
Brown and Heim reported that such reactions are often incomplete and side reactions take
place resulting in C-N cleavage to the corresponding alcohol.3 However, Newman and
Fukunaga8 carried out amide reductions using LiAlH4/THF or ether (reflux or room
temperature) with case. They suggested that the reduction of amides to primary amines
proceeds via a base-induced dehydration of the amide to a nitrile, which is then reduced
to the amine. This is different from Adkins’4 original proposal, and was confirmed by
detecting the nitrile intermediate and hydrogen evolution from two sterically different

compounds, triisopropylacetamide and benzamide. The reduction of

triisopropylacetonitrile to the amine was found to be quite slow, but gave almost

quantitative yield, whereas the reduction of benzonitrile was very fast with only trace

amounts detected (Figure 1.8).

gm “r ‘3 —* ““98

triisopropylacetamide triisopropylacetonitrile 2.2,2-triisopropylethylamine
N _
O S :
benzamide benzonitrile Benzylamine

Figure 1.8 Reduction of triisopropylamide and benzamide to a primary amine via a
nitrile intermediate

For the above reductions, a few possible mechanisms were considered. The
amide can undergo nucleophilic attack by a hydride ion or form an amide salt by an acid-
base reaction. The amide salt can further react through another acid-base reaction
forming a double salt, with the evolution of two equivalents of H2. The authors believe
the amide forms the double salt to yield a nitrile (Figure 1.9). The other possibility is that
the amide salt can be reductively attacked to form an aldehyde-ammonia intermediate,

which is then dehydrated forming an aldimine salt.

 

G) G)
O . ,NHS . INH3 9
/ll\ L'A'H‘ i Li® +H2 —->L'NH‘ i 21.59 +H2 —. R—ZN triage-NH“ 2Li©

Figure 1.9 Proposed mechanism of amide reduction to a nitrile with LiAlH4

1.2.1.2 8H3

Borane complexes are often used over LiAlH4 due to their electrophilic nature and
mild reducing conditions which are rapid, quantitative, and clean. Selectively for amides
over nitro, halogen, ester, sulfone and carbamate groups is often achieved using borane
complexes.

As mentioned above, the reduction of primary, secondary and tertiary amides
(both aliphatic and aromatic) derivatives in the presence of other less reactive functional
groups was achieved using BH3/THF from 0-25°C.3‘9‘l0 Early studies by Brown and
Heim9 reported the ease of using borane over LiAlH4, with yields of 94 and 98% for the
reduction of N,N-diethylpivalamide to N,N-diethylneopentylamine and N,N-

diisopropylbenzamide to N,N-diisopropylbenzylamine, respectively (Figure 1.10).

O

XkNm 8H3 XNA

K THFll K o
1hr. 94 /°

N,N-diethylpivalamide N,N-diethylneopentylamine

(3% —1- ©0698,

N,N-diisopropylbenzamlde N,N-diisopropylbenzylamlne

Figure 1.10 Reduction of N,N-diethylpivalamide and N,N-diisopropylbenzamide to
their corresponding amines with BH3/T HF (0-25°C)

Additional reductions by Brown and Heim9 were carried out with BH3 (refluxing
THF) for 1hr to give the amines shown in Figure 1.11. Under these conditions, the

primary amides, hexanamide and benzamide, were both reduced to their corresponding

10

amines in 87% yield. Secondary amines, N-methylhexanamide and N-methylpivalamide,
were converted in 98 and 83% yields. The selective reduction of the amide carbonyl of
N,N-dimethyl-p-nitrobenzylamide, was obtained in 97% yield.

The authors concluded that primary amides rapidly form a complex with borane,
but are reduced slowly, compared to secondary and tertiary amides. Primary aliphatic
amides were also reduced faster than primary aromatic amides. The authors rationalized
the outcome in terms of H2 evolution from the amide-borane complex, to form a poorly

reactive iminoxyborane species.

9 \
/\/\/\ HZN I O G) N‘—
Q \~
/N ’I
O
87% 87% 98% 97%
H l
98% XWHz Xk KM
I 33% 83% 92%

95%

Figure 1.11 Amine yields from amide reduction with BH3/T HF (0-25°C)

In later studies by Brown and Heim,3 steric and electronic effects were evaluated
and reaction rate measurements showed that increasing branching on the (rt-carbon
increases the rate of reduction. They also concluded that aliphatic amides are reduced

faster than aromatic amides (Figure 1.12).
o o o o
Z/‘LN/ > Y‘kN/ > /\)l\N/ > glhll/
l I |

Figure 1.12 Relative rate of reduction of or substituted N,N-dimethylamides with
BH3/'T HF (25°C)

11

Faster reduction rates were found with N,N-diethylbutyramide and N,-
butyrylpyrrolidine compared to N,N-dimethylbutyramide and N,N-
diisopropylbutyramide (Figure 1.13), due to steric effects of the substitutuents on

nitrogen.

AiN .AJOLNJ , AiN/ ,’ Ajax
0 2 . A

Figure 1.13 Rate of reduction of N substituted N,N-dialkylbutyramides with BH3/T HF at
25°C

Polar substituents on N,N-dimethylbenzamide did not influence the rate of
reduction. However, ortho substitution of the benzamide did slow the reaction rate

significantly, suggesting difficulty in forming the tetrahedral intermediate due to steric

effects (Figure 1.14).

O O O O O
(3%“ = Din“ = 0*?” = Dir 1561’
CI 02N H3CO

Figure 1.14 Rates of reduction of substituted N,N-dimethylbenzamides with BH3/T HF at
25°C

Though amides are typically less reactive than esters toward nucleophilic
reductants, BH3/THF effected selective reduction of a primary and secondary amides in

the presence of an ester as shown by Komet and others10 (Figure 1.15). Despite the close

proximity. of the amide and the ester in ethyl N,N-diethyloxamate, . the amide was

12

selectivity reduced to N,N-diethylaminoacetate (94%) and 5% of the amino alcohol.
Ethyl N,N-diethylsuccinamate gave the selectively reduced product, ethyl 4-
diethylaminobutanoate in only 36% yield. The reduction of a benzylic amide, para
substituted with a methyl ester gave a 66% yield of the amino ester. The higher yield of
the amine is believed due to the slower reaction rate of esters adjacent to aromatic

systems.

0 ON/\ BH3ITHF DYNA N/\

ethyl N,N-diethyloxamate 94% 5%

+

 

o
0w
N BH3/THF o l/ (

ethyl N,N-diethylsuccinamate 36% 7.3%
l
o o I
O O
BH3rl’ HF
HN
Y HN\/
O
66%
4-Acetylamino-benzoic
‘ acid methyl ester
0 0
,NR ,NR
0 N N
R R
R=Me. El or phenyl 57 %

R'=Et or Me

Figure 1.15 Representative selective amide reduction in the presence of an ester

(13an (0-25°C))

13

The reduction of methyl hippurate gave a mixture of products consisting of 2-
benzylaminoethanol (85%), methyl 2-benzylaminoacetate (11%) and 4% of an
unidentified product (Figure 1.16). The authors believe a cylic intermediate is formed in
which the ester carbonyl oxygen forms a coordinate convalent bond with boron,
activating the carbonyl of the ester (Figure 1.17) which results in the higher yield of the

amino alcohol.

0

O HN
O HN
H
w ————»BH3/THF v0 +
_. 13 W «We
methyl hippurate 2-benzylaminoethanol 2-benzylaminoacetate

85% 1 1%

Figure 1.16 Reduction of methyl hippurate with BH3/T HF (O-25°C)

O
l |
H2890

Figure 1.17 Proposed intermediate of the corrdinate covalent bond of an ester oxygen
with boron

1.2.1.3 NaBH4

Selective reduction of an amide to an amine has been reported using various
combinations of reagents with NaBH4. Examples including NaBI-14/CH3SO2OH,l3
I~313o+131=.{/NaBH.,‘6'17 POClg/NaBH4,” PC15/NaBI-I4,‘8 NaBHngyme,” and P4S10/Raney
Ni'9 will be discussed. Other systems such as NaBH4/C0C12,l5 and PC15/NaBH4,18 have
been used to successfully reduce nitriles, amides and nitro compounds to their

corresponding primary amines in both hydroxylic and non-hydroxylic solvents.

l4

Warm and others13 carried out similar reductions to those reported by Brown and
Heim,3 but with NaBH4/CH3302OH in DMSO. They reported that excess NaBH4
increased product yields which ranged from 60-88% for primary benzylic amides. This
combination of reagents, however, is not necessary selective for amides, and was shown
to also reduce carboxylic acids, esters and C=C double bonds. Carbon-halogen bonds
were undisturbed under these conditions. Esters are typically reduced much slower than
carboxylic acids under BH3/T HF conditions, but were equally reduced under the present
conditions. No reduction occurred if either the ester or acid carbonyl group was adjacent
to an aromatic ring. The only conclusions made were that the reduction of the amide was
carried out by activation of the carbonyl group with the acid, followed by nucleophilic
attack by the hydridic protons of BH4'.

Umino and others” reported a simple, highly selective reagent used to reduce

' Using sodium

both primary and secondary amides to their corresponding amines.l
acyloxyborohydrides NaBH3(OCOR) in either THF or dioxane, provided the amine salt
in 82-90% yield. Tertiary amines were reduced only in 20% yield. Figure 1.18 shows

the amines studied and their corresponding yields.

15

H°Co NH Ph \1
UV 2 PhA NH2 \ N /\ N
H 3C0 H ph’ \/

71% 76% 88% 89%
H3CO / /\ /
WT PhAu/ Ph T
H300 ‘
27% 83% 35%

©NSV\© “Or“

93% 20%

OCH3 OCHa
H H
WNW/fl _. D/VNWQ/
H3CO 0 H3CO

82%

Figure 1.18 Amine yield from the reduction of amides with NaBH3(OCOR) in THF or
dioxane

Raucher and Klein found that thioamides are more reactive than their
corresponding amides toward alkylation with Et3O+BF4_/CH2C12, which can then be
reduced to an amine with NaBH4/MeOH.16 A variety of amides, primary, secondary,
tertiary, benzamides, acetanilides, and lactams, were reduced to their amines via the

thioamide intermediate in this manner (Figure 1.19),

Et

0 N2P2$4 S + . SI
,lL __. A Em ear, —i}aH-B§ir* RANRR.
R NR.R' R NR'R' CHZCIZ R 5R.R. 3

thioamide

Figure 1.19 Thioamide reduction to an amine with Et30+BF4_/CH2C12 and NaBHJMeOH

16

The procedure was found to be selective for the thioamide over conjugate double
bonds, esters, nitro groups and sulfonamides. Reactions typically ran 30—60 min. at
ambient temperature, with yields ranging from 70-100% (Figure 1.20).

/\ Ph Ph Ph
Ph N(CH3)2 \/\ N(CH3)2 \/\ NHCH3 \/\ NH2 9N0: PhA NICHaiz

93% 100% 98% 88% 100%
O ph 0 Ph Ph
phdLN \/\N /[j\ J I /l
N /\N
H H
93% 70%
Ph Ph
0 J J 0 Ph Ph
d —’ O A“ ’T" A”)
90% 80%

%9

\‘_Q
0 w
\ \
N N
l C02M9 l 002Me
Ivittc,,H,,so2 MeC5H4802

94%

Figure 1.20 Amine yields from the reduction of its thioamide with Et3O+BF4./CH2C12
and NaBH4/MeOH

Borsh ran similar reductions of secondary amides using Et3O+BF4—/NaBH4 in
ethanol but directly formed the iminium ether tetrafluoroborate intermediate (20hrs,
25°C), which was reduced (20hrs. 25°C) to the amine in essentially quantitative yield

(Figure 1.21).17

0 Et o‘aF ‘ o’Et
3 4 NaBH /\
—____+ e 4 o 1-
ANR'R' CH2CI2 RANR'R' BF4 CH3CH20R R NRR
(D

Figure 1.21 Reduction of an amide to an amine via direct formation of an imino ether
fluoborate intermediate

A two step mechanism was suggested as shown for N-ethylbenzamide in Figure
1.22, for secondary and tertiary amides, lactams, and ethers. Primary amides were not
reduced under these conditions, but instead underwent a dehydration type reaction to

form a nitrile by eliminating ethanol.

CE! 9 G H H
g a... (05: a...
C5H5 NEl CGHS H ga CeHsANEl fast C5H5 NEI

Figure 1.22 Mechanism of reduction of imino ether fluoborate intermediate to an amine
with BH4’

Selective reduction of N-mono- and disubstituted amides and lactams with
preservation of ester, nitrile and olefinic groups was achieved using POClg/glyme and
NaBH4 at room temperature, 15-60 min., as reported by Kuehne and Shannon.‘4 These
studies are similar in selectivity to those by Komet and others using BH3/THF (0-25°C),
for the selective reduction of amides in the presence of esters. Kuehne and Shannon
proposed that diborane is generated with the addition of N3BH4 to the imino
intermediate. Figure 1.23 shows the reaction sequence of an amide or lactam to an imino
intermediate. Either the O-phosphoryl derivative or the corresponding imino chloride

can be formed in the reaction medium. In order to avoid the formation of diborane and

18

still achieve direct reduction of the amide, NaBH4 was added as an ethanolic solution to
the imino derivative. Figure 1.24 shows the reactions in which the amide was selectively

reduced.

0 J): Ye /\ R"
° 6) o I
R" R”

imino derivative
X and/or Y=OPOC12
and/or Cl

Figure 1.23 Reaction sequence of an amide or lactam reduction to an amine via an imino
derivative

cc“
NH —> |
d I 76%
C N
A NVN __. NV

66%

0 o
o N 71%
éNMofi G MOW

0. o 0‘7,” .
é? 7 50/.

