. oaU-V .. .
... 4,..."

4.4M! ~

‘ .1: 2...: I
IL 513.}:
o 3.3..
.7: In}...
. .5 .
4

, ‘ _ an»? F.
. . , «ﬂawwumnw
an. nu}. ' a... . . . , . a... It.
0; 53.. x! .55. K... . . . .
. run-45‘353. . .
51.05.52.316... . .
an...» Andi»? ,

 

 

ILS‘ \!
5.0:... v ~
iii:

i235.
R‘vgil

as .0. fag-lift”;
um ”Mg

rlvun‘.%mlz

33" , 1.531096.“

5‘!

’9‘}:
,r Ilkv nil]. ,
u :3. ’1. 1:91:03
i»? ax: I‘ie :2‘5311 g u
.x} 21022;»

 

 

 

LIBRARY

Michigan State
University

 

 

 

This is to certify that the
dissertation entitled

CATALYTIC MEMBRANES PREPARED BY ADSORPTION
OF POLYELECTROLYTE/METAL NANOPARTICLE FILMS IN
POROUS SUPPORTS

presented by

DAVID M. DOTZAUER

has been accepted towards fulﬁllment
of the requirements for the

 

 

 

 

Ph.D. degree in Chemistry
/ ~ \
Major Professor’s SigWe
Date

MSU is an Afﬁrmative Action/Equal Opportunity Employer

 

ln-I-a-I-u-Q-I-0-I-I-v-|-I--o-o-o-o-I-'-I-I-I---0-I-I-I-l-l—l-I-I-o-v-l-l-n-O-I-l-l-'-l-O-D-U-I-D-o-I-I-l-I-I---—i— - A

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

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5/08 KIIProjIAcc8-PreleIRClDateDue.indd

 

CATALYTIC MEMBRANES PREPARED BY ADSORPTION OF
POLYELECTROLYTE/METAL NANOPARTICLE FILMS IN POROUS
SUPPORTS
By

David M. Dotzauer

A DISSERTATION
Submitted to
Michigan State University
in partial fulﬁllment of the requirements
for the degree of
DOCTOR OF PHILOSOPHY
Chemistry

2009

ABSTRACT
CATALYTIC MEMBRANES PREPARED BY ADSORPTION OF

POLYELECTROLYTE/METAL NANOPARTICLE FILMS IN POROUS
SUPPORTS

By
David M. Dotzauer

Well-deﬁned metal nanoparticles are attractive for catalytic applications
because of their high speciﬁc surface area and unique electronic properties.
However, the high surface energy of these particles often leads to their
aggregation so the nanoparticles must be capped with a stabilizer or immobilized
on a support. This dissertation demonstrates metal nanoparticle immobilization
in porous membranes via layer-by-layer (LbL) adsorption of polyelectrolyte/metal
nanoparticle ﬁlms and the potential advantages of these membranes for catalytic
applications including reduction of nitro compounds, hydrogenation, and wet air
oxidation for wastewater treatment. The immobilization method utilizes
electrostatic interactions to facilitate deposition of oppositely charged
polyelectrolytes and nanoparticles, and controlled synthesis of the nanoparticles
prior to their deposition allows tuning of the nanoparticle size and composition for
. Speciﬁc catalytic processes. Nanoparticles can also be formed after ﬁlm
deposition by incorporating metal ions into the polyelectrolyte ﬁlm and later
reducing these metal ions to form nanoparticles with diameters of 2-4 nm.

LbL deposition of polyelectrolyte/Au nanoparticle ﬁlms in porous alumina
and polymer substrates yields catalytic membranes with a high density of well-
separated Au nanoparticles in the membrane pores. These nanoparticles

catalyze the reduction of nitroaromatic compounds by sodium borohydride with

rate constants that are the same as those for nanoparticles in solution.
Moreover, the membranes selectively catalyze the reduction of nitro groups in
compounds containing other reducible functionalities such as cyano, chloro, and
styrenyl moieties. These membranes are particularly attractive for controlling
product distributions through variation of solution ﬂuxes.

Membrane catalysts also facilitate contact between reactants in gas/liquid
reactions by operating as ﬂow-through or interfacial contactors. LbL deposition
of polyelectrolyte/Pd nanoparticle ﬁlms in ﬂat membranes yields ﬂow-through
contactors that effectively hydrogenate allyl alcohol to 1-propanol when ﬂowing
Hz-saturated solutions through these membranes. Hydrogen reacts completely
after one pass of the solution through the modiﬁed membranes, and thus the
reaction is limited by the solubility of H2 in the reactant solution. The interfacial
contactor conﬁguration, in which a tubular membrane separates the gas and
liquid phases, provides enhanced contact between the immobilized catalytic
nanoparticles and the gas/liquid interface. Control over the catalyst location in
such membranes is crucial for efﬁcient use of noble metals. LbL adsorption of
polyelectrolyte/Pt nanoparticle ﬁlms in tubular ceramic membranes allows
deposition of the catalytic nanoparticles only near the interior of tube, where the
gas/liquid interface is typically located. In wet air oxidation of formic acid, acetic
acid, and phenol, tubular membranes modiﬁed by LbL deposition showed 2 to 10
times higher speciﬁc activities than similar membranes modified by conventional
impregnation techniques, presumably due to. controlled deposition of the Pt

nanoparticles in the membrane.

 

To my wife, Heather, for all her love and support

 

ACKNOWLEDGEMENTS

l’d ﬁrst like to thank my advisor, Dr. Merlin Bruening, for being a great
advisor over the last four years. He has offered guidance when needed, but has
also forced me to think and work independently so that I may learn. I appreciate
his patience, when things were not working, his lack of patience, when things
were working, and his positive encouragement throughout the entire process.

I would like to express gratitude to past and present members in the
Bruening group for their help and suggestions in the last four years. I am
especially grateful to Dr. Somnath Bhattacharjee, Dr. Jinhua Dai, Dr. Lei Sun,
and Dr. Lu Ouyang for their direct involvement in my research projects. I would
also like to thank Ya Wen, Stephanie Koszalka, and Katie Worley for their
contributions to this work.

I want to thank Dr. Sylvain Miachon for being my co-advisor and giving me
the opportunity to work with him at the Institute for Researches on Catalysis and
the Environment in Lyon, France. He was a remarkable colleague and will be
greatly missed.

Thank you to the MSU Department of Chemistry, the National Science
Foundation, and the American Chemical Society for funding my research and
providing me with many great opportunities during my time at Michigan State.

Lastly, I want to give a special thank you to my family for all of their
support through the years and especially my wife, Heather, for her continuous

encouragement and understanding.

TABLE OF CONTENTS

 

 

 

List of Tables ix
List of Figures xi
List of Schemes xix
List of Abbreviations xx

 

Chapter One: Introduction

 

1.1 Introduction to Catalysis

 

1.2 Nanoparticles and Catalysis

 

1.3 Common Metal Nanoparticle Supports

 

1.3.1 Metal Oxide Supports
1.3.2 Carbonaceous Supports

 

1.3.3 Polymer Supports

 

1.3.4 Dendrimer-Encapsulated Metal Nanoparticles
1.4 Porous Membranes as Catalyst Supports
1.4.1 Membrane Extractors

 

1.4.2 Membrane Distributors
1.4.3 Membrane Contactors
1.4.3.1 lnterfacial Contactors

 

1.4.3.2 Flow-Through Contactors
1.5 Formation of Metal Nanoparticles

 

1.5.1 lmpregnation

 

1.5.2 Deposition Precipitation

 

1.5.3 Coprecipitation

 

1.5.4 Chemical Vapor Deposition

 

1.5.5 Synthesis of Unsupported Metal Nanoparticles

 

1.5.6 Surface Modiﬁcation with Preformed Metal Nanoparticles

 

1.6 Layer-by-Layer Assembly of Polyelectrolyte Multilayer Films

 

1.7 Scope of This Work

 

1.8 References

 

Chapter Two: Catalytic Reduction of Nitroaromatic Compounds with

 

Au-nanoparticle Containing Membranes
2.1 Introduction

 

2.2 Experimental Methods

 

2.2.1 Materials

 

2.2.2 Modification of Aluminum Wafers with Polyelectrolyte/AuNP Films ,,,,,,,

2.2.3 Modification of Aluminum Oxide Powder with Polyelectrolyte/AuNP
Films

vi

48
48
51
51
52

53

v

 

 

 

 

 

 

2.2.4 Modiﬁcation of Disk-Shaped Membranes with Polyelectrolyte/AuNP

Films 54
2.2.5 Characterization of Polyelectrolyte/AuNP Films 55
2.2.6 Catalytic Reactions with As-Prepared Au Nanoparticles and Au

Nanoparticles Supported on Alumina Powder 56
2.2.7 Catalytic Reactions with Au Nanoparticle Containing Membranes ,,,,,,,,,,,,, 58
2.3 Results and Discussion 58
2.3.1 Characterization 58
2.3.1.1 Au Nanoparticle Characterization 58
2.3.1.2 Characterization of Polyelectrolyte/AuNP Films on Aluminum ,,,,,,,,,,,,,, 60
2.3.1.3 Characterization of Polyelectrolyte/AuNP Films in Membranes ,,,,,,,,,,,, 64

2.3.2 Catalytic Reduction of 4-Nitrophenol with Au Nanoparticles in Solution_,_,72
2.3.3 Catalytic Reduction of 4-Nitrophenol with Membrane-Supported Au

 

 

 

 

 

 

 

 

 

 

 

Nanoparticles 75
2.3.3.1 4-Nitrophenol Reduction as a Function of Film Composition ,,,,,,,,,,,,,,,,, 79
2.3.3.2 Comparison of the Catalytic Activities of Au Nanoparticles in

Alumina, Nylon, and Polycarbonate Membranes 82
2.3.4 Effect of Substituent Groups on the Reduction of Nitroaromatic

Compounds 84
2.3.4.1 Reduction of Compounds with Single Reaction Products ,,,,,,,,,,,,,,,,,,,,,, 84
2.3.4.2 Reduction of Compounds with Multiple Reaction Products ,,,,,,,,,,,,,,,,,,, 86
2.3.4.3 Effect of Flux on Product Distributions 89
2.3.4.4 Selectivity in Reactions with Multiple Reducible Functional Groups_"_91
2.3.4.5 Catalytic Reduction of Nitrocyclohexane 92
2.3.4.6 Effect of Temperature on Reaction Rates and Product

Distributions 93
2.3.5 Catalyst Stability 95
2.4 Conclusions 98
2.5 References 99
Chapter Three: Hydrogenation Reactions with Pd-Containing Catalysts,105
3.1 Introduction 105
3.2 Experimental Methods 107
3.2.1 Materials 107

 

3.2.2 Modiﬁcation of Alumina Powder with Polyelectrolyte/PdNP Films ,,,,,,,,,,,, 108
3.2.3 Modiﬁcation of Alumina Membranes with Polyelectrolyte/PdNP Films,_,,111

 

 

 

 

3.2.4 Characterization 112
3.2.5 Hydrogenation Reactions with Pd-Containing Alumina Powder ,,,,,,,,,,,,,,, 113
3.2.6 Hydrogenation Reactions with Pd-Containing Membanes 114
3.3 Results and Discussion 115
3.3.1 Characterization of Polyelectrolyte/PdNP Films 115
3.3.2 Hydrogenation of Allylic Compounds with Pd-Containing Alumina

Powder Catalysts 119

 

3.3.3 Hydrogenation of Allylic Compounds with Pd-Containing Membranes 127
3.3.4 Hydrogenation of Nitroaromatic Compounds with Pd-Containing
Membranes 132

vii

3.4 Conclusions
3.5 References

 

4 Chapter Four: Catalytic Wet Air Oxidation with Pt-Containing
Membranes

 

4.1 Introduction

 

4.2 Experimental Methods

 

4.2.1 Materials

 

4.2.2 Modiﬁcation of Aluminum Oxide Powder
4.2.3 Modification of Disk-Shaped Alumina Membranes
4.2.4 Modification of Tubular Ceramic Membranes

 

4.2.4.1 Method 1 - LbL Adsorption with Ex-situ Nanoparticle Formation
[PAA/PAH/PtNP]1
4.2.4.2 Method 2 — LbL with ln-situ Nanoparticle Formation
[PAA/PEl-Pt(0)]1

 

4.2.4.3 Method 3 — LbL with ln-situ Nanoparticle Formation [Pt(0)/PEl]2m,

4.2.4.4 Method 4 — Evaporation/Recrystallization/Reduction

 

4.2.4.5 Method 5 - Anionic impregnation/reduction

 

4.2.5 Characterization

 

4.2.6 Catalytic Reactions

 

4.3 Results and Discussion

 

4.3.1 Characterization of Polyelectrolyte/PtNP Films
4.3.2 Wet Air Oxidation Catalyzed by Pt Nanoparticles on Alumina
Powder

 

4.3.3 Oxidation of Formic Acid with Disk-Shaped Membranes

 

4.3.4 Wet Air Oxidation with Tubular Membranes
4.3.4.1 Wet Air Oxidation of Formic Acid

 

4.3.4.2 Wet Air Oxidation of Acetic Acid

 

4.3.4.3 Wet Air Oxidation of Phenol

 

4.3.5 Conventional Reactions with Pulverized Tubular Membranes ,,,,,,,,,,,,,,,

4.4 Conclusions
4.5 References

 

Chapter Five: Conclusions and Future Work

 

5.1 Conclusions

5.2 Future Work

5.2.1 Modiﬁcation of Polymeric Hollow Fiber Membranes for Catalytic
Reactions

5.2.2 Catalytic Selectivity with Pd-Containing Membranes

 

5.2.3 Catalytic Selectivity as a Function of Nanoparticle Size

5.2.4 Electrocatalytic Hydrogenation with Pd-Containing Membranes ,,,,,,,,,,,

5.3 Summary
5.4 References

 

viii

134
136

139
139
143
143
144
145
147

147

148
149
150
151
151
152
154
154

156
158
161
162
166
168
172
174
176

181
181
184

184
185
188
192
198
199

LIST OF TABLES

Table 2.1 Au loading, percent 4-nitrophenol reduction, and rate constant vs. the
number of PAH/AuNP bilayers contained in PAA/[PAH/AuNP]x ﬁlms deposited in
porous alumina membranes employed for ﬂow-through reactions. 81

Table 2.2 Rate constants for reduction of nitrophenol and nitroaniline. The
reactions were catalyzed by PAA/PAH/AuNP ﬁlms in porous alumina membranes
(kmm) and on alumina powder (ksluny). 85

Table 2.3 Product distributions in membrane-catalyzed reductions of
nitroaromatic compounds at two different solution ﬂuxes. The membranes
consisted of porous alumina modiﬁed with PAA/PAH/AuNP ﬁlms. 86

Table 2.4 Product distributions in the membrane-catalyzed reduction of
nitroaromatic compounds containing other reducible functional groups. The
reactions employed a gold nanoparticle-modiﬁed alumina membrane and a ﬂux
of 0.015 mUcmz-sec. 92

Table 2.5 Product distributions in the membrane-catalyzed reduction of selected
nitroaromatic compounds at different temperatures. The catalytic membranes
consisted of alumina modiﬁed with a PAA/PAH/AuNP ﬁlm, and the ﬂux was 0.07
mL/cmz-sec. 95

Table 3.1 Pd content of several different Pd catalysts used for hydrogenation
reactions. 1 18

Table 3.2. TOFs and TOF ratios (selectivities) for unsaturated-alcohol
hydrogenation catalyzed by Pd catalysts that contained Pd nanoparticles with
different diameters. Values in parentheses are the sum of TOFs for
hydrogenation and isomerization. 120

Table 3.3. TOFs and TOF ratios (selectivities) for unsaturated-alcohol
hydrogenation catalyzed by Pd catalysts that contained PAA/[PAH/PdNP]x ﬁlms
with 1, 2, and 3 PAH/PdNP bilayers. Values in parentheses are the sum of TOFs
for hydrogenation and isomerization. 122

Table 4.1 Pt contents in tubular membranes. 156

Table 4.2 Catalytic activities of several tubular membranes in phenol oxidation

 

 

at room temperature with 4 bar air overpressure. 171

Table 4.3 Catalytic activities of several tubular membranes in phenol oxidation

with a feed temperature of 60 °C with 4 bar air overpressure. 171
ix

 

Table 4.4 Catalytic activities in formic acid oxidation for tubular membranes
used as interfacial contactors and as powders in conventional stirred tank
reactors. 173

LIST OF FIGURES

Figure 1.1 Potential energy diagram of a catalyzed and uncatalyzed reaction in
which A and B react to form a single product, C. Figure adapted from
Chorkendorff et al. 1

Figure 1.2 Schematic diagram of metal nanoparticles stabilized by a)
polyelectrolytes b) a polymer microsphere and c) ion exchange resin. ,,,,,,,,,,,,,,,,,,, 8

Figure 1.3 Chemical structure of a ﬁrst generation a) PPI dendrimer and b)
PAMAM dendrimer. Both dendrimers are commonly used for'encapsulation of
metal nanoparticles. 9

Figure 1.4 Schematic diagram showing the synthesis of metal nanoparticles
encapsulated in a dendrimer structure. 9

Figure 1.5 Schematic diagram of membranes operated as a) extractors, b)
distributors, c) interfacial contactors, and d) ﬂow-through contactors. The ﬁrst
three conﬁgurations involve tangential ﬂow along the membrane surface, while
the last conﬁguration uses dead-end ﬂow through the membrane pores. ,,,,,,,,,,,,, 12

Figure 1.6 Schematic diagram showing the ﬂow of reactants through a
membrane operated as a ﬂow-through contactor, and around a catalyst bead
from a traditional ﬁxed bed reactor. 17

Figure 1.7 Schematic diagram of the electric double-layer structure surrounding
a nanoparticle stabilized with charged ligands. In this ﬁgure mercaptoundecanoic
acid is the stabilizing ligand. 24

Figure 1.8 Variation of Au nanoparticle Size with molar ratio of the initial
concentration of sodium citrate and Au salt. 25

Figure 1.9 Schematic drawing of Iayer—by—layer deposition by alternating
adsorption of polyanions and polycations. (The rinsing between each deposition
step is not shown.) These ﬁlms can incorporate a wide variety of functional
materials including metal nanoparticles and enzymes. 29

Figure 1.10 FESEM image of a cross-linked [PAA/PAH]3 ﬁlm that was deposited
in the interior pores of an anodized alumina membrane (0.2 pm pore size). The
alumina membrane was dissolved prior to imaging the polymer nanotubes. ,,,,,,, 32

Figure 2.1 Schematic diagram of the deposition of polyelectrolyte/AuNP ﬁlms on
alumina powder. 53

Figure 2.2 Experimental apparatuses used for performing catalytic reactions
with membranes at a) low ﬂow rates, and b) higher ﬂow rates. 58

xi

Figure 2.3 UV—Vis spectrum of a 0.12 mM Au nanoparticle solution Showing a
maximum absorbance at 517 nm. Molarity is given with respect to Au atoms.,_59

Figure 2.4 a) TEM image of citrate-stabilized AuNPS on a carbon—coated copper
grid. b) Histogram of AuNP diameters taken from several TEM images. The
average particle diameter is 12 i 1 nm. 59

Figure 2.5 Structures of polyelectrolytes used in this work. 60

Figure 2.6 Ellipsometric thicknesses for multilayered [PEI/PAA]x (squares) and
[PAA/PAH]x (circles) ﬁlms deposited on AI-coated Si wafers. 62

 

Figure 2.7 Specular reﬂectance UV-Vis Spectra of an Al-coated Si wafer before
(light gray line) and after (dark gray line) deposition of a PAA/PAH layer and
subsequent adsorption of Au-nanoparticles (black line). 63

Figure 2.8 FESEM images of Al-coated Si wafers after deposition of a)
PAA/PAH b) PAA/[PAH/AuNPh c) PAA/[PAH/AuNP]2 and d) PAA/[PAH/AuNP]3
ﬁlms. 64

Figure 2.9 FESEM images of the a) Skin layer and b) support layer of a
commercially available anodized alumina membrane. 64

Figure 2.10 FESEM images of cross-sections of anodized alumina membranes
3) before and b) after modiﬁcation with a PAA/PAH/AuNP ﬁlm. 66

Figure 2.11 Plot of absorbance at 518 nm vs. volume eluted during deposition of
Au nanoparticles in a PAA/PAH-modiﬁed alumina membrane. The Au
concentration in the nanoparticle feed solution was 0.045 mM, which
corresponds to an absorbance of 0.163 at 518 nm. 67

Figure 2.12 FESEM images of cross sections of anodized alumina membranes
after modification with a) PAA/[PAH/AuNP]2 and b) PEI/AuNP ﬁlms. ,,,,,,,,,,,,,,,,,,,, 68

Figure 2.13 Plot of pure water ﬂux and estimated ﬁlm thickness vs. number of
PAH/AuNP bilayers during modiﬁcation of an alumina membrane. The pure
water ﬂux was measured at a constant pressure of 1.4 bar. The values for ﬂux
and ﬁlm thickness after deposition of PAA are displayed as 0 PAH/AuNP
bilayers. 69

Figure 2.14 Top-down FESEM images of unmodiﬁed a) track-etched
polycarbonate and b) nylon membranes. 71

Figure 2.15 Cross-sectional FESEM images of polymer membranes modiﬁed
with PSS/PAH/AuNP ﬁlms. a) polycarbonate and b) nylon. 71

xii

Figure 2.16 UVNis spectra demonstrating the catalytic reduction of 3 mL of 0.1
mM 4-nitrophenol in 10 mM NaBH4 to form 4-aminophenol over time. Spectra
were collected in 4 minute time intervals after addition of 20 uL of as-prepared
AuNPs. 73

Figure 2.17 Plot of ln(A/Ao) vs. time for the reduction of 4-nitrophenol with
NaBH.; catalyzed by Au-nanoparticles in solution. A is the absorbance of the
solution at 400 nm during the reaction and A0 is the initial absorbance of the
solution. [4-nitrophenol] = 0.085 mM, [NaBH4] = 12 mM, and [Au] = 5.9 x 104
mM. 75

Figure 2.18 UV-VIS absorbance spectra of a 0.1 mM solution of 4-nitrophenol
containing 10 mM NaBH4 before (solid black line) and after (solid gray line)
passing through a membrane modified with a PANPAH/AuNP ﬁlm at a ﬂux of
0.29 mL/cmz-Sec. The spectrum of a 0.1 mM 4-aminophenol solution containing
10 mM NaBH4 is also shown as a reference (dotted black line that overlaps the
gray line). 76

Figure 2.19 Plot of 4-nitrophenol reduction percentage vs. ﬂow rate of a 0.4 mM
4-nitrophenol solution through a porous alumina membrane containing a
PAA/PAH/AuNP ﬁlm. [NaBH4] = 20 mM. 76

Figure 2.20 Effect of NaBH4 concentration on the percent reduction of 4-
nitrophenol achieved by ﬂowing NaBH4/4-nitrophenol solutions through an
alumina membrane modiﬁed with a PAA/PAH/AuNP ﬁlm. Conditions: [4-
nitrophenol] = 0.8 mM, ﬂux = 0.28 mUcmZ-sec. 79

Figure 2.21 Comparison of 4-nitrophenol reduction in membranes modiﬁed by
adsorption of PANPAH/AuNP and PAA/PAH/AuNP/PAH ﬁlms when [4-
nitrophenol] = a) 0.4 mM, and b) 0.8 mM. Other conditions: [NaBH4] = 20 mM,
ﬂux = 0.28 mL/cmz-sec. 80

Figure 2.22 Plot of percent conversion vs. ﬂux for membranes containing 1 and
2 PAH/AuNP bilayers. The curves represent a ﬁrst—order reaction model with
rate constants (kmem) of 0.0018 cm/sec for each membrane. (Rate constants are
normalized to the nanoparticle surface area in each membrane.) Conditions: [4-
nitrophenol] = 0.4 mM, [NaBH4] = 20 mM 81

Figure 2.23 Plot of conversion vs. a) ﬂux and b) residence time multiplied by the
AuNP surface area per membrane pore volume (tmsA/Vpoms) in the reduction of 4-
nitrophenol to 4-aminophenol using membranes modiﬁed with a
polyanioanAH/AuNP ﬁlm. In both plots, 0 = polycarbonate, [II = alumina, A=
nylon. Conditions: [4-nitrophenol] = 1 mM, [NaBH4] = 100 mM 83

xiii

Figure 2.24 Slurry reaction profiles for reduction of a) 4-nitrophenol, b) 4-
nitroaniline, c) 4-nitrotoluene, d) 2-nitrotoluene, e) nitrobenzene, and f) 2-nitro-m—
xylene with NaBH4 using alumina powder modiﬁed with a PAA/PAH/AuNP ﬁlm as
a catalyst. For each reaction, circles represent the starting nitro compound,
triangles represent the amine product, squares represent the nitroso product, and
diamonds represent the azoxy product. 88

Figure 2.25 Plot of efﬂuent composition as a function of a) ﬂux and b) residence
time multiplied by nanoparticle surface area per pore volume (tmyA/Vpoms) in the
reduction of 2-nitrotoluene using AuNP-containing membranes. In both
diagrams: A = amine, E] = nitroso, O = nitro, solid black line = nylon membrane,
dotted black line = alumina membrane, and solid gray line = polycarbonate
membrane. The inset in Figure b shows an expanded plot of the data at the right
of this ﬁgure. 90

Figure 2.26 Slurry reaction proﬁles and corresponding Arrhenius plots for the
reduction of a,b) 4-nitrophenol, c,d) 4-nitrotoluene, and e,f) 1,4-
chloronitrobenzene by NaBH4 at different temperatures. The reaction was
catalyzed by alumina powder modiﬁed with a PAA/PAH/AuNP ﬁlm. The lines in
a, c, and e are ﬁts of the data to equation 2.7 in the text. 94

Figure 2.27 Plot of percent reduction of 4-nitrophenol vs. volume eluted through
an alumina membrane modiﬁed with a PAA/PAH/AuNP ﬁlm. Conditions: [4-
nitrophenol] = 2.0 mM, [NaBH4] = 100 mM, ﬂux = 0.17 mL/cmz-sec. The plot
contains a compilation of 5 sequential passages of 500 mL of the reaction
solution. 96

Figure 2.28 Cross-sectional FESEM images of membranes modiﬁed with
PAA/PAH/AuNP films a) before and b) after ﬂowing a solution of 1 mM 4-
nitrophenol and 100 mM NaBH4 through the membrane for 2 hours at 0.030
mL/cmZ-sec. 97

Figure 3.1 Schematic diagram showing the formation of [PAA-Pd(0)/PEI]3 ﬁlms
on alumina powder. Rinses are not shown. 109

Figure 3.2 Schematic diagram showing the formation of PAA/[PAH/PdNP]x ﬁlms
on alumina powder. Rinses are not Shown. 110

Figure 3.3 Schematic diagram of the apparatus used for slurry hydrogenation
reactions with Pd-containing alumina powder catalysts at room temperature and
atmospheric pressure. 113

Figure 3.4 Schematic diagram of the apparatus used for ﬂow-through
membrane hydrogenation reactions with Pd-containing alumina membranes.,_115

Figure 3.5 UV-Vis spectrum of a 0.10 mM Pd nanoparticle solution showing a
peak absorbance at 250 nm. Molarity is given with respect to Pd atoms. ,,,,,,,,, 117

xiv

Figure 3.6 a) TEM image of citrate-stabilized Pd nanoparticles on a carbon-
coated copper grid. b) Histogram of Pd nanoparticle diameters taken from
several TEM images. The average particle diameter was 7.2 :I: 1.2 nm. ____________ 117

Figure 3.7 Plot of Pd loading vs. number of PAH/PdNP bilayers in catalysts
containing PAA/[PAH/PdNP]x ﬁlms. 118

Figure 3.8 Percent hydrogenation of a) allyl alcohol and b) B-methallyl alcohol
during competitive hydrogenation of these substrates using a [PAA-Pd(0)/PEI]3
catalyst prepared with 8 mM Pd(ll). The initial concentration of each substrate
was 25 mM. After 6 min, allyl alcohol was consumed due to hydrogenation and
formation of ~20% propionaldehyde. In (D), lines represent slopes before and
after reaction of allyl alcohol. 125

Figure 3.9 Percent hydrogenation of a) allyl alcohol and b) crotyl alcohol during
competitive hydrogenation of these substrates using a [PAA-Pd(0)/PEI]3 catalyst
prepared with 8 mM Pd(ll). The initial concentration of each substrate was 25
mM. After 6 min, most of the allyl alcohol was consumed due to hydrogenation
and formation of ~20% propionaldehyde. In (b), lines represent slopes before
and after reaction of allyl alcohol. 126

Figure 3.10 a) Conc. of 1-propanol formed vs. initial H2 sparging time in
membrane-catalyzed hydrogenation of 5 mM allyl alcohol. The ﬂux through the
Pd nanoparticle-modiﬁed alumina membrane was 0.023 mL/cmZ-Sec. Circles
represent the concentration of 1-propanol in the permeate solution and squares
represent the concentration of 1-propanol in the feed solution. The plot in b)
represents the difference in the 1-propanol concentration in the permeate and the
feed. 128

Figure 3.11 Concentration of 1-propanol formed in membrane-catalyzed
hydrogenation with several initial allyl alcohol concentrations. Solutions were
sparged with H2, and the ﬂux through the Pd nanoparticle-modiﬁed alumina
membrane was 0.023 mUcmz-sec. 129

Figure 3.12 Cumulative concentration of 1-propanol formed in membrane-
catalyzed hydrogenation of a 20 mM allyl alcohol solution after multiple passes
through an alumina membrane modiﬁed with a PAA/PEI/PdNP ﬁlm. Solutions
were sparged with H2 for 2 min before each pass, and the ﬂux through the
membrane was 0.023 mL/cmZ-sec. 130

Figure 3.13 Concentration of 1-propanol formed during membrane-catalyzed
hydrogenation as a function of ﬂux through an alumina membrane modiﬁed with
a PAA/PEl/PdNP ﬁlm. The initial allyl alcohol concentration was 10 mM, and
solutions were sparged with H2 for 2 min before a Single pass through the
membrane. 131

XV

Figure 3.14 Concentration of 1-propanol formed in membrane-catalyzed
hydrogenation with several initial allyl alcohol concentrations in ethanol.
Solutions were sparged with H2, and the ﬂux through the PAA/PEl/PdNP-
modiﬁed alumina membrane was 0.023 mL/cmz-sec. 131

Figure 3.15 Decrease in 4-nitrophenol concentration in membrane-catalyzed
hydrogenation with several initial 4-nitrophenol concentrations. Solutions were
sparged with H2, and the ﬂux through the Pd nanoparticle-modiﬁed alumina
membrane was 0.023 mL/cmZ-sec. 133

Figure 4.1. Schematic diagram Showing the interfacial contactor conﬁguration in
a tubular membrane and its application to wet air oxidation. 141

Figure 4.2 Schematic diagram of the apparatus used for depositing
polyelectrolyte/metal nanoparticle ﬁlms in tubular ceramic membranes. The
pressurized solution ﬂows though the membrane pores in an inside-out
conﬁguration. 147

Figure 4.3 Schematic diagram of modiﬁcation of membrane surfaces (including
pores) using LbL deposition of PAA/PAH/PtNP ﬁlms. 148

Figure 4.4 Schematic diagram of modiﬁcation of membrane surfaces (including
pores) using LbL deposition of PAA/PEl-Pt(ll) ﬁlms followed by reduction with
NaBH4. 149

Figure 4.5 Schematic diagram of modification of membrane surfaces (including
pores) using LbL deposition of [Pt(lV)/PEI]2 ﬁlms followed by reduction with
NaBH4.. 150

Figure 4.6 Schematic diagram of modiﬁcation of membrane surfaces (including
pores) using evaporation/recrystallization of H2PtCl6 followed by reduction with
H2 at 200 °C. 150

Figure 4.7 Schematic diagram of modiﬁcation of membrane surfaces (including
pores) using anionic impregnation with H2PtCI6 followed by reduction with H2 at
200 °C. 151

Figure 4.8 Schematic diagram of the wet air oxidation apparatuses that employ
a) disk-shaped membranes as ﬂow through contactors and b) tubular catalytic
membranes as interfacial contactors. 153

Figure 4.9 TEM images of a) mercaptosuccinic acid-stabilized Pt nanoparticles
and b) citrate-stabilized Pt nanoparticles on carbon-coated copper grids. The
inset in each image shows the nanoparticle size distribution determined from
several high resolution TEM images. 154

xvi

Figure 4.10 Low (a) and high (b) magniﬁcation TEM images of Pt nanoparticles
immobilized in the pores of an alumina membrane modiﬁed with a
PAA/PAH/PtNP ﬁlm. The Pt nanoparticles contained in the ﬁlm are citrate-
stabilized. 155

Figure 4.11 Change in formic acid concentration in membrane-catalyzed
oxidation with several initial formic acid concentrations. Solutions were sparged
with O2, and the ﬂux through the porous alumina membrane modiﬁed with citrate-
stabilized Pt nanoparticles (method 1) was 0.023 mL/cmZ-sec. 158

Figure 4.12 Cumulative change in formic acid concentration in membrane-
catalyzed oxidation after multiple passes through alumina membranes modiﬁed
with either Pt, Au, Pd, or Ru nanoparticles. Solutions were sparged with O2, and ..
the ﬂux through the metal nanoparticle-containing membranes was
0.023 mUcmz-sec. 159

Figure 4.13 Change in formic acid concentration during membrane-catalyzed
oxidation as a function of ﬂux through a Pt nanoparticle-modiﬁed disk-shaped
alumina membrane prepared by method 1. The initial formic acid concentration
was 10.8 mM, and solutions were sparged with 02 before passing through the
membrane. 161

Figure 4.14 Normalized rate of formic acid oxidation vs. air or N2 overpressure
for a tubular membrane modiﬁed with a PAA/PAHPtNP ﬁlm. Normalization was
performed with respect to the area of the internal wall of the membrane. __________ 163

Figure 4.15 Normalized rate of formic acid oxidation vs. air overpressure for Pt-
containing membranes prepared by the methods shown in Figures 4.3-4.7.
Normalization was performed with respect to a) the area of the internal wall of the
membrane and b) the amount of Pt in the membrane. 164

Figure 4.16. Normalized rates of membrane-catalyzed acetic acid oxidation vs.
air overpressure for different membrane types at a) room temperature and b) 60
°C. Oxidation rates were normalized to the amount of Pt in the membrane.._,,168

Figure 5.1 Compounds that could be used to probe steric effects in
hydrogenation as a function of nanoparticle size. 189

Figure 5.2 Synthesis of relatively large nanoparticles in dendrimers using
multiple cycles of metal-ion loading and reduction. Higher generation dendrimers
could be employed in the actual nanoparticle formation. 190

Figure 5.3 Experimental setup for examining ﬂow-through electrocatalytic
hydrogenation in Pd nanoparticle-containing membranes. Hydrogen is
electrochemically generated in the conductive, nanoparticle-containing
membrane. The hydrogenation of allyl alcohol to 1-propanol is shown as an
example reaction. 194

xvii

Figure 5.4 Cross-sectional FESEM images of a) a bare silver membrane and b)
a silver membrane coated with a mercaptoundecanoic acid/PEI/PdNP ﬁlm. ,,,,, 195

xviii

LIST OF SCHEMES

Scheme 1.1 Series of reactions believed to occur during the citrate reduction of

 

 

AuCl3 to form Au nanoparticles. 25
Scheme 2.1 Formation of functionalized anilines from the parent nitro compound
via hydrogenation with multiple reduction pathways. 48
Scheme 2.2 Au-catalyzed reduction of 4-nitrophenol by NaBH4 72

 

Scheme 2.3 Au-catalyzed reduction of nitroaromatic compounds by NaBH4 ,,,,, 84

Scheme 2.4 Reduction of nitrocyclohexane catalyzed by a Au nanoparticle-
containing alumina membrane with a solution ﬂux of 0.020 mL/cmZ-sec. _____________ 93

Scheme 3.1 Pd-catalyzed hydrogenation of allyl alcohol, B-methallyl alcohol,
and crotyl alcohol. 119

Scheme 3.2 Pd-catalyzed hydrogenation of 4-nitrophenol to 4-aminophenol,,.132
Scheme 4.1 Catalytic wet air oxidation of formic acid. 157

Scheme 4.2 Proposed mechanism for catalytic wet air oxidation of acetic acid.
Adapted from Duprez et al. 167

Scheme 4.3 Proposed pathway for wet air oxidation of phenol, adapted from
Devlin et al. 169

xix

LbL

PAA

 

PSS

 

PEI

 

PAH

 

PVP

 

PAMAM

PPI

 

MSA

 

NP
AuNP
PdNP
PtNP

CVD

 

TEM

 

FESEM

 

GC

 

GC-MS

 

TOC

HPLC

 

 

LIST OF ABBREVIATIONS

layer-by-Iayer

poly(acry|ic acid)

poly(styrene sulfonate)
poly(ethylenimine)

poly(allylamine hydrochloride)
poly(vinylpyrrolidone)

poly(amidoamine)

poly(propylenimine)

mercaptosuccinic acid

nanoparticle

gold nanoparticle

palladium nanoparticle

platinum nanoparticle

chemical vapor deposition

transmission electron microscopy
ﬁeld-emission scanning electron microscopy
gas chromatography

gas chromatography-mass spectrometry
total organic carbon

high performance liquid chromatography

ion chromatography

g-I

lCP-OES ,,,,,,,,,,,,,,,,,,,, inductively coupled plasma-optical emission Spectroscopy

 

 

 

 

FAAS ﬂame atomic absorption spectroscopy
FAES ﬂame atomic emission spectroscopy
SAM self assembled monolayer

UV-Vis ultraviolet-visible

TOF turnover frequency

 

Chapter One: Introduction

1.1 Introduction to Catalysis

Catalysts are used in roughly 90% of chemical processes, so it is essential
to produce catalytic materials that are efﬁcient, reusable, and broadly applicable
to a wide range of processes.1 In general, the purpose of a catalyst is to make
chemical processes more efﬁcient by accelerating the rate at which chemical
reactions proceed. Catalysts do not change the thermodynamics of a reaction,
but rather they change the kinetics of a reaction. Hence they are only useful for
reactions that are thermodynamically favorable. Typical catalytic reactions
between two molecules involve three steps: binding of reactants to the catalyst,
reaction, and separation of the product from the catalyst.2 AS ﬁgure 1.1 Shows,
by incorporating a catalyst, the reaction has become more complex, but much

more energetically favorable.

  

Potential energy

3
‘-

Binding Reaction Separation

 

Ir

Reaction coordinate

Figure 1.1 Potential energy diagram of a catalyzed and uncatalyzed reaction in
which A and B react to form a single product, C. Figure adapted from
Chorkendorff et al.2

Catalysts are frequently classiﬁed as enzymatic, homogeneous, or
heterogeneous. Enzymatic catalysts are the most speciﬁc of the three types;
they allow biological reactions to occur at the rates necessary for living systems
to survive.3 Homogeneous catalysts reside in the same phase as the reactants,
and are often comprised of single molecules. Conversely, heterogeneous
catalysts are solids, usually in the form of supported transition metals, which
catalyze reactions of molecules in liquid or gaseous phases. As a result,
heterogeneous catalysis takes place at an interface between a gas or liquid and
a solid surface, where many different types of active Sites may be involved in the
reaction. The heterogeneity of the solid often leads to decreased selectivity.

Because homogeneous catalysts are most often individual molecules, the
reactions typically take place on a single active site. Common advantages of
homogeneous catalysts relative to heterogeneous systems include: higher
selectivity and reproducibility, milder reaction conditions, ready chemical
modiﬁcation, and greater efﬁciency due to participation of all metal atoms in the
catalytic reaction.3 However, homogeneous catalysts are also more expensive,
less robust, and more difﬁcult to separate from reaction mixtures than
heterogeneous catalysts. Thus, heterogeneous catalysts are still more widely
used for industrial applications, while homogeneous catalysts are primarily
employed for ﬁne chemical synthesis.

Much of the ongoing research in heterogeneous catalysis focuses on
bridging the gap between homogeneous and heterogeneous catalysts and

overcoming the disadvantages described above. Speciﬁcally, interest in using

II:‘"-'““-'ﬁ

novel nanoparticles for catalysis has dramatically increased in the last decade
because these materials have extremely high surface area to volume ratios that
lead to efﬁcient use of expensive metals. Many recent studies show that the
properties of nanoparticles can be tuned to achieve high selectivity and efﬁcient
use of precious metals."’6 Thus, metal nanoparticle catalysts are increasingly
being used in ﬁne chemical synthesis and other reactions that historically relied
on homogeneous catalysts.7'8

This dissertation discusses catalysis by supported noble metal
nanoparticles with a speciﬁc focus on the use of porous membranes and
polyelectrolyte multilayer ﬁlms for nanoparticle immobilization. These catalysts
exhibit excellent activities in a number of catalytic applications including
reduction, hydrogenation, and oxidation reactions. The remainder of this chapter
provides a review of recent work in the ﬁelds of nanoparticle catalysis and
catalytic membranes to show the signiﬁcance of this work. We ﬁrst describe the
most common catalyst supports and discuss the advantages and disadvantages
of each material. We then discuss the beneﬁts of using porous membranes as
catalyst supports and describe the common types of catalytic membranes. A
thorough discussion of traditional methods for in-situ and ex-situ incorporation of
metal nanoparticles in porous supports follows, with special attention to formation
of polyelectrolyte/metal nanoparticle ﬁlms. Finally, a brief overview of the studies

described in each of the remaining chapters is provided.

:r“ -‘3"1"'-

1.2 Nanoparticles and Catalysis

Metal nanoparticles were likely observed over 2500 years ago. For
example, colloidal gold was used as early as the 5th century BC. for coloring
ruby glass and other ceramic materials.9 In 1857, Faraday reported the
formation of deep red solutions of colloidal gold by reduction of AuCI4‘ using
phosphorus in CS2.10 However, in the last 20 years, the study of metal
nanoparticles has increased dramatically due to their unique size-related optical
and electronic properties.”13 This interest is especially apparent in the ﬁeld of
catalysis because metal nanoparticles often Show large increases in activity
compared with their bulk metal counterparts.

Initial reports of nanoparticle catalysts include the use of silver
nanoparticles for photography and platinum nanoparticles for the decomposition
of hydrogen peroxide in the 1800’s.14 However, it wasn’t until the work of Bond
and SerrrIon15'16 in the 1970’s and Harutaﬂ'19 in the 1980’s that researchers
began showing interest in controlled nanoparticles for catalysis. This increased
interest is partly due to the fact that these works Showed that gold nanoparticles
could be highly active catalysts even though bulk gold is typically an ineffective
catalyst.”21 Haruta22 and Goodman6 later demonstrated that TiO2-supported Au
nanoparticles Showed increased activity in CO oxidation when decreasing the
nanoparticle diameter from 6 nm to 3 nm, but then showed decreased activity
with nanoparticle diameters lower than 3 nm. This discovery led to an explosion

of interest in trying to understand the mechanisms of reactions at nanoparticle

surfaces and in developing new methods to produce highly active and selective
nanoparticle catalysts?”25

From the numerous studies published on catalytic nanoparticles, it is clear
that the most active nanoparticles have sizes on the order of 1 to 10 nm.
However, the high surface energy of nanoparticles with diameters in this low nm
range often leads to aggregation, which decreases the catalyst activity.”28
Stabilization of metal nanoparticles typically occurs through immobilization on a

suitable support or addition of a stabilizing agent.

1.3 Common Nanoparticle Supports

The most effective catalyst supports should have high surface areas to
facilitate loading of the catalyst. Thus, highly porous materials such as ceramics,
metal oxides, carbonaceous materials, polymers, and a number of other
materials are commonly used as catalyst supports. The properties of the support
material often inﬂuence both the preparation of nanoparticle catalysts as well as

their reactivity. Each support offers its own advantages for speciﬁc applications.

1.3.1 Metal Oxide Supports

Metal oxides are one of the most widely employed catalyst supports
because they offer high thermal and chemical stability. When they are in the
form of zeolites, they also have a well-deﬁned pore structure and a high surface

area. The surfaces of metal oxides can be easily functionalized, which simpliﬁes

the procedure for catalyst deposition. The most common metal oxides used as
catalyst supports are silica, alumina, titania, ceria, and zirconia.

Silica is commonly used in. catalytic reactions that require mild
temperatures, <300 °C, but at higher temperatures, it is less stable and can
produce volatile hydroxides. Synthesis of silica materials occurs via sol-gel
synthesis (composed of a network of silica spheres) or ﬂame hydrolysis (fumed
silica). The fumed Silica has higher mechanical strength and can have surface

areas as high as 300 m2/g.29

Alumina offers much higher thermal and
mechanical stability than silica, and is thus the most widely used metal oxide in
catalysis. Alumina exists in a number of forms, but the two most common are v-
alumina and a-alumina. v—Alumina is a highly porous, amorphous material that
offers surface areas as high as 300 m2/g and pore sizes as small as 5 nm,
whereas a-alumina is a nonporous, crystalline solid that has a relatively low
surface area (3-5 m2/g). However, d-alumina is highly stable even at
temperatures as high as 1200 °C.2

Silica and alumina are typically classiﬁed as inert materials, but other
metal oxide supports such as titania, ceria, and zirconia Show reactive properties

in certain reactions.”32

Thus, these latter materials have the potential to
improve catalytic activity. For example, TiO2 by itself exhibits excellent
photocatalytic properties, but when it is used as a support for noble metal or
transition metal nanoparticles, it shows enhanced activity and an increased

effective wavelength range that extends into the visible region.32

1.3.2 Carbonaceous Supports

Carbon materials offer a variety of advantages as catalyst supports.
Some carbon supports have surface areas as high as 1500 mz/g and pore Sizes
less than 1 nm, but graphitic carbon has a relatively low surface area. Activated
carbon materials containing functionalized surfaces can be prepared through
several methods including ozone, plasma, or acid or base treatment.33 This
functionalization on the surface of carbon leads to more stable interactions
between the catalytic nanoparticles and the support.

Most commercial carbon catalysts such as Pd/C or PVC use charcoal
supports, and numerous studies examined these catalysts in hydrogenation,
Suzuki, Heck, and C-C bond forming reactions.”38 Much of the current research
with carbon supported catalysts focuses on developing novel carbon materials

3942 Carbon nanotubes are

such as carbon nanotubes as catalyst supports.
unique substrates for nanoparticle immobilization because of their high surface
area, unique physical properties and morphology, and high electrical
conductivity. Moreover, their small Size and hollow geometry facilitates the

formation of small nanoparticles (typically 1 - 4 nm), which is ideal for catalysis.

1.3.3 Polymer Supports

AS an alternative to traditional metal oxides and carbon, polymeric
materials are attractive as nanoparticle supports because of their versatility.
Many polymeric materials contain heteroatoms that can form complexes with

metal nanoparticles, and the ﬂexible structure of the polymers often makes them

particularly effective at stabilizing metal nanoparticles and preventing
aggregation. Furthermore, the variety of functional groups available in polymeric
materials is virtually unlimited. Some of the most common types of polymers
used for nanoparticle immobilization include water soluble polyelectrolytes such

as poly(vinylpyrrolidone) (PVP) and poly(ethylenimine) (PEI),43’45

polymeric
- 46,47 - - 48-50 -
mIcrospheres, and Ion exchange reSIns. FIgure 1.2 shows examples of

each of these Structures.

 

a) b) c)

Figure 1.2 Schematic diagram of metal nanoparticles stabilized by a)
polyelectrolytes b) a polymer microsphere and c) ion exchange resin.

Polyelectrolyte-stabilized nanoparticles are typically soluble in water and
thus act similarly to homogeneous catalysts. The difﬁculty with using these
systems lies in catalyst recovery and reuse, but changing solution conditions can
sometimes force polyelectrolytes to precipitate out of solution.51 Polymeric
microspheres and ion exchange resins act Similarly to traditional heterogeneous
supports and are much easier to separate from the reaction mixture for catalyst
recovery. However, polymer stabilizers suffer from a lack of stability at high

temperature.

1.3.4 Dendrimer-Encapsulated Metal Nanoparticles

Dendrimers represent a special class of polymers that have a very uniform
size distribution, as their polydispersity is ~ 1.52 They are well-deﬁned on a
molecular level and often contain heteroatoms that can bind metal ions for
nanoparticle synthesis. For example, poly(propylenimine) (PPI) and
poly(amidoamine) (PAMAM) dendrimers contain multiple amine groups (Figure
1.3), and each tertiary amine can complex with metal ions. Subsequent
reduction of these ions with NaBH., leads to the formation of metal nanoparticles

that are encapsulated within the dendrimer (Figure 1.4).

H2N NH2
o’lNH HNLO
H2NV8 ~N NchﬁvNH2

\—\
HzN“ NHerHz ”W H

H

HZNJ"N h/NWNHZ H )1”th H H

H2N
W H

HNz a) ”2 H2
Figure 1.3 Chemical structure of a ﬁrst generation a) PPI dendrimer and b)

PAMAM dendrimer. Both dendrimers are commonly used for encapsulation of
metal nanoparticles.

W
Metalion $% < NaBH4

,, 765555

Figure 1.4 Schematic diagram showing the synthesis of metal nanoparticles
encapsulated in a dendrimer structure.

 

In the past decade, a number of groups exploited the inherent properties
of dendrimer-encapsulated nanoparticles for highly effective catalysis.53'57 The
well-deﬁned structure of dendrimers allows the formation of metal nanoparticles
with very narrow size distributions.53'58'59 The caged structure of the dendrimers
serves as a barrier to limit aggregation of particles, but it can also be used to
selectively allow only certain molecules or certain functionalities to come in
contact with the nanoparticles, thus making the catalyst more selective.6°'61

Dendrimer—encapsulated particles can be recovered and reused for additional

catalytic reactions, but recovery often requires nanoﬁltration.62

1.4 Porous Membranes as Catalyst Supports

The most effective catalyst supports have a high surface area to promote
catalyst loading and contact between the catalyst and the reactants. As
mentioned above, high surface areas result from a highly porous structure with
relatively small pore sizes, but small pores frequently exhibit high mass transport
resistances during the catalytic reaction. In heterogeneous catalysis, there are
generally ﬁve steps for every reaction: 1) transport of reactants to the catalyst, 2)
adsorption of reactants on the catalyst, 3) reaction at the catalyst surface, 4)
desorption of products from the catalyst, and 5) transport of products away from
the catalyst. The middle three steps are chemical in nature and deﬁne the rate of
the catalytic reaction under ideal circumstances where there is no mass transport
limitation. However, diffusion of the reactants into and out of the pores of the

support may be slow in comparison to the reaction rate at the catalyst surface. In

10

this case, the overall mass transport will limit the reaction rate, and such
limitations are especially apparent for liquid phase reactions where the diffusion
of reactants is slower than in the gas phase.

Several different reactor conﬁgurations are used for heterogeneous
catalysis. Gas-phase reactions typically use ﬁxed-bed or ﬂuidized-bed reactor
conﬁgurations. Fixed-bed reactors consist of catalyst particles packed in a
column through which the gaseous reactants ﬂow, while ﬂuidized-bed reactors
consist of ﬁne catalyst powders that appear as a “ﬂuid” because the gas is
ﬂowing upward through the tube with sufﬁcient velocity to keep the particles in
continuous motion. Liquid phase and gas/liquid reactions typically consist of a
trickle-column, batch, or stirred tank reactor. The trickle-column reactor is similar
to the reactors for gas phase reactions except the liquid phase passes downward
through the column, and the gas phase either ﬂows concurrently with the liquid
down the bed or counter-currently up the bed. Batch and stirred-tank reactors
are much more common in industry and require suspension of powder catalyst in
the liquid by shaking or stirring and incorporation of gas in the case of gas/liquid
reactions. All of these conﬁgurations can suffer from mass-transport limitations.