 

Figure 1.24 Selective reduction of amides over olefin, nitrile and esters groups with
1)POCl3/glyme and 2)NaBH4/EtOH

l9

Steric and electronic effects on reaction rates were also studied (Figure 1.25).
The authors found that 01 substitution decreases reaction rate, and secondary amides react
faster than tertiary amides. In particular, aliphatic amides reacted faster than benzamides,
which reacted faster than anilides. The addition of a base, pyridine, increased the rate of

the formation of the imino derivative, especially among amides with no 01 proton.

est err or be m:

89% 88% 87% 84% 70%

0%.- 0%.. 0“?” Oct

80% 85% 86% 83%

Figure 1.25 Amine yields from their corresponding amide reduced with POCl3/glyme
and NaBHr/EtOH

In a different study by Rahman and others, the 01 carbon of amides or imides was
activated via a Vilsmeier complex, and reduced to an amine with NaBHr/EtOH.18
Isolated yields ranged from 75-90% for secondary and tertiary amides and primary

amides formed the corresponding nitrile quantitatively (Figure 1.26).

20

CI
\
N \LOCJL. N/\__N_‘£'_"_L.O|
l\ @I\ \
/ I
R

R

Vilsmeier complex

yield
R = O-Cl 80%
= m-Br 85%
= p-Cl 78%
= p-NOz 85%
= H 90%

0 Cl
PCIS / NaBH4
N l \ ———- Ne I \ ——> N \
H k H k

R R R
Vilsmeier complex

yield
R = o-Cl 90%
= m—Br 86%
= p°N02 770/0
= O-N02 800/0

N
PCI5
NH? NaBH,
Figure 1.26 Amide reduction to an amine via a Vilsmeier intermediate

Bazant and others reported reduction of lactams, aldehydes, ketones, oximes, acid
anhydrides, lactones, and imides using NaAlH2(OCH2CH2OCH3)2 in benzene (100°C).20
Of particular interest, e-caprolactam was reduced to hexamethylenimine in 82% yield and
N-methylsuccinimide was reduced to N-methylpyrrolidine in 92% yield (Figure 1.27).

No chemoselectivity studies were carried out.

21

0

HM D -—-> HN D
82%

c-caprolactam hexamethylenimine

‘ l
«are <5

N-methylsuccinimide N-methylpyrrolidine

92%

Figure 1.27 Lactam reduction to an amine with NaAlH2(OCH2CH2OCH3)2 in benzene
(100°C) '

Satoh and others reported the reduction of nitro compounds, nitriles and amides in
dioxane ( 100°C) or methanol (30°C), with NaBH4/CoCl2.IS Nitro compounds require
higher reaction temperatures than nitriles, and primary amides reacted much faster than
secondary amides. Figure 1.28 shows the amides reduced; tertiary amides were

unreactive under these conditions.

0 o o o
NH2 N/ A/lL HSAI/lLNHr‘t
H NH2 NH
2

60% 3' 30% 70% 40%

Figure 1.28 Amides reduced with NaBH4/C0Cl2 in dioxane or methanol (100°C)

Zhu and others‘2 reported the reduction of primary aromatic amides using
NaBH4/diglyme at high temperature. They found that amides (primary and secondary)
were first dehydrated to their nitriles with evolution of one mole H2 upon treatment with
NaBH4/diglyme at 162°C. The nitriles were reduced to the amines with the addition of

LiCl. Their conclusions are similar to Newman and Fukunaga’s (LiAlHrt/THF),8 with

22

amide reduction proceding through a nitrile intermediate. The mechanism proposed by

Zhu and others is shown in Figure 1.29.

o e 0A)? (a?

o
R'J‘NH2 R/KQ

 

 

 

 

NH2 R '1“ R \Né)
H H
R=aromatic
@ G)
,BH3 O’BH3
e 0“H + H
+ N ‘——-— ‘ . <1——-—— 2
[ 81131011) R/ R/LN RAN“
other roducts
p RANH

Figure 1.29 Proposed mechanism of amide reduction with NaBH4/glyme

Bhandari and others” reported the reduction of lactams to their corresponding
amines by Na metal/n-propanol (reflux, 1hr.). Yields ranged from 70% for a five
membered ring to 72 and 65% for six and seven membered rings, respectively (Figure
1.30). When LiAlHr was used, yields decreased 5-10%, and no reduction was achieved

with either H2/Pd or H2/N i (R).

23

O O

H
N O
NH N
70% 67% HO

O O OMe 80%
72% 75% N

0 3
O N

UH m
N
NH2 H

65% 68% 73%

Figure 1.30 Amides reduced using Na metal/n-propanol

1. 2. 2 Catalytic Hydrogenation

As cited in Adkins and Wojcik,22 and as early as 1906, Mailhe and others had
reported the hydrogenation of acetamide and propionamide to ethyl and dimethyl amines,
and dipropylamines, respectively, in the vapor phase over a nickel catalyst (Figure 1.31).
No additional reaction conditions, yields or purities were reported. These experiments by

Mailhe, unfortunately, have never been reproduced with yields higher than 1%.4

0
H2 \/\
\/U‘NH2 T” NH2 *

Figure 1.31 First reported hydrogenation of an amide by Mailhe

24

Beginning in the 1930’s additional cases of catalytic hydrogenation of amides to
amines were published. Winans and Adkins6 first reported the catalytic hydrogenation of
a number of nitrogen compounds over nickel. As shown in Figure 1.32, hydrogenation
converts benzamide to a mixture of the primary amine, benzyl amine, and its dimer,
dibenzylamine. It is believed that the conversion of benzamide involves an imine
intermediate which converts to the primary amine in the presence of the an anhydrous

solvent.‘5

0 NH;
1500 psi H2 Ni
200°C
0 NH2 HO NH2

NH NH; NH
/ /
1500 psi H2 Ni +
200°C '

Figure 1.32 Hydrogenation of benzamide over Ni (1500 psi H2, 200°C)

Adkins and co-workers continued to explore hydrogenation experiments in which
an amide was reacted exclusively with hydrogen22 over copper chromite catalysts." 5’ ”'25
Wojcik and Adkins4 found that reductions using H2 were incomplete when performed
under aqueous conditions due to the hydrolysis of the starting amide to form an acid and

ammonia. The formation of ammonia is undesirable since it may deactivate most

catalysts. Alternative reaction conditions using dioxane were chosen to reduce

25

hydrolysis, over ZCUO'CI'203 under 1470.4400 psi H2 at 175-250°C.22 In the work of
Wojcik and Adkins,4 both carbon-oxygen and carbon-nitrogen hydrogenolysis were
observed. Primary amides of lauric, heptanoic and tetrahydrofuric acids gave an
average yield of 88.5% to their corresponding amine plus water; and yields for the
substituted amides were lower ranging from 54-72% (Figure 1.33). Secondary cleavage
products from nitrogen-carbon bond hydrogenolysis formed predominantly with

secondary and tertiary amines.

O

WNW
lauramrde O/ENH:
O

HZNJW/ ot— —pheny|butyramide tetrahydrofuranamide

he tanam'de
P I O

WN

O O
WNW WNW H
H H

Figure 1.33 Amides reduced over 2CuO-Cr2O3/dioxane (1470-4400 psi H2, 175-250°C)

D’Ianni and Adkins23 looked specifically at the effect of hydroxyl group position
of amide hydrogenation over 2CuO'Cr203 at 250-260°C (4 hrs.) and over Raney Ni at
225°C (5-8 hrs.) in dioxane. Raney Ni showed no activity for most of the amides studied,
whereas 2CuO-Cr2O3 gave appreciable yields. Five hydroxyamides were studied, with
the hydroxyl group placed in the or, B, y, 5, 8 positions with respect to the amide. The

corresponding hydroxyamines were formed in yields of 51, 0, 79, 76, and 60%,
26

respectively, over 2CuO-Cr2O3. The low yield of the B-hydroxybutyramide producing the
hydrogenolysis product (78% alkylpiperidine) was predicted as seen in previous
experiments with oxygen and nitrogen B to each other. Another type of hydrogenolysis
of the C-N bond was also observed producing glycol in each of the reaction mixtures of
the five hydroxyamides. The authors believe the hydroxylamine is formed by the
interaction of the amine and glycol.

Surprisingly, hydrogenation of B-hydroxybutyramide over Ni (R) yielded 86% 4-
amino-butan-Z-ol. The rate of the Ni-eatalyzed reaction of hydroxyamides was higher
than for amides with no hydroxyl group. This was shown by competitive hydrogenation
of caproamide and S-hydroxycaproamide. After 8 hrs, 83% of caproamide was still
present, compared to only 14% B-hydroxycaproamide. In the case of hydroxydiamides
and lactams, mixtures of cleavage products were formed over 2CuO-Cr203.

Sauer and Adkins also reported the hydrogenation of several amino or carbethoxy
pyrrolidones, piperidones, or quinolones over 2CuO-Cr203/dioxane, 1470-4400 psi H2, at
elevated temperatures of ZOO-260°C.25 The focus of their experiments was to compare
the relative reactivity of amido and carbethoxy pyrrolidones, piperidones, or quinolones.
Three competing reactions were described: the reduction of an ester to a hydroxyl methyl
group, reduction of an amide carbonyl to a methylene group, and hydrolysis of a
hydroxyl to methyl. In the case of pyrrolidones, the esters of 5-carbethoxypyrroline-2
and 5-amylcarbamylpyrrolidone-2 were reduced to their alcohols in 93 and 68% yield,
respectively with no reaction occurring at the amide carbonyl (Figure 1.34). When

comparing B-carbethoxypyrrolidones, the ester carbonyl was again preferentially

27

hydrogenated to the hydroxyl group (50% yield) over the amide. Extending the time of

the reaction did allow the carbonyl of the lactam to be hydrogenated to methylene.

O

0
fig}: Q
0\ OH

O
93%
O O

NH Q}:
NH NH
° C O
68%

( NR fiR extended QR
ime HO
O
HO
O

R=n-amy| or phenethyl

Figure 1.34 Ester reduction over 2CuO'Cr203/dioxane (1470-4400 psi H2, ZOO-260°C) in
the presence of an amide

B-carbethoxypiperidones also reacted in which the ester hydrogenated first to the
hydroxyl (24%), then to the fully reduced cyclic amine (Figure 1.35). In the case of a
diamide, the lactam carbonyl hydrogenated first, followed by reduction of the side chain

amide with subsequent cleavage to the hydroxyl group.

28

Figure 1.35 Lactam hydrogenation over 2CuO-Cr2O3/dioxane (1470-4400 psi H2, 200-
260°C)

Hydrogenation of quinolones occurs at the amide and ester groups forming the
primary alcohol (24%), which is then converted to a methyl group (29%). When the ester
was 8— from the nitrogen, it was cleaved by hydrogenolysis or retroaldol, and the

carbonyl amide was reduced to the amine (85%) (Figure 1.36).

\/O O OH I
: E113% : :N: N
H H H

24% 29%

O J __.
o (11 CO
M N o N
N O H H
H
85%

Figure 1.36 Hydrogenation of quinolones over 2CuO-Cr2O3/dioxane (1470-4400 psi H2,
ZOO-260°C)

29

1.2.2.1 Mechanisms for Catalytic Hydrogenation

The transalkylation of secondary and tertiary amines with alcohols over copper-
chromium catalysts was investigated by Adkins and coworkers.“ 26 They determined
that the catalyst dehydrogenates the alcohol toan aldehyde or ketone, which reacted with
the amine to form a gem-amino alcohol, which was then hydrogenated to the amine

(Figure 1.37), presumably via an imine intermediate.

0
RzHC_OH Calalysl JL + H2
R R
R'NH2
R R H2 R
X —’ R'HN—( t H20
R'HN OH R

Figure 1.37 Transalkylation of an amine with alcohol over 2CuO-Cr2O3/H2

It is also possible to dehydrogenate the starting amine, which reacts with the

alcohol to form an amino alcohol, then cleaves into an ether and ammonia (Figure 1.38).

H2N—\ Catalyst HN=\ + H2
R ———> R
R'OH
R “2
H2N—< ' "—’ R—\ ' + NH3
OR' OR

Figure 1.38 Formation of ether and ammonia via an amino alcohol over 2CuO-Cr2OflI2

30

A possible side reaction, alkyl group exchange among amines, may occur via the
following route (Figure 1.39) which starts with the initial dehydrogenation of the amine

to an imine intermediate, followed by amine addition across the C=N double bond.

H2N—\ Catalyst HN=\ + H2
R

__..__. R

Ax ——’ RAAR + NH;

Figure 1.39 Side reaction forming a secondary amine via an imine intermediate over
2CuO°Cr203/H2

The formation of secondary amines through a nitrile intermediate, which is

reduced to an imine (Figure 1.40), was also considered.

 

N=--R H2 HN=\
R
H2
HN=\
NH
R 2 H2 /\ /\
HZNAR ——’R/\N R #13.th + NH:
H

Figure 1.40 Secondary amine formation via nitrile over 2CuO-Cr2O3/H2

To test the reactivity of nitriles, n-butyl cyanide and n-hexyl cyanide were
hydrogenated in ammonia over Raney nickel at 125°C. The primary amines formed in

90-95%, and the secondary amines formed in less than 5% (Figure 1.41).26

31

_ Ni (R) H /\
N=——R —L. HN R + NHa
W7 ”ZN/kl, 2

F1=n-hexyl or n-butyl

Figure 1.41 Nitrile hydrogenation over Ni (R)/NH3 ( 125°C) to an amine

The transalkylation of tertiary amides with alcohols to form amines over copper-
chromium barium oxide (I)/dioxane was investigated by Schneider and others (Figure
1.42).24 The major product formed during the reduction of lauramide was the alcohol,
carried out under 4500—12000 psi H2 (250°C). Decreased temperatures (175°C) favored
alcohol formation. When an alcohol was used as the solvent, an intermediate ester was
converted into an amine as the major product under the hydrogenation conditions stated.

Water as a solvent deactivated the catalyst, and little hydrogenation took place.

 

O copper-chromium OH
barium oxide (1) +
M1011 2 4500-12000 psr W WNHZ
9 2 d' 9
toxane
major product
0 copper-chrornium 0 %
barium oxide (l) A NH H H
M So 0121450042000 psi? “"20 ”'M /\ W0" —’* “It/Wt»
9 O NH o
‘ /\ 9
OH

Figure 1.42 Transalkylation of tertiary amides with alcohols to form amines over copper-
chromium barium oxide (I)/dioxane (4500-12000 psi H2, 250°C)

1.2.2.2 Bimetallic Catalyst Systems
The use of bimetallic catalyst systems has proven to be advantageous, enabling
selectivity for amide reduction and less harsh conditions (lower pressures and

temperatures), then with simpler catalyst sytems.

32

The bimetallic catalytic hydrogenation of amide derivatives was canied by
Hirosawa and others, concluding that synergestic effects of metals are quite usefirl, and
can be carried out under relatively mild reaction conditions, 1500 psi H2 (170°C) in 1,2-
dimethoxyethane (DME).27 Metals from . Groups 8 to 10, late transition metals
Rh6(CO)16 or Ru;(CO)r2, and Group 6 or 7 early transition metal complexes, Re2(CO)to,
W(CO)6, or Mo(CO)6 were found to catalyze reduction of N-acetylpiperidine to N-
ethylpiperidine in yields ranging from 50-98%. Alone, Rh6(CO)16 afforded 19% yield of
N-ethylpiperidine. The most effective combinations were rhodium-rhenium (96%),
rhodium-molybdenum (98%), and ruthenium-rhenium (96%). Synergistic effects of
rhenium, molybdenum, or tungsten with rhodium or ruthenium were observed in both
homo- and heterogeneous systems. Several amides were examined over Rh6(CO)16-
Re2(CO)to metal complexes at varied catalyst loading ratios. Tertiary amides were
reduced with amine yields ranging from 70-92%. Increased catalyst loading and reaction
time showed increased yields. Secondary amides were reduced to their corresponding
amines with yields ranging from 82-90%, with 3% mol catalyst loading at 170-180°C.
Primary amides, on the other hand, were converted to the secondary amine.

When evaluating the reactivity of bimetallic catalysts on lactams, it was found
that ring size affects reactivity. Slightly higher reaction temperatures were needed for 5-
valerolactam than for s-caprolactam, (160 versus 180°C) and m-heptalactam required
higher catalyst loading (3 mol% versus 1 mol %) than 5-valerolactam and s-caprolactam.
Substitution at the nitrogen atom of the lactam decreased overall activity. These
bimetallic systems have been shown to be effective toward the hydrogenation of

carboxylic acids, but not esters. Selectivity studies have not been carried out as of yet.

33

Kuwano and others described a rhodium complex catalyzed reaction with Ph2SiH2
for the selective reduction of tertiary amides at room temperature in THF.28 N,N-
dibenzylacetamide was reduced to dibenzylethylamine with various rhodium metal
complexes. RhH(CO)(PPh3)3/Ph2SiH2 gave a 94% yield, and RhH(PPh3)4/Ph2SiH2
(93%). Rhodium complexes without a hydride ligand, [Rh(COD)2]'2BF4PPh3,
RhCl(PPh3)3 and RhCl3-H2O, all in Ph2SiH2, gave yields of 836, 93, and 93% respectively,
but were very slow. When the hydride equivalent was varied, Ph2SiH2, PhSiH3 and
Ph3SiH with RhH(CO)(PPh3)3 gave yields varying from 49, 90, and 0%, respectively.
Even when sterically hindered amides were employed, amine yields were still high.
However, reaction rate slowed down.

Chemoselectivity studies of amide reduction was carried out as well. It was found
that RhH(CO)(PPh3)3 catalyzed reductions with Ph2SiH2 selectively reducing amides to
their amines without reacting with bromo, ester or epoxy groups. Amides with either
double or triple bonds, and a-active methylenes gave mixtures of products.

The proposed mechanism by Kuwano of the reduction of an amide with the

dialkylsilane is shown in Figure 1.43.

34

/\ Rth

R NR'2 ' R"2HSiH
Reductive Elimation Oxidative Addition
l
LnRh’fR Lnrlh-SiHR"2
NR'2
Carboxamide bond
insertion
Rth + R"2HSiOSiHR"2 0
if OSiHR"2 R NR'z
H LnRh R
Lth-SiHR"2 NR2

C-O bond cleavage. rapid H transfer

Figure 1.43 Proposed catalytic cycle of an amide reacting with dialkylsilane

Later studies by Igarashi and others reported metal complex catalyzed amide
reductions with hydrosilane (H2SiPh2 or H3SiPh) to be an effective means of producing

29 Monohydrosilanes

tertiary, secondary and primary amines at elevated temperatures.
were used as the reducing agent for N-acetylpiperidine with a variety of Group 7-10
transition metals (Mn, Re, Ru, Os, Rh, Ir, Pd and Pt). Highest yields were found using
HSiEt3/033(CO)12/Et2NH (99.8%), HSiMe2Ph/RuCl2(CO)2(PPh3)2/Etl (90.2%),
HSiMe2(OEt)/RuCl2(CO)2(PPh3)2 (93.1%), HSiMe(OEt)2/RuCl2(CO)2(PPh3)2 (90.8%),
and HSiMe(OEt)3/[RuCl2(CO)3]2/Etl/Et2NH (86.1%) at 100°C, in toluene. A variety of

lactams were also reduced in yields ranging from 86-99%.

35

1.3 Relevant Laboratory Studies

In our laboratory, working under alternative ‘green’ conditions plays a major role.
In particular, reactions compatible with biomass-based substrates are desired, to open
paths to value-added based products as an alternative to petroleum based products. The
catalytic hydrogenation of organic acids to their corresponding alcohols has been studied
quite extensively under aqueous conditions over ruthenium catalysts. More recently

hydrogenation of sugars, and now amides, has also been carried out.

1.3.] Catalytic Hydrogenation of Amino Acids
1.3.1.1 L-Alanine

Studies by Jere and others30‘3|

evaluated the catalytic hydrogenation of L-alanine
to L—alaninol over 5% Ru/C (1000 psi H2, 100°C) in excess H3PO4/H20r In these studies,
it was found that the stereochemistry at the a—carbon was retained. Reduction
selectivities and enantiometric excesses for L-alaninol were found >95%, and >99%,
respectively. The presence of acid was found to be necessary in order for hydrogenation
to take place; presumably the acid keeps the amino acid carboxylate in its fully

protonated form. Together with isotopic labeling and kinetics data, the following

mechanism of hydrogenation of L-alanine to L-alaninol was proposed (Figure 1.44).

o “2 OH “20 o “2 H H
”3 S “3 2' H3 5 Ha
v. OH 360:4 v. H ,. OH
H‘ H‘ H‘ H\\
@NHa GaNl-la ®NH3 ©NH3
gem-diol aldehyde amino-alcohol

Figure 1.4-4 Reaction pathway of L-alanine to L-alaninol hydrogenation

36

1.3.1.2 Glycylglycine

From the work by Jere and others, the next step was to run these reactions on
peptides. The simplest peptide, glycylglycine, was explored briefly in our laboratory
under the same conditions: 5% Ru/C (1000 psi H2, 100°C) in excess H31>o./Hzo.32
Isolated yields of the resulting products were detected using NMR, which show that the
reaction mixture consisted of the starting material (50%), 2% of its cyclic anhydride (2,5-
piperazincdionc), hydrolysis product (glycine, 25%), and the product of glycine

hydrogenation (ethanolamine, 20%) (Figure 1.45).

 

O
H QCa 1000 psi H2 H 09 G) 00
(D N 5%) RU/C G) N NH OH HN
NH3/\'r \Ao > NHa/Y \AO + S/Y +H2N/\/ + NH
0 H3P04/H20 o 0
100°C, 24 hrs 0
50% 25% 20% 2%
glycylglycine glycylglycine glycine ethanolamine 2.5-piperazinedione

Figure 1.45 Catalytic hydrogenation of glycylglycine with H3P04/H20 (1000 psi H2,
100°C, 1 g 5% Ru/C)

Subsequent studies were carried out to stabilize the intermediate and prevent
hydrolysis using metal complexes. Glycylglycine under hydrogenation conditions of
1000 psi H2 (100°C), in the presence of the mildly acidic zinc phosphate (ZnHPO4)/HzO
gave similar results as previously mentioned recovering 49% of the starting material,
1.5% 2,5-piperazinedione, 34% ethylamine, 24% glycine and 20% ethanolamine at 24
hours. Glycinamide was reacted over 5% Ru/C (1000 psi H2 100°C) with

Ca(H2PO4)2/HzO, producing 72% ethanolamine, 14% ethylenediamine, 9% glycine and

37

2% ethylamine (Figure 1.46). Although the desired ethylenediamine product was

produced, the reaction was still predominantly going through hydrolysis.

G

 

 

G
H @ O /\/OH
<9 NH H N
Mia/fir” o " mor ” 2
O .
50% 24% 21 °/o
1000 PsiH . ,
H 09 5% Ru/c2 Q'YCY'Q'YC'“e glycme ethanolamine
(9 N >
NH3/\l(l)/ O ZnHPO4/H20 0
10 ° ,24h
0 C rs + HN + H2N/\
NH
glycylglycine O -
1.6% 3%
2,5-piperazinedione ethylamine
1000 Psi H2 (9 09
I.-'2N/\ll/ 2 > NH3/\/ 2 + NH3/\ll/ + HzN/VO + H2N/\
O 08(H2PO4)2/H20 O
O
“0 93'“ 14% 9% 72% 2%
'H3PO4
glycinamide

Figure 1.46 Catalytic hydrogenation (1000 psi H2, 100°C, 5% Ru/C) of glycylglycine
(ZnHPO4/H2O) and glycinamide (Ca(H2PO4)2/H2O)

Next, hydrogenation of glycylglycine in the presence of
Zn[HN(C2H4NH2)2][HN(CH2COO)2] and Ca(H2PO4)2, was carried out at 130°C, 500 psi
H2 for 42 hrs. At 23 hrs, NMR analysis showed only slight conversion to the alcohol,
glycylaminoethanol among a complicated mix of products with overlapping peaks.

In another attempt to prevent hydrolysis, the solvent was changed to a non-
aqueous system. In the studies by Redko,32 the ionic liquid (Me3N-HCOOH) was chosen
assuming formic acid could play a role as the reducing agent, and still be separated easily
from the final product. However, it can also react with primary and secondary amines

forming an amide. It can also serve as a buffer preventing protonation of the

38

oligopeptide; thus preventing hydrogenation. Glycylglycine and Me3N-HCOOH were
refluxed for 4 hours at which time catalyst (5% Ru/C) was added and the reaction ran an
additional 4 hours. Results by NMR showed partial conversion to 2,5-piperazinedione
and glycine, but no hydrogenation products.

Polyphosphoric acid (H6P4013) was also studied as a dehydrating agent to prevent
hydrolysis of glycylglycine and also acidic enough to protonate the amide and accelerate
the reaction. Under hydrogenation conditions of 660 psi H2, 100°C, 5% Ru/C, the
reaction was run for 45 hrs. The only product detected using NMR and confirmed by
HPLC was 2,5-piperazinedionc.

Subsequent studies with 2,5-piperazinedione, under similar acidic conditions gave
primarily the desired cyclic reduction product piperazine (25%), ethanolamine (30%),

and ethylenediamine (20%) (Figure 1.47).

 

“J5 2923:: H 09 e 09 EH
(9 u 6) OH HN
NH3/\"/N O fiA NHB/firNAO + NH3/\g/ + HzN/V + Hr

 

NH
O H3PO4/H20 0
100°C. 24 hrs 0
50% 25% 20% 2%
glycylglycine glycylglycine glycine ethanolamine 2,5-piperazinedione
o H
1000 psi H2 (9 09 N
HN NH 5% NC > "Hit/Hr + H2N/\/OH + HzN/\ + HzN/\’NHZ + E 1
Hr H3PO4/H20 0 ii
0 110°C. 24 hrs negligible
but detectable 29% 19% 20% 25%
2,5—piperazinedione glycine ethanolamine ethylamine ethylenediamine piperazine

Figure 1.47 Catalytic hydrogenation of 2,5-piperazinedione with H3PO4/H2O (1000 psi
H2, 100°C, 1 g 5% Ru/C)

39

1.3.1.2.] Hydrolysis of Glycylglycine

One of the biggest challenges in Redko’s studies was the hydrolysis of the amide
under aqueous conditions. Mechanistic studies in the area of peptide hydrolysis have
been found to very complicated due to the simultaneous reactions occurring in solution.
Hydrolysis of the peptide bond, internal aminolysis, cyclization of the dipeptide to the
diketopiperazine, and decomposition of product amino acids must be considered. Recent
studies”36 by Steinberg and Radzicka describe the rate constants and mechanistic theory
of peptide hydrolysis and diketopiperazine formation.

The results indicate equilibrium between glycylglycine and 2,5-piperazinedione,
which occurs more rapidly than the hydrolysis to glycine in neutral solutions at 25°C
(Figure 1.48).”36 The hydrolysis rate of peptide bonds is slow at ambient temperatures
and the formation of the diketopiperazine is the major product at increased temperatures.
The decomposition of tripeptides by Steinberg34 supported the formation of the
diketopiperazine as the major product during the decomposition of peptides, via the

internal aminolysis reaction.

0

 

o
(lLNH 9 9 e
+ H O __ HSN\/IL O ——’ 2 0 ——>2C02 + 2CH3NH2
my 2 {I mtg/I
o
2,5-piperazinedione glycylglycine glycine
Figure 1.48 Equilibrium between 2,5-piperazinedione and glycylglycine at 25°C
One of the other most challenging portions of Redko’s studies proved to be the

limitations of product analysis. Quantitative NMR analysis was limited,32 by the

presence of paramagnetic metal ion impurities leached from the stainless steel reactor. In

40

addition, the NMR spectra of the compounds studied have many overlapping proton
peaks making it difficult to distinguish among products. The presence of acid in solution
was found to corrode the reactor, increasing dissolved metal ion concentrations and
increasing difficulty with product analysis. In response to the difficult product analysis
of amide hydrogenation, a direct method was solely needed. Such a method, which
analyzes all possible hydrogenation and hydrolysis products quantitatively using High
Performance Liquid Chromatography (HPLC) was developed, and the subject of the next

chapter.