Porous membranes, which can consist of virtually any ceramic or polymer
material, offer a support conﬁguration that should provide several advantages
over the conventional catalyst supports.63 One of the motivations for using
membranes as catalyst supports is the ability to catalyze reactions and perform
separations at the same time. Also, membrane reactions can run continuously

because the catalyst does not need to be separated from the reaction products,

11

unlike batch and stirred tank reactors. The internal pores of the membrane also
provide a large surface area that permits a high loading of the catalytic material.
Depending on the conﬁguration of membrane reactors, they may provide a
number of other advantages as well. Catalytic membranes are typically divided
into three categories: extractors, distributors, and contactors (Figure 1.5), which

are discussed below.

a) b) c) d)
A A+B A+B

 

 

   

A+B §—’C A+B-’C(+B)-*D A+B—>0 A+B->0

Figure 1.5 Schematic diagram of membranes operated as a) extractors, b)
distributors, c) interfacial contactors, and d) ﬂow-through contactors. The ﬁrst
three conﬁgurations involve tangential ﬂow along the membrane surface, while
the last conﬁguration uses dead-end ﬂow through the membrane pores.

1.4.1 Membrane Extractors

Membranes are generally not thought of as catalytic materials because
they are widely used for separation processes such as reverse osmosis for water
treatment, gas separations to produce high purity gases, and a number of other
applications. However, in the late 1960’s, researchers began trying to exploit the
excellent separation capabilities of membranes for catalytic applications. The
earliest work in this area employed membrane extractors to remove one or more
of the products from the reaction to shift the equilibrium and increase conversion.

When reactions are limited by equilibrium, the membrane can continuously and

12

selectively remove the product to shift the equilibrium (Figure 1.5a). An
additional beneﬁt of membrane extractors is the ability to decrease undesirable
side reactions. After removal from the reaction mixture, the product can no
longer react with starting materials to form unwanted byproducts. If the
compound being removed is the end product, additional separation and
puriﬁcation steps will be less complicated and require less energy.

Many of the initial studies with membrane extractors involved beneﬁcial
hydrogen removal in reactions such as hydrocarbon dehydrogenation?“65 water-
gas shift,66 and methane steam reforming reactions.”68 These reactions were
attractive targets because of the relative ease of selectively removing hydrogen
by diffusion through Pd-based membranes or by Knudsen diffusion through

membranes with large pore sizes.”70

Though removal of H2 is the most
common application of membrane extractors, membranes that selectively

remove H2O or 02 have also been exploited.“72

1.4.2 Membrane Distributors

Distributor membrane reactors provide controlled addition of a reactant to
the reaction mixture to limit side reactions and provide higher selectivity for the
desired product. This may occur by simply controlling the amount of a reactant
that is introduced on one side of the membrane, or by selectively allowing one
component from a mixture to pass through the membrane and undergo reaction.
As an example, figure 1.5b demonstrates the reaction of two components, A and

B, to yield a product, C without the continued formation of D. As component B

13

permeates thrOugh the membrane, it reacts with component A on the other side
of the membrane to form product C. The reaction of B creates a concentration
gradient within the membrane, which is the driving force for the continued
permeation of B and subsequent reaction with A. Because B reacts with A
immediately upon permeating through the membrane, very little side reactions
occur to produce D. The most common example of this type of reaction is the
partial oxidation of hydrocarbons.73 In these reactions, 02 reacts more easily
with the partially oxidized Species than the starting material. By controlling the
addition of 02 (B) to the hydrocarbon stream (A), complete oxidation can be

limited.

1.4.3 Membrane Contactors

The newest form of catalytic membrane reactor is the membrane
contactor, where the membrane acts as a support that brings the reactants into
contact with the catalyst. In the case of extractors and distributors, the
membrane is often catalytically inert and simply acts as a support with a ﬁxed
bed of catalyst on one side of the membrane. In that case, the membrane is
solely responsible for the separation step, while the ﬁxed catalyst bed is
responsible for the reaction. Membrane contactors typically have catalyst
deposited within the membrane pores, making them catalytically active, but they
usually do not perform a separation function. Although they do not fulﬁll to
requirements to be catalytic membrane reactors in the traditional sense

(combining separation and reaction), most researchers in the ﬁeld still

14

characterize them as a form of catalytic membrane reactor.74 Membrane
contactors are further classiﬁed into two categories in which the membrane acts
as an interface between multiple phases (Figure 1.5c) or the membrane is
utilized for dead-end ﬂow where all the reactants pass through the catalytically

active membrane pores (Figure 1.5d).

1.4.3.1 lnterfacial Contactors

lnterfacial contactors, which are also known as catalytic diffusers, use the
membrane as a catalyst support to facilitate contact between reactants that are in
multiple phases. The membrane can separate the two phases and cause
enhanced contact between the catalyst and the reactants. The most common
application for interfacial contactors is gas/liquid contacting for reactions such as
hydrogenations or oxidations. A gas/liquid contactor typically has the gas on the
side of the membrane that has larger pores and the liquid on the side that
contains more of the catalyst, which is usually the skin layer. The gas and liquid
solutes then diffuse to the catalyst surface and react. Flow of the liquid and gas
facilitates transport Of reactants and removal of products from the reaction zone.
This conﬁguration was successfully applied to oxidize organic acids to CO2 and
H2075 and short chain alkanes to oxygenates.76 A number of studies looked at
the hydrogenation of unsaturated substrates including cinnemaldehyde,77
methylenecyclohexane,78 and sunﬂower seed oil,79 with excellent success, in

some cases achieving high selectivity for the desired product. There has also

15

been Signiﬁcant interest in using interfacial contactors for removal of nitrates from
drinking water by hydrogenation.7°'8°'81

The interfacial contactor conﬁguration is also beneﬁcial for forming contact
between two liquid phases. In processes such as solvent extraction, a
membrane separates an aqueous phase from an organic phase.82 Studies by
Guibal and Vincent also used this conﬁguration for contacting two reactants that
are in the samephase but are separated by the membrane. These studies
utilized chitosan hollow ﬁber membranes for nitrophenol reduction”85 and
chlorophenol dehalogenation86 by sodium formate. In these reactions, a solution
of sodium formate ﬂows by one side of the membrane while solutions of
nitrophenol or chlorophenol ﬂow by the opposite side of the membrane, and the
concentration of sodium formate is much higher than the reduced species. The
reactants diffuse into the pores from each Side of the membrane and come into
contact at the catalyst surface. However, if the reaction has fast kinetics, mass

transport limitations may occur. Use of a ﬂow-through mode can overcome mass

transport limitations.

1.4.3.2 Flow-Through Contactors

Membranes operated as ﬂow-through contactors act similarly to
conventional ﬁxed bed reactors. In traditional fixed bed reactors, the reactants
need to diffuse into the pores of the support to react with the catalyst, but in ﬂow-
through membrane contactors, convective mass transport rapidly brings the

reactants to the active surface of the immobilized catalyst (Figure 1.6). If the

16

pores are sufficiently small, radial diffusion will not limit reactions even at high
ﬂow rates where the contact time is very short. In this way, the pores of the
membrane act as microreactors in which the reaction conversion can be
controlled by simply adjusting the ﬂow rate.87 Flow-through contactors also have
the advantage of constantly removing the products from the reaction zone. AS a
result, there will be less competition between products and reactants for active
sites on the catalyst and a decreased possibility for the products to undergo side

reactions or poison the catalyst surface.

   

Reaction Mixture Reaction Mixture
Membrane Su pport Conventional Catalyst
Suppon

Figure 1.6 Schematic diagram showing the ﬂow of reactants through a
membrane operated as a ﬂow-through contactor, and around a catalyst bead
from a traditional ﬁxed bed reactor. '

The ﬂow-through contactor conﬁguration allows ﬁne control over the
residence time of reactants inside the membrane, which directly affects
conversion. Even a single pass through the membrane yields nearly complete

conversion in reactions such as volatile organic carbon (VOC) combustion or

17

photocatalytic oxidation of organic compounds.”90

However, a majority of
previous experiments with ﬂow-through contactors passed the reactant mixture
through the membrane multiple times to obtain complete conversion. Many of
these studies focused on gas-phase hydrogenation or oxidation reactions.91
Flowing the gaseous mixture through the membrane very rapidly allows very
short contact time, which results in partial hydrogenation of compounds such as
acetylene, propyne, butadiene, and hexadiene.92'95

The ﬂow-through conﬁguration is also beneﬁcial for gas/liquid reactions
where a limited amount of gas is able to dissolve in the liquid phase. Thus with
each pass, the liquid solution can again be resaturated with the gas. Schmidt
and coworkers found that this strategy afforded high selectivity in the partial
hydrogenation of a number of unsaturated organic compounds including
cyclooctadiene, 1-octyne, phenyl acetylene, and geraniol.”97 Others had similar
success with partial hydrogenation of sunﬂower oils’a'100 and u-methylstyrene101
and hydrogenation of nitrate for water denitriﬁcation.81'102 The overall success of

membrane reactors in restricting the extent of reaction could make them quite

valuable in a number of applications.

1.5 Formation of Metal Nanoparticles

Nanoparticles can be integrated into porous membranes via a number of
different techniques, and this section describes the most common strategies for
incorporating metal nanoparticle into porous supports. As ﬁgure 1.4 shows,

reduction of metal ions with sodium borohydride to form the zerovalent metal is

18

one of the common methods for the preparation of supported metal
nanoparticles. In general, formation of supported metal nanoparticles occurs by
either in-situ or ex—situ methods. We ﬁrst discuss in-situ formation in which
nanoparticles are formed directly on the support, and then we present ex-situ
methods that involve the formation of nanoparticles prior to immobilization on the
porous support.

The ﬁrst step in the in-situ formation of nanoparticles is binding of a metal
precursor to the support via covalent, electrostatic, or other adsorptive
interactions with the surface. Next a calcination step is sometimes required to
remove any excess solvent. Finally, reduction of the metal precursor to the
zerovalent state results in the formation of metal nanoparticles. This typically
involves the use of a reducing agent such as H2 at high temperatures or sodium
borohydride. The size and shape of the nanoparticles depend on the amount of
metal precursor that is initially bound to the support, and the morphology (ie. pore
size and shape) of the support.

A variety of in-situ methods can be used to form supported metal
nanoparticles. The most common methods are impregnation, deposition
precipitation, coprecipitation, and chemical vapor deposition, which are described

in detail below. However, other physical means to form supported nanoparticle

3 104,105

catalysts include electrochemical reduction,10 sonochemical reduction,

106.107

microwave irradiation, and laser ablation,108 but this is by no means an

exhaustive list.

19

1.5.1 lmpregnation
lmpregnation techniques are probably the most widely used methods for

109‘” and the methods

forming catalytic metal nanoparticles on solid supports,
are often used to form metal nanoparticles in ceramic membranesm‘113 In these
techniques, the initial step Simply involves “wetting” the solid support with a
solution containing the metal precursor. Evaporation of the solvent leads to
deposition of the metal precursor on the surface of the support, and subsequent
calcination and reduction steps result in the formation of dispersed nanoparticles.
The size of the nanoparticles can be controlled to some extent by varying the
concentration of the metal precursor solution used to impregnate the support.
This also controls the overall loading of metal on the support.

Various interactions can occur between the support surface and the metal
precursor to facilitate adsorption. For example, supports composed of metal
oxides like silica and alumina contain hydroxyl groups that serve as sites that
anchor the metal precursor. These surface hydroxyl groups can also provide
surface charge, depending on the pH of the solution and isoelectric point of the

support. Thus ionic precursors may bind to the oppositely charged Sites of the

support, and this form of impregnation is also called ionic impregnation.

1.5.2 Deposition Precipitation
Similar to lmpregnation, deposition precipitation involves precipitation of a
metal hydroxide or carbonate species on the surface of a support. This occurs

by ﬁrst dissolving a metal precursor in solution and then adding base to adjust

20

 

the pH of the solution and form a precipitate. The support surface provides
nucleation sites for the deposition of the precipitate. Calcination and reduction of
the metal precipitate yields supported metal nanoparticles.“4'115 This method
can be problematic as it is essential to have efficient stirring to make sure the pH
is homogeneous throughout the solution and the support and metal precursor are
well-distributed. Hence the method is difﬁcult to apply in porous membranes.
Control over nanoparticle size is often difﬁcult as deposition of the metal salt on
the support is not homogeneous.116 This inhomogeneity also leads to the

formation of larger aggregates of particles in some locations.

1.5.3 Coprecipitation
Unlike impregnation and deposition precipitation techniques where the
metal precursor is deposited onto the support, coprecipitation involves

7 In this method,

precipitation of the metal and support at the same time.11
deposition of the metal during formation of the support and subsequent
calcination and reduction allow the metal to deposit in the porous stmcture of the

support. A number of studies looked at the coprecipitation of metals during sol-

109,114,118 9

gel synthesis of Silica and during carbon activation.11 Notably, co-
precipitation with Ag yielded a silica support with evenly dispersed 3.5 nm
diameter Ag nanoparticles, while impregnation yielded a support with- unevenly
distributed 15-18 nm diameter Ag nanoparticles.118 Though coprecipitation can

lead to a high dispersion of catalyst in the support materials, the incorporated

21

metal may also lead to decreased thermal and mechanical stability of the
support, and some of the metal may not be accessible for catalysis.
Coprecipitation was recently used to form polymer-metal nanoparticle
composites for catalysis.120 In some cases the nanoparticles tend to form
aggregates during reduction,121 but studies by Mallick et al.122'123 found that using
Pd acetate as the oxidizing agent during polymerization of anthranilic acid or
aminobenzoic acid resulted in formation of ~2 nm diameter Pd nanoparticles that
were well-distributed and stabilized within the polymer matrix. These polymer-Pd
nanoparticle composites showed high activity in ethylene hydrogenation
reactions. Similar methods were used to incorporate metal nanoparticles into
polymeric membranes by adding metal precursors to polymer solutions and

reducing the metals during membrane synthesis.95"9""'12“'125

Incorporation of
metal nanOparticles during polymer membrane synthesis may lead to some
heterogeneity in pore structure, but the main disadvantage of using
coprecipitation to incorporate metal nanoparticles into the support material is that
some of the nanoparticle surface iS not accessible for catalysis, and the overall
catalytic activity decreases. This is often a concern when using expensive

metals that need to be utilized efﬁciently, so this process is more beneﬁcial for

incorporation of relatively cheap catalyst materials.
1 .5.4 Chemical Vapor Deposition

Chemical vapor deposition (CVD) is attractive for producing well-dispersed

catalysts with a narrow size distribution. In general, CVD involves the deposition

22

of a metal precursor from the vapor phase onto a support under high vacuum
conditions in the presence of a reducing agent. The main advantage of this
method lies in the potential to control the nanoparticle size for a given metal
loading by adjusting the amount of precursor deposited on the surface prior to
reduction. 126 Compared to the conventional wet impregnation methods, the
absence of solvent favors the uniform distribution of the metal precursor inside
the pores of the support and eliminates the drying step during which the metal
can redistribute.127 CVD can modify a wide variety of support materials including
carbon nanoﬁbers.12‘5'128 This technique is also an excellent method for forming a
thin Pd layer on the surface of membranes that serve as extractors for
hydrocarbon dehydrogenation70 or interfacial contactors for hydrogenation

reactions.80

1.5.5 Synthesis of Unsupported Metal Nanoparticles

To date, one of the biggest limitations of using in-situ methods to form
supported metal nanoparticles is the difﬁculty of producing metal nanoparticles
with a uniform size distribution. Ex-situ preparation of metal nanoparticles that
are stable in solution has the potential to yield much more uniform particle
shapes and sizes. In solution, the nucleation and growth of nanoparticles can
occur freely without the vagaries characteristic of heterogeneous processes.
A variety of chemistries are available for stabilizing metal nanoparticles in
solution. These unsupported metal nanoparticles typically form by the reduction

of metal ions in the presence of a ligand or other stabilizing agent that can bind to

23

the surface of the newly formed particle and terminate growth. The stabilizing
agent offers control over nanoparticle size and polydispersity. Figure 1.7
demonstrates that a charged stabilizing ligand on the surface of the nanoparticles
produces an electric double layer that causes particle-particle repulsion and
minimizes particle agglomeration.129 Choice of ligand is important in nanoparticle
synthesis as it provides control over surface functional groups and the solvent
compatibility of the nanoparticles. Citrate, thiols, amines, and phosphines are the

most common stabilizing agents used in metal nanoparticle synthesis.

Metal Ligand Stern Diffuse Slipping
core shell layer layer plane

 

   

® Solvated cation

O Q G Solvated anion

O
G O Solvent Molecule

Figure 1.7 Schematic diagram of the electric double-layer structure surrounding
a nanoparticle stabilized with charged ligands. In this ﬁgure mercaptoundecanoic
acid is the stabilizing ligand.

Citrate reduction, which was pioneered by Turkevich in the 1950’s, yields
nearly monodisperse, water soluble Au nanoparticles that can range in size from

8 to 100 nm.‘3°'132 This technique utilizes citrate as both the reducing agent and

24

the stabilizing agent. Upon reducing the Au(lll) to Au(0) (scheme 1.1), excess
citrate stabilizes the Au nanoparticles to terminate particle growth.133 Thus, a
layer of citrate anions forms at the surface of the nanoparticles, leading to a net

negative charge that allows the nanoparticles to remain separated.

O O O

l l I
('OCCH2)20(OH)CO —' ('OCCH2)2C=O + CO2 + H+ + 2e' (1)
AuCl3 + 2e" —> AuCl + 2Cl' (2)
3AuCl —> 2Au(0) + AuCl3 (3)

Scheme 1.1 Series of reactions believed to occur during the citrate reduction of
AuCl3 to form Au nanoparticles.

100 - .
OFrens

E 80 - lFreund
5 0 OTurkevich
s. 60 ‘
s
gw- .0
CU
- e o
P 20 ' Io _ e g

0 I I

 

 

0 1 2 3 4 5 6

Citrate/Au Ratio
Figure 1.8 Variation of Au nano article size with molar ratio of the initial
concentration of citrate and Au salt.1 1‘34””

Variation of the molar ratio of the citrate to Au in the solution affords ﬁne
control over nanoparticle diameter (Figure 1.8).131'133'136 Moreover, the citrate

method is quite adaptable as it has been used for formation of aqueous solutions

0f Pd,47’137'138 Pt,139'140 141,142

and Ag nanoparticles. Although the citrate reduction

method is simple and versatile, a number of other methods have been developed

25

to improve nanoparticle stability, change surface functional groups, and produce
nanoparticles with smaller diameters.

In the early 1990’s, Brust and coworkers developed a nanoparticle
synthesis in which a dodecanethiol stabilizer was added to the metal precursor
solution during reduction by NaBH4.143'144 This method initially yielded organic-
soluble Au nanoparticles with diameters in the range of 2 to 8 nm. Subsequent
studies with other thiol capping agents produced particles with a wider range of

5

sizes and surface functionalities.14 Kimura and Kunitake demonstrated that

careful selection of thiol ligands can also yield water-soluble nanoparticles.“‘6'147
This method was later extended to the synthesis of thiol-stabilized Pt
nanoparticles.148

When nanoparticles are coated with strong stabilizing agents such as
thiols, additional surface functionalization can be very difﬁcult, and catalytic
activity will also be low. These drawbacks of thiol stabilizers led researchers to
look to alternative stabilizing ligands such as amines. Amine-stabilized
nanoparticles can be prepared by simply substituting an appropriate amine for
the thiol.149 This method has been used for synthesis of Au, Pt, and Ag

nanoparticles in organic solvents,"’°'151

and a Similar method for synthesizing
water-soluble Au nanoparticles has also been demonstrated.152 In some cases
amines can be used as both the reducing and stabilizing agent during
nanoparticle synthesis.153 Triphenylphosphine is another stabilizing ligand that

has advantages over thiol ligands because it is a weak stabilizing ligand and can

26

easily undergo ligand-exchange reactions to functionalize the nanoparticle
surface.1"""156

All of these methods of nanoparticle synthesis involve the generation of
metal nanoparticles from molecular precursors in the presence of reducing
agents and stabilizing agents. Other methods of nanoparticle fabrication include
laser ablation and sonochemical, electrochemical, and microwave syntheses.145
In all cases, the resulting nanoparticles are soluble in a speciﬁc solvent and can
be used directly to perform catalytic reactions. Thus the nanoparticles in these
systems act as semi-homogeneous catalysts and demonstrate some of the
advantages of molecular catalysts even though they are actually solid materials.
As we mentioned previously, the difﬁculty in using the as-prepared nanoparticles
directly for catalysis lies in the catalyst recovery step. Immobilization of the metal

nanoparticles onto a suitable support, which is presented in the next section,

facilitates catalyst recovery and reuse.

1.5.6 Surface Modification with Preformed Metal Nanoparticles

The ability to synthesize metal nanoparticles prior to immobilization on a
support affords much more control over nanoparticle size and surface
functionality than the methods for in-situ formation (sections 1.5.1 to 1.5.4) of
supported metal nanoparticles. Ex-situ formation can lead to the development of
more selective or reactive catalysts that can be tuned to a speciﬁc process.
Assembly of metal nanoparticles on surfaces is often driven by electrostatic

interactions or covalent bonding between the chemical functionalities of a metal

27

nanoparticle and a surface. Chemical modiﬁcation of the substrate may be
needed to create interactions with the metal nanoparticles, and such modiﬁcation
can be achieved through a variety of methods. Surfaces modiﬁed with amine- or
carboxylate-terminated self assembled monolayers (SAMs) of functionalized
thiols are capable of capturing metal nanoparticles capped by various types of
charged stabilizing ligands. The extent of nanoparticle adsorption varies with the
strength of the interaction between the nanoparticles and the surface.
Silanization is another common method for surface functionalization, and
Chaudret and Schmid reacted porous alumina membranes with alkoxysilanes
and subsequently adsorbed several noble metal clusters or colloids.157 These
membranes catalyzed the gas phase hydrogenation of 1,3-butadiene and CO
oxidation.91 '92 Sehayek and coworkers modiﬁed porous alumina membranes with
aminopropyltrimethoxysilane prior to deposition of citrate-stabilized Au and Pd
nanOparticles. However, this procedure led to aggregated nanoparticles that
formed nanotube structures (the intended result of that study) rather than the
well-separated nanoparticles that are efﬁcient catalysts.‘38'158.

The methods of nanoparticle immobilization discussed above lead to the
immobilization of a monolayer of metal nanoparticles. But in some catalytic
applications it would be useful to increase the nanoparticle loading by
immobilizing multiple layers of nanoparticles. The next section introduces the

method of layer-by-layer (LbL) assembly as a simple and effective method for

incorporating metal nanoparticles into catalyst supports.

28

1.6 Layer-by-Layer Assembly of Polyelectrolyte Multilayer Films

The LbL method, which was pioneered by Decher in the early 1990’s,
typically involves alternating adsorption of polycations and polyanions and can be
simply carried out on nearly any substrate that will support adsorption of an initial
layer of polymer.159'162 The adsorption of the ﬁrst polymer layer is usually driven
by strong electrostatic attraction between the charged surface and the initial
polyelectrolyte. The many charged groups on the polyelectrolyte can
overcompensate the charge on the support surface to present an excess charge
for subsequent adsorption of an oppositely charged polyelectrolyte. The excess
surface charge of each layer also acts to repel the adsorption of additional
similarly charged polyelectrolyte molecules, limiting deposition to a single “layer".
Subsequent addition of an oppositely charged polyelectrolyte results in the
formation of one bilayer on the support surface (Figure 1.9). This process can be

repeated multiple times to form ﬁlms with the desired thickness.

A n A

: Polycation Polyanion Polycation Nanoparticle Polycation Enzyme
men 0 O

b

      

 

“who

 

Figure 1.9 Schematic drawing of layer-by-layer deposition by alternating
adsorption of polyanions and polycations. The rinsing between each deposition

step is not shown. These ﬁlms can incorporate a wide variety of functional
materials including metal nanoparticles and enzymes.

 

29

The adsorption of a polyelectrolyte on a surface is an entropy driven
process.163 For each polyelectrolyte that adsorbs on the surface, many counter
ions are released into solution. Thus, adsorption of a Single polyelectrolyte layer
should release a sufﬁcient number of counterions to cause a large increase in
entropy.

The properties of polyelectrolyte multilayer ﬁlms can easily be tailored to fit
speciﬁc applications by changing the pH, salt concentration, or solvent
composition of polyelectrolyte deposition solutions. For films containing weak
polyelectrolytes, changing the solution pH greatly affects ﬁlm thickness because
the degree of protonation (and therefore charge density) will change according to
the solution pH.164 For example, Rubner and coworkers determined that the
thickness of ﬁlms containing poly(acry|ic acid) (PAA) and poly(allylamine
hydrochloride) (PAH) is highly dependent on pH. When deposition solutions had
a pH of 7, the thickness of 1 layer was ~5 A, whereas the layer thickness
increased to 80 A when the solutions were adjusted to a pH of 5.165 Likewise,
changes in ionic strength can lead to dramatic changes in ﬁlm thickness. When
the polyelectrolyte solution has a high salt concentration, the salt ions screen the
charges on the polyelectrolyte to give a more coiled polyelectrolyte structure that
forms a thicker ﬁlm.16"* ‘67 Conversely, decreasing the ionic strength leads to less
charge screening, and the polyelectrolytes stretch out and form relatively thin
ﬁlms with few loops and coils.