41

CHAPTER TWO
ANALYTICAL METHOD DEVELOPMENT
2.] Introduction

The preliminary portion of this project was to develop a single quantitative
method using High Performance Liquid Chromatography (HPLC) for the routine
determination of specific amides, polyamides, their corresponding amines, and other
related compounds (Figure 2.1) which may be present under the planned aqueous-phase
catalytic hydrogenation conditions. High Performance Liquid Chromatography is often
used because of its high sensitivity, reproducibility, and rapid analysis. However, the
broad range of polarity and charge among the compounds analyzed in this work makes
direct HPLC quantization difficult both in terms of separations and detector type. Simple
amines and amino acids, typically present in their ionic or zwitterionic forms, are often
analyzed by a combination of either pre- or postcolumn derivatization and reverse-phase
column chromatography, with ultra violet (UV) or fluorescence detection. However,
derivatization fails for most uncharged amides or carboxylic acids; and consequently
limits its use in this work. Ionic species, on the other hand, under aqueous conditions are
quantified using ion-chromatography. This type of separation is ideal for quantifying the
amine products of interest and the other related side products.

With the above complexity in mind, mixed-mode separation was chosen to
identify and quantify the compounds of interest. The mixed-mode column selected is
specific for retaining basic compounds by a cation—exchange mechanism while retaining
neutral compounds in a reverse-phase mechanism. In this particular case, the net charge

of the compounds must be manipulated to allow for complete baseline separation of the

42

compounds of interest (Figure 2.2). This is done by simply changing the pH of the

mobile phase under gradient conditions, and therefore limits the use of refractive index

0
O
H2N\)l\ OH H N H2N OH
H2N\/ll\ N/Y 2 \/\OH \/\N/\/
OH H 0 H

glycine 759/mol glycylglyclne 132 g/mol ethanolamlne 61 g/mol amlnoethylethanolamlne 104 g/mol
O H N OH

H2N\/\NH H N\)l\ OH 2 \AHW

2 2 N/\/ O

H

ethlyenedlamine GOQ/mol glycylaminoethanol 1 1eg/mol N-(2-amlnoethyl)glyclne 118 giant

0 O

(“\NH #NH (\NH
+1ij HN\/l HNQ
O

2,5-plperazlnedlone 114 g/mol plperazlne 2-one toog/rnol piperazine 869/mol

Figure 2.1 Amino acids and corresponding amines of interest

(RI) detection. As a result, the evaporative 1i ght-scattering detection (ELSD), a technique
almost as versatile and general as R1 was selected. This type of detector functions on the
differences in volatility between the solvent and the sample. The key operation is the
nebulization of volatile effluent from the column into a fine mist. The mist is carried
through a heated drift tube which evaporates the mobile phase, leaving behind only non-
volatile solute particles. The solute particles are carried at high speed through a beam of
light; an off-axis photomultiplier then detects the scattered light. The amount of light

scattered is dependent upon size and number of particles and is proportional to

43

concentration. The key is that all solutes must have a lower volatility than the mobile

phase.

2.2 Experimental Procedures

2.2.1 Materials

Complete separation and quantitative analysis of all samples was performed on a
Primesep® 100 mixed-mode column (3.2x150 mm, particle/pore: Sum/100A) (SIELC
Technologies, Chicago, IL). A guard column (3.2x10 mm, particle/pore: 511ml 100A) was
also in place prior to the analytical column (SIELC Technologies, Chicago, IL).
Acetonitrile used to make the mobile phase was HPLC grade, (Sigma-Alrich, St. Louis,
MO). Trifluoroacetic acid (TFA) (99%) (Si gma-Alrich, St. Louis, MO) was stored under
dry house nitrogen following each use. A Milli-Q ultrapure water purification system
(Millipore, Billerica, MA) was used as the sole source of water. Additional filtration of
the mobile phase upon entering the instrument was carried out using a Corning top filter
nylon membrane (70mm membrane diameter, pore size 0.20am) (Fisher Scientific,
Pittsburgh, PA).

The chemical standards, glycine, glycylglycine, and 2,5-piperazinedione were
purchased from Si gma-Alrich (St. Louis, MO), piperazine from Lancaster Synthesis Inc.,
(Pelham, NH), piperazine-Z-one from Alfa-Aesar, (Ward Hill, MA), and N-(2-
aminoethyl)glycine from Toronto Research Chemicals Inc. (North York, Ontario,
Canada). Glycylaminoethanol is not commercially available.

All chromatography was performed using a Waters 600E HPLC pump (Waters,
Milford, MA). Both UV 490E (Waters, Milford, MA), and ELSD 800 (Alltech, Chicago,

IL) detection were carried out. The ELSD was used for all quantification and the UV

44

was used to monitor the gradient flow and equilibration confirmation. The evaporation
tube temperature was set at 60°C and the nebulizer inlet gas (N2) pressure set at 3.00 bar.
All data acquisition and analyses were carried out using Star Chromatography

Workstation (Version 6.20) (Varian Inc., Palo Alto, CA).

2.2.2 Methods

Running conditions of the mobile phase consisted of 10% acetonitrile in water
with a T FA buffer gradient from 0.05% to 0.5% at a constant flowrate of lmL/min. The
HPLC was set at initial gradient conditions (10% acetonitrile in water, 0.05% TFA,
(reservoir A)) for approximately one hour to allow the column to equilibrate. The
gradient elution began with reservoir A at injection and proceded from 0% to 100%
reservoir B (10% acetonitrile in water, 0.5% TFA) over 7.5 min. Reservoir B was held
for 5 min and then the gradient returned to initial conditions (reservoir A) over 3 min and
held until the column equilibrated, approximately 10 minutes. Small peaks (< IOmV) at
9.5 and 14 min are systematic peaks resulting from the change in acid concentration and
do not interfere with the compounds of interest. Mobile phase and water were injected
as standards to account for any additional peak area interference. Control samples of
water were run in the reactor and analyzed to identify any metal peaks present in the
HPLC.

Important Note: For this experimental design, it was concluded that the source of
water must be chosen carefully. By using an ion-exchange exchange column with such
low sensitivity, any unpurified water increased baseline noise. The Milli-Q ultrapure
water was found to keep the most stable baseline and reproducible results. Proper storage
of TFA under nitrogen storage was also found essential for reproducible results. Lastly,

45

it is highly suggested to prepare fresh mobile phase just prior to use, but always keep the
pump running.

Calibration Setup: Calibration curves were prepared from a mixed standard
sample set, by using the dilutions (0.003, 0.005, and 0.007M) from a prepared stock
solution (0.01M). One sample set consisted of 2,5—piperazinedione, ethanolamine, and
piperazine. The second set consisted of glycine, glycylglycine and piperazine-Z-one.
The stock solutions and each subsequent dilution were prepared using water.

Four point (0.003, 0.005, 0.007 and 0.010M) linear calibration curves were used
for the quantification of products in the parametric study. An initial single calibration
method was created and used in all experimental analyses for the parametric study.
Mixed standard solutions of the compounds, prepared initially and kept at 0°C, were
injected each day to check for system performance and sample stability.

For the desorption study, four point (0.003, 0.005, 0.007 and 0.010M) quadratic
calibration curves were used for all analysis. A new calibration method was created for

each individual study, with standards prepared fresh each day.

2.3 Results & Discussion

The analytical method development for the detection and quantification of
amides, amines, amino alcohols, and amino acids proved to be challenging, yet
necessary. The method developed is a direct method, with no derivatization needed,
which detects all the compounds of interest for this study in less than 25 min. Nearly
complete baseline separation is shown in a sample chromatogram (Figure 2.2) of a mixed

standard at 0.007M. Retention times (min.) are as follows: 2,5-piperazinedione 1.0,

46

glycine 3.2, glycylglycine 4.3, ethanolamine 4.7, piperazine-2ch 5.2, N-(2-
aminoethyl)glycine 9.2, ethylenediamine 10.2, AEEA 10.6, and piperazine 10.8. All
peaks on the chromatogram have been accounted for. A small peak (<10mV) is present
just prior to glycylglycine at 4.2 min. This was concluded to be the detection of metals
from the reactor, an interpretation supported by l) the observation of this peak in HPLC
analyses of water that been heated in the Parr reactor with catalyst and hydrogen alone,
but no substrate, and 2) the absence of observable 1H NMR peaks, other than that for

HOD, in the blank sample.

 

mVolts
700 1

d ethylenediamine
2,5-piperazinedione (10.2)
500 _, (1.0)

ethanolamine aminoethylethanolamine (AEEA)
40° (47) . .
N-(2-aminoethyl)glycme

300 glycylglycine (9.2) piperazine (10.8)
d (4.3)

 
     

piperazine-Z-one
(3.2) (5.2)

 

 

 

 

 

rim-"m'“*"""t2';5‘*““" “155 L;_5 I10,0 '12.5 15.0

Time (minutes)

 

 

 

Figure 2.2 Sample chromatogram of amides, corresponding amines, and aminoalcohols

The initial calibration method created used a linear fit. However, over time (7
months) the ELSD response at lower concentrations (0.003 and 0.005M) was found to be

non-linear. At these low concentrations the analyte may not form droplets readily, and

47

thus may produce a weak signal resulting in a non-linear response. A non-linear
calibration curve (quadratic fit) might be better for future experimental analyses.

In addition, the standard solutions used for the analysis of the parametric study
were unstable over time, even when refrigerated. In standard set 1, the concentration of
the 2,5-piperazinedione slowly decreased, and glycylglycine was observed and its
concentration increased slowly. The concentrations of ethanolamine, piperazine (set 1),
glycine, glycylglycine, and piperazine-Z-one (set 2) stayed constant. There was no
indication of 2,5-piperazinedione formation in set 2. Therefore, it is suggested that
standard mixes be prepared daily for any quantitative analysis.

Figure 2.3 shows a sample chromatogram of the separation of a selection of
amino acids under the same HPLC running conditions as stated previously. Full
separation of asparagine, proline, valine, cysteine, isoleucine and arginine was achieved.
Glycine, glutamine and threonine co-elute at 2.7 min, while lysine and histidine co—elute
at 8.8 min. Serine was not detected under these conditions. Further evaluation and

manipulation of this method is underway to further separate all amino acids.

48

 

mVolt;

 

 

 

 

 

 

 

 

 

Time (minutes)

l-glutamine .
g (2194) l proline (3.255)
500
l
‘ l-arginine
400 I-asparagine l- isoleucine . . . (10'0673)
a (2.603) I (6. 7116) histidine
| l- valine (88760)
300 I (4. 767)
l
200 ‘ -lcystein
(4.1914)
100
0 l 1 Ill 1 1- 1') TI T—T
l l ' 1
2.5 5.0 7.5 10.0 12.5

 

Figure 2.3 Sample Chromatogram of amino acid separation

49

 

CHAPTER THREE

CATALYTIC HYDROGENATION OF AMIDES TO AMINES

3.1 Introduction

The primary goal of this project is to optimize heterogeneous catalytic reduction
of an amide under aqueous conditions. It is the latter aspect of such a ‘green’ process
that poses the greatest technical challenge. Amide hydrolysis in water may be
competitive with or faster than hydrogenation of the amide group itself.7 Thus, the
targeted amines are formed within a mixture of products also comprising amino acids,
which themselves can be reduced to the corresponding amino alcohols. Such a reduction
of a monomeric amino acid, L-alanine, has been explored in great detail in our
laboratory. Under hydrogenation conditions (1000 psi H2, 100°C, 5% Ru/C in excess
acid) L—alanine was reduced to 2-aminol-l-propanol in 91% yield. In these studies, it
was found that the stereochemistry at the oc-carbon was retained. Selectivities and
enantiometric excesses for L-alaninol were found >95%, and >99%, respectively.” 31
These studies confirmed that the carboxylic acid must be protonated to undergo complete
reduction, as expected from simple thermochemical and pKa arguments.

From the work by Jere and others, the next step was to attempt hydrogenation of
peptides. Hydrogenation of the simplest peptide, glycylglycine, was explored briefly in
our laboratory32 under the same conditions, (1000 psi H2, 100°C, 5% Ru/C in excess
acid). As summarized in Chapter 1, a mixture of products was seen via 1H NMR

detection, showing 50% of unreacted starting material, 2% of its cyclic anhydride, 2,5-

piperazinedione, hydrolysis to glycine (25%), and hydrogenation to ethanolamine (20%).

50

Subsequent studies with 2,5-piperazinedione, under similar acidic conditions gave 25%
of the cyclic reduction product, piperazine, 29% ethanolamine, and 20% each of

ethylenediamine and ethylamine (Figure 3.1).

 

 

0
H 06 1000 psi H2 H 09 (a 09
Q N 5% Flu/C (9 OH HN
NH3/\n’ A0 : NH3/fivago + NH3/\n/ + HzN/V + NH
0 H3POJH20 0 , 0
100°C. 24 hrs 0
50% 25% 20% 2%
glycylglycine glycylglycine glycine ethanolamine 2,5-piperazinedione
0 H
1000 psi H2 6) 09 N
W NH “RU/C : ”Ha/\ll/ + H2N/\/OH + H2N/\ + HzN/VNHZ * E :l
Kn” H3PO4/H20 0 a
0 110°C, 24 hrs negligible
but detectable 29% 19% 20% 25%
2,5-piperazinedione glycine ethanolamine ethylamine ethylenediamine piperazine

Figure 3.1 Catalytic hydrogenation of glycyl glycine and 2,5-piperazinedioine with
H3PO4/H2O (lOOOpsi H2, 100°C, 1 g 5% Ru/C)

In order to understand and control these competing hydrolysis and hydrogenation
reactions further, a systematic experimental design was used to optimize conditions for
aqueous amide reduction. The starting material chosen in this study, 2,5—piperazinedione,
is a cyclic dipeptide formed from two residues of glycine, the simplest amino acid.