The LbL method is not limited to Simple polymers, but can be used for

174,175

many charged species including biomolecules,168'173 dendrimers, carbon

30

176-178 179-182 183-189

nanotubes, metal oxide particles, and metal nanoparticles, (see

Figure 1.9). Metal nanoparticles can even be incorporated into multilayer ﬁlms

by in-situ methods of reduction,5"9“"191

which typically involve the addition of a
metal salt to the polyelectrolyte during deposition and subsequent reduction of
the metal ions to form nanoparticles stabilized by the polyelectrolyte ﬁlm. While
most LbL applications primarily use electrostatic attraction as the driving force for

ﬁlm formation, there are many other interactions that can initiate deposition,

192-195 182,196,197

including hydrogen bonding and covalent bonding. In general,
there simply needs to be some form of interaction between the materials that are
incorporated into the thin ﬁlm structure.

Given the vast range of materials that can form multilayer ﬁlms, LbL
deposition is an excellent way to produce ﬁlms that can be tuned to speciﬁc
processes, such as catalysis.“85'1“"200 Although LbL assembly is most easily
applied to ﬂat surfaces, it can also occur in porous and irregular structures.

Thus, LbL assembly of ﬁlms containing preformed metal nanoparticles is an ideal

method for modifying porous supports to make highly effective catalysts.

1.7 Scope of this work

Adaptation of Layer-by-layer adsorption to membrane modiﬁcation occurs
by exposing porous membranes to alternating polyanion and polycation
solutions, with water rinses after each deposition step. Previous and on-going
research in the Bruening group involves ﬁlm deposition primarily on top of the

support?“203 In such work, only the top portion of the membrane is in contact

31

with the polyelectrolyte solutions, and because the membranes have small
surface pores (~20 nm), the solution does not enter into the pores and thus

modiﬁes only the top of the membrane.

 

Figure 1.10 F ESEM image of a cross-linked [PAAlPAH]3 film that was deposited
in the interior pores of an anodized alumina membrane (0.2 pm pore size). The
alumina membrane was dissolved prior to imaging the polymer nanotubes.

The deposition of polyelectrolyte films in the interior of membranes simply
requires the ﬂow of the polyelectrolyte solutions through membranes with
relatively large pores (typically 200 nm or larger). Figure 1.10 shows
polyelectrolyte tubes that were formed by depositing a PAA/PAH ﬁlm in porous
alumina, crosslinking via amide formation,204 and dissolving the alumina205 in 3 M
sodium hydroxide. The image clearly shows the formation of hollow polymeric
tubes that still allow convective ﬂow of solution through the membrane. Al and
Caruso separately demonstrated the formation of similar polyelectrolyte tubes

even without cross-linking of polyelectrolytes.206207

32

This dissertation discusses the use of LbL deposition to modify porous
membrane and powder supports with metal nanoparticle-containing thin ﬁlms,
and the subsequent use of these supports for heterogeneous catalysis. The next
three chapters each focus on a speciﬁc application. Chapter 2 presents the use
of Au nanoparticle-containing membranes for catalyzing the reduction of
nitroaromatic compounds by sodium borohydride. These studies show the effect
of functional group, solution ﬂow rate, and membrane structure on conversion in
membrane reactions. This chapter also presents the characterization of Au
nanoparticle-containing polyelectrolyte ﬁlms on aluminum wafers, alumina
powder, and in porous membranes. Chapter 3 discusses the application of Pd-
containing supports for hydrogenation of a variety of substrates and
demonstrates the use of ﬂat disk-shaped membranes as ﬂow-through contactors
for gas/liquid reactions. Experiments with Pd-containing alumina powder
catalysts Show the effect of nanoparticle size on reaction selectivity. Chapter 4
discusses the use of Pt nanoparticle-containing supports for the wet air oxidation
of several model compounds and demonstrates the use of tubular ceramic
membranes as interfacial contactors. This work compares LbL deposition of
catalytic nanoparticles with catalytic membranes prepared by other procedures.
Chapter 5 discusses some conclusions and future directions for this work as well
as some of the challenges that should be addressed. Overall, this dissertation
Shows that the LbL method yields catalytically active support materials that

effectively catalyze reduction, hydrogenation, and oxidation reactions.

33

1.8 References

(1) Adams, C. J. North American Catalysis Society,
h_ttp:l/www.nacat_soc.orgledu info.asp Accessed June 4, 2009.

(2) Chorkendorff, l.; Niemantsverdriet, J. W. Concepts of Modern Catalysis
and Kinetics; Wiley-VCH: Weinheim, Germany, 2003.

(3) Swiegers, G. F. Mechanical Catalysis: Methods of Heterogeneous,
Homogeneous, and Enzymatic Catalysis; John Wiley & Sons, Inc: Hoboken,
New Jersey, 2008.

(4) Corma, A.; Serna, P. Science 2006, 313, 332-334.

(5) Kidambi, 8.; Dai, J.; Li, J.; Bruening, M. L. J. Am. Chem. Soc. 2004,
126, 2658-2659.

(6) Valden, M.; Lei, X.; Goodman, D. W. Science 1998, 281, 1647-1650.

(7) . Augustine, R. L. Heterogeneous Catalysis for the Synthetic Chemist;
Marcel Dekker, Inc.: New York, New York, 1996.

(8) Astruc, 0.; Lu, P; AranzaeS, J. R. Angew. Chem. Int. Ed. 2005, 44,
7852-7872.

(9) Daniel, M.-C.; Astruc, D. Chem. Rev. 2004, 104, 293-346.
(10) Faraday, M. Philos. Trans. R. Soc. London 1857, 147, 145-181.

(11) Zhao, J.; Bradbury, C. R.; Fermin, D. J. J. Phys. Chem. C 2008, 112,
6832-6841.

(12) Zhou, W. P.; Lewera, A.; Larsen, R.; Masel, R. |.; Bagus, P. S.;
Wieckowski, A. J. Phys. Chem. B 2006, 110, 13393-13398.

(13) Akamatsu, K.; Deki, S. J. Mater. Chem. 1997, 7, 1773-1777.
(14) Schmid, G. Clusters and Colloids; Wiley-VCH: Weinheim, Germany, 1994.

(15) Bond, G. C.; Sermon, P. A.; Webb, 6.; Buchanan, D. A.; Wells, P. B.
J. Chem. Soc, Chem. Commun. 1973, 444-5.

(16) Bond, G. C.; Sermon, P. A. Gold Bulletin 1973, 6, 102-105.

34

(17) Haruta, M.; Kobayashi, T.; Sano, H.; Yamada, N. Chem. Lett. 1987,
405-8.

(18) Haruta, M.; Tsubota, S.; Kobayashi, T.; Kageyama, H.; Genet, M. J.;
Delmon, B. J. Catal. 1993, 144, 175-192.

(19) Haruta, M.; Yamada, N.; Kobayashi, T.; lijima, S. J. Catal. 1989,
115, 301-309.

(20) Hammer, B.; Norskov, J. K. Nature 1995, 376, 238-240.
(21) Haruta, M. Chem. Rec. 2003, 3, 75-87.

(22) Bamwenda, G. R.; Tsubota, S.; Nakamura, T.; Haruta, M. Catal.
Lett. 1997, 44, 83-87.

(23) Shiju, N. R.; Guliants, V. V. Appl. Catal., A 2009, 356, 1-17.
(24) Corma, A.; Garcia, H. Top. Catal. 2008, 48, 8-31.

(25) Zecchina, A.; Groppo, E.; Bordiga, S. Chem. Eur. J. 2007, 13, 2440-
2460.

(26) Mirkin, C. A. Small 2005, 1, 14-16.

(27) Doyle, A. M.; Shaikhutdinov, S. K.; Jackson, S. D.; Freund, H.-J.
Angew. Chem. Int. Ed. 2003, 42, 5240-5243.

(28) Schmid, G.; Maihack, V.; Lantermann, F.; Peschel, S. J. Chem.
Soc, Dalton Trans. 1996, 589 - 595.

(29) Jacobsen, H.; Kleinshmit, P. In Handbook of Heterogeneous Catalysis;
Ertl, G., Knbzinger, H., Weitkamp, J., Eds.; VCH: Weinheim, Germany, 1997.

(30) Qian, K.; LV, 8.; Xiao, X.; Sun, H.; Lu, J.; Luo, M.; Huang, W. J. Mol.
Catal. A: Chem. 2009, 306, 40-47.

(31) Boaro, M.; Vicario, M.; Llorca, J.; De Leitenburg, C.; Dolcetti, G.;
Trovarelli, A. Appl. Catal., B 2009, 88, 272-282.

(32) Han, F.; Kambala, V. S. R.; Srinivasan, M.; Rajarathnam, D.; Naidu,
R. Appl. Catal., A 2009, 359, 25-40.

35

(33) Tasis, D.; Tagmatarchis, N.; Bianco, A.; Prato, M. Chem. Rev. 2006,
106, 1105-1136.

(34) Toebes, M. L.; Van Dillen, J. A.; De Jong, K. P. J. Mol. Catal. A:
Chem. 2001, 173, 75-98.

(35) Leblond, C. R.; Andrews, A. T.; Sun, Y.; Sowa, J. R., Jr. Org. Lett.
2001, 3, 1555-1557.

(36) Kohler, K.; Heidenreich, R. G.; Krauter, J. G. E.; Pietsch, J. Chem.
Eur. J. 2002, 8, 622-631.

(37) Toshima, N.; Shiraishi, Y.; Teranishi, T.; Miyake, M.; Tominaga, T.;
Watanabe, H.; Brijoux, W.; Bonnemann, H.; Schmid, G. Appl. Organomet.
Chem. 2001, 15, 178-196.

(38) Reetz, M. T.; Schulenburg, H.; Lopez, M.; Spliethoff, B.; Tesche, B.
Chimia 2004, 58, 896-899.

(39) Wildgoose, G. G.; Banks, C. E.; Compton, R. G. Small 2006, 2, 182-
193.

(40) Pan, X.; Bao, X. Chem. Commun. 2008, 6271-6281.

(41) Wang, C.; Guo, 8.; Pan, X.; Chen, W.; Bao, X. J. Mater. Chem.
2008, 18, 5782-5786.

(42) Endo, M.; Kim, Y. A.; Ezaka, M.; Osada, K.; Yanagisawa, T.;
Hayashi, T.; Terrones, M.; Dresselhaus, M. S. Nano Lett. 2003, 3, 723-
726.

(43) Sun, X. P.; Zhang, Z. L.; Zhang, B. L.; Dong, X. D.; Dong, S. J.;
Wang, E. K. Chin. Chem. Lett. 2003, 14, 866-869.

(44) Tsunoyama, H.; lchikuni, N.; Tsukuda, T. Langmuir 2008, 24, 11327-
11330.

(45) Tsunoyama, H.; Sakurai, H.; Negishi, Y.; Tsukuda, T. J. Am. Chem.
Soc. 2005, 127, 9374-9375.

(46) Liu, W.; Yang, X.; Huang, W. J. Colloid Interface Sci. 2006, 304,
160-165.

36

(47) Dokoutchaev, A.; James, J. T.; Koene, S. C.; Pathak, S.; Prakash,
G. K. S.; Thompson, M. E. Chem. Mater. 1999, 11 , 2389-2399.

(48) Jana, S.; Ghosh, S. K.; Nath, 8.; Pande, S.; Praharaj, 8.; Panigrahi,
S.; Basu, S.; Endo, T.; Pal, T. Appl. Catal., A 2006, 313, 41-48.

(49) Shi, F.; Zhang, 0.; Ma, Y.; He, Y.; Deng, Y. J. Am. Chem. Soc.
2005, 127, 4182-4183.

(50) lshida, T.; Haruta, M. Angew. Chem. Int. Ed. 2007, 46, 7154-7156.
(51) Wittmer, J.; Johner, A.; Joanny, J. F. J. Phys. II 1995, 5, 635-654.

(52) Tomalia, D. A.; Naylor, A. M.; Goddard, W. A. Angew. Chem. 1990,
102, 119-57.

(53) Kim, Y.-G.; Oh, S.-K.; Crooks, R. M. Chem. Mater. 2004, 16, 167-
172.

(54) Niu, Y.; Crooks, R. M. Comptes Rendus Chimie 2003, 6, 1049-1059.
(55) Balogh, L.; Tomalia, D. A. J. Am. Chem. Soc. 1998, 120, 7355-7356.
(56) Esumi, K.; lsono, R.; Yoshimura, T. Langmuir 2004, 20, 237-243.

(57) Hayakawa, K.; Yoshimura, T.‘; Esumi, K. Langmuir 2003, 19, 5517-
5521.

(58) Zhao, M.; Crooks, R. M. Adv. Mater. 1999, 11 , 217-220.

(59) Zhao, M.; Sun, L.; Crooks, R. M. J. Am. Chem. Soc. 1998, 120,
4877-4878.

(60) Niu, Y.; Yeung, L. K.; Crooks, R. M. J. Am. Chem. Soc. 2001, 123,
6840-6846.

(61) Oh, S.-K.; Niu, Y.; Crooks, R. M. Langmuir 2005, 21, 10209-10213.

(62) Andres, R.; De Jesus, E.; Flores, J. C. New J. Chem. 2007, 31,
1161-1191.

(63) Armor, J. N. Catal. Today 1995, 25, 199-207.

37

(64) Yildirim, Y.; Gobina, E.; Hughes, R. J. Membr. Sci. 1997, 135, 107-
115.

(65) Collins, J. P.; Schwartz, R. W.; Sehgal, R.; Ward, T. L.; Brinker, C.
J.; Hagen, G. P.; Udovich, C. A. Ind. Eng. Chem. Res. 1996, 35, 4398-
4405.

(66) Uemiya, 8.; Sato, N.; Ando, H.; Kikuchi, E. Ind. Eng. Chem. Res.
1991, 30, 585-589.

(67) Shu, J.; Grandjean, B. P. A.; Kaliaguine, S. Appl. Catal., A 1994,
119, 305-325.

(68) HSiung, T. H.; Christman, D. 0.; Hunter, E. J.; Homyak, A. R. AIChE
J. 1999, 45, 204-208.

(69) Dalmon, J.-A. In Handbook of Heterogeneous Catalysis; Ertl, G.,
Knozinger, H., Weitkamp, J., Eds.; Wiley-VCH: Weinheim, Germany, 1997; Vol.
3, p 1387-1398.

(70) Dittmeyer, R.; HOllein, V.; Daub, K. J. Mol. Catal. A: Chem. 2001,
173, 135-184.

(71) Lipnizki, F.; Field, R. W.; Ten, P.-K. J. Membr. Sci. 1999, 153, 183-
210.

(72) Dixon, A. G. Specialist Periodical Reports: Catalysis 1999, 14, 40-
92.

(73) Marcano, J. G. 8.; Tsotsis, T. T. Catalytic Membranes and Membrane
Reactors; Wiley-VCH: Weinheim, Germany, 2002.

(74) Dittmeyer, R.; Svajda, K.; Reif, M. Top. Catal. 2004, 29, 3-27.

(75) Miachon, 8.; Perez, V.; Crehan, G.; Torp, E.; Raeder, H.; Bredesen,
R.; Dalmon, J. A. Catal. Today 2003, 82, 75-81.

(76) Espro, C.; Arena, F.; Frusteri, F.; Parmaliana, A. Catal. Today 2001,
67, 247-256.

(77) Pan, X. L.; Liu, B. J.; Xiong, G. X.; Sheng, 8. 8.; Liu, J.; Yang, W. S.
Catal. Lett. 2000, 66, 125-128.

38

(78) Bottino, A.; Capannelli, G.; Comite, A.; Di Felice, R. Desalination
2002, 144, 411-416.

(79) Veldsink, J. W. J. Am. Oil Chem. Soc. 2001, 78, 443-446.

(80) Daub, K.; Wunder, V. K.; Dittmeyer, R. Catal. Today 2001, 67, 257-
272.

(81) Reif, M.; Dittmeyer, R. Catal. Today 2003, 82, 3-14.

(82) Gabelman, A.; Hwang, 8.-T. J. Membr. Sci. 1999, 159, 61-106.
(83) Guibal, E.; Vincent, T. J. Environ. Manage. 2004, 71, 15-23.

(84) Vincent, T.; Guibal, E. Environ. Sci. Technol. 2004, 38, 4233-4240.
(85) Vincent, T.; Guibal, E. Langmuir 2003, 19, 8475-8483.

(86) Vincent, T.; Spinelli, 8.; Guibal, E. Ind. Eng. Chem. Res. 2003, 42,
5968-5976.

(87) Westermann, T.; Melin, T. Chem. Eng. Process. 2009, 48, 17-28.

(88) Pina, M. P.; Menéndez, M.; Santamaria, J. Appl. Catal., B 1996, 11 ,
L19-L27.

(89) Maire, A. J.; Lau, w. N; Lee, c. Y.; Yue, P. L.; Chan, 0. K.; Yeung,
K. L. Chem. Eng. Sci. 2003, 58, 959-962.

(90) Tsuru, T.; Kan-No, T.; Yoshioka, T.; Asaeda, M. Catal. Today 2003,
82, 41-48.

(91) Kormann, H.-P.; Schmid, G.; Pelzer, K.; Philippot, K.; Chaudret, B.
Z. Anorg. Allg. Chem. 2004, 630, 1913-1918.

(92) Pelzer, K.; Philippot, K.; Chaudret, B.; Meyer-Zaika, W.; Schmid, G.
Z. Anorg. Allg. Chem. 2003, 629, 1217-1222.

(93) Lange, C.; Storck, 8.; Tesche, B.; Maier, W. F. J. Catal. 1998, 175,
280-293.

(94) Lambert, C.; Gonzalez, R. Catal. Lett. 1999, 57, 1-7.

39

(95) GrOschel, L.; Haidar, R.; Beyer, A.; Coelfen, H.; Frank, B.;
Schomacker, R. Ind. Eng. Chem. Res. 2005, 44, 9064-9070.

(96) Schmidt, A.; Haidar, R.; Schomacker, R. Catal. Today 2005, 104,
305—312.

(97) Schmidt, A.; Wolf, A.; Warsitz, R.; Dittmeyer, R.; Urbanczyk, D.;
Voigt, l.; Fischer, G.; Schomacker, R. AIChE J. 2008, 54, 258-268.

(98) Schmidt, A.; Schomacker, R. J. Mol. Catal. A: Chem. 2007, 271,
192-199.

(99) Bengtson, G.; Fritsch, D. Desalination 2006, 200, 666-667.
(100) D. Fritsch, G. B. Adv. Eng. Mater. 2006, 8, 386-389.

(101) Purnama, H.; Kurr, P.; Schmidt, A.; Schomacker, R.; Voigt, l.; Wolf,
A.; Warsitz, R. AIChE J. 2006, 52, 2805-2811.

(102) llinich, O. M.; Gribov, E. N.; Simonov, P. A. Catal. Today 2003, 82,
49-56.

(103) Dominguez-Dominguez, 8.; Arias-Pardilla, J.; Berenguer-Murcia, A.;
Morallon, E.; Cazorla-Amoros, D. J. Appl. Electrochem. 2008, 38, 259-268.

(104) Haas, l.; Gedanken, A. Chem. Commun. 2008, 1795-1797.
(105) Gedanken, A. Ultrason. Sonochem. 2004, 11 , 47-55.

(106) Glaspell, G.; Hassan, H. M. A.; Elzatahry, A.; Abdalsayed, V.; EI-
Shall, M. 8. Top. Catal. 2008, 47, 22-31.

(107) Glaspell, G.; Fuoco, L.; El-Shall, M. 8. J. Phys. Chem. B 2005, 109,
1 7350-17355.

(108) Savastenko, N.; Volpp, H. R.; Gerlach, 0.; Strehlau, W. J. Nanopart.
Res. 2008, 10, 277-287.

(109) Barau, A.; Budarin, V.; Caragheorgheopol, A.; Luque, R.;
Macquarrie, D. J.; Prelle, A.; Teodorescu, V. 8.; Zaharescu, M. Catal. Lett.
2008, 124, 204-214.

(110) Campelo, J. M.; Lee, A. F.; Luque, R.; Luna, D.; Marinas, J. M.;
Romero, A. A. Chem. Eur. J. 2008, 14, 5988-5995.

40

(111) Perego, 0.; Villa, P. Catal. Today 1997, 34, 281-305.
(112) Uzio, D.; Miachon, 8.; Dalmon, J.-A. Catal. Today 2003, 82, 67-74.

(113) Perez, V.; Miachon, 8.; Dalmon, J.-A.; Bredesen, R.; Pettersen, G.;
Raeder, H.; Simon, C. Sep. Purif. Technol. 2001, 25, 33-38.

(114) Akolekar, D. B.; Bhargava, 8. K.; Foran, G.; Takahashi, M. J. Mol.
Catal. A: Chem. 2005, 238, 78-87.

(115) Centeno, M. A.; Portales, C.; Carrizosa, l.; Odriozola, J. A. Catal.
Lett. 2005, 102, 289-297.

(116) Date, M.; Okumura, M.; Tsubota, 8.; Haruta, M. Angew. Chem. Int.
Ed. 2004, 43, 2129-2132.

(117) Tsubota, 8.; Haruta, M.; Kobayashi, T.; Ueda, A.; Nakahara, Y. Stud.
Surf. Sci. Catal. 1991, 63, 695-704.

(118) Liu, H.; Ma, D.; Blackley, R. A.; Zhou, W.; Bao, X. Chem. Commun.
2008, 2677-2679.

(119) Bekyarova, E.; Kaneko, K. Adv. Mater. 2000, 12, 1625-1628.

(120) Xu, C.; Ouyang, C.; Jia, R.; Li, Y.; Wang, X. J. Appl. Polym. Sci.
2009, 111, 1763-1768.

(121) Hasik, M.; Turek, W.; Nyczyk, A.; Stochmal, E.; Bernasik, A.;
Sniechota, A.; Soltysek, A. Catal. Lett. 2009, 127, 304-311.

(122) Mallick, K.; Mondal, K.; Witcomb, M.; Scurrell, M. J. Mater. Sci.
2008, 43, 6289-6295.

(123) Scalzullo, 8.; Mondal, K.; Witcomb, M.; Deshmukh, A.; Scurrell, M.;
Mallick, K. Nanotechnology 2008, 19, 075708/1-075708/8.

(124) Meyer, D. E.; Bhattacharyya, D. J. Phys. Chem. B 2007, 111, 7142-
7154.

(125) Meyer, D. E.; Wood, K.; Bachas, L. G.; Bhattacharyya, D. Environ.
Prog. 2004, 23, 232-242.

(126) Serp, P.; Kalck, P.; Feurer, R. Chem. Rev. 2002, 102, 3085-3128.

41

(127) Okumura, M.; Nakamura, 8.; Tsubota, 8.; Nakamura, T.; Azuma, M.;
Haruta, M. Catal. Lett. 1998, 51, 53-58.

(128) Liang, C.; Xia, W.; Soltani-Ahmadi, H.; Schlueter, O.; Fischer, R. A.;
Muhler, M. Chem. Commun. 2005, 282-284.

(129) Laaksonen, T.; Ahonen, P.; Johans, C.; Kontturi, K.
ChemPhysChem 2006 7, 2143 - 2149.

(130) Kimling, J.; Maier, M.; Okenve, B.; Kotaidis, V.; Ballot, H.; Plech, A.
J. Phys. Chem. B 2006, 110, 15700-15707.

(131) Turkevich, J.; Stevenson, P. C.; Hillier, J. Discussions of the Faraday
Society 1951, No. 11, 55-75.

(132) Turkevich, J.; Stevenson, P. C.; Hillier, J. J. Phys. Chem. 1953, 57,
670-673.

(133) Kumar, 8.; Gandhi, K. 8.; Kumar, R. Ind. Eng. Chem. Res. 2007, 46,
3128-3136.

(134) Frens, G. Nature Physical Science 1973, 241, 20-22.
(135) Freund, P. L.; Spiro, M. J. Phys. Chem. 1985, 89, 1074-1077.

(136) Chow, M. K.; Zukoski, C. F. J. Colloid Interface Sci. 1994, 165, 97-
109.

(137) Turkevich, J.; Kim, G. Science 1970, 169, 873-9.

(138) Sehayek, T.; Lahav, M.; Popovitz-Biro, R.; Vaskevich, A.;
Rubinstein, l. Chem. Mater. 2005, 17, 3743-3748.

(139) Turkevich, J.; Miner, R. 8., Jr.; Babenkova, L. J. Phys. Chem. 1986,
90, 4765-7.

(140) Brugger, P. A.; Cuendet, P.; Graetzel, M. J. Am. Chem. Soc. 1981,
103, 2923-7.

(141) Pillai, Z. 8.; Kamat, P. V. J. Phys. Chem. B 2004, 108, 945-951.
(142) Lee, P. C.; Meisel, D. J. Phys. Chem. 1982, 86, 3391-3395.

42

(143) Brust, M.; Walker, M.; Bethell, D.; Schiffrin, D. J.; Whyman, R. J.
Chem. Soc, Chem. Commun. 1994, 801-2.

(144) Brust, M.; Fink, J.; Bethell, D.; Schiffrin, D. J.; Kiely, C. J. Chem.
Soc., Chem. Commun. 1995, 1655-1657.

(145) Dahl, J. A.; Maddux, B. L. 8.; Hutchison, J. E. Chem. Rev. 2007,
107, 2228-2269.

(146) Chen, 8.; Kimura, K. Langmuir1999, 15, 1075-1082.

(147) Yonezawa, T.; Sutoh, M.; Kunitake, T. Chem. Lett. 1997, 619-621.
(148) Chen, 8.; Kimura, K. J. Phys. Chem. B 2001, 105, 5397-5403.
(149) Leff, D. V.; Brandt, L.; Heath, J. R. Langmuir 1996, 12, 4723-4730.
(150) Jana, N. R.; Peng, X. J. Am. Chem. Soc. 2003, 125, 14280-14281.
(151) Hiramatsu, H.; Osterloh, F. E. Chem. Mater. 2004, 16, 2509-2511.

(152) Aslam, M.; Fu, L.; Su, M.; Vijayamohanan, K.; Dravid, V. P. J. Mater.
Chem. 2004, 14, 1795 - 1797.

(153) Newman, J. D. 8.; Blanchard, G. J. Langmuir 2006, 22, 5882-5887.

(154) Weare, W. W.; Reed, 8. M.; Warner, M. G.; Hutchison, J. E. J. Am.
Chem. Soc. 2000, 122, 12890-12891.

(155) Warner, M. G.; Reed, 8. M.; Hutchison, J. E. Chem. Mater. 2000, 12,
3316-3320.

(156) Brown, L. O.; Hutchison, J. E. J. Am. Chem. Soc. 1999, 121, 882-
883.

(157) Braunstein, P.; Kormann, H. -P. Meyer-Zaika, W.; Pugin, R; Schmid,
G. Chem. Eur. J. 2000, 6, 4637-4646.

(158) Lahav, M.; Sehayek, T.; Vaskevich, A.; Rubinstein, l. Angew. Chem.
Int. Ed. 2003, 42, 5576-5579.

(159) Decher, G.; Schlenoff, J. B. Multilayer Thin Films: Sequential Assembly of
Nanocomposite Materials; Wiley-VCH: Weinheim, Germany, 2003.

43

(160) Decher, G.; Schmitt, J. Thin Solid Films 1992, 210/211, 831-835.
(161) Benkirane-Jessel, N.; Lavalle, P.; Ball, V.; Ogier, J.; Senger, B.;
Picart, C.; Schaaf, P.; Voegel, J.-C.; Decher, G. Macromol. Eng. 2007, 2,
1249-1305.

(162) Decher, G. Science 1997, 277, 1232-1237.

(163) Arys, X.; Jonas, A. M.; Laschewsky, A.; Legras, R. In Supramolecular
Polymers; Ciferri, A., Ed.; CRC Press: 2005, p 651-710.

(164) Rubner, M. F. In Multilayer Thin Films; Decher, G., Schlenoff, J. B., Eds.
2003, p 133-154.

(165) Shiratori, 8. 8.; Rubner, M. F. Macromolecules 2000, 33, 4213-4219.

(166) Ruths, J.; Essler, F.; Decher, G.; Riegler, H. Langmuir 2000, 16,
8871-8878.

(167) Kurth, D. G.; Volkmer, D.; Klitzing, R. V. In Multilayer Thin Films; Decher,
G., Schlenoff, J. B., Eds.; Wiley-VCH: Weinheim, Germany, 2003.

(168) Porcel, C. H.; lzquierdo, A.; Ball, V.; Decher, G.; Voegel, J. C.;
Schaaf, P. Langmuir 2005, 21 , 800-802.

(169) Picart, C.; Lavalle, P.; Hubert, P.; Cuisinier, F. J. G.; Decher, G.;
Schaaf, P.; Voegel, J. C. Langmuir 2001, 17, 7414-7424.

(170) Ladam, G.; Schaaf, P.; Cuisinier, F. J. G.; Decher, G.; Voegel, J.-C.
Langmuir 2001 , 1 7, 878-882.

(171) Cavalieri, F.; Postma, A.; Lee, L.; Caruso, F. ACS Nano 2009, 3,
234-240.

(172) Johnston, A. P. R.; Read, E. 8.; Caruso, F. Nano Lett. 2005, 5, 953-
956.

(173) Wang, Y.; Caruso, F. Chem. Mater. 2005, 17, 953-961.

(174) Bergbreiter, D. E.; Franchina, J. G.; Kabza, K. Macromolecules
1999, 32, 4993-4998.

(175) Zhao, M.; Liu, Y.; Crooks, R. M.; Bergbreiter, D. E. J. Am. Chem.
Soc. 1999, 121, 923-930.

(176) Srivastava, 8.; Kotov, N. A. Acc. Chem. Res. 2008, 41, 1831-1841.

(177) Qin, 8.; Qin, D.; Ford, W. T.; Zhang, Y.; Kotov, N. A. Chem. Mater.
2005, 17, 2131-2135.

(178) Lee, 8. W.; Kim, B.-S.; Chen, 8.; Shao-Hom, Y.; Hammond, P. T. J.
Am. Chem. Soc. 2009, 131, 671-679.

(179) Lee, D.; Gemici, 2.; Rubner, M. F.; Cohen, R. E. Langmuir 2007, 23,
8833-8837.

(180) Lee, D.; Rubner, M. F.; Cohen, R. E. Nano Lett. 2006, 6, 2305-2312.

(181) Lee, D.; Omolade, D.; Cohen, R. E.; Rubner, M. F. Chem. Mater.
2007, 19, 1427-1433.

(182) lchinose, l.; Kawakami, T.; Kunitaki, T. Adv. Mater. 1998, 10, 535-
539.

(183) Chen, H.; Dong, S. Talanta 2007, 71, 1752-1756.

(184) Chirea, M.; Garcia-Morales, V.; Manzanares, J. A.; Pereira, C.;
Gulaboski, R.; Silva, F. J. Phys. Chem. B 2005, 109, 21808-21817.

(185) Chirea, M.; Pereira, C. M.; Silva, F. J. Phys. Chem. C 2007, 111,
9255-9266.

(186) Yang, M.; Yang, Y.; Yang, H.; Shen, G.; Yu, R. Biomaterials 2005,
27, 246-255.

(187) Srivastava, 8.; Ball, V.; Podsiadlo, R; Lee, J.; Ho, P.; Kotov, N. A. J.
Am. Chem. Soc. 2008, 130, 3748-3749.

(188) Malikova, N.; Pastoriza-Santos, l.; Schierhorn, M.; Kotov, N. A.; Liz-
Marzan, L. M. Langmuir 2002, 18, 3694-3697.

(189) Lee, 0.; Cohen, R. E.; Rubner, M. F. Langmuir 2005, 21 , 9651-9659.

(190) Joly, 8.; Kane, R.; Radzilowski, L.; Wang, T.; Wu, A.; Cohen, R. E.;
Thomas, E. L.; Rubner, M. F. Langmuir 2000, 16, 1354-1359.

(191) Del, J.; Bruening, M. L. Nano Lett. 2002, 2, 497-501.

45

(192) Yang, S. Y.; Lee, D.; Cohen, R. E.; Rubner, M. F. Langmuir 2004,
20, , 5978-5981 .

(193) Kim, B.-S.; Park, 8. W.; Hammond, P. T. ACS Nano 2008, 2, 386-
392.

(194) Delongchamp, D. M.; Hammond, P. T. Langmuir 2004, 20, 5403-
5411.

(195) Bergbreiter, D. E.; Tao, G.; Franchina, J. G.; Sussman, L.
Macromolecules 2001, 34, 3018-3023.

(196) Sun, J.; Wu, T.; Liu, F.; Wang, Z.; Zhang, X.; Sheri, J. Langmuir
2000, 16, 4620-4624.

(197) Kohli, P.; Blanchard, G. J. Langmuir 2000, 16, 8518-8524.

(198) Kidambi, 8.; Bruening, M. L. Chem. Mater. 2005, 17, 301-307.
(199) Yu, A.; Liang, Z.; Caruso, F. Chem. Mater. 2005, 17, 171-175.
(200) Bhattacharjee, 8.; Bruening, M. L. Langmuir 2008, 24, 2916-2920.
(201) Malaisamy, R.; Bruening, M. L. Langmuir 2005, 21, 10587-10592.

(202) Stanton, B. W.; Harris, J. J.; Miller, M. D.; Bruening, M. L. Langmuir
2003, 19, 7038-7042.

(203) Bruening, M. In Multilayer Thin Films; Decher, G., Schlenoff, J. B., Eds.
2003, p 487-510.

(204) Harris, J. J.; Derose, P. M.; Bruening, M. L. J. Am. Chem. Soc. 1999,
121, 1978-1979.

(205) Martin, C. R. Science 1994, 266, 1961-6.

(206) AI, 8.; Lu, G.; He, 0.; Li, J. J. Am. Chem. Soc. 2003, 125, 11140-
11141.

(207) Liang, Z.; Susha, A. 8.; Yu, A.; Caruso, F. Adv. Mater. 2003, 15,
1849-1853.

46

Some of the text and ﬁgures contained in chapter two appeared in the following
publications.

Dotzauer, D. M.; Dai, J.; Sun, L.; Bruening, M. L. Nano Lett. 2006, 6, 2268-2272.
Dotzauer, D. M.; Bhattacharjee, 8.; Wen, Y.; Bruening, M. L. Langmuir 2009, 25,
1865-1871.

47

Chapter 2: Catalytic Reduction of Nitroaromatic Compounds with Au-
Nanoparticle Containing Membranes
2.1 Introduction
The reduction of aromatic nitro compounds to the corresponding amines is
important in synthetic chemistry because aromatic amines are frequent
intermediates in the production of pharmaceuticals, agrochemicals, polymers,
dyes, and other ﬁne chemicals.145 A variety of methods exist for nitro group

reduction, including direct hydrogenation,7'12 catalytic transfer hydrogenation,13“5

l17,18 19-21

and photocatalytic,16 electrochemica, enzymatic, and hydride reduction.22
In each case, transformation to the amine likely occurs via formation of nitroso
and hydroxylamine intermediates.17 Furthermore, under basic conditions,
additional azo, azoxy, and hydrazo intermediates may also form (Scheme 2.1).23

NHOH

:20 9:1» O @5726“ O:
1-.../

Scheme 2.1 Formation of functionalized anilines from the parent nitro compound
via hydrogenation with multiple reduction pathways.

48

If reactions of these intermediates occur too slowly, they may accumulate.”26

Thus, it is important to develop reactions that lead to rapid formation of amine
compounds.

Hydrogenation of simple aromatic nitro compounds readily occurs with
traditional commercial catalysts and can often be carried out easily at large
scales.24 However, the selective hydrogenation of nitro groups in the presence of
other reducible functional groups is more challenging. Reduction of these other
functional groups can occur during any one of the intermediate steps in the
transformation to the amine. Therefore, effective nitro-hydrogenation catalysts
should be active toward each step in the nitro reduction but inactive toward
reduction of other functional groups such as C=C, CEC, CEN, C=O, and CI.

Recently, a number of studies focused on the development of noble metal
nanoparticle catalysts and methods that can provide high selectivity for the amine
product while maintaining high yields. Studies by Serna et al. Show that titania-
supported AuNPs provide selectivity for nitro group reduction in the presence of
C=C, C=O, and CEN functional groups.”29 Chen et al. also used supported
AuNPs to demonstrate selective reduction for a number of nitroaromatic

compounds.30

Gold is sometimes chosen as a catalyst material because it can
effectively catalyze the reduction of the nitro group, albeit under relatively high
temperature, whereas it is much less active than traditional platinum or palladium
catalysts in the hydrogenation of other reducible functional groups.31 This may

be partly because N02 adsorbs molecularly to the surface of Au and

dissociatively to Pt or Pd surfaces.32 Nevertheless, the high

49

temperature/pressure and relatively long reaction times for nitro hydrogenation
with gold catalysts limit the application of these materials.27

Carbon nanofiber-supported Pt and Pd nanoparticles catalyze selective
hydrogenation of aromatic nitro groups in halogenated compounds, but these
catalysts are less selective in the presence of other reducible functional groups.33
Ranu et al. use Cu nanoparticles and ammonium formate in ethylene glycol for
reduction of aromatic nitro compounds with high chemoselectivity. However,
these reactions also take place under high temperature (120 °C) and long
reaction times (8-12 m.”

A number of studies also examine the reduction of nitroaromatic
compounds by sodium borohydride in reactions facilitated by nanoparticles in

“'36 and on polymeric supports.3M4 These studies mainly discuss the

solution
catalytic reduction of 4-nitrophenol, however, and do not Show whether these
catalyst systems can selectively reduce nitro groups in compounds containing
other reducible functional groups. More recent research investigated the use of
Au45 and Ni46 clusters as chemoselective catalysts in the reduction of
nitroaromatic compounds with sodium borohydride, but these clusters are difﬁcult
to prepare.

This chapter examines the use of Au nanoparticle-containing membranes
as catalysts for the reduction of aromatic nitro groups. We begin by discussing
the preparation and characterization of Au nanoparticle-containing ﬁlms. The

catalytic reduction of 4-nitrophenol to 4-aminophenol by sodium borohydride was

used as a test reaction to investigate the kinetics of membrane reactions.

50

Controlling the ﬂow rate and therefore residence time in membrane reactions
allows excellent control over conversion. Remarkably, Au nanoparticles
contained in membranes exhibit the same activity as unsupported Au
nanoparticles in solution. Furthermore, additional studies looking at the reduction
of nitro groups in the presence of a variety of reducible and non-reducible
functional groups demonstrate the ability to control product formation in

membrane reactions.

2.2 Experimental Methods
2.2.1 Materials

Anodisc aluminum oxide membranes (25 mm disks with 0.2 pm diameter
pores) and Nuclepore track-etched polycarbonate membranes (25 mm disks with
0.2 pm diameter pores) were purchased from Whatman. Nylon membranes (25
mm disks with nominal 0.45 pm diameter pores) were purchased from Millipore.
Nitroaromatic compounds, HAuCl4, sodium citrate, sodium borohydride,
aluminum oxide (100 mesh), poly(allylamine hydrochloride) (PAH, MW = 15,000),
poly(ethylenimine) (PEI, branched, MW = 25,000), and poly(sodium styrene
sulfonate) (P88, MW = 70,000) were reagent grade and used as received from
Sigma-Aldrich. Poly(acrylic acid) (PAA, MW = 5,000) was obtained from
Polysciences. All solutions were prepared with deionized water (18.2 MO)

obtained from a Milli-Q (Millipore) puriﬁcation system.

51

2.2.2 Modiﬁcation of Aluminum Wafers with Polyelectrolyte/AuNP Films

Gold nanoparticles with an average diameter of 12 nm were prepared
according to the citrate reduction method described in the literature.“48 In a 250
mL round bottom ﬂask equipped with a condenser, 100 mL of aqueous HAuCI4
(1.0 mM) was heated to boiling under vigorous stirring, and 10 mL of a heated
sodium citrate (38.8 mM) was added to this solution under continued stirring.
Immediately after adding the sodium citrate, the solution turned from faint yellow
to colorless. After 15 seconds, the color changed to light blue, then dark blue,
and ﬁnally to red after about 45 seconds. The mixture was boiled while stirring
for 15 minutes to allow the reaction to go to completion, and then removed from
heat and stirred for an additional 10 minutes. The resulting AuNP solutions were
stored in an amber bottle at 5 °C until use.

Al-coated Si wafers (200 nm Al on 8i(100) wafers) were ﬁrst cleaned in a
UV/ozone cleaner (Boekel) for 15 min. For ﬁlm formation, a bare aluminum
wafer was immersed in an aqueous solution containing 0.02 M PAA and 0.5 M
NaCl (pH adjusted to 4.5 with 1.0 M NaOH) for 5 min (molarities of
polyelectrolytes are always given with respect to the repeating unit). The
aluminum substrate was rinsed with deionized water for 1 min before exposure to
a solution containing 0.5 M NaCl and either 0.02 M PAH (pH adjusted to 4.5 with
0.1 M HCI) or 0.02 M PEI (pH adjusted to 9.0 with 1.0 M HCI) for 5 min, followed
by another water rinse for 1 min. The substrate was then immersed for 5 min in

the as-prepared AuNP solution for adsorption of a layer of AuNPs and ﬁnally

52

rinsed for 1 min in deionized water. Alternating exposure to polycation and AuNP

solutions was repeated until the desired number of bilayers was deposited.

2.2.3 Modification of Aluminum Oxide Powder with Polyelectrolyte/AuNP
Films

To prepare heterogeneous catalysts for use in a conventional batch
reactor, alumina powder was modified with AuNPs using LbL deposition of a
polyelectrolyte/AuNP ﬁlm as shown in Figure 2.1. This catalyst preparation
procedure was developed from a previous method for in-situ preparation of Pd
nanoparticles in polyelectrolyte multilayer ﬁIms.49-5° In a 100-mL ﬂask, 7.5 g of
100 mesh alumina was added to 50 mL of a pH 4.5 solution containing 0.02 M
PAA and 0.5 M NaCl, and then stirred vigorously for 10 min. The alumina was
then allowed to settle and the supernatant was decanted. To remove excess
polyelectrolyte, the alumina particles were washed 4 times by adding 50 mL of
deionized water, stirring, letting the alumina settle, and decanting the water.
Next, 50 mL of a pH 4.5 solution containing 0.02 M PAH and 0.5 M NaCl was
added to the alumina, and this mixture was stirred for 10 min, after which the

particles were washed as described above.

 

Figure 2.1 Schematic diagram of the deposition of polyelectrolyte/AuNP ﬁlms on
alumina powder.

To deposit the AuNPs, 15 mL of the as-prepared AuNP solution was
diluted to 60 mL with deionized water, and this solution was added to the
PAA/PAH-modiﬁed alumina and stirred for 10 min. The alumina was then
allowed to settle and rinsed 4 times with water as described above. After
decanting the ﬁnal water rinse, the resulting powder was dried under vacuum for
8 hours to remove residual water. The ﬁnal gold loading was 0.4 mg Au per g of
alumina powder as determined by ﬂame atomic absorption spectroscopy (FAAS,

Varian SpectrAA-200).

2.2.4 Modification of Disk-Shaped Membranes with Polyelectrolyte/AuNP
Films

Anodized alumina membranes were pretreated in a UV/ozone cleaner for
15 min prior to modiﬁcation with polyelectrolyte films. During ﬁlm formation, the
membranes were placed in an Amicon 8010 ultraﬁltration cell, and all solutions
were passed through the membranes using a peristaltic pump (Cole Parmer
Masterﬂex C/L) attached to the permeate side of the Amicon cell. For alumina
membrane modiﬁcation, solutions of PAA (10-15 mL), water (40 mL), PAH (10-15
mL), water (40 mL), citrate-stabilized AuNPs, and water (40 mL) were
sequentially passed through the membranes at a ﬂux of ~ 0.02 mUcmZ-sec to
deposit a PAA/PAH/AuNP ﬁlm. Additional PAH/AuNP bilayers could be added
by simply alternating passage of the PAH and AuNP solutions with rinsing in
between. The polyelectrolyte deposition solutions had the same composition as

those used for modiﬁcation of alumina powder. Polycarbonate and nylon

membrane modiﬁcation involved the same process, except a solution containing
0.02 M PSS (pH not adjusted) and 0.5 M NaCl was used to deposit the precursor
polyanion layer because P88 likely adheres better than PAA to polymeric

membranes.51

For each membrane type, the as-prepared AuNP solution was
diluted by a factor of 10 to 20 and passed through the membrane until the AuNPs
were no longer adhering to the polyelectrolyte layer, i.e., the solution exiting the
membrane exhibited the same red color as the solution entering the membrane.

For some experiments, PEI solution was used in place of PAH solution during the

deposition process.

2.2.5 Characterization of Polyelectrolyte/AuNP Films

Citrate-stabilized AuNPs were characterized by UV-Vis spectrophotometry
(Perkin Elmer Lambda 40) and transmission electron microscopy (JEM-2200FS
microscope). TEM samples of Au nanoparticles were prepared by diluting the
as-prepared AuNP solution by a factor of 10 and placing a 2 uL drop of the
resulting solution on a carbon-coated copper TEM grid and letting the solvent
evaporate. The thicknesses of PAA/[PAH/AuNP]x and PAA/[PEl/AuNP]x ﬁlms on
Al wafers were determined by ellipsometry using a J. A. Woollam model M-44
rotating analyzer ellipsometer. Au deposition was also conﬁrmed by specular
reﬂectance UV-Vis spectroscopy.

Aluminum wafer and membrane samples were imaged using a Hitachi 8-
4700 ll ﬁeld-emiSSion scanning electron microscope (FESEM). Prior to imaging,

membrane samples were fractured (alumina was broken with tweezers,

55

polycarbonate was simply torn, and nylon was fractured in liquid nitrogen with
tweezers) to reveal a cross-section and coated with 5 to 10 nm of Au with a
Pelco SC-7 sputter coater. No sample preparation was necessary for imaging of
AuNP-containing ﬁlms on Al wafers. To determine the AuNP content in
membrane and alumina-powder samples, aqua regia (3 parts HCI, 1 part HNO3)
was used to dissolve the Au, and these solutions were diluted and analyzed by

FAAS.

2.2.6 Catalytic Reactions with As-Prepared Au Nanoparticles and
Nanoparticles Supported on Alumina Powder

Homogeneous catalytic reactions with the as-prepared AuNPs allowed
comparison of the unsupported AuNP activity with the activities of alumina
powder- and membrane-supported AuNPs. In a typical reaction, 20 (IL of the as-
prepared AuNP solution was added to 3 mL of a solution containing 0.1 mM nitro
compound and 10 mM NaBH4 in a quartz cuvette. Samples were analyzed by
UV-Vis spectrophotometry either by scanning a range of wavelengths during
regular time intervals or monitoring the absorbance at a specific wavelength
continuously. In some cases, the reactions took place in a reaction ﬂask with 20
to 50 mL volumes under continuous stirring, and sample aliquots were collected
and analyzed immediately by UV-Vis.

In a typical slurry reaction using alumina powder-supported AuNPs, 75 mg
of catalyst was added to 60 mL of an aqueous solution containing 0.1 to 5 mM

nitro compound and a 100-fold excess of NaBH4, and the mixture was stirred

56

vigorously at room temperature to suspend the catalyst. The use of a high
excess of NaBH4 ensures that its concentration remains essentially constant
during the reaction, which allows the assumption of pseudo ﬁrst order kinetics
with respect to the nitro compound. Samples of the reaction mixture were
collected at speciﬁc time intervals and ﬁltered through cotton prior to analysis by
UV-Vis spectrophotometry and/or gas chromatography-mass spectrometry (GC-
MS, Hewlett-Packard 618008 GCD system with a HP-5MS column). For GC-
MS analysis, 3 mL of sample was extracted three times into 1 mL portions of
diethyl ether, which were then combined. In all reactions, the MS response was
assumed to be the same for reactants, intermediates, and products. (This
assumption was conﬁrmed in the case of nitrobenzene, aniline, and
nitrosobenzene, but due to the limited availability of other nitroso species it was
not conﬁrmed with other compounds.) The ratios of peak areas from the
chromatograms were used to determine the percent of each substrate in the
reaction mixture. Identiﬁcation of intermediate species was conﬁrmed by
comparison with a mass spectrum library. Azoxy intermediates appeared at low
levels in some reactions and small amounts of azo and azoxy species may have
formed in other cases,”53 but these compounds might not readily elute from the
column. To prove that the extraction was efﬁcient at collecting all reaction
products, the aqueous portion that remained after ether extraction was analyzed

by UV-Vis spectrophotometry from 200 to 600 nm.

57

2.2.7 Catalytic Reactions with Au Nanoparticle-Containing Membranes

To perform reduction reactions with the AuNP-containing membranes, a
room-temperature, aqueous solution containing 0.1 to 5 mM nitro compound and
a 100-fold excess of NaBH4 was passed through the membrane using either a
peristaltic pump (Figure 2.2a) or pressure supplied by nitrogen gas on the feed
side of the membrane (Figure 2.2b). Samples were analyzed by UV-Vis
spectrophotometry and/or GC-MS to determine the extent of the reaction as
described above. In the experiments that examined the effect of temperature on
membrane reactions, the solution was first heated in a silicon oil bath or cooled in
an ice bath to achieve the desired temperature, and then the solution passed
through the membrane as in a typical experiment. From the beginning to the end

of these experiments, the solution temperature ﬂuctuated by only 1 to 2 °C

a) i / Membrane
U

  
  
     

Membrane

  

 

I
Feed :
solution I

 

Figure 2.2 Experimental apparatuses used for performing catalytic reactions
with membranes at a) low ﬂow rates, and b) higher ﬂow rates.

2.3 Results and Discussion

2.3.1 Characterization

2.3.1.1 Au Nanoparticle Characterization

58

0.3 -

0.25 -

0.15 -

Absorbance

0.1 '

 

 

o I l I I I I

400 450 500 550 600 650 700
Wavelength (nm)

Figure 2.3 UV-Vis spectrum of a 0.12 mM Au nanoparticle solution showing a
maximum absorbance at 517 nm. Molarity is given with respect to Au atoms.

20 b)
15
10

Number

0
8 910111213141516
Nanoparticle Diameter (nm)

 

Figure 2.4 a) TEM image of citrate-stabilized AuNPs on a carbon-coated copper
grid. b) Histogram of AuNP diameters taken from several TEM images. The
average particle diameter is 12 i 1 nm.

Au nanoparticles prepared according to the citrate reduction method48
exhibited a bright red color with an absorbance maximum at 517 nm (Figure 2.3),
which suggests a AuNP size of 10 to 15 nm.5‘"55 A red shift in the absorbance

maximum would indicate that some nanoparticle aggregation occurred or that the

59

Size of the nanoparticles was larger than expected.56 Further analysis by TEM
showed a monodisperse size distribution with an average nanoparticle diameter

of 12 1 1 nm (Figure 2.4).

2.3.1.2 Characterization of Polyelectrolyte/AuNP Films on Aluminum

000' H" "W
Poly(acrylic acid) NH: " Cl
Poly(allylamine hydrochloride)
PAH

$03 Na"
Poly(styrene sulfonate)
PSS

H2
Branched Poly(ethylenimine) |\
PEI NH3
4-

Figure 2.5 Structures of polyelectrolytes used in this work.