Figure 3.2 shows possible pathways for 2,5-piperazinedione conversion under
hydrogenation conditions. The desired route is sequential hydrogenation of 2,5-
piperazinedione via the piperazine-2-one intermediate to piperazine. Competing
hydrolysis leads to ring opening of 2,5-piperazinedione to glycylglycine, a reversible
process that favors the cyclic form at higher temperatures.”36 Glycylglycine itself
undergoes further hydrolysis to glycine monomers. Glycylglycine may also undergo

reduction in two ways: (1) at the amide, to form N-(2-aminoethyl)glycine, and (2) at the

51

carboxylate, to form glycylaminoethanol; the combination of these two reduction modes
would form N-(2-aminoethyl)ethanolamine. Finally, as found by Jere, monomeric
glycine may be reduced to ethanolamine. However, this reduction and presumably the
one in (1) above require the protonated form of the carboxylate moiety; in the absence of

1.3031

added acid, carboxylate reduction should be minima Thus, the main products

downstream from 2,5-piperazinedione hydrolysis should simply be glycylglycine and

 

 

 

glycine.
O O
2H2 2H2
(LLNH 5 (“‘NH 5 (3) ‘/\NH
HN HN HN
1% H20 \J H2O \J
O
2,5-piperazinedione piperazine-Z-one piperazine
(a (+3 ‘
G H (9 2H2 N.(2-aminoethyl)glycine$2H2 H G)
o N\[(\NH3 Y Y Ho/VNV‘NHa
glycylglycine \HONNrfiHb N-(2-aminoethyl)-ethanolamine
0
glycylaminoethanol
H201 1120-»)
G (9 2H
0 H20
glycine ethanolamine

Figure 3.2 Possible reaction pathways of 2,5—piperazinedione

Despite the apparent simplicity of the scheme in Figure 3.2, several of the
products may be formed via multiple paths, and even in processes not explicitly
represented here. There are two modes of cleavage of the inferred 1,1-aminoalcohol

intermediate formed by addition of H2 across the C=O moiety of an amide. The desired

52

path (a) releases water and forms the imine whose subsequent hydrogenation leads to the
target amine. Path (b), however, is a C-N bond cleavage process, which produces the
aldehyde and liberates the amine (or ammonia) that was bound in the amide.

One of the most challenging aspects of our preliminary amide studies was product
analysis. Quantitative NMR analysis was limited,32 by the presence of paramagnetic
metal ion impurities leached from the stainless steel reactor. In addition, the NMR
spectra of the compounds studied have many overlapping proton peaks making it difficult
to distinguish among products. In response to the difficulties of NMR-based product
analysis of amide hydrogenation, a direct method, High Performance Liquid
Chromatography (HPLC), was developed to analyze all possible hydrogenation and
hydrolysis products quantitatively (See Chapter 2).

Control of selectivity is critical to the success of any practical process. Among
these many possible pathways, a multiparameter reaction optimization was set up using a
Box-Behnken experimental design. In this approach, reaction variables are
systematically varied between selected boundary values. The treatment combinations are
at the midpoints of edges of the process space and at the center. The geometry of this
design can be envisioned as a sphere occupying a cube, with the center (0,0,0) reactions
at values corresponding together with a base set of conditions. This suggests the surface
of the sphere within the process space protrudes through each face with the surface of the
sphere tangential to the midpoint of each edge of the space. These designs are rotatable
and require 3 levels of each factor. For this experiment, three reaction conditions were
chosen; temperature, pressure, and catalyst loading (X1,X2,X3) as independent variables,

respectively. The following parameter set points were used: Temperature: 90, 110, and

53

130°C; Hydrogen pressure: 500, 1000, and 1500 psi; Catalyst loading: 0.3, l, and 3 g
(dry weight basis) of 5% Ru/C.

Results from this survey of parameters space (e.g. yields, selectivities, reaction
time to completion, costs, etc.) are then fit to an idealized empirical quadratric functional
form which describes their local dependenceon reaction conditions and reveals trends in
product yields and optimum values. Figure 3.3 shows a representation of the Box-
Behnken design. From these results, evaluations of the product yield, selectivity and

conversion functions were carried out and the optimal conditions determined.

 

 

 

 

 

 

 

 

A
Catalyst Loading 5% Flu/C
(0.3. 1. 3 g) '
i.
e
o
i ,1 ' e
.3" 0
3",
‘ .I' . ’
If
Temperature f, Pressure
(90, 110, 13090) (500, 1000, 1500 psi H2)

Figure 3.3 Box-Behnken Model

54

3.2 Experimental Procedures

3.2.] Materials
The starting substrate, 2,5-piperazinedione, celite, sodium hydroxide

(NaOH) and analysis standards, glycine, and glycylglycine were purchased from Sigma-
Alrich (St. Louis, MO), piperazine from Lancaster Synthesis Inc., (Pelham, NH),
piperazine-2-one from Alfa-Aesar, (Ward Hill, MA), and N-(2-aminoethyl)glycine from
Toronto Research Chemicals Inc. (North York, Ontario, Canada). Glycylaminoethanol is
not commercially available.

A Milli-Q ultrapure water purification system (Millipore, Billerica, MA) was used
as the sole source of water. Whatman #1 filter (42.5-mm cut to 15-mm) were purchased
from (Whatman Co., Maidstone, England). Ultra high purity H2 (AGA 99.999%) was
used without further purification.

Syringe filters (Millex-GV 0.22 am PVDF 33 mm non-sterile, Millipore,
Billerica, MA) were used to filter all samples from the reactor prior to analysis using the
HPLC. The catalyst used in these experiments was 5% ruthenium metal on activated
carbon support (5% Ru/C) (Johnson Matthey-Tennessee, Sevierville, TN). All pH
measurements were taken on a digital Chemcadet pH meter (Model 5984-50) (Cole-
Palmer Instruments Co., Chicago, IL).

All reactions were carried out in a 300 mL batch reactor (Series 4560, Parr
Instrument Company, Moline, Illinois) made of 316 stainless steel, equipped with a
magnetic driven impeller, liquid sampling condenser, gas sampling vent for real-time

analysis, and a heating mantle for controlling the temperature of reactor (:1: 1°C).

55

3.2.2 Methods

Catalyst Reduction: Prior to the start of each reaction, the catalyst (2.0136g wet
basis is equal to 1.000g dry basis) was reduced under the following conditions. The
appropriate amount of catalyst was placed into the reactor, which was then sealed, purged
three times with hydrogen, and finally pressurized to 200 psi H2. The temperature was
then set to 250°C. The reduction was carried out overnight for an average of 14 hrs, at
which time the reactor was cooled to room temperature and the remaining hydrogen was
vented.

Sample preparation: All reactions were carried out using freshly prepared 0.1M
substrate (2,5-piperazincdione) loading, dissolved in water. At room temperature, the
reactor vessel was charged with the starting substrate, and purged with hydrogen (200-
300 psi) three times. The reactor temperature was set to the desired temperature (90,
110, or 130°C). Once the reactor reached the desired temperature, the vessel was
pressurized with H2 to the desired pressure (500, 1000, 1500 or 1800 psi), and the stirring
set to 1000 rpm. A portion of the charged substrate (~2 mL) always remained in the
sidearm of the charge vessel and was used as a control for the initial concentration of
substrate, which did not come in contact with either hydrogen or catalyst.

Samples (~2mL) were taken through the sampling condenser attached to the
reactor at zero minutes, defined as the time in which hydrogen is introduced into the
system at the desired temperature, every fifteen minutes up to one hour, and every hour
thereafter up to either six or twenty four hrs. Each sample taken from the reactor was
filtered using a syringe filter to remove suspended catalyst particles, had its pH recorded,

was diluted (20-fold) to lie within the center of the calibration curve, and analyzed

56

immediately by HPLC (See Chapter 2). No calibration for ethylenediamine or AEEA
was performed so only qualitative results are reported.

Product Evaluation: For each individual experiment, reaction conversion (Eq. 1),
selectivity (Eq. 2) and product yields (Eq. 3) were calculated using the following
equations: V

% conversion = moles of reactant consumed x 100 (Eq. 1)
moles reactant fed

 

 

 

 

% selectivity = moles of product of interest = % conversion x _1_ (Eq. 2)
moles reacted % yield 100
% yield = moles of product formed x 100 (Eq. 3)

moles reactant fed

Equation 4 was used for regression analysis;
Y: b0 + b|X1 + b2X2 + b3X3 + b12X1X2 + b13X|X3 + b23X2X3 + qulz + b22X22 + b33X32 (Eq. 4)

where Y=yield, xl=pressure, X2=temperature and x3=catalyst loading

Control Experiments: Control experiments were carried out to verify the effect of
reaction parameters and intermediate product fate. A typical reaction with 2,5—
piperazinedione (0.1M) in the absence of either hydrogen or catalyst was carried out.
The reactions run in the absence of hydrogen were run with helium (1000 psi). The
catalyst was reduced as previously stated with hydrogen, but purged with helium before
charging with starting material.

Glycylglycine, piperazine-Zone and piperazine were run individually under the

following hydrogenation conditions: 1000 psi H2, 110°C, 1 g 5% Ru/C with 0.1M

57

substrate loading for six hrs. Glycylglycine and piperazine were also run with 3 g 5%
Ru/C loading.

Desorption of products: Using a Pasteur pipette, (3-5 mL at a time), the entire
reaction solution was slowly introduced into a half inch glass column with a fitted frit,
layered with filter paper (Whatman #1) glass wool and Celite (~1.5 g). House N2 was
used to push the remaining liquid through the column. Once all the catalyst was filtered
through the reaction solution, the column was eluted with hot water (90—95°C) (two times
the volume of reaction solution), followed by 1M NaOH, (volume equivalent to the
reaction volume). All filtrate fractions were clear, and therefore not filtered. Filtrate
fractions from the initial filtration and hot water washes were diluted (20x) with water.
Base washes were diluted (20x) with 10% acetonitrile (0.05% TFA) (mobile phase).

Each fraction was analyzed immediately using HPLC.

3.3 Results & Discussion

3.3.1 Box-Behnken Model

Following completion of the experiments based on the Box-Behnken model,
regression analysis predicted increased pressure, temperature and catalyst loading as the
optimal conditions for high yield of piperazine. This prediction is based from the center
of the cube using 1000 psi H2, 110°C and 1 g catalyst loading (5% Ru/C) as the base
conditions. For maximum yield at six hrs, the predicted conditions were 1000 psi H2,
111°C, and 2.7 g 5% Ru/C for 10% piperazine yield or at 1800 psi H2, 114°C, 3.8 g 5%
Ru/C, 29% piperazine yield. Figure 3.4 shows the plot of experimental yield versus

predicted yield at six hrs, (80% R2 value).

58

 

 

 

 

 

 

 

 

 

Experimental Yield

 

 

 

 

 

Predicted Yield

 

 

 

 

Figure 3.4 Regression analysis of predicted %yield versus experimental % yield

Experimental yields from the Box-Behnken model indicated lower temperature
with higher pressure and catalyst loading increased product yield. Results of the
experiments conducted using the Box-Behnken model at six hrs are shown in Table 3.1.
Optimizing for piperazine selectivity rather than yield may have been a better choice,
even if reaction rate decreased. Therefore, additional experiments were carried out
beyond the experimental design of the Box-Behnken model to verify the true optima

along with the predicted trend of laboratory data.

3.3.1.1 Eflect of Temperature

The effect of temperature at 1500 psi H2 and 3 g Ru/C is shown in entries 12, 14
& 17 (Table 3.1, Figure 3.5 A). As the temperature is increased from 90 to 110C, the
yield of piperazine increases from 8 to 23%. Increasing temperature from 110 to 130°C
decreases the yield to 15%. The effect of temperature on the rate of conversion is shown

in Figure 3.5 B. Overall, increased temperature increases the rate of conversion.

59

3. 3.1.2 Effect of Pressure

The experimental results from the Box-Behnken model at six hrs show that
increasing pressure may not necessarily increase yield (Table 3.1, Entries 14 and 19,
Figure 3.6). At both 1500 and 1800 psi H2, 110°C, 3 g 5% Ru/C, only 23% piperazine
forms. Figure 3.7 shows the effect of pressure (1000, 1500 and 1800 psi H2) on all
product yields at 90°C, 3 g 5% Ru/C. At 1000 and 1500 psi H2, piperazine-Z-one has the
highest yield overall, 6% at six hrs. At 1800 psi H2, piperazine yield is equal to
piperazine-Z-one at 4 hrs. but continues to increase, as piperazine-Z-one yield decreases,
reflecting hydrogenation of piperazine-Z-one to piperazine. At 1000 and 1500 psi H2,
glycylglycine forms at the same rate of piperazine-Z-one, but at elevated pressure (1800
psi H2), glycineglycine continues to form at 5 hrs. whereas piperazine-2-one yield starts

to decrease. In addition, ethanolamine is detected (2%) only at the elevated pressure.

3.3.1.3 Efiect of Catalyst Loading

Increased catalyst loading increased product yields at six hrs. For example,
piperazine yield increased from <1 (0.3 g Ru/C) to 7.5% (3 g 5% Ru/C) at 1000 psi H2,
90°C, and from 3.6 (l g 5% Ru/C) to 8.2% (3 g Ru/C) at 1500 psi H2, 90°C (Table 3.1.
Entries 5 & 6 and 11 & 12 respectively). Entries 13, 16 and 17, (Figure 3.8 A, B, and C)
show an increase in piperazine yield from 8 to 12 to 16% as catalyst loading is increased
from 0.3, to 1 and then 3 g 5% Ru/C at 1500 psi H2, 130°C. All intermediate product
yields are shown; at higher catalyst loading, piperazine is primarily formed, whereas

glycylglycine is the primary product formed at the lower catalyst loadings.