This study employed PAA or P88 as the polyanion and PAH or PEI as the
polycation (Figure 2.5) for multilayer ﬁlm formation because these are the most
frequently used polyelectrolytes in layer-by-layer ﬁlm formation.57451 The
polyelectrolytes in a particular ﬁlm have a large impact on the amount of
nanoparticle adsorption that takes place. For example, ﬁlms containing branched
PEI should be thicker and thus have more sites for nanoparticle adsorption than

ﬁlms containing PAH. This increased thickness with the PEI-containing ﬁlm is

60

likely due to the highly branched structure of PEI compared to the linear structure
of PAH. The polyanion also plays an important role in nanoparticle adsorption.
In studies involving polymeric substrates, P88 is the best choices for the
precursor polyelectrolyte because it adheres well to polymer surfaces. It PAA is
used instead, almost no polyelectrolyte adsorption takes place, and this leads to
very thin ﬁlms that would not bind much Au.

Prior to depositing polyelectrolyte/AuNP ﬁlms in membranes, we
deposited these coatings on Al-coated Si wafers for characterization. We ﬁrst
used ellipsometry to determine the thickness of [PAA/PAH]x and [PEI/PAA]x ﬁlms.
In the case of [PEI/PAA]X, PEI was deposited at a pH of 9 because the
aluminum/aluminum oxide surface is negatively charged at this pH,62 so it is not
necessary to start with an initial PAA layer. In contrast, we deposited PAH at a
pH of 4.5, so adsorption of [PAA/PAH]x ﬁlms began with PAA. At this deposition
pH, the PAH is partially protonated yielding ﬁlms with thicknesses of several nm,
which are desired for this application.57 Ellipsometry demonstrates uniform,
stepwise growth of both types of ﬁlms (Figure 2.6). The thickness increase after
the addition of each bilayer, with the exception of the initial bilayer, was
approximately 4 nm and 5 nm for PAA/PAH and PEI/PAA ﬁlms, respectively,

which agrees with previous reports.“-63

61

25.0 -

 
 
  

 

 

A I PEI/PAA
E 20.0 - e PAA/PAH
f.- —Linear(PEl/PAA)
3 15.0 . —-Linear(PAA/PAH)
=3
E 10.0 1
I-
E
E 5.0 1
0.0 1 I l I
0 1 2 3 4

Number of Bilayers

Figure 2.6 Ellipsometric thicknesses for multilayered [PEI/PAA]x (squares) and
[PAA/PAH]x (circles) ﬁlms deposited on Al-coated Si wafers.

Determination of the ellipsometric thickness of thin ﬁlms containing non-
transparent Au nanoparticles is inherently much less accurate than measurement
of polymer ﬁlms which have a uniform refractive index. Thus, the deposition of
Au nanoparticles was conﬁrmed with UV-Vis and SEM analysis of the
PAA/[PAH/AuNP]x ﬁlms. Specular reﬂectance UV-Vis spectrophotometry of ﬁlms
on Al-coated Si wafers Showed that a wafer coated with a PAA/PAH bilayer had
almost no absorbance from 400 - 700 nm (Figure 2.7). On the other hand, a
wafer coated with a PAA/PAH/AuNP ﬁlm showed a peak absorbance at 517 nm,
which is consistent with the UV-Vis spectrum of the as-prepared Au
nanoparticles (Figure 2.4). Additionally, F ESEM images of Al wafers coated with
PAA/[PAH/AuNP]x ﬁlms demonstrate that Au nanoparticles deposited uniformly
on the Al wafer and remained well-separated (Figure 2.8). Images of ﬁlms

containing 2 and 3 PAH/AuNP bilayers (Figure 2.8c and 2.8d) show an increased

62

density of Au nanoparticles in the ﬁlms, conﬁrming that the LbL modiﬁcation
increased the Au loading with each deposition step. In the ﬁlm containing 3
PAH/AuNP bilayers (Figure 2.8d), the Au nanoparticles appeared to form chains
of 3 to 5 nanoparticles. However, since these images represent dry ﬁlms, it is
possible that the nanoparticles were still well-separated when the ﬁlms were
immersed in solution. When in solution, polyelectrolytes .often undergo
swelling,59 which allows the AuNPs to have some mobility. The AuNPs may also

have been somewhat separated in the z direction.

°-°5 ' — PAA/PAH/AuNP
0.04 ., — PAA/PAH
3 ——-- Bare Aluminum
5003-
E
in 0.02 '
.o
<
0.01 -

 

.anmm

400 450 500 550 600 650 700
Wavelength (nm)

Figure 2.7 Specular reﬂectance UV-Vis Spectra of an Al-coated Si water before
(light gray line) and after (dark gray line) deposition of a PAA/PAH layer and
subsequent adsorption of Au nanoparticles (black line).

63

     

.,, ,. ‘ . ioOnIit
Figure 2.8 FESEM images of AI—coated Si wafers after deposition of a)
PAA/PAH b) PAA/[PAH/AuNPj1 c) PAA/[PAH/AuNP]2 and d) PAA/[PAH/AuNP]3
films.

2.3.1.2 Characterization of Polyelectrolyte/AuNP Films in Membranes

 

commercially available anodized alumina membrane.

64

We chose to ﬁrst modify anodized alumina membranes because they
have a well-deﬁned cylindrical pore structure, are mechanically stable, and can

dams The anodized alumina membranes used in this

be easily functionalize
work were 60 pm thick with a 5 pm thick skin layer. The skin layer was highly
branched and had pores with ~200 nm diameters (Figure 2.9a). The support
layer of the membranes contained slightly larger pores (~230 nm diameter) that
were cylindrical through the length of the membrane (Figure 2%). During
modiﬁcation of the membranes, we ﬂowed solutions from the support side of the
membrane to the skin layer side because the larger pore sizes in the support
were less susceptible to cake layer formation on the membrane surface.

Alumina membranes should be positively charged when exposed to the
citrate-stabilized AuNP solution (pH 6), but no visible adsorption of AuNPs
occurred when ﬂowing the Au nanoparticle solution through an untreated alumina
membrane. Deposition of the AuNPs in the membrane required adsorption of an
underlying PANPAH bilayer. PAA was chosen as the polyanion layer because in
addition to electrostatic interactions with the positively charged alumina surface,
the carboxylate groups could also coordinate strongly with the aluminum oxide.
Subsequent addition of a polycation layer provided the highly positively charged
surface needed for adsorption of citrate-stabilized nanoparticles. Furthermore,
the amine groups of PAH and PEI most likely coordinated with AuNPs to

enhance the adsorption of the particles onto the membrane surface. Figure 2.10

shows FESEM images of an alumina membrane before and after coating with a

65

PAA/PAH/AuNP ﬁlm. The images clearly show the presence of the adsorbed Au

nanoparticles.

 

Figure 2.10 FESEM images of cross—sections of anodized alumina membranes
a) before and b) after modiﬁcation with a PAA/PAH/AuNP ﬁlm.

After deposition of 1 PANPAH bilayer, adsorption of AuNPs yielded a red
membrane, and the ﬁrst 90 mL of a red 0.045 mM AuNP solution (concentration
is given with respect to gold atoms) exited a PAA/PAH-modiﬁed membrane as a
colorless liquid. UV-Vis spectrophotometry of the AuNP solution exiting the
membrane demonstrated that >99% of the nanoparticles adsorbed to the
membrane during modiﬁcation. Once most of the adsorption sites became ﬁlled
with AuNPs, the particles passed through the membrane, and the concentration
of AuNPs in the permeate solution increased, as shown by UV-Vis
spectrophotometry (Figure 2.11). Typically during membrane modiﬁcation, the
nanoparticle deposition step was considered complete once the permeate

solution exhibited a red color that was easily visible.

66

0.1

0.08 -

0.06 .

0.04 1

Abs. at 518 nm

0.02 -

 

 

0 20 40 60 80 100 120
Volume Eluted (mL)

Figure 2.11 Plot of absorbance at 518 nm vs. volume eluted during deposition of
Au nanoparticles in a PAA/PAH-modiﬁed alumina membrane. The Au
concentration in the nanoparticle feed solution was 0.045 mM, which
corresponds to an absorbance of 0.163 at 518 nm.

Dissolution of the adsorbed nanoparticles in aqua regia and analysis of
this solution using FAAS revealed that the gold loading in porous alumina
modiﬁed with one layer of PAA/PAH/AuNP was (4.9 :l: 0.2) x 10'6 mol of Au per
membrane. (The exposed surface area of the top of the 60 um-thick membranes
was 2.8 cmz.) Assuming a porosity of 30% and a pore diameter of 0.23 pm, this
gold loading corresponds to 5.5 x 10'9 mol Au per cm2 .of internal (pore) surface
area. For a nanoparticle diameter of ~12 nm, this value is equivalent to 620
nanoparticles/ umz, which is reasonably consistent with the density of ~500
nanoparticles/ um2 seen in SEM images such as Figure 2.10b.

Multiple PAH/AuNP bilayers were deposited by continued alternating
adsorption of PAH and AuNPs. For membranes modiﬁed with 2 and 3 total

PAH/AuNP bilayers, the gold loading was 8.3 :l: 0.2 umol and 11.4 :I: 0.6 pmol of

67

Au per membrane, respectively. Enhanced Au loading with adsorption of more
layers conﬁrms the layer-by-Iayer deposition. The FESEM image in Figure 2.12a
corroborates the higher loading of AuNPs in the membrane with an additional
PAH/AuNP bilayer. Porous alumina membranes were also modiﬁed through
direct adsorption of PEI/AuNP bilayers (Figure 2.12b). In this case, the pH of the
PEI deposition solution was 9, which allowed the polycation to adsorb directly to
the alumina surface without the need for adsorption of a precursor PAA layer.
Gold coverages were 9.2 :l: 0.9, 17 i 2, and 24 :l: 3 umol per membrane for
alumina modiﬁed with 1, 2, and 3 PEI/AuNP bilayers, respectively. Deposition of
PEI at pH 9 on a PAA precursor layer resulted in a surface that allowed similarly

high AuNP adsorption.

i‘
.’
'1
S
5,
i

. hep-y)

    

'O-co‘W-y

- i
? .
Figure 2.12 FESEM images of cross sections of anodized alumina membranes
after modiﬁcation with a) PAA/[PAH/AuNP]2 and b) PEI/AuNP ﬁlms.

One concern with LbL formation of polyelectrolyte/nanoparticle multilayer
films in porous membranes is the possibility of plugging the pores. To conﬁrm
that the pores were still open after addition of each layer, we measured the ﬂux

of pure water through the membrane at a constant pressure. Furthermore, the

68

thickness of each layer was estimated by using the Hagen-Poiseuille equation for

laminar ﬂow through a tube with a circular cross-section, equation 2.1:

_77r‘A_P (21)
8;] / '

 

where Q is the volumetric ﬂow rate, r is the membrane pore radius, ,u is the
viscosity of water, AP is the pressure drop across the membrane, and l is the
membrane thickness. Hence, by assuming that each pore is a cylindrical tube
extending across the length of the membrane with an initial pore diameter of 230
nm, the decrease in ﬂux and thus pore radius could be related to an increase in
ﬁlm thickness.

Interestingly, during deposition of each layer, the ﬂux declined signiﬁcantly
after PAH deposition, and recovered slightly after AuNP deposition (Figure 2.13).

This change in ﬂux is likely due to the fact that the polyelectrolyte multilayer ﬁlms

 
 
 

 

0.40 50
E
’3‘ 0.30 4° f;
‘I’ in
"E 30 g
x
3 020 5 .9
.C
.5. 20 I—
x E
5 01° to s
0.00 0

 

 

Number of PAH/Au-NP Bilayers

Figure 2.13 Plot of pure water ﬂux and estimated ﬁlm thickness vs. number of
PAH/AuNP bilayers during modiﬁcation of an alumina membrane. The pure
water ﬂux was measured at a constant pressure of 1.4 bar. The values for ﬂux
and ﬁlm thickness after deposition of PAA are displayed as 0 PAH/AuNP

bilayers.

69

undergo more swelling when the terminating layer is PAH. When terminating
with a AuNP layer, the ﬂux most likely increased because the PAH was bound to
the nanoparticles and underwent less swelling into the solution. The overall
estimated thickness of a PAA/[PAH/AuNP]3 ﬁlm was approximately 20 nm, which
is consistent with the results seen for ellipsometric thickness measurements on
Al-coated Si wafers.

To demonstrate the versatility of the LbL technique for membrane
modiﬁcation, we also deposited polyelectrolyte/AuNP ﬁlms in polycarbonate track
etched membranes containing nominal pore Sizes of 0.2 pm and nylon
membranes containing nominal pore sizes of 0.45 pm. Figure 2.14 shows
FESEM images of the tops of unmodiﬁed polycarbonate and nylon membranes.
Like the 60 pm thick porous alumina, track-etched polycarbonate membranes
(Figure 2.14a) contain cylindrical pores, but the polycarbonate membrane
thicknesses (~10 pm) and pore densities (10-15%) are relatively low. AS a result,
less gold was immobilized within the polycarbonate membranes. In contrast, the
nylon membranes (Figure 2.14b) are thicker (150 um) than alumina membranes
and have a high pore density (~70%) in their sponge-like structure. In spite of
the different structures, however, FESEM images Show that the LbL technique
deposited a dense layer of well-separated metal nanoparticles in both

polycarbonate and nylon membranes (Figure 2.15).

70

      
  

 

Figure 2.14 Top-down FESEM images of unmodiﬁed a) track-etched
polycarbonate and b) nylon membranes.

Figure 2.15 Cross-sectional FESEM images of polymer membranes modiﬁed
with PSS/PAH/AuNP ﬁlms. a) polycarbonate and b) nylon.

   

The amount of gold deposited in a polyanion/PAH/AuNP layer is 0.81, 4.9,
and 9.1 umoles for polycarbonate, alumina, and nylon membranes, respectively,
and these values reﬂect the different surface areas of the membranes.
Calculated porous surface areas of polycarbonate and alumina membranes are
62 and 890 cm2/ membrane, assuming pore radii of 100 and 115 nm and
porosities of 0.1 and 0.3 for polycarbonate and alumina, respectively. In the case

of the nylon membrane, N2 adsorption measurements showed a surface area of

71

1900 cmzl membrane. From these surface areas, the gold loading was 2.6, 1.1,
and 0.95 119! cm2 for polycarbonate, alumina, and nylon membranes,
respectively. The polycarbonate had a somewhat higher loading than the other
membranes, which might be due to differences in the interactions between the
polyelectrolyte ﬁlm and the substrate, or an underestimation of the polycarbonate

membrane surface area.

2.3.2 Catalytic Reduction of 4-Nitrophenol with Au Nanoparticles in
Solution

Reduction of 4-nitrophenol to 4-aminophenol in the presence of NaBH4
(Scheme 2.2) provides a rapid, easily characterized reaction for examining the
catalytic activity of nanoparticles immobilized in membranes. The reduction does
not occur at a measurable rate in the absence of a catalyst .and is readily
monitored with UVNis spectrophotometry. In basic solution, the 4-nitrophenolate
ion shows a strong absorbance maximum at 400 nm, and the rate of reduction is
easily monitored by the disappearance of this peak (Figure 2.16). The gradual

appearance of a peak at 298 nm shows the corresponding production of 4-

aminophenol.
N02 NH2
0 +Na+BH.,-—Al‘—. O +Na+BO2'+ H2
OH OH

Scheme 2.2 Au-catalyzed reduction of 4-nitrophenol by NaBH4.

72

 

 

  

 

1.6 ~ Decreasing
4-nitrophenol
1.4 -
1.2 1
e
2
g 1.0]
8 0.8
2 06- Increasing
4-aminophenol
0.4 . T
250 300 350 400 450 500
Wavelength (nm)

Figure 2.16 UVNis spectra demonstrating the catalytic reduction of 3 mL of 0.1
mM 4-nitrophenol in 10 mM NaBH4 to form 4-aminophenol over time. Spectra
were collected in 4 minute time intervals after addition of 20 uL of as-prepared
AuNPs.

As a control experiment, the as-prepared citrate-stabilized Au
nanoparticles (without immobilization in a ﬁlm or on a support) were used for the
catalytic reduction of 4-nitrophenol. In these experiments, the NaBH4 was in
>100 fold excess of 4-nitrophenol to ensure that the reaction followed pseudo-

ﬁrst order kinetics with respect to the 4-nitrophenol concentration according to

equation 2.2:
d0
7’— = —/(appC (2.2)

where C is the molar concentration of 4-nitrophenol at time t, and kapp is the

apparent ﬁrst order rate constant for the reaction.

73

To facilitate comparison of different nanoparticle systems, the batch
reaction rate constants were normalized to the total surface area of the

nanoparticles in the system divided by the solution volume according to equation

2.3:

dC

_ = —knorm AC (23)
0'! I/

so/n
where A is the total surface area of nanoparticles in the system in cm2 and V30," is
the volume of the solution in cm3 used in the reaction. Previous batch reactions
that employed Pt, Au, or Ag nanoparticles dispersed in solution produced
normalized rate constants, know, ranging from 0.0005 to 0.055 cm/sec, and

3638.41.43.67 These values were

showed some variation with particle size.
calculated from the apparent rate constants reported in the literature. The total
surface area was calculated from the given diameters assuming that the
nanoparticles were spherical, and volumes were determined from the
experimental parameters.

For our reactions with AuNPs in solution, the solution was continuously
stirred, and the rate constant was determined by collecting samples every 4
minutes and measuring the absorbance at 400 nm. The plot in Figure 2.17
conforms well to a linear ﬁt suggesting that the reaction was indeed ﬁrst order
with respect to the concentration of 4-nitrophenol. The citrate-stabilized AuNPs
used in this work exhibited a normalized rate constant, know, of 0.014 :I: 0.002

cm/sec. Thus, the activities of the citrate-stabilized AuNPs were in the same

range as other nanoparticle systems previously reported in the literature.

74

    
 

y = -0.0258x

 

 

-0.9 - - -
0 5 10 15 20 25 30 35
Time (min)

- i 1 j

Figure 2.17 Plot of ln(A/Ao) vs. time for the reduction of 4-nitrophenol with
NaBH4 catalyzed by Au-nanoparticles in solution. A is the absorbance of the
solution at 400 nm during the reaction and A0 is the initial absorbance of the
solution. [4-nitrophenol] = 0.085 mM, [NaBH4] = 12 mM, and [Au] = 5.9 x 10"
mM.

2.3.3 Catalytic Reduction of 4-Nitrophenol with Membrane-Supported Au
Nanoparticles

To the best of our knowledge, this was the ﬁrst investigation of the
reduction of 4-nitrophenol via ﬂow-through catalytic membranes. and Au
nanoparticle-containing membranes showed remarkable conversions of 4-
nitrophenol to 4-aminophenol at high ﬂow rates. Passage of a solution
containing 0.1 mM 4-nitrophenol and 10 mM NaBH4 through a PAA/PAH/AuNP
modiﬁed alumina membrane at a rate of 0.29 mUcmZ-sec resulted in reduction of
greater than 99% of the 4-nitrophenol (Figure 2.18). Similarly high conversions
occurred with 4-nitrophenol concentrations as high as 5 mM, provided that

NaBH4 was in at least a 50-fold excess. At such high conversions, the reaction

75

was essentially mass-ﬂow limited even at these high ﬂuxes. The solution ﬂux of
0.29 mL/cmz-sec corresponds to a residence time within the membrane of only

6.1 msec and a linear velocity of 0.98 cm/sec.

   
 
    

2.5 .,
—4-Nitrophenol
' 2 . ——4-Aminophenol
----- Red uced 4-Nitrophenol
1.5 -

1.

Absorbance

0.5 ‘

 

 

0 1 I .
250 300 350 400 450 500

Wavelength (nm)

T

Figure 2.18 UV-VIS absorbance spectra of a 0.1 mM solution of 4-nitrophenol
containing 10 mM NaBH4 before (solid black line) and after (solid gray line)
passing through a membrane modiﬁed with a PAA/PAH/AuNP ﬁlm at a ﬂux of
0.29 mL/cmZ-sec. The spectrum of a 0.1 mM 4-aminophenol solution containing
10 mM NaBH4 is also Shown as a reference (dotted black line that overlaps the
gray line).

 

 

100 7 I Experimental
—Fit
C
.2
:2 98 -
2
C
o
0
g 96 ‘
94 l l l I l
0.1 0.2 0.3 0.4 0.5 0.6

Flux (lecmZ-sec)
Figure 2.19 Plot of 4-nitrophenol reduction percentage vs. ﬂow rate of a 0.4 mM

4-nitrophenol solution through a porous alumina membrane containing a
PAA/PAH/AuNP ﬁlm. [NaBH4] = 20 mM.

76

At higher ﬂuxes, conversion declined Slightly as would be expected. Still,
as Figure 2.19 shows, conversion was greater than 95% at ﬂuxes as high as 0.44
mLIcmz-sec (residence time of 4.1 msec). The maximum ﬂux in this system was
practically limited to about 0.53 mL/cmZ-sec because higher applied pressures
resulted in membrane fracture.

The data in ﬁgure 2.19 conform well to a simple ﬁrst-order kinetic model of
the reaction, in which we assume that the membrane acts as a plug ﬂow reactor.
In a plug of solution ﬂowing through a porous membrane, reduction of 4-

nitrophenol occurs with pseudo ﬁrst-order kinetics according to equation 2.4,

d0 ,U A
_d;—g- = _kmem T— Cp/ug (2'4)

pores
where Cpmg is the molar concentration of 4-nitrophenol in the plug of solution at a
given time, t, kmem is the normalized, pseudo ﬁrst order heterogeneous rate
constant, A is the surface area of the nanoparticles in the membrane, and Moon,s
is the volume of the pores in the membrane. The use of nanoparticle surface
area and membrane pore volume in the equation allowed the direct comparison
between kmem and km", for solution reactions.

Integration of the rate law in equation 2.4 yields the fractional conversion
of the nitro group as a function of residence time in the membrane, tres, as shown

in equation 2.5.

—k t A] (25)

conversion =1— exp [ﬁﬂi—

pores

Noting that residence time is equal to Ic/F, where I is membrane thickness, 5 is

membrane porosity, and F is the solution ﬂux through the membrane; and that

77

VW...s is Vmems where Vmem is the volume of the membrane including porous and

nonporous regions, equation 2.5 can be rewritten as Shown in equation 2.6,

conversion = 1— ex M (2.6)
Vmem F

in quantities that are relatively easy to determine. We calculated the nanoparticle
surface area, A, from the gold loading in the membrane, assuming that the
nanoparticles had a diameter of 12 nm with all of the surface area accessible.
The ﬁt of equation 2.6 to the data in Figure 2.19 was excellent and yielded a rate
constant (kmem) for this system of 0.018 :I: 0.002 cm/sec, which was essentially
the same within experimental uncertainty as the value for nanoparticles in
solution (0.014 :I: 0.002 cm/sec). This shows that immobilization of nanoparticles
did not restrict access to catalytic sites or alter catalytic activities.

Most reductions were performed with a 50 to 100-fold excess of NaBH4
with respect to 4-nitrophenol to allow the concentration of BH4' to remain
essentially constant throughout the reaction. When the NaBH4 concentration
was lower than a 50-fold excess, the reaction conversion became dependent on
both the 4-nitrophenol and NaBH4 concentrations. In Figure 2.20, all parameters
in the reduction were held constant except for the NaBH4 concentration, which
was in 63, 50, 35, and 25-fold excess with respect to the 4-nitrophenol
concentration of 0.8 mM. The ﬁgure Shows that only about 90% of 4-nitrophenol
was reduced with a 25-fold excess of NaBH4, while at a 50-fold excess of NaBH4,
> 99% of the 4-nitrophenol was reduced. Excesses of NaBH4 > 50-fold did not

signiﬁcantly enhance the fraction of 4-nitrophenol reduced.

78

 

 

 

 

 

 

100 1 g n a e e e

A A A A e 50 mM
U A ‘ ‘
g 95 I 40 mM
.5
6'2 ' A 28 mM
, 90 - P O Q
,\ O

o O 20 mM
85 f f I l I
0 50 100 150 200 250
Volume Eluted (mL)

Figure 2.20 Effect of NaBH4 concentration on the percent reduction of 4-
nitrophenol achieved by ﬂowing NaBH4/4-nitrophenol solutions through an
alumina membrane modiﬁed with a PAA/PAHIAuNP ﬁlm. Conditions: [4-
nitrophenol] = 0.8 mM, ﬂux = 0.28 mLIcmz-sec.

2.3.3.1 4-Nitrophenol Reduction as a Function of Film Composition

Adding a PAH capping layer to adsorbed ﬁlms had little impact on the
reactivity of underlying nanoparticles. For reductions using a 4-nitrophenol
concentration of 0.4 mM and NaBH4 concentration of 20 mM, membranes
terminated with PAH gave essentially the same percent reduction of 4-
nitrophenol as membranes terminated with a AuNP layer (Figure 2.21). Since
the percent reduction was >99% in both cases, membranes with and without a
PAH capping layer were also used for reductions with only a 25-fold excess of
NaBH4. In this experiment, membranes coated with a PAH capping layer still
Showed similar activity to membranes without a capping layer, which supports
the hypothesis that the polyelectrolyte ﬁlm does not hinder access of the

reactants to the active sites of the AuNPs.

79

 

 

10° ' l t t e t t e
'u 95 -
g 6
3 0 ‘
P 90 - t .
9 6
‘E * -
..\ Oa) Capped with PAH
85 - Ia) Capped with Au-np
Ab) Capped with PAH
Cb) Capped with Au-np
80 l i I I l
0 50 100 150 200 250

Volume Eluted

Figure 2.21 Comparison of 4-nitrophenol reduction in membranes modiﬁed by
adsorption of PAA/PAH/AuNP and PAA/PAHIAuNP/PAH ﬁlms when [4-
nitrophenol] = a) 0.4 mM, and b) 0.8 mM. Other conditions: [NaBH4] = 20 mM,
ﬂux = 0.28 mLIcmz-sec.

Increasing the number of AuNPs adsorbed in the membrane should
improve conversion in the membrane reactor. However, since conversion of 4-
nitrophenol to 4-aminophenol was nearly 100% even with a Single layer of
adsorbed nanoparticles, we initially decreased reaction rates to study the effect
of additional PAH/AuNP bilayers on the reaction. To do this, we reduced the
initial NaBH4 excess from 50-fold to 25-fold relative to 4-nitrophenol. Table 2.1
shows that under these conditions at a ﬂux of 0.30 mL/cmZ-sec, conversion
increased from 89 to 98% on going from 1 to 3 adsorbed PAH/AuNP bilayers in
the membrane. This supports the hypothesis that a higher Au loading leads to

higher activity.

80

Table 2.1 Au loading, percent 4-nitrophenol reduction, and rate constant vs. the
number of PAH/AuNP bilayers contained in PAA/[PAH/AuNP]X ﬁlms deposited in
porous alumina membranes employed for ﬂow-through reactions.

 

Number of Au Loading % of 4-Nitrophenol Rate Constant (kmm)

 

Bilayers (pmoles) Reduceda (cm/sec)”
1 4.9 :I: 0.2 88.8 :I: 0.9 0.018 t 0.002
2 8.3 :I: 0.2 94.1 :I: 0.3 0.018 :I: 0.002
3 11.4 :I: 0.6 98.3 :I: 0.5 -c

 

a[4-nitrophenol] = 0.8 mM, [NaBH4] = 20 mM, ﬂux = 0.30 mL/cmz-sec.
b[4-nitrophenol] = 0.4 mM, [NaBH4] = 20 mM, ﬂux varied between 0.20 mUcmz-
sec and 0.55 mL/cmz-sec.

cThe high conversion did not allow determination of a rate constant.

 

 

 

 

100 -
98 -
'0
O
O
3
g 95 -
0:
.\°
94 -
92 . . a
0.1 0.3 0.5

Flux (mL/cmz-sec)

Figure 2.22 Plot of percent conversion vs. ﬂux for membranes containing 1 and
2 PAH/AuNP bilayers. The curves represent a ﬁrst-order reaction model with
rate constants (kmem) of 0.0018 cm/sec for each membrane. (Rate constants are
normalized to the nanoparticle surface area in each membrane.) Conditions: [4-
nitrophenol] = 0.4 mM, [NaBH4] = 20 mM.

Unfortunately, the use of a 25-fold excess of NaBH4 does not always

result in ﬁrst-order kinetics. However, even with a 50-fold excess of NaBH4,

81

measurement of percent conversion vs. ﬂow rate showed that the overall activity
of the membrane improves on going from 1 to 2 PAH/AuNP bilayers (Figure
2.22). Furthermore, the calculated rate constant, kmem, which was normalized to
the total surface area of the nanoparticles in the membrane, was unchanged by
depositing an additional PAH/AuNP layer (Table 2.1). This suggests that kmm
was not affected by the number of layers deposited and that interior
nanoparticles were readily accessible to both 4-nitrophenol and BH4'. Because
of very high conversiOns, we could not determine the value of kmem for
membranes containing 3 PAH/AuNP bilayers.

Membranes modiﬁed with PEI/AuNP ﬁlms showed catalytic activities
similar to those of membranes containing PAA/PAH/AuNP coatings. The
reduction of 4-nitrophenol was greater than 99% even at ﬂuxes of 0.30 mLIcmz-
sec, and the % conversion increased as a function of layer number for PEI/Au NP
ﬁlms when using a 25-fold excess of NaBH4. However, the higher gold loading
and consequently higher conversions with PEI/AuNP coatings made it difﬁcult to

determine rate constants for membranes modiﬁed with these ﬁlms.

2.3.3.2 Comparison of the Catalytic Activities of Nanoparticles in Alumina,
Nylon, and Polycarbonate Membranes

To further demonstrate the generality of LbL polyelectrolyte/metal
nanoparticle adsorption for forming catalytic membranes, we compared the
catalytic activity of nanoparticles in alumina, nylon, and polycarbonate

membranes. As Figure 2.23a shows, at sufﬁciently low ﬂow rates all three types

82

of membranes catalyzed reduction of more than 99% of the 4-nitrophenol in a
solution containing 1 mM 4-nitrophenol and 100 mM NaBH4. Comparison of the
membranes at identical ﬂuxes was not possible because the low porosity and
relatively low burst pressure (0.7 atm) of the polycarbonate membranes greatly
limited ﬂux through these membranes. In contrast, because the nylon
membranes are much more porous than polycarbonate membranes, higher ﬂow
rates were obtained in the reduction reactions. In fact, P88/PAHIAuNP modiﬁed
nylon membranes reduced more than 99% of 4-nitrophenol in a 1 mM solution at
ﬂuxes as high as 0.70 mLIcmz-sec. However, nylon membranes also have a
lower burst pressure (2.3 atm) than alumina membranes (4.3 atm), which still
limits the amount of pressure that can be applied.

As Figure 2.23a shows, the conversion of 4-nitrophenol to 4-aminophenol
decreased with increasing solution ﬂux for all three types of porous membrane
substrates coated with a AuNP-containing ﬁlm. If it were reasonable to compare

conversions at a given ﬂux, they would decrease in the order nylon > alumina >

100 % 13%: AAAAA a) 100-OA W b)
93.0 .

 

 

 

 

C

.2 c

2 9 98 -

9 I: g I:

g 96 - D E 96 - D

e 0

°\ 94 . . . . . .\° 94 . . .
0 0.2 0.4 0.6 0.8 1 600 400 200 0

Flux (m Llcmzsec) truANpores (sec*cm2Icm3)

Figure 2.23 Plot of conversion vs. a) ﬂux and b) residence time multiplied by the
AuNP surface area per membrane pore volume (th/Vpores) in the reduction of 4-
nitrophenol to 4-aminophenol using membranes modiﬁed with a
polyanion/PAH/AuNP ﬁlm. In both plots, O = polycarbonate, El = alumina, A=
nylon. Conditions: [4-nitrophenol] = 1 mM, [NaBH4] = 100 mM.

83

polycarbonate. Because the membranes had different thicknesses and gold
loadings, however, an appropriate comparison of the catalytic activities of
different membranes needed to account for the membrane geometry and
reaction kinetics.

Replotting the conversions in Figure 2.23a as a function of the product of
residence time and nanoparticle surface area per membrane pore volume,
th/Vpoms, which is equivalent to AI/(VmemF), shows that the three membranes
behaved similarly (Figure 2.23b). Put another way, ﬁtting of the data in Figure
2.23a to equation 2.5 gives values of kmem that differ by less than 20% among the
three membranes. The km”, values are 0.017, 0.015, and 0.018 cm/sec for
alumina, track-etched polycarbonate, and nylon membranes, respectively. Thus,
the catalytic activity of the nanoparticles in the different membranes is quite
similar. Nylon membranes give the highest conversions because they are the
thickest of the three substrates, and conversions with alumina are higher than
with polycarbonate membranes because the polycarbonate is thinner and also

has a lower porosity than alumina.

2.3.4 Effect of Substituent Groups on the Reduction of Nitroaromatic
Compounds

2.3.4.1 Reduction of Compounds with Single Reaction Products

R R
Au \
N02 + NaBH4 ———> I —NH2
/

R = H, OH, NHZ, CH3, CN, Cl, CH=CH2
Scheme 2.3 Au-catalyzed reduction of nitroaromatic compounds by NaBH4.

84

Table 2.2 Rate constants for reduction of nitrophenol and nitroaniline. The
reactions were catalyzed by PAA/PAH/AuNP ﬁlms in porous alumina membranes
(kmem) and on alumina powder (ksruny).

 

Compound km (cm/sec) km”, (cm/sec)

 

4-Nitrophenol 0.018 :I: 0.003 0.020 :I: 0.004
2-Nitrophenol 0.015 t 0.003 0.016 :I: 0.003
4-Nitroaniline 0.023 :t 0.004 0.022 :I: 0.005
2-Nitr0aniline 0.018 :I: 0.004 0.015 t 0.003

 

One of the goals of this work was to determine the effects of substituent
composition and position on the reduction of nitrobenzene derivatives. To do
this, we reduced the compounds in Scheme 2.3 using alumina membranes
modiﬁed with 1 layer PAA/PAHIAuNP ﬁlms. The reduction of 2- or 4-nitrophenol
gave the corresponding amine as the only observed product, so kmm was
determined by measuring nitro group conversion as a function of ﬂux and ﬁtting
the data to equation 2.5. The same was true for the reduction of 2- and 4-
nitroaniline. Table 2.2 shows that the values of kmem were nearly the same for
these 4 compounds. Thus, the hydroxyl and amino groups inﬂuenced the nitro
reduction to nearly the same degree, and the effect of substituent position (ortho
versus para) was also small.

For comparison, we performed conventional Slurry reactions with catalytic
nanoparticles on micron-sized alumina. For these reactions, the pseudo ﬁrst
order rate constant, ksmy, was calculated by modeling conversion versus time, t,

with equation 2.7.

— It At
conversion = 1— exp[—i’-"-’-7—] (2.7)

salmon

85

As Table 2.2 shows, within experimental error, the rate constants for the slurry
and membrane reactions were the same. Thus the nanoparticle catalytic

properties are independent of the substrate on which they are deposited.

2.3.4.2 Reduction of Compounds with Multiple Reaction Products

Table 2.3 presents the product distributions for the membrane-catalyzed
reduction of several substituted nitroaromatic compounds. In contrast to hydroxyl
or amino functionalized nitroaromatics, reduction of 2- and 4-nitrotoluene yielded

a mixture of the amino and nitroso forms of each compound.

Table 2.3 Product distributions in membrane-catalyzed reductions of
nitroaromatic compounds at two different solution ﬂuxes. The membranes
consisted of porous alumina modiﬁed with PAA/PAH/AuNP ﬁlms.

 

 

 

Compound Flux = 0.015 mLIcmz-sec Flux = 0.15 mLIcmz-sec

% N02 % NOa %NH2 % N02 % NOa %NH2
4-Nitrophenol <1 - >99 1 - 99
2-Nitrophenol <1 - >99 1 - 99
4-Nitroaniline <1 - >99 1 - 99
2-Nitroaniline <1 - >99 1 - 99
4-Nitrotoluene <1 23 76 3 44 53
2-Nitrotoluene <1 19 80 2 40 58
Nitrobenzene 3 24 73 4 47 49
2-Nitro-m-xylene 2 5 93 5 23 72

 

aUncertainties in these values are about 2%.

Slun'y reactions provided a better window into reduction kinetics because
the reaction occurred on a longer time scale due to the large solution volume.
Figure 2.24 shows the temporal composition of slurry mixtures during the

reduction of several of the substituted nitroaromatic compounds. For 4-

86

nitrophenol (ﬁgure 2.24a) and 4-nitroaniline (ﬁgure 2.24b), the reduction of the
nitro species was accompanied by the corresponding formation of the amino
species. Similar plots were obtained in the reduction of 2-nitrophenol and 2-
nitroaniline. For 4-nitrotoluene (Figure 2.240), the initial rapid formation of both
the amino and nitroso species followed by slow reduction of the nitroso
compound to the amine suggests that there were two pathways to the amine in
this reaction. Similar trends occurred for 2-nitrotoluene (Figure 2.24d) and
nitrobenzene (Figure 2.24e). The most likely explanation for the absence of the
nitroso product after reduction of hydroxyl or amino functionalized nitroaromatics
is that the electron-donating hydroxyl and amino substituents inhibited reduction
to the nitroso functionality. The overall reduction of the nitro group was actually
slower for 4-nitrophenol than 4-nitrotoluene, perhaps because of inhibition of
nitrosophenol formation in reduction of 4—nitrophenol. It is possible that the
hydroxyl and amino groups increased the rate of reduction of the nitroso group
(so nitroso compounds are not observed) while decreasing the overall rate of
nitro reduction, but this seems improbable.

In both slurry (Figure 2.24f) and membrane (Table 2.3) reductions, 2-nitro-
m-xylene showed less formation of the nitroso compound than 2- and 4-
nitrotoluene, presumably because of additional electron donation by the second
methyl group. However, the overall conversion of the nitro compound was lower
for 2-nitro-m-xylene than for 2- and 4-nitrotoluene, which may be partly due to

steric hindrance caused by the two methyl groups.

87

The formation of nitroso intermediates is consistent with a hydrogenation
pathway where nitroso and hydroxylamine species form in a direct pathway to

the amine.17 Other studies have proposed several pathways in which the nitroso

  
  
 
  

 
  
  
 

 

 

   

 

 

 

 

 

   

 

1.0 1.0 b)
E 0.8 ' g 0.8 "
g 0.6 - g 0.6 -
o 0.4 - ,5 0.4 -
5 5 02 -
o 0.2 ' o .
0.0 . . , 0.0 r .
0 5 10 15 20 0 5 10 15 20
Time (min) Time (min)
1.0 c)
’5‘ 0.8
g 0.6
g 0.4
o 02
0.0
0 5 10 15 20 0 5 10 15 20
Time (min) Time (min)

 

 

 

 

 

 

0 10 20 30 0 5 10 15 20 25
Time (min) Time (min)

 

l +Nitro +Amino -I- Nitroso +Azoxy

 

Figure 2.24 Slurry reaction proﬁles for reduction of a) 4-nitrophenol, b) 4-
nitroaniline, c) 4-nitrotoluene, d) 2-nitrotoluene, e) nitrobenzene, and f) 2-nitro-m-
xylene with NaBH4 using alumina powder modiﬁed with a PAA/PAH/AuNP ﬁlm as
a catalyst. For each reaction, circles represent the starting nitro compound,
triangles represent the amine product, squares represent the nitroso product, and
diamonds represent the azoxy product.

88

species is not an intermediate in the reaction, but rather a product that may result
from side reactions.5’2'53 Litvic’: et al. also observed the formation of nitroso,
hydroxylamino, and diazo intermediates in the reduction of several nitroaromatic
compounds with Raney nickel and sodium borohydride.46 In the membrane
systems, the increased amounts of the nitroso compound at shorter reaction
times (higher ﬂow rates) suggest that it is an intermediate to the formation of the
amine rather than a side product; however, the data for slurry reactions provide
evidence for an additional, more rapid pathway to the amine that does not go

through a nitroso intermediate.

2.3.4.3 Effect of Flux on Product Distributions

Table 2.3 indicates that solution ﬂux, which is inversely related to
residence time within the membrane, can be used to control conversion in
membrane reactions. As the solution ﬂux through nanoparticle-modiﬁed alumina
membranes increased, the ratio of the amine to nitroso product decreased. To
further investigate this trend, we determined the product distribution for reduction
of 2-nitrotoluene as a function of the ﬂux through nylon, polycarbonate, and
alumina membranes. Figure 2.25a Shows the percentage of each compound in
the product stream for all three membrane types. Similar to Figure 2.23, when
product composition was plotted as a function of the residence time multiplied by
the AuNP surface area per pore volume (tresA/Vpores), the product compositions

from the different membranes were comparable (Figure 2.25b).

89

 

0'8 - QN
,an .

g \ . s‘~A- N
g 0.6 5 I 3\, . & ....... a. .......... ﬂ

' ' " “warn—a...
g 1" \_ ......... ﬂg" g
0 0I4 I r iﬁ/ '75 ....... ﬂ-
0 «‘-

 

 

 

 

 

 

 

 

 

0 0.1 0.2 0.3
. Flux (mLIcmz-sec)
1 I N b)
0.8 - 0.8 - ARA
A 0.5 ”ﬂavﬁ—we, GA 7A-Ng‘)
E 0.6 " 0.4 . . - L L13 , «.1
5 W -.. H
,5 0.2 I a;
g 0.4 ‘ 0 ‘-e‘; F: F. "’g3r 1‘:
o 2000 1500 1000 500 0 ,0" L
0.2 - In»?
f‘
0 -—ﬁ . Wigs.“
25000 20000 15000 10000 5000 0

t...*AIVpo,., (sec*cm2Icm3)

Figure 2.25 Plot of efﬂuent composition as a function of a) ﬂux and b) residence
time multiplied by nanoparticle surface area per pore volume (thl/Vpoms) in the
reduction of 2-nitrotoluene using AuNP-containing membranes. In both
diagrams: A = amine, III = nitroso, O = nitro, solid black line = nylon membrane,
dotted black line = alumina membrane, and solid gray line = polycarbonate
membrane. The inset in Figure b shows an expanded plot of the data at the right
of this ﬁgure.

90

Most importantly, Figure 2.25 demonstrates the control over product
composition provided by variation of ﬂux, or alternatively residence time.
Remarkably, Figure 2.253 Shows that the yield of 2—aminotoluene was 94% at
low ﬂuxes through nylon membranes (solid black line) and only 34% at relatively
high ﬂuxes through polycarbonate membranes (solid gray line). Moreover, the
yield of the nitroso compound was higher at low values of tmsA/Vpoms, as one
would expect, and the conversion to the nitroso compound ranged from 5% at
high values of treeA/Vpores to 61% at the lowest values of tmsA/Vpoms. Increased
ﬂuxes (or lower residence times) might produce even larger fractions of the
nitroso compound; however, the amount of unreacted starting material might also
increase at these higher ﬂuxes. Unfortunately, ﬂuxes are currently limited by the
strength of the membranes. When rate constants for the formation of an
intermediate and for the subsequent conversion to product are even more
disparate, it should be possible to obtain the intermediate in even higher yield.
Overall, ﬁne control over residence time makes membranes attractive for

selective product formation.

2.3.4.4 Selectivity in Reactions with Multiple Reducible Functional Groups
We investigated catalytic reduction of cyano, chloro, and styrenyl
functionalized nitroaromatic compounds to determine whether AuNPs can
selectively catalyze nitro group reduction in the presence of other reducible
functional groups. As Table 2.4 shows, the reaction with NaBH4 resulted in

reduction of 2% or less of cyano, styrenyl, and chloro groups at conversions of

91

Table 2.4 Product distributions in the membrane-catalyzed reduction of
nitroaromatic compounds containing other reducible functional groups. The
reactions employed a gold nanoparticle-modiﬁed alumina membrane and a ﬂux
of 0. 015 mL/cm2 -sec.

R NaBHR
I :—N02 '_> :—NH2 :_NO :—NH2
a/

 

 

Compound (R) % a % b % c % d
4-Nitrobenzonitrile (CN) <1 >97 - 2
2-Nitrobenzonitrile (CN) <1 >98 - 1

1,4-Chloronitrobenzene (Cl) 2 92 6 -
1,2-Chloronitrobenzene (Cl) 2 93 5 -
3-Nitrostyrene (-CH=CH2) 1 85 13 <1

 

*R = reduced form of R-group.

98% or more of the nitro groups to amino or nitroso products. In the catalytic
reduction of 2- and 4-nitrobenzonitrile, no nitroso product formed, whereas
reductions of 1,2-chloronitrobenzene, 1,4-chloronitrobenzene, and 3-nitrostyrene
gave small amounts of the nitroso compound, but the amine was still the major

product at this ﬂux. Thus, selective reduction of the nitro group is feasible.

2.3.4.5 Catalytic Reduction of Nitrocyclohexane

68-69 without

The ability to reduce both aromatic and aliphatic nitro groups
changing the catalyst system is important in synthetic organic chemistry, but is
often quite difﬁcult to achieve. To determine whether membrane-catalyzed
reduction by NaBH4 could be extended to aliphatic nitro groups, we attempted
reduction of nitrocyclohexane. Passage of a solution of nitrocyclohexane and

sodium borohydride through a gold nanoparticle-containing alumina membrane

at 0.020 mUcmz-sec yielded greater than 95% reduction of the nitro group, but

92

the product was a mixture of cyclohexanone oxime and cyclohexanamine

‘ tr 6

b (73%) c (27%)
Scheme 2. 4 Reduction of nitrocyclohexane catalyzed by a Au nanoparticle-
containing alumina membrane with a solution ﬂux of 0. 020 mUcm2 -sec.

(Scheme 2.4).

2.3.4.6 Effect of Temperature on Reaction Rates and Product Distributions
Temperature provides another variable for controlling reaction rates and

product distributions. Panigrahi, et al. demonstrated that when catalyzed with 20

nm Au nanoparticles, the rate of reduction of 4-nitrophenol with sodium

67 We tested the effect of temperature

borohydride exhibits Arrhenius behavior.
on the reduction of 4-nitrophenol, 4-nitrotoluene, and 1,4-chloronitrobenzene in
slurry reactions catalyzed by alumina powder modiﬁed with AuNP-containing
ﬁlms. The results of these experiments were consistent with the data of
Panigrahi, et al. and are shown in Figure 2.26.

Though it appears that the reductions of 4-nitrotoluene and 1,4-
chloronitrobenzene were slightly faster than the reduction of 4-nitropheneol, it
should be noted that during reduction of 4-nitrotoluene and 1,4-
chloron'rtrobenzene, nitroso intermediates appeared. In the case of 1,4-
chloronitrobenzene, a small amount of the azoxy species also formed. The

complete conversion to the amine was actually faster for 4-nitrophenol. For each

compound, the reactions showed Arrhenius behavior (Figures 2.26b, d, and f).

93

The activation energies determined from these plots were 27, 62, and 59 kJ/mol
for 4-nitrophenol, 4-nitrotoluene, and 1,4-Chloronitrobenzene respectively, but in
the case of reduction of 4-nitrotoluene and 1,4-chloronitrobenzene, these

activation energies represent a composite value for formation of different reaction

   
   
   

products.
1 : 2:223:55. a) 0-0 b)
E 0_8 A MdegreesC '1 .0
E. 0.5 ’ 64mm“ -2.0 y = 6.89 - 3240x

 

 
   

 

 

 
 
     

 

E
0
d 3
g ”-4 g -4.0
o 0.2 e; -5.0
0 E -6.0 . . . .
0 5 1o 15 20 0.0025 0.003 0.0035 0.004
Time (min) 1IT WK)
1 c) A 0.0 d)
E 0'8 'E' -2.0 y=20.9-7510x
E. 0.6 E .40 o
<5 0.4 o '
5 o 2 : 338233ch 52‘ -6.0
U - O 63deqreesC E .80
0 I T I l I - ' I I I
0 1o 20 3o 40 0.0025 0.003 0.0035 0.004
time (min) 1IT(1IK)
1 e) a 0.0 f)
A o 4degreesC O _ _
E 3.: = 5333338 3 4.0 y—20.4 7080x
‘: - E
E 0-4 g -4.0
o 0.2 E
0 -6.0 . . . .
0 1o 20 30 40 0.0028 0.0032 0.0036 0.004
Time (min) 1IT(1IK)

Figure 2.26 Slurry reaction proﬁles and corresponding Arrhenius plots for the
reduction of a,b) 4-nitrophenol, c,d) 4-nitrotoluene, and e,f) 1,4-
chloronitrobenzene by NaBH4 at different temperatures. The reaction was
catalyzed by alumina powder modiﬁed with a PAA/PAH/AuNP ﬁlm. The lines in
a, c, and e are fits of the data to equation 2.7 in the text.

94

We also examined the membrane-catalyzed reductions of 4-nitrophenol,
4-nitrotoluene, and 1,4-chloronitrobenzene at three different temperatures with
the solution ﬂux kept constant at 0.07 mL/cmZ-sec. As Table 2.5 shows,
reduction of 4-nitrophenol proceeded to >99% conversion to the amine even at 4
°C, so the effect of temperature on the membrane-catalyzed reduction of this
compound could not be determined. For 4-nitrotoluene and 1,4-
chloronitrobenzene, temperature does not affect the fraction of nitro groups
reduced, but the ratio of amino to nitroso products increases with increasing
temperature. This again suggests that the nitroso compounds are converted to

the amine as the reaction proceeds.

Table 2.5 Product distributions in the membrane-catalyzed reduction of selected
nitroaromatic compounds at different temperatures. The catalytic membranes
consisted of alumina modiﬁed with a PAA/PAHIAuNP ﬁlm, and the ﬂux was 0.07
mL/cmz-sec.

 

 

 

 

Compound T(°C) % N02 % NO % NH:

4 <1 0 >99

4-Nitrophenol 24 <1 0 >99
44 <1 0 >99
4 1 39 60

4-Nitrotoluene 24 1 31 68
44 1 23 76
4 1 27 72

1 ,4-Chloronitrobenzene 24 1 23 76

44 1 18 81

 

2.3.5 Catalyst Stability
In addition to high activity, stability is also vital for developing useful

catalytic membranes. Dry membranes modiﬁed with AuNPs could be stored in

95

an ambient environment for several months without negative effects on their
activity. Figure 2.27 demonstrates that membranes are also reasonably stable
under ﬂow of a solution containing 4-nitrophenol and NaBH4. The percent
reduction remains essentially constant, even though over 140,000 membrane

volumes have passed through the system.

3 99.3 lira’hrwe’...‘

 

 

99 . . r r v
0 500 1000 1500 2000 2500
Penneate Volume (mL)

Figure 2.27 Plot of percent reduction of 4-nitrophenol vs. volume eluted through
an alumina membrane modiﬁed with a PAA/PAH/AuNP ﬁlm. Conditions: [4-
nitrophenol] = 2.0 mM, [NaBH4] = 100 mM, ﬂux = 0.17 mUcmz-sec. The plot
contains a compilation of 5 sequential passages of 500 mL of the reaction
solution.

The data in Figure 2.27 show a turnover number >1000 moles of 4-
nitrophenol reduced per mole of Au contained within the membrane. After
attainment of this high turnover number, 4-nitrophenol conversion at a ﬂux of
0.17 mUcmZ-sec decreased only marginally, from 99.8% to 99.7%. These
results indicate that with > 50-fold excess of NaBH4, minimal fouling occurred on
the surface of the AuNPs, and little or no Au was leached from the membrane.