60

Further increases in catalyst loading from 3 g to 6 g 5% Ru/C, do not show an
increase in piperazine yield, but similar trends in product formation, at a slower rate
(Figure 3.9). Piperazine-Z—one, on the otherhand, did form faster with increased catalyst
loading (12% with 6 g 5% Ru/C) versus <10% with 3 g 5% Ru/C. With the increased
catalyst loading piperazine-2-one yield dropsrapidly at 2 hrs., whereas with 3 g 5% Ru/C

the amount of piperazine-Z-one yield is steady

61

85883on 03233936 c .8 .BmoNEE c .ofiozmsu . 3 33830 0.3 3:53—

oENEomEHm .ocoimefinfloaaumm .ocucfiocafiou/Nm
.oEobw—zobwug .059:qu AGE ocofiocfifiofimfiN $2.0 wEm: 2036.80 :05an @882.” 05 S E: 203 80:03.. =<

B: c 3 ocomvofiufiomfififi .No .530sz 05 New Boa—08m 2a No 22% “cocoon N6. oBaH

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

me so: was. 98 .3. _v 92 v m EN 82 .2
R 8% 2 we 3 o 3 o m 8 82 e:
R ‘ so: v.2 v.2 _v V 4H v m SN 82 c:
3 3e «.2 oi 3 V GS ca N SN 82 e2
8 so: 98 SN em Nv M2: v e EN 82 2
. 2. so: 3N EN 3 NV 2: _v m SN 82 c:
on $0 3; ow as NV 98 _v .3 c2 82 MN
3 G8 we S E _v E _v m 8 82 .2
R N N5 3 em on o S. o N 8 82 :
2 so: 3 I o _v 3. 3. m 3N 82 2
S 30 we 3 S o me o N SN 82 so
3 38 we 3. we NV a: NV 3 Q2 82 m
we 30 3. E 3. o 3 o 8 SN 82 e
8 68 ed 2 M: o N: c m 8 82 c
mm 8: S _v N: o N: o 3 ea 82 m
mm :5 we 3 o.» o 2: N N o2 Be a.
R 38 S 3 3 o oNN o m SN 8m m
a 3.8 St 3 V o 3. o he EN 8n &
R as 3 _v Z c 3 o N 8 8n N
3.3—am 5.— 2.3553 NR;
arse: .N 558.8 s N E 5. we 0 N3 DEN CL N NNN Nae .N bee
e... 8 22» ee .355...ch ”€338 geese:

 

 

 

 

 

fin 93.;

62

Figure 3.5 The effect of temperature (90, 110 and 130°C), on yield (A) and conversion
(B) of piperazine from 2,5-piperazinedione at 1500 psi H2, 3 g 5% Ru/C

Results are expressed as the average of duplicate (90°C) and quadruplicate (110 and
130°C) measurements. Error bars correspond to : SD

90°C (+), 110°C (+), 130°C (+)

63

 

 

 

30%

 

25%2 ,,L_-.,..__. LC , "Lififldm (_-_. __.“--.C

20% 4W .— ,-, , - , ,,,_,___ -

15%, - . ,- .5,_ WC. . .. . __ -2 _ - .. __- i ._--.--_N

°/o Yield

 

1 0% 1

 

5% 1

 

 

 

 

00/0 ' V Y I i 7 I T

0 60 120 180 240 300 360
Time (min.)

 

 

 

 

 

100%— —-~——-- ---— - - j... 4.

 

30% 4 , do--- _— __-W — _._.--.

 

600/0 i _. C M ,H____- .__,, . ,. -._,,,_ _.__H-.__-A _ .21-..- ._-..-7. , ._.-_

 

 

 

% Conversion

400/01“ __W,_ ,H-. ,__,. , ,, wfiwm,c,,,z -—

 

20% , -_ » u ,, » h__-___,_-.,-_ 2._,ML,LH._-__

 

 

 

00/0 I r 1 fl 1 T
0 60 120 180 240 300 360
Time (min.)

 

 

 

Figure 3.5

Figure 3.6 The effect of pressure (500, 1500, and 1800 psi H2) on % yield of piperazine
from 2,5-piperazinedione at 110°C, 3g 5% Ru/C

Results 1500 psi H2 are expressed at the average of quadruplicate measurements, and
1800 psi are the average of duplicate measurements. Error bars correspond to i SD

500 (+), 1500 (+), and 1800 (+) psi H2

65

 

% Yleld

 

30%

25%

20%

15% —~— ,

1 0%

5%

0%

 

1

l

 

 

 

 

l

180

Time (min.)

 

 

Figure 3.6

66

 

 

 

 

 

Figure 3.7 Product yields from the conversion of 2,5-piperazinedione at (A) 1000 (B)
1500 and (C) 1800 psi H2, 90°C, 3g 5% Ru/C

Results are expressed as the average of (B) duplicate and (C) triplicate measurements;
Error bars correspond to 2: SD. ((A) n=1)

piperazine (—O—)s Piperazine-Z-one (—I—)s glycyl glycine (-x—). glycine (+),
ethanolamine (+)

67

 

 

 

 

14%

1 2%

1 0%

8%

°/o Yield

6%

4% 1

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2%
0%
0 60 1 20 180 240 300 360
Time (min.)
14% 1 —-——- »— -- - — -- - — — - - -—-——--—-—-—------ 1'
B
12% 1— — » - -—~ - *1 11 —-- 11 --—-———
10% "
'0
'5 8% . ~ -
5
.\° 6% «— .— _— — Mm -__
40/o ., __._ ___ __ _. _ -
2% 11 1 -- 1-1-1~—.,—--1-- - /l
0% i — - 4: = =
O 60 120 180 240 300 360

Time (min.)

 

 

4%

2%

Oo/o

 

 

 

 

0 60

120

180 240

Time (min.)

300

 

360

 

Figure 3.7

68

 

 

 

Figure 3.8 The effect of catalyst loading (A) 0.3 g, (B) 1 g, (C) 3 g 5% Ru/C on all
product yields at 1500 psi H2, 130°C

Results are expressed as the average of (B) triplicate and (C) quadruplicate
measurements; Error bars correspond to i SD. ((A) n=1)

piperazine (+), piperazine-Z-one (—l—), glycylglycine (-x-), glycine (+),
ethanolamine (+)

69

 

 

250/0 ,W__W_ . . _ . .,._._...___-_-.._-_....___________...._.-.

 

2096 wn— — »—~ .-. ~~ ._,,- e~~-~.~ee»e~eeexe~en- ~e~u—-~ ,,,.,WW

315% __.._ . WW h- . . . , WWAW W .- . ... . -_W
o //
°\ 100/0 .. . _ - ._.._ , e ,- ..- _.-..__,_. . . .

 

 

 

 

 

 

 

 

 

 

see _ ._

0‘70 4 — v - : _ , — W

0 60 120 180 240 300 360
T1nne(nnth
25°/o T
B

20%- NN _ __WWDW-W_

i152“ h -H _ .. - W - W____---

.6

° 10°/o 4* _._, ,_ .7 .- WW

 

 

 

5 °/o

 

 

 

 

 

 

 

 

 

 

0%,
120 180 - 240 300 360
'Thrua(nnhto

 

 

 

 

1 4%:

1 2%

1 0%

8%

6%

961fleki

4%

2%:

 

0%)
0 60 120 180 240 300 360
'Thnn:(nnhm)

 

Figure 3.8

70

 

 

 

Figure 3.9 The effect increased catalyst loading on product yields on from (A) 3 g 5%
Ru/C to (B) 6 g 5% Ru/C at 1500 psi H2, 110°C

(A) Results are expressed as the average of quadruplicate measurements; Error bars
correspond to :t SD. ((8) n=1)

piperazine (+), piperazine-Z-one (—I-—), glycylglycine 61—), glycine (+ ),
ethanolamine (+)

71

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

A 1
25%
20%
E
.2 o,
>_15/
o\°
10%
5% 1&7/ 1 1 ‘i
0% 1 N.- = = = a T:
0 60 120 180 240 300 360
Time (min.)
25%
B

 

 

 

 

Time (min.)

 

 

 

 

 

 

 

Figure 3.9

72

 

3. 3.2 Beyond the Boundaries of Box-Behnken Model

In order to truly verify which set of conditions is optimal, experiments that lie
beyond the parameters were repeated for an extended time (24 hrs). This was done to
ensure that both 2,5-piperazinedione and piperazine-2—one, the partially reduced cyclic
product, react to complete conversion to the desired product, piperazine. It also reflects
the realization that piperazine was being destroyed at were otherwise fast conditions, and
the fact that amide hydrolysis showed substantial temperature sensitivity. Table 3.2
shows the results at 100% conversion; a lower constant temperature (90°C) and higher
catalyst loading (3 g 5% Ru/C) increases product yield from 28 to 37% at 1500 and 1800
psi, respectively (Entries 1 & 3). The change of temperature at 1500 psi H2, did not
show any increase in piperazine yield, 28%. However, increasing temperature at 1800
psi H2, decreases piperazine yield from 37% to 26% (Entries 3 & 4). The conditions at
1800 psi H2, 90°C and 3 g catalyst (5% Ru/C) loading have been found to give the
highest product yield, 37% (Entry 3).

Figure 3.10 is sample chromatogram series showing the evolution of piperazine
formation over time at optimal conditions. Peaks at 4.2 and 9.2 are systematic peaks
from the presence of metals in the reactor and changes in the acid gradient in the mobile
phase.

Figures 3.11 and 3.12 show the effect of temperature at 1500 psi H2, 3 g 5% Ru/C
and 1800 psi H2, 3 g 5% Ru/C on all product yields. At both pressure levels, the rate of
piperazine formation is slower at the lower temperature (90°C). At 1500 psi H2 and
110°C, the formation of piperazine levels off at ~27% at 12 hrs., but at 90°C, the
formation of piperazine is still increasing. This is also true with 1800 psi H2. At 90°C,
piperazine formation is still increasing at 24 hrs., but at 110°C, the yield of piperazine

73

starts to decrease at 12 hrs. The formation of ethylenediamine was detected, likely due
to piperazine cleaving at elevated temperatures. No yield for ethylenediamine is
reported.

It is also interesting to observe the fate of the intermediate products.
Glycylglycine forms at the same rate as piperazine at 1500 psi H2, 110°C for the first six
hrs, but then decreases. It would be reasonable to expect the yield of glycine would start
increase at this point, but only 2% glycine formed over the course of 24 hrs. Piperazine-
2-one at 1500 H2, 110°C, also formed at the same rate as piperazine, but after 4 hrs., the
yield dropped from 7% yield to < 1% at 24 hrs. When the temperature was lowered to
90°C, piperazine-Z-one formed at a faster rate than glycyl glycine for 5 hrs., and leveled
off at 7% yield over the 24 hrs. Glycylglycine yield slowly increase to ~12% at 12 hrs.
and then leveled off.

At the higher pressure and temperature levels (1800 psi H2, 110°C), glycylglycine
showed the same trend. At the lower temperature, (1800 psi H2, 90°C), glycylglycine
quickly formed a maximum yield of 14% at 5.5 hrs. and decreased to ~10% over the 24
hrs.

From these results, we can see the effects of temperature and H2 pressure on the
rate of formation of not only piperazine, but also the intermediate products. Control
experiments were run at the intermediate pressure and temperature conditions (1000 psi

H2, 110°C) to monitor intermediate reactions at a slower reaction rate.

74

 

Table 3.2

 

Reaction Conditions

% yield (24 hrs)

 

 

 

 

 

 

 

entry P (psi) H2 T(°C) Ru/C (g) G GG EA P2 P

1" 1500 90 3 <1 9.47 <1 <1 27.8
2 1500 110 3 1.9 5.6 2.4 0 28.1
3‘ I800 90 3 1.3 9.9 2.1 0 37.0
4" 1800 110 3 5.2 1.9 1.7 0 26.2

 

 

 

 

 

 

 

Table 3.2 Percent yield of all products for the reduction of 2,5-piperazinedione at 24 hrs.

All reactions were run at the stated reaction conditions using 0.1M 2,5-piperazinedione
EA=ethanolamine,

(PD); G=glycine,
P=piperazine

GG=glycyl glycine,

Results are expressed as a duplicate, or b triplicate measurements

75

P2=piperazine-2-one,

 

 

piperazine .

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

(10.6)
2,5-piperazinedione
250 (1.1) .
glycylglycrne
(4.6)
200 - ethanolamine
‘ (4.8)
glycrne piperazine 2-one
150_ (3:2) ,1 .1514) .,-”__ X, l _ W
, x . ii ‘ . / 1‘
, I A" A A 4‘” 2 . ‘ 'II- ' , 24
. 4/ ‘ ' 126—. . .4 I] , l"' 3 .. 12
1001- x i ' ' V . / 'j ‘0. '2 ' ' ' 10
. _, _ W. ' l _ , . j ._ .,,../..W_r.. _j/w , 1 1 1; l‘ , __ 8
. . . __ .X ] .2. ,.,1 6
' F1. / __.L / 1V / 5
50 ~ .. [fivh 7 J... I rt / 4
- _-_, . W - m... 3
T iT/yrm x ._ 1-; a A LI) [/1 "x 2
0 73' ""7 f 7 j 1
A..-/._. . 242 2.7 WW :/ “W ”Li; :“',i:,Wz;:.z _ 0
1,, I / , / / I/ / I
mVolts 2.0 5.0 8.0 11.0 12.5

 

Minutes

 

 

Figure 3.10 Representative chromatogram series of 2,5-piperazinedione hydrogenation
to piperazine at 1800 psi H2, 90°C, 3 g 5% Ru/C

Retention times (min.) are as follows: 2,5-piperazinedione 1.1, glycine 3.2,
g1ycylg1ycine4.6, ethanolamine 4.8, piperazine-Z-one 5.4, piperazine 10.6. Peaks at 4.2
and 9.2 min are systematic impurities.

76

Allvoooa $556.02 A «V nag—05:8 A * V ofiobw A x v 3633.83» A I. ofiYm-ofinflafi A o. oEnEomE

. Cu: .92 9 23:5... 2: 55 5:55 0.8 cohowno
8: 23 “ohm .Dm TI. 9 neon—mote“. was ._otm ”9:082:32: 659 23:96 no owfio>a 05 as 38230 8a 3.3.3—

03. s. w m .£ .2. 82 a 2% s .5 2338...... .8 .85 ._.m as»...

 

 

£52....
3.1 own. com. omo. com ova on. com om.q own ovm om. o
....... . q...... . _ a: ‘11) .1 . O\Oo
............. zummmmmmmmm m2... 2.. ........ ‘\_‘
....... .5.-. \W ‘xu...

 

 

 

 

 

 

 

 

 

 

 

 

 

 

V\ ..
.y _ _ . . voo- Ax6 _.
.Xii.....1wn... *m F%
x ........ K .....o .
a... $8“
. m...
.xbm
obom
okmm

 

 

 

 

 

$9.

 

 

77

d

gouache". ......a
n. ..obm ...:oEoSmao .m
E _Bim o
...... 6.8.. 28...... ...... Oceanofiwmfimfim ”swag... .9. €3.85 .3 .u
. .0 o 82:. on. ma 9
@0323... 0.3 333
m

0.3. ....n m a
m : .... 8.: ._. _._.... .... .5 8388...... .o .8...
m N; 25m...

 

 

 

 

'..II
..ICI
..-

W .............. +5 .. Ina“! «m
u‘ o
‘

.. ......» 050—.

 

 

 

 

\ h . .... $8
\\ { ...... . ......
. smm

.Y .......... \
\. ........................... . -.........
w .. $0....
$8

$0..

 

 

 

PIGIA %

 

 

 

 

 

 

 

 

 

 

 

 

 

78

3.3.3 Control Experiments

A series of control experiments was conducted to help verify intermediate product
formation, and the dependence of this process on either hydrogen or Ru catalyst (Figure
3.13). Control experiments were run under either of the following conditions; no catalyst
or no hydrogen, and using either glycyl glycine, piperazine-Z-one, normally intermediates,
or piperazine as the starting substrate.

In the absence of either hydrogen or Ru catalyst, 2,5-piperazinedione underwent
hydrolysis to 2.5% glycylglycine at (1000 psi H2, 110°C, 0 g 5% Ru/C) or (1000 psi He,
110°C, 1 g 5% Ru/C) at six hrs. Glycylglycine was the only product formed under each
of these conditions. In both cases, 2,5-piperazinedione conversion averaged 20% and
overall material balances were >80%.

Control experiments with glycylglycine (1000 psi H2, 110°C, 1 g 5% Ru/C)
showed hydrolysis to glycine (4% yield), and 8% ethanolamine at six hrs. The
conversion of glycylglycine at six hrs was 43%, and the overall material balance was
68%. When the catalyst loading was increased to 3 g 5% Ru/C, glycine and
ethanolamine formed at the same rate, with equal yields (12%) at six hrs. The overall
glycylglycine conversion was 78%, and reaction material balance was 48%. The
formation of ethanolamine is surprising, since acidic conditions are favored for the
hydrogenation of amino acids. The pH was monitored over the course of the reactions,
starting at 7 and increasing to pH 8.0 or 9.0 at 6 hrs for 1 and 3 g 5% Ru/C, respectively.

Piperazine was run under hydrogenation conditions of 1000 psi H2, 110°C, 1 and
3 g 5% Ru/C. No reaction products were detected with 1 g catalyst loading. However,
24% conversion was found, and an overall material balance of 78%. At increased
loading, 3 g 5% Ru/C, ethylenediamine (no yield reported) was the only product detected

79

at six hrs Piperazine conversion was 35%, and overall material balance was 65%. The
pH remained constant (pH 11) in both cases.

When piperazine-Z-one was reacted under hydrogenation conditions, 1000 psi H2,
110°C, 1 g 5% Ru/C, the only products detected were 20.7 % yield piperazine and 8.9 %
N-(2-aminoethyl)glycine at six hrs Piperazine-Z-one conversion at six hrs was 56%, and
overall material balance was 78%. The hydrolysis of piperazine-2-one to N-(2-
aminoethyl)glycine was detected in the control experiments only when using piperazine-

2-one as the starting material. The pH steadily increased from 8-10 over six hrs.

 

 

 

 

 

 

 

 

1000 psi H2
0 o g 5% Flu/C 0
110°C, 6 hrs. 6)
5;, .....W.-
OR 1000 psr He H O
o 1 g 5% Ru/C
2,5-piperazinedione 110°C. 6 hrs. glycylglycine 2.5 %
1000 psi H2, 0
1 g 5% Ru/C 9 H2N\/\
= H + OH
110°C. 6 hrs. 3 \/U\oe
O o o
(*3 C9 4 /o 8 /o
H3N\/U\N/firo glycine ethanolamine
H .
O 1000 psr H2,
glycylglycine 3 9 5% Flu/C : 12% 12%
110°C. 6 hrs.
(\NH 1000 pSl H2, OH
HM") 1 9 5% Flu/C: HN/W + H2N\/\N/\n/
110°C. 6 hrs. K/NH H o
O
piperazine 2-one 20.7% 8.9%
piperazine N-(2-Aminoethyl)glycine
1000 psi H2, H N
2 \/\
HN/\| 3 g 5% Ru/C _ "“2
NH 110°C, 6 hrs.w ethylenediamine
i erazine
p p 1000 psi H2,
1 9 5% Flu/L no products deteted
110°C. 6 hrs.

Figure 3.13 Product formation from control experiments

80

Gas chromatographic analysis of reactions run at optimal conditions did not detect
low molecular weight volatiles (CH4, NHg, etc.). The pH was monitored over the course
of all reactions. The evolution of piperazine formation from 2,5-piperazinedione by the

increase of pH from 7 to 11, at 1800 psi, 90°C, 3g Ru/C is shown in Figure 3.14.

 

11 —-——

 

 

10.5--~~---~~ - -— —» ~ -~

 

 

 

 

 

 

'7 - - - - A- _._ ._....— -

6.5 1 — ~ — ~ ~ —— ~ — - -— —-—- ——-- —

6 r r T r r I r r I r 1 r
O 120 240 360 480 600 720 840 960 1080 1200 1320 1440

 

 

Time (min.)

 

Figure 3.14 Change of pH over time from the conversion of 2,5-piperazinedione to
piperazine at 1800 psi, 90°C, 3g 5% Ru/C

Results are given as the average of triplicate experiments; Error bars correspond to :1: SD.

3.3.1.4. Preliminary Studies on the Efiect of Carbon Support

It is necessary to point out that the overall material balance was not closed. In all
the experiments carried out, the initial concentration of the starting material (2,5-
piperazinedione, glycylglycine, piperazine-Z-one or piperazine) decreased at time zero
(t=O), or in other words, the nominal % conversion increased. Between the initial

recording and t=0, the reactor was charged with the substrate and heated to the desired

81

temperature, at which time hydrogen was introduced. It is likely that during this time the
substrate adsorbed to the carbon support.

The observed concentration decrease, (or conversion increase), of 2,5-
piperazinedione at t=0, changed with catalyst loading. The conversion at t=0, for 2,5-
piperazinedione (0.1M) is <1 (0.3g Ru/C), 25 (1.0 g Ru/C), and 40% (3.0 g Ru/C) at
1500 psi H2, 130°C. At increased catalyst loading (6.0 g Ru/C), the conversion of 2,5-
piperazinedione (0.1M) is 40%, at 1500 psi H2, 110°C. Figures 3.15 and 3.16 show 2,5-
piperazinedione conversion over time of at the different catalyst loadings. At t=0, the
three catalyst loadings increase the apparent initial conversion by 25 and 41%, when
ultimately no reaction has yet taken place; thus, % conversion should equal zero.

In Figure 3.16 the level of substrate loading can be compared with catalyst
loading. Even when twice the amount of substrate is charged, with equal amount of
catalyst, % conversion at t=0 differs by 15%. At equal substrate loading, with 3 and 6 g
5% Ru/C, % conversion at t=0 is practically equal. This suggests that if 2,5-
piperazinedione is adsorbing to C, it may be a fully saturated with 3 g 5% RulC.

From the control experiments, <1 and 11% conversion at t=0, was noticed for
glycylglycine with l and 3 g 5% Ru/C, respectively. At t=0, % conversion for
piperazine-Z-one was 18%, and piperazine was 12 and 30% with l g and 3 g 5% Ru/C,

respectively.

82

Figure 3.15 The effect of catalyst loading on initial % conversion of 2,5-piperazinedione
at 1500 psi H2, 130°C, with 0.3, 1, and 3 g 5% Ru/C

Results for catalyst loadings of 1 g Ru/C are expressed at the average of triplicate
measurements, and catalyst loadings of 3 g 5% Ru/C are the average of quadruplicate
measurements; Error bars correspond to :tl SD. (Catalyst loading of 0.3 g 5% Ru/C,
n=l)

0.3 g (+), 1 g (-I- ), and 3 g (+) 5% Ru/C

Figure 3.16 The effect of catalyst loading on initial % conversion of 2,5-piperazinedione
(0.1 and 0.05M) at 1500 psi H2, 110°C, with 3 and 6 g 5% Ru/C

Results for catalyst loadings of 1500 psi H2, 110°C, and 3 g 5% Ru/C at 0.1M substrate
loading are expressed at the average of quadruplicate measurements; Error bars
correspond to i1 SD. (Catalyst loading with 6 g 5% Ru/C with 0.1M substrate loading,
n=1, and 3 g 5% Ru/C with 0.05M substrate loading, n=1)

0.3 g 5% Ru/C with 0.1M substrate loading (+), 6 g 5% Ru/C with 0.1M substrate
loading (+), and 3 g 5% Ru/C with 0.05M substrate loading (+)

83

 

 

1 00%

 

 

80%—5 ~

ran

60% -

% Convers
8
°\°

20% - -

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0% .
o 50 100 150 200 250 300 350 400
Time (min.)
100% 4
80% — , ————
C
.2
2 60% — »
0
>
C
8 40% 1
.\°
20%.--- -_.__.,,,-_.__,___ ,, ——— ~ ——~ --_ * ~~——
0% l l l l I l l
o 50 100 150 200 250 300 350 400
Time (min.)
Figure 3.16

84

 

 

From these observations, it was necessary to determine if the starting material,
2.5-piperazinedione, or any of the intermediate products is absorbed within the pore
structure of the activated carbon. A series of absorption/desorption experiments were
carried out with 2.5—piperazinedione and piperazine.

Under reaction conditions (1800 psi H2, 90°C, 3 g 5% Ru/C) 2,5-piperazinedione,
on average % conversion at t=0 was 40%. A typical experiment was carried out, with
samples only taken at t=0 and t=24 hrs. No starting material recovered from desorption
washes, but an additional 0.01 glycine, 0.04 glyclyglycine, 0.01 ethanolamine, and 0.90
mmol piperazine were recovered. This increases the overall piperazine yield 10%.
Bench top control experiments with equal mixtures of 2,5-piperainedione and piperazine
(0.05M each) with 0.75 g 5% Ru/C (unreduced) was equilibrated overnight at room
temperature. Approximately 50% of each absorbed onto the carbon support at
equilibrium. Desorption washes recovered 84% 2,5-piperazinedione and 74%
piperazine. Individual studies on the remaining compounds were not carried out, but are
suggested. Detailed studies on the effect of the carbon support, and perhaps the
comparison of other supports for each of the compounds studied in this study should also
be carried out. The question of how much trapped material continues to react is

undetermined, and supports future research for this portion of the study.

3. 3.4 Reaction pathways for 2,5 -piperazinedione hydrogenation to piperazine

Previous studies suggest that amides simply hydrolyze under aqueous conditions,7
and that the equilibrium between glycylglycine and 2,5-piperazinedione is shifted toward

the latter at high temperatures.36

We have found through a rigorous, systematic study,
that 2,5-piperazinedione can be reduced under aqueous catalytic hydrogenation

85

conditions. It is important to point that, that the reduction of an amide has been achieved
under relatively mild and more importantly ‘green’ working conditions. The results from
this study complement and in some cases may even replace the existing classical
approach of amide reductions using organic solvents and costly metal hydrides.

From our results, we observed both hydrolysis and hydrogenation products of 2,5-
piperazinedione. In order to understand this competing reaction sequence we must first
understand the mechanism. Our results confirm hydrolysis of both 2,5-piperazinedione
and glycylglycine under aqueous conditions. In the absence of hydrogen (1000 psi He),
at 110°C with l g 5% Ru/C, hydrolysis of 2,5-piperazinedione to 20% glycylglycine was
detected at 6hrs. At 1000 psi H2, 110°C in the absence of catalyst, 20% glycylglycine
was the only product detected at six hrs This verifies that the ruthenium catalyst is
essential for hydrogenation. This conclusion is also supported by the results found when
comparing different levels of catalyst loading.

The level of hydrogen pressure also plays a critical role, with increased hydrogen
pressures favoring hydrogenation over hydrolysis, presumably due to increasing

37 At the lower

hydrogen concentration in aqueous media at increased pressures.
hydrogen pressures, hydrolysis was the primary reaction.

Temperature was also found to play a role in the rate and stability of piperazine
formed. At elevated temperature (110°C) the yield of piperazine decreased (See Figure
3.13) and ethylenediamine was detected (no yield recorded). From the control
experiments, the cleavage of piperazine to ethylenediamine was the only product

detected. The formation of ethylenediamine was observed in previous studies in our

laboratory.32

86

In order to gain insight in the mechanism of the hydrogenation of 2,5-
piperazinedione, hydrogenation of several amides was modeled via ab initio methods in
both gas phase and solution; results are summarized in Figure 3.17. As expected,
hydrogenation is more favored in solution. The initial hydrogenation step is slightly
harder for amides than acids, but the second step is more favorable for amides than for
the corresponding acid. Cyclic piperazinedione was only slightly easier to hydrogenate
than the acylic amide. The dehydration of secondary acyclic amides is favored over C-N
cleavage (+6.7 versus +13.l kcal/mol) to the aldehyde.

The cyclic anhydride, 2,5-piperazinedione, undergoes hydrogenation to
piperazine via a route that is shown in Figure 3.18. The formation of piperazine-Z-one is
believed to be by a Schiff base intermediate which can either undergo hydrogenation
further to the final product, piperazine, or hydrolyze to form N-(2-aminoethyl)glycine.
The hydrolysis product of piperazine-Z-one was not detected over the course of the
reaction, but only in the control experiments. It is possible that further C-N hydrolysis of
N-(2-aminoethyl)glycine may lead to the trace amount of the ethanolamine detected. It is
still undetermined at this point.

Figure 3.19 shows the proposed hydrolysis of 2,5-piperazine to glycylglycine,
which can further cleave to glycine. Glycylaminoethanol, N-(2-aminoethyl)glycine and
N-(Z-aminoethyl)-ethanolamine were not formed at all over the series of reactions carried
out.

The result of changes in initial conversion of the starting material at t=0, may be
attributed to the overall low material balance. It is possible the starting material, or

intermediate compounds are absorbing in the activated carbon catalyst support. Further

87

quantitative investigation of each products absorption to the carbon support is necessary
to determine exactly how much of the material is adsorbed, and whether adsorbed
material remains accessible for reaction. These assessments of overall material
availability during the reaction may help solidify the correct mechanistic picture for the

formation of piperazine.

88

5.5.8 E Ba 8...... 8w 5 assessmeac 02:8 .8. Ear—8:88 882.89 .0 $.33: E...- 0.53".

89

8.8-Kam- 8...+.m.m.+ 8.2 a... 8.9.3.... 8+...»- 2.8.3.8...
8.8-ZS- 8...... 8.8.. 8.-. to- 8.8.3.... 8+:.+8 68.8538...
8.8.5.... 8.8+. 3+ 2.18.? calm..- 8.~+.m..+ fuzz ...:0
8.5.3“- 3.9-5.2+ 2.8-. 8.8. 8.8-.....- 8.38.?- fz £0
8.810....- .....+. ...? 8.18..- 8.8-K...- c.o+:....+ :0 ...-.0
Arts .5 3.3 is 3.3 .5. 3.3 :2 3.3 .5. :x c
Ix m x m 1x a XI IX :0 m
...j 1.4 11.811.318.11 ..
...: on: NI of 1x:
8.9-3.9- .10
3.3.? m
Arms-Ki.- forz ”Io
8.01.8.5- N...z ..xo
8.38.?- IO .10
z a 2:34... 5. ....

I
I 1
xx: +
.../n4“ . J IXI + 4‘

o
N:

ocofio:_§uo&m-m.m mo noun—homage»: 88 @3353 5:82 38:55 afim DEE-m

89028 9:85 89: 828.8 85

328.0ch .ocm£8c_Eu_>o>_o

 

O
NIZ\/\OI QQQ>8_O NIZ/\=\s.sZ\/\OI
9.09 2-0 I
A8598 Hos

282235855982

sz/\/ o:
2
Q

@9320
neon 2-0

N
I

 

0
Z
I

I?
2
£7;\_

0“:-

 

+
5F
—’O

oco-chfiSan

 

 

 

 

 

 

O O O O O 0
W52: WFz: -: sz: WFz: _\Fz.._ Krzx A
. N - _V #2 N :2 :2 +1
Iz/L I zflL +1 :mflL .o. @101 I @I/o 14%
O I Q
oEEmfiocmS£m mcfifioaa ace-mmcfifioaa 05.35359de
O
O
ZNI A 929362 2: - \FZI
NIz\/\ 8656 IO . Q: 4 5333052.. 525 I2
IZ

O

90

eco€o=_nfloma-m.m .6 Ema—Bu»: new @8353 .53an 382 36 Sam

Emmi ocfimacgm-:>£8c_&9mv-z

 

 

osozzsfioctaé-z 8558228393
0
m N
86.2% 6: NT_z/\/z\/\o: Alnl :z/\/2\J.\o: + Iz/\=/z\/\o:
I I O I
620208 9:385 89: o\o :95 VT
32:90:26 @5038 SEBEB
O O O O
om Xfiz: WFz:
o
mIZ\/\OI +081 .mI I%\/n_ud\ $9.993 ANIz

z: TF2...
A ...-... .X ...- .a l .2
282$ oo :0 :0 9

8829328 ocsuocfiewaa-md

ZI
O ONI I231
O

‘l

o 2,392; o
NIZ /\=/Z OI UCOD 2-0 _\=/
I

 

91

CHAPTER FOUR

SUMMARY AND CONCLUSIONS

Optimized operating conditions for the hydrogenation of 2,5-piperazinedione to
piperazine have been established. Experimental conditions at 1800 psi H2, 90°C and 3 g
catalyst (5% Ru/C) loading have been found to give the highest piperazine yield, 37%, at
24 hrs. Not only are the hydrogenation product yields of higher than that of hydrolysis,
the reaction was completed under ‘green’ conditions. One of biggest challenges in this
study is overcoming the competing hydrolysis reaction. It has been shown that the
hydrogenation takes place as the primary reaction with the formation of the intermediate
piperazine-Z-one before the hydrolysis product glycylglycine. With the optimized
conditions established, a more detailed mechanistic study can be completed.

It is important to note that optimization of the analytical methods continued
throughout the course of the hydrogenation experiments. Careful evaluation of the
analytical results concluded that all data are, in fact, valid for the hydrogenation
experiments. This is based on the use of one single calibration method created and used
for all of the data interpretations. More importantly, a portion of the actual substrate
charged into the reactor was saved and used for all calculations of the parametric study.
Having this “control” sample and accounting for the stability of the desired product,

piperazine, all conclusions remain as stated.

92

CHAPTER FIVE

FUTURE RESEARCH
Some suggestions for future research are listed below:

1. Optimized conditions for the reduction of the cyclic glycine anhydride have been
established. These conditions should now be carried out on linear peptides (e.g.
glycylglycine, alanylalanine, glycylalanine etc.)

2. Ruthenium catalyst was supplied in its pre-rcduced form and needed to be fully
reduced (14-16hrs.) before running actual experiments. A brief study on the
effect of catalyst reduction time should be carried out in order to make research
time more efficient.

3. Only heterogeneous ruthenium catalyst was used for this study. A survey of other
metal catalysts (e.g. Re, Rh, Ni, Mo, Pd, Cu) should be sought out (heterogeneous
and homogeneous). In addition, synergistic metal effects should be explored.

4. Activated carbon support was used in this study for the metal catalyst. This
support was found superior for the catalytic hydrogenation of carboxylic acids in
our laboratory. However, initial concentration of the starting material was found
to decrease before the reaction conditions started. In addition, overall material
balances were not closed. Absorption studies of each of the compounds involved
in the reaction should be carried out on the activated carbons. A survey of

catalyst supports may also be beneficial and help close the material balance.

93

5. Continued analytical method development for linear peptides should be carried

out for immediate qualitative and quantitative product analysis.

94

LIST OF REFERENCES

l. Challis, B. C.; Challis, J. A., The Chemistry of Amides. Interscience: New York,
1970; p 731-857.

2. Gaylord, N. 6., Reduction with Complex Metal Hydrides. Intersciences
Publishers, Inc.: New York, 1956; p 544-642.

3. Brown, H. C.; Heim, P., Selective Reductions.18. Fast Reaction Of Primary,
Secondary, And Tertiary Amides With Diborane - Simple, Convenient Procedure For
Conversion Of Amides To Corresponding Amines. Journal Of Organic Chemistry 1973,
38, (5), 912—916.

4. Wojcik, B.; Adkins, H., Catalytic hydrogenation of amides to amines. Journal Of
The American Chemical Society 1934, 56, (7), 2419-2424.

5. Oeda, H., Catalytic hydrogenation of amides of alpha-hydroxy acids. Bulletin of
the Chemical Society of Japan 1937, ( 12), 121-127.

6. Winans, C. F.; Homer Adkins, Alkylation of Amines as Catalyzed by Nickel.
Journal of the American Chemical Society 1932, (54), 306-312.

7. Zabicky, J., The Chemistry of Amides. Interscience: New York, 1970.

8. Newman, M. S.; Fukunaga, T., The Reduction Of Amides To Amines Via Nitriles
By Lithium Aluminum Hydride. Journal Of The American Chemical Society 1960, 82,
(3), 693-696.

9. Brown, H. C.; Heim, P., Diborane As Mild Reducing Agent For Conversion Of
Primary Secondary + Tertiary Amides Into Corresponding Amines. Journal Of The
American Chemical Society 1964, 86, (17), 3566-&.

10. Komet, M. J .; Thio, P. A.; Tan, S. 1., Borane Reduction Of Amido Esters. Journal
Of Organic Chemistry 1968, 33, (9), 3637-&.

11. Umino, N.; Iwalcuma, T.; Nobuo Itoh, Sodium Acyloxyborohydride as New
Reducing Agents. 1. Reduction of Carboxamides to the corresponding Amines.
Tetrahedron Letters 1976, (10), 763-766.

12. Zhu, H.-J.; Lu, K.-T.; Sun, G.-R.; He, J.-B.; Li, H.-Q.; Charles U. Pittman, J.,
Reduction of Amides with NaBH4 in diglyme at 162C. New J Chem 2003, 27, 409413.

13. Wann, S. R.; Thorsen, P. T.; Kreevoy, M. M., Reduction Of Carboxylic-Acid

Derivatives By Bh4- In Acidic Dimethylsulfoxide. Journal Of Organic Chemistry 1981,
46, (12), 2579-2581.

95

14. Kuehne, M. 13.; Shannon, P. J ., Reduction Of Amides And Lactams To Amines
By Reactions With Phosphorus Oxychloride And Sodium-Borohydride. Journal Of
Organic Chemistry 1977, 42, (12), 2082-2087.

15. Satoh, T.; Suzuki, S.; Suzuki, Y.; Miyaji, Y.; Imai, Z., Reduction Of Organic
Compounds With Sodium Borohydride-Transition Metal Salt Systems. 1. Reduction Of

Organic Nitrile, Nitro And Amide Compounds To Primary Amines. Tetrahedron Letters
1969, (52), 4555-&. '

l6. Raucher, S.; Klein, P., A Convenient Method For The Selective Reduction Of
Amides To Amines. Tetrahedron Letters 1980, 21, (42), 4061-4064.

17. Borch, R. F., A New Method for the Reduction of Secondary and Tertiary
Amides. Tetrahedron Letters 1968, (1), 61-65.

.18. Rahman, A.; Basha, A.; Waheed, N.; Ahmed, S., New Method For Reduction Of
Amides To Amines With Sodium-Borohydride. Tetrahedron Letters 1976, (3), 219-222.

19. Komfeld, E. C., Raney Nickel Hydrogenolysis Of Thioamides - A New Amine
Synthesis. Journal Of Organic Chemistry 1951, 16, (1), 131-138.

20. Bazant V; Capka M; Cemy M; Chvalovsky V; Kochloefl K; Kraus M; Malek J,
Properties of Sodium-bis-(2-methoxyethoxy)aluminiumhydride. 1. Reduction of some
Organic Functional Groups. Tetrahedron Letters 1968, (29), 3303-3306.

21. Bhandari, K.; Sharma, V. L.; Chatterjee, S. K., A convenient method for the
reduction of amides to their corresponding amines. Division of process development,
Central Drug Research Institute, India 1990.

22. Adkins, H.; Wojcik, B., Hydrogenation of amides to amines. Journal Of The
American Chemical Society 1934, 56, 247-247.

23. D'Ianni, J. D.; Homer Adkins, Hydrogenation of Hydroxyamides- Journal of the
American Chemical Society 1939, (61), 1675-1681.

24. Schneider, H. J .; Adkins, H.; McElvain, S. M., The Hydrogenation Of Amides
And Ammonium Salts - The Transalkylation Of Tertiary Amines. Journal of the
American Chemical Society 1952, 74, (17), 4287-4290.

25. Sauer, J. C.; Adkins, H., The selective hydrogenation of substituted amides.
Journal Of The American Chemical Society 1938, 60, 402-406.

26. Schwoegler, E. J .; Homer Adkins, Preparation of Certain Amines. Journal of the
American Chemical Society 1939, (61), 3499-3502.

96

27. Hirosawa, C .; Wakasa, N.; Fuchikami, T., Hydrogenation of amides by the use of
bimetallic catalysts consisting of Group 8 to 10, and Group 6 or 7 metals. Tetrahedron
Letters 1996, 37, (37), 6749-6752.

28. Kuwano, R.; Takahashi, M.; Ito, Y., Reduction of amides to amines via catalytic
hydrosilylation by a rhodium complex. Tetrahedron Letters 1998, 39, (9), 1017-1020.

29. Igarashi, M.; Fuchikami, T., Transition-metal complex-catalyzed reduction of
amides with hydrosilanes: a facile transformation of amides to amines. Tetrahedron
Letters 2001, 42, (10), 1945-1947.

30. Jere, F.; Jackson, J .; Miller, D. J ., Kinetics of the Aqueous-Phase Hydrogenation
of L-Alanine to L-Alaninol. Industrial Engineering and Chemical Research 2004, 43,
3297-3303.

31. Jere, F. T.; Miller, D. J .; Jackson, J. E., Stereoretentive C-H bond activation in the
aqueous phase catalytic hydrogenation of amino acids to amino alcohols. Organic Letters
2003, 5, (4), 527-530.

32. Redko, M., Unpublished results. Michigan State University 2004-2005.

33. Steinberg, S.; Bada, J. L., Diketopiperazine Formation During Investigations Of
Amino-Acid Racemization In Dipeptides. Science 1981, 213, (4507), 544-545.

34. Steinberg, S. M.; Bada, J. L., Peptide Decomposition In The Neutral Ph-Region
Via The Formation Of Diketopiperazines. Journal Of Organic Chemistry 1983, 48, (13),
2295-2298.

35. Radzicka, A.; Wolfenden, R., Rates of uncatalyzed peptide bond hydrolysis in
neutral solution and the transition state affinities of proteases. Journal Of The American
Chemical Society 1996, 118, (26), 6105-6109.

36. Li, J .; Brill, T. B., Spectroscopy of hydrothermal reactions. 27 . Simultaneous
determination of hydrolysis rate constants of glycylglycine to glycine and glycylglycine-
diketopiperazine equilibrium constants at 310-330 degrees C and 275 bar. Journal Of
Physical Chemistry A 2003, 107, (41), 8575-8577.

37. Zhang, Z. 6.; Jackson, J. E.; Miller, D. J ., Aqueous-phase hydrogenation of lactic
acid to propylene glycol. Applied Catalysis A-General 2001, 219, (1-2), 89-98.

97

IIIIIIIIIIIIIIIIIIIIIIIIIIIIII

”ll-Willi“Willi””ill-“1’"