Larger turnover numbers may be more easily achieved with higher

96

concentrations of 4-nitrophenol and NaBH4, but the high pH (>10) of such
solutions results in slow dissolution of the alumina support and may also remove
the polyelectrolyte ﬁlm.”72 Additionally, some studies suggest that the catalytic
activity of AuNPs in NaBH4 reductions decreases slightly over time, perhaps
because of catalyst poisoning.73

Analysis of the amount of Au contained in membranes before and after
using them in the reduction of 4-nitrophenol for several hours at high flow rates,
showed no signiﬁcant decrease (<5%) in gold content. Furthermore, FESEM
images of membranes that were used for 2 hours in the reduction of 4-
nitrophenol at a ﬂux of 0.030 mUcmz-sec show that the nanoparticles were still
well separated from one another and did not leach from the membrane during the
reaction (Figure 2.28). However, we do have some preliminary evidence that at
. higher ﬂuxes, catalytic activity decreases over time and a cake layer can form on

the surface of the membrane. Investigations into gentler reducing agents will be

beneficial for future work with these catalytic membranes.

    

Figure 2.28 Cross-sectional FESEM images of membranes modiﬁed with
PAA/PAH/AuNP ﬁlms a) before and b) after ﬂowing a solution of 1 mM 4-
nitrophenol and 100 mM NaBH4 through the membrane for 2 hours at 0.030
mL/cm -sec.

97

2.4 Conclusions

Overall, membranes modiﬁed with Au nanoparticles in multilayer
polyelectrolyte ﬁlms present a stable system that has great potential for
catalyzing fast reactions. The polyelectrolyte multilayer provides a convenient
platform for immobilizing active, accessible catalytic nanoparticles in a variety of
porous supports including anodized alumina, track-etched polycarbonate, and
nylon. The ability to vary the type and amount of polyelectrolyte and/or catalytic
material makes this method even more versatile. Membranes modified by LbL
deposition of polyelectrolyte/AuNP ﬁlms show high catalytic activity in the
reduction of nitroaromatic compounds with sodium borohydride. A comparison of
the three membrane types as well as alumina powder modiﬁed with
polyelectrolyte/AuNP ﬁlms shows that the catalytic activity is independent of the
substrate. The membranes also catalyze selective reduction of the nitro group in
the presence of other reducible functional groups, and the extent of formation of
nitroso products depends on ring substituents. An important asset of the
membrane catalysts is the possibility of varying ﬂux to control product
distribution, as demonstrated by the formation of increased levels of nitroso
intermediates at high ﬂux. The following chapters will discuss the use of LbL
formation of polyelectrolyte/nanoparticle films with other noble metal
nanoparticles, additional support materials, and catalytic reactions with milder

oxidizing or reducing agents.

98

2.5 References

(1) Ono, N. The Nitro Group in Organic Synthesis; Wiley-VCH: New York,
2001.

(2) Downing, R. S.; Kunkeler, P. J.; Bekkum, H. V. Catal. Today 1997,
37, 121-136.

(3) Adams, J. P. J. Chem. Soc, Perkin Trans. 1 2002, 2586-2597.

(4) Jacobsen, M. F.; Moses, J. E.; Adlington, R. M.; Baldwin, J. E. Org.
Lett. 2005, 7, 641-644.

(5) Tafesh, A. M.; Weiguny, J. Chem. Rev. 1996, 96, 2035-2052.

(6) Brickner, S. J.; Hutchinson, D. K.; Barbachyn, M. R.; Manninen, P.
R.; Ulanowicz, D. A.; Garmon, S. A.; Grega, K. C.; Hendges, S. K.; Toops,
D. S. J. Med. Chem. 1996, 39, 673-9.

(7) Figueras, F.; Coq, B. J. Mol. Catal. A: Chem. 2001, 173, 223-230.

(8) Pietrowski, M.; Zielinski, M.; Wojciechowska, M. Catal. Lett. 2009,
128, 31-35.

(9) Motoyama, Y.; Kamo, K.; Nagashima, H. Org. Lett. 2009, 11, 1345-
1348. '

(10) Zheng, Y.; Ma, K.; Wang, H.; Sun, X.; Jiang, J.; Wang, C.; Li, R.; Ma,
J. Catal. Lett. 2008, 124, 268-276.

(11) Maity, P.; Basu, 8.; Bhaduri, S.; Lahiri, G. K. Adv. Synth. Catal.
2007, 349, 1955-1962.

(12) Du, Y.; Chen, H.; Chen, R.; Xu, N. Appl. Catal., A 2004, 277, 259-
264.

(13) Wen, H.; Yao, K.; Zhang, Y.; Zhou, Z.; Kirschning, A. Catal.
Commun. 2009, 10, 1207-121 1.

(14) Saha, A.; Ranu, B. J. Org. Chem. 2008, 73, 6867-6870.
(15) Adger, B. M.; Young, R. G. Tetrahedron Lett. 1984, 25, 5219-22.

99

(16) Wang, H.; Yan, J.; Chang, W.; Zhang, Z. Catal. Commun. 2009, 10,
989-994.

(17) Haber, F. 2. Elektrochem. Angew. Phys. Chem. 1898, 4, 506-13.

(18) Noel, M.; Ravichandran, C.; Anantharaman, P. N. J. Appl.
Electrochem. 1995, 25, 690-8.

(19) Li, F.; Cui, J.; Qian, X.; Zhang, R.; Xiao, Y. Chem. Commun. 2005,
1901-1903.

(20) Li, F.; Cui, J.; Qian, X.; Zhang, R. Chem. Commun. 2004, 2338-
2339.

(21) Blackie, J. A.; Turner, N.; Wells, A. Tetrahedron Lett. 1997, 38,
3043-3046.

(22) Rahaim, R. J., Jr.; Maleczka, R. E., Jr. Org. Lett. 2005, 7, 5087-
5090.

(23) Blaser, H.—U. Science 2006, 313, 312-313.

(24) Blaser, H. U.; Siegrist, U.; Steiner, H.; Studer, M. In Fine Chemicals
Through Heterogeneous Catalysis; Sheldon, R. A., Bekkum, H. v., Eds.; Wiley-
VCH: Weinheim, Germany, 2001, p 389-406.

(25) Baumeister, P.; Studer, M.; Roessler, F. In Handbook of Heterogeneous
Catalysis; Ertl, G., Kndzinger, H., Wieitkamp, J., Eds; Wiley-VCH: Weinheim,
Germany, 1997, p 2186-2209.

(26) Siegrist, U.; Baumeister, P.; Blaser, H.-U.; Studer, M. Chem. Ind.
(Dekker) 1998, 75, 207-219.

(27) Corma, A.; Serna, P. Science 2006, 313, 332-334.

(28) Boronat, M.; Concepcion, P.; Corma, A.; Gonzalez, 8.; Illas, F.;
Serna, P. J. Am. Chem. Soc. 2007, 129, 16230-16237.

(29) Corma, A.; Serna, P. Nature Protocols 2006, 1, 2590-2595.
(30) Chen, Y.; Qiu, J.; Wang, X.; Xiu, J. J. Catal. 2006, 242, 227-230.

(31) Rylander, P. Catalytic Hydrogenation in Organic Syntheses; Academic
Press: New York, 1979.

100

(32) Lu, X.; Xu, X.; Wang, N.; Zhang, Q. J. Phys. Chem. A 1999, 103,
1 0969-1 0974.

(33) Takasaki, M.; Motoyama, Y.; Higashi, K.; Yoon, S.-H.; Mochida, I.;
Nagashima, H. Org. Lett. 2008, 10, 1601-1604.

(34) Pradhan, N.; Pal, A.; Pal, T. Langmuir 2001, 17, 1800-1802.
(35) Pradhan, N.; Pal, A.; Pal, T. Colloids Surf, A 2002, 196, 247-257.

(36) Ghosh, S. K.; Mandal, M.; Kundu, S.; Nath, S.; Pal, T. Appl. Catal., A
2004, 268, 61-66.

(37) Hwang, C.-G.; Kim, S.-H.; Hoon, O. J.; Kim, M.-R.; Choi, S.—H. J.
Ind. Eng. Chem. 2008, 14, 864-868.

(38) Jana, S.; Ghosh, S. K.; Nath, S.; Pande, S.; Praharaj, S.; Panigrahi,
S.; Basu, S.; Endo, T.; Pal, T. Appl. Catal., A 2006, 313, 41-48.

(39) Liu, W.; Yang, X.; Huang, W. J. Colloid Interface Sci. 2006, 304,
160-165.

(40) Lu, Y.; Spyra, P.; Mei, Y.; Ballauff, M.; Pich, A. Macromol. Chem.
Phys. 2007, 208, 254-261.

(41) Esumi, K.; lsono, R.; Yoshimura, T. Langmuir 2004, 20, 237-243.

(42) Hayakawa, K.; Yoshimura, T.; Esumi, K. Langmuir 2003, 19, 5517-
5521.

(43) Mei, Y.; Sharma, G.; Lu, Y.; Ballauff, M.; Drechsler, M.; Irrgang, T.;
Kempe, R. Langmuir 2005, 21, 12229-12234.

(44) Mei, Y.; Lu, Y.; Polzer, F.; Ballauff, M.; Drechsler, M. Chem. Mater.
2007, 19, 1062-1069.

(45) Kundu, S.; Lau, S.; Liang, H. J. Phys. Chem. C 2009, 113, 5150-
5156.

(46) Pogorelié, l.; Filipan-Litvié, M.; Merkas, S.; Ljubic, G.; Cepanec, l.;
Litvic, M. J. Mol. Catal. A: Chem. 2007, 274, 202-207.

(47) Grabar, K. C.; Freeman, R. G.; Hommer, M. B.; Natan, M. J. Anal.
Chem. 1995, 67, 735-43.

101

(48) Turkevich, J.; Stevenson, P. C.; Hillier, J. Discussions of the Faraday
Society 1951, No. 11 , 55-75.

(49) Kidambi, S.; Bruening, M. L. Chem. Mater. 2005, 17, 301-307.

(50) Kidambi, S.; Dai, J.; Li, J.; Bruening, M. L. J. Am. Chem. Soc. 2004,
126, 2658-2659.

(51) Malaisamy, R.; Bruening, M. L. Langmuir 2005, 21, 10587-10592.

(52) Corma, A.; Concepcion, P.; Serna, P. Angew. Chem. Int. Ed. 2007,
46, 7266-7269.

(53) Gelder, E. A.; Jackson, S. D.; Lok, C. M. Chem. Commun. 2005,
522-524.

(54) Akamatsu, K.; Deki, S. J. Mater. Chem. 1997, 7, 1773-1777.
(55) Link, 8.; El-Sayed, M. A. J. Phys. Chem. B1999, 103, 4212-4217.

(56) He, Y. 0.; Liu, S. P.; Kong, L.; Liu, 2. F. Spectrochim. Acta, Part A
2005, 61, 2861-2866.

(57) Shiratori, S. S.; Rubner, M. F. Macromolecules 2000, 33, 4213-4219.

(58) Hong, S. U.; Miller, M. D.; Bruening, M. L. Ind. Eng. Chem. Res.
2006, 45, 6284-6288.

(59) Miller, M. D.; Bruening, M. L. Chem. Mater. 2005, 17, 5375-5381.

(60) Stanton, B. W.; Harris, J. J.; Miller, M. D.; Bruening, M. L. Langmuir
2003, 19, 7038-7042.

(61) Dai, J.; Bruening, M. L. Nano Lett. 2002, 2, 497-501.

(62) Robinson, M.; Pask, J. A.; Fuerstenau, D. W. J. Am. Ceram. Soc.
1964, 47, 516-520.

(63) Harris, J. J.; Bruening, M. L. Langmuir 2000, 16, 2006-2013.

(64) Jani, A. M. M.; Anglin, E. J.; Mcinnes, S. J. P.; Losic, D.; Shapter, J.
G.; Voelcker, N. H. Chem. Commun. 2009, 3062 - 3064.

102

(65) Braunstein, P.; Kormann, H.-P.; Meyer-Zaika, W.; Pugin, R.; Schmid,
G. Chem. Eur. J. 2000, 6, 4637-4646.

(66) Lahav, M.; Sehayek, T.; Vaskevich, A.; Rubinstein, l. Angew. Chem.
Int. Ed. 2003, 42, 5576-5579.

(67) Panigrahi, S.; Basu, S.; Praharaj, 8.; Pande, 8.; Jana, S.; Pal, A.;
Ghosh, S. K.; Pal, T. J. Phys. Chem. C 2007, 111 , 4596-4605.

(68) List, 8.; Pojarliev, P.; Martin, H. J. Org. Lett. 2001, 3, 2423-2425.

(69) Sasai, H.; Tokunaga, T.; Watanabe, 8.; Suzuki, T.; ltoh, N.;
Shibasaki, M. J. Org. Chem. 1995, 60, 7388-9.

(70) Dubas, S. T.; Schlenoff, J. B. Macromolecules 2001, 34, 3736-3740.
(71) Kharlampieva, E.; Sukhishvili, S. A. Langmuir 2003, 19, 1235-1243.

(72) Kohli, P.; Wharton John, E.; Braide, 0.; Martin Charles, R. Journal of
Nanoscience and Nanotechnology 2004, 4, 605-10.

(73) Wee, J.-H.; Lee, K.-Y.; Kim, S. H. Fuel Process. Technol. 2006, 87,
811-819.

103

Some of the results and ﬁgures contained in chapter three appeared in the
following publication.

Bhattacharjee, S.; Dotzauer, D. M.; Bruening, M. L. J. Am. Chem. Soc. 2009,
131, 3601-3610

104

Chapter Three: Hydrogenation Reactions with Pd-Contalning Catalysts
3.1 Introduction

The previous chapter showed that catalytic membranes effectively
catalyze reactions in which all reactants are in the liquid phase; however, many
industrial applications involve gas/liquid reactions such as hydrogenations or
oxidations of substrates in the liquid phase. The preference to use
heterogeneous catalysts in industry leads to the difﬁcult task of maintaining close
contact between three phases (solid catalyst/liquid/gas) within the catalytic
reactor. This multiphase catalysis poses many challenges in designing the
proper reactor system. Important factors to consider in designing an appropriate
reactor include the rate of the reaction, mass transfer of reactants, ease of scale-
up, and cost of materials. This chapter describes hydrogenation reactions with
slurry reactors in which the catalyst is immobilized on an alumina powder support
and ﬂow-through membrane reactors where the catalyst is immobilized in the
pores of ﬂat disk-shaped alumina membranes. These initial studies examine
reactions where the gas is simply bubbled into the solvent. Chapter four
discusses gas/liquid interfacial contactors.

In this research area, we are particularly interested in hydrogenation
reactions because they occur on short time scales and are widely used in the
reduction of oleﬁns and a number of other functional groups.1 Previous studies
in the Bruening group focused on the hydrogenation of allyl alcohols containing
several different alkyl substituents in the position a to the double bond.“ These

studies showed that Pd nanoparticles embedded in polyelectrolyte multilayer

105

films catalyze hydrogenation of smaller molecules at a faster rate than molecules
containing bulkier alkyl substituents, presumably because of faster transport of
the smaller molecules to the catalyst surface. As an extension of this work, we
were interested in looking at the effect of nanoparticle size on activity and
selectivity in these reactions. Control over particle size is often necessary for
creating highly active nanoparticle catalysts for hydrogenation of alkeness'9 or

10'" and conjugated alkenes.12 Moreover,

partial hydrogenation of allqrnes
nanoparticle catalysts showed high intermolecular selectivities in the
hydrogenation of alkenes13 and allylic alcohols,”15 but these studies did not look
at selectivity as a function of particle size. Investigations of catalyst selectivity as
a function of particle size should provide insight into whether the enhanced
catalyst activity for certain compounds occurs due to steric effects, increased
catalyst surface area, or changes in the nanoparticle electronic properties. The
initial studies in this chapter use catalysts comprised of alumina powder modiﬁed
with polyelectrolyte/PdNP ﬁlms to explore the effect of nanoparticle size on
reactivity/selectivity in slurry reactions.

The second part of this chapter focuses on hydrogenation reactions in
catalytic membranes. Slurry reactions often require fast stirring to maintain
contact between gas, liquid, and catalyst in the reaction system, but diffusion of
the reactants to the catalyst surface may still limit the reaction rate in the case of
fast kinetics. Flow-through catalytic membrane contactors can overcome this

diffusion limitation and also avoid the need to separate the catalyst from the

reaction mixture. By forcing a premixed reactant stream containing the liquid and

106

the dissolved gas through the membrane, efficient contact occurs between the
reactants and the catalyst on the pore walls. When membrane pore diameters
are small, the diffusion time is negligible compared to reaction times, and the
reaction rate depends on either kinetics or the rate of mass ﬂow into the
membrane. As discussed in Chapter 2, membrane contactors also allow
variation of the contact time between the reactants and the catalyst, which may
allow control over selectivity or product distribution. Moreover, catalytic
membrane reactors could provide easier scale-up for industrial-scale reactions
than conventional slurry reactors. This work uses membranes containing PdNPs
for the hydrogenation of a number of substrates including allyl alcohol and

several nitroaromatic compounds.

3.2 Experimental Methods
3.2.1 Materials

PEI (MW = 25,000 Da), PAA (MW = 90,000 Da), PAH (Mw = 15,000 Da), a-
alumina (100 mesh, ~100 pm particle size), 5% Pd on alumina, allyl alcohol, 2-
buten-1-ol (mixture of isomers), 2-methyl-2-propen-1-ol, isobutyraldehyde,
nitrobenzene, 4-nitrophenol, 4-nitrotoluene, potassium tetrachloropalladate(ll),
potassium hexachloropalladate(lV), and sodium borohydride were purchased
from Aldrich. PAA (Mw = 5,000 Da) was obtained from Polysciences. and
butyraldehyde and propionaldehyde were purchased from Fluka. 1-Butanol (J.T.
Baker), n-propyl alcohol (Columbus Chemical), and isobutyl alcohol (Spectrum)

were used as received. Anodisc aluminum oxide membranes (25 mm disks with

107

 

0.2 pm diameter pore sizes) were acquired from Whatman. Aqueous solutions
were prepared with deionized water (18.2 MO cm, Milli-Q puriﬁcation system),

and hydrogen (99.9%) was obtained from AGA gases.

3.2.2 Modification of Alumina Powder with Polyelectrolyte/PdNP Films
Previous studies in the Bruening group involved the preparation of
[PAA/PEl-Pd(0)]x ﬁlms on alumina powder.“ Variation of the concentration of
Pd in the polyelectrolyte deposition solution afforded control over the size of the
Pd nanoparticles. However, the presence of Pd precursors at concentrations
>10 mM in PEI solutions results in ﬂocculation of the polyelectrolyte. Thus, this
research incorporated the Pd precursor into the PAA solution to form [PAA-
Pd(0)/PE|]3 ﬁlms on alumina (Figure 3.1). In this modiﬁed procedure, 15 g of
alumina was alternately suspended for 10 min in a 100 mL solution that
contained 0.02 M PAA (Mw = 90,000, pH adjusted to 4.0 with 0.1 M NaOH) and
1, 4, or 8 mM K2PdCI4, and for 10 min in a 100 mL solution that contained PEI (1
mg/mL, pH adjusted to 9.0 with 0.1 M HCI). After the deposition of each
polyelectrolyte layer, the alumina was rinsed four times by suspension in 100 mL
of deionized water, stirring for 5 min, allowing the alumina to settle, and
decanting the supernatant. After deposition of three PAA-Pd(ll)/PEI bilayers,
reduction to Pd(0) occurred by stirring the modiﬁed alumina powder in 100 mL of
freshly prepared 0.1 M NaBH4 for 30 min, followed by washing with four 100-mL

aliquots of deionized water as described above. After decanting the ﬁnal water

108

rinse, the resulting powder was dried under vacuum for 8 hours to remove

residual water.

(2)

  

(1) ,
0.02 M PAA-Pd(ll) - ' 1 mg/mL PEI
10 min 3. 10 min
—_> ‘
W ' W

0: Pd(ll)

 

- ‘ 4’s?!
,‘4‘5‘551. 36$- 2',
«fess-«EMMA»:

.".‘,’ ~thr-u’
' ,J; . .
'

Repeat
(1 ,2,1 ,2) ""
—

 

J C
Figure 3.1 Schematic diagram showing the formation of [PAA-Pd(0)/PEI]3 ﬁlms
on alumina powder. Rinses are not shown.

Alternatively, alumina powder was also modiﬁed with preformed Pd
nanoparticles using the same procedure described in chapter 2 for formation of
polyelectrolyte/AuNP ﬁlms on alumina. In a 100-mL ﬂask, 7.5 g of 100 mesh
alumina was added to 50 mL of a pH 4.5 solution containing 0.02 M PAA (Mw =
5,000) and 0.5 M NaCl, and then stirred vigorously for 10 min. The alumina was
allowed to settle and the supernatant was decanted. To remove excess
polyelectrolyte, the alumina particles were washed 4 times by adding 50 mL of
deionized water, stirring, letting the alumina settle, and decanting the water.
Next, 50 mL of a pH 4.5 solution containing 0.02 M PAH and 0.5 M NaCI was

added to the alumina, and this mixture was stirred for 10 min, after which the

109

particles were washed as described above. To deposit the Pd nanoparticles, 50
mL of as-prepared Pd nanoparticle solution was added to the PAA/PAH-modiﬁed
alumina and stirred for 10 min. The alumina was then allowed to settle and
rinsed 4 times with water as described above. Additional PAH/PdNP bilayers
were added by alternating deposition of PAH and PdNPs in the manner
described above with rinsing between each layer. After decanting the final water
rinse, the resulting powder was dried under vacuum for 8 hours to remove
residual water. Figure 3.2 shows the overall procedure for forming
PAA/[PAH/PdNP]x ﬁlms on alumina powder.

The ~7nm diameter preformed PdNPs were prepared according to the
citrate reduction method.16 In a 250 mL volumetric ﬂask equipped with a
condenser, 100 mL of aqueous K2PdCI6 (0.393 mM) was heated to boiling under
vigorous stirring, and 12 mL of a heated 1% sodium citrate solution was added to

the Pd solution under continued stirring. After adding the sodium citrate, the

 

Figure 3.2 Schematic diagram showing the formation of PAA/[PAH/PdNP]x ﬁlms
on alumina powder. Rinses are not shown.

110

solution turned from yellow to brown. The mixture was boiled while stirring for 4
hours, during which the solution gradually changed from light brown to dark
brown in color. The solution was then removed from the heat and cooled to room
temperature. The resulting Pd nanoparticle solution was stored in an amber

bottle at 5 °C until use.

3.2.3 Modification of Alumina Membranes with Polyelectrolyte/PdNP Films
Alumina membranes were pretreated in a UV/ozone cleaner for 15 min
and then modified with polyelectrolyte/PdNP ﬁlms by the same method described
in chapter 2 for formation of polyelectrolyte/AuNP ﬁlms in alumina membranes.
During ﬁlm formation, the membranes were placed in an Amicon 8010
ultraﬁltration cell, and all solutions were passed through the membranes using a
peristaltic pump (Cole Parmer Masterﬂex C/L) attached to the permeate side of
the Amicon cell. Brieﬂy, solutions of PAA (10-15 mL), water (40 mL), PEI (10-15
mL), water (40 mL), citrate-stabilized Pd-nanoparticles, and water (40 mL) were
sequentially passed through the membranes at a ﬂux of ~ 0.02 mL/cmZ-sec to
deposit a PAA/PEI/PdNP ﬁlm. The PAA solution composition was the same as
that used for modiﬁcation of alumina powder with PAN[PAH-PdNij films, and
the PEI solution concentration was 0.02 M in 0.5 M NaCl with the pH adjusted to
9.0 with 1 M NaOH . For PdNP deposition, the as-prepared PdNP solution was
diluted by a factor of 5 and passed through the membrane until the PdNPs were

no longer adhering to the polyelectrolyte layer, i.e., the solution exiting the

111

membrane exhibited the same brown color as the solution entering the

membrane.

3.2.4 Characterization

Preformed citrate-stabilized PdNPs were characterized by transmission
electron microscopy (TEM, JEM-2200FS microscope) to determine their
approximate size and shape. TEM Samples were prepared by diluting the as-
prepared PdNP solution by a factor of 5 with water and placing a 2 pL drop of this
mixture on a carbon-coated copper grid. Pd-containing alumina membrane
samples were imaged using ﬁeld-emission scanning electron microscopy
(FESEM) with a Hitachi S-4700 II microscope. Prior to imaging, membrane
samples were fractured to reveal a cross-section and coated with 5 to 10 nm of
Au with a Pelco SC-7 sputter coater.

The amounts of palladium in modiﬁed alumina powder catalysts were
determined by ﬂame atomic emission spectroscopy (FAES) or inductively
coupled plasma optical emission spectroscopy (lCP-OES, Varian 710-ES).
Standard solutions were prepared by dissolving K2PdCl4 in 2% HNOa, and
sample solutions were prepared by adding the desired amount of alumina
powder catalyst to 2 mL of aqua regia, agitating the mixture by sonication,
removing the supernatant (the d-alumina support does not dissolve in aqua
regia), and diluting the supernatant with deionized water prior to elemental
analysis. Determination of Pd content in alumina membranes occurred similarly

by dissolving the Pd from the alumina membranes with 2 mL aliquots of aqua

112

 

regia and diluting those solutions with deionized water prior to elemental

analysis.

3.2.5 Hydrogenation Reactions with Pd-Containing Alumina Powder
Catalytic hydrogenation reactions were carried out in a 100-mL, three-
neck, round-bottomed ﬂask (Figure 3.3). A desired amount of alumina-supported
catalyst was suspended in 25 mL of deionized water and bubbled with H2 for 30
min prior to adding 25 mL of a 50 mM aqueous solution of the allylic substrate.
Hydrogen was constantly bubbled into the solution, which was stirred vigorously
during the reaction. Different amounts of catalyst were used for hydrogenation
reactions because each powder has a different wt% of Pd. The values are
shown in Table 3.1. Sample aliquots were removed from the reaction ﬂask at
specific time intervals, filtered through cotton, and then diluted by a factor of 3

prior to GO analysis (Shimadzu GC-17A equipped with an RTx-BAC1column).

Liquid phase
reactants with
dispersed catalyst

 

 

 

\_J
Figure 3.3 Schematic diagram of the apparatus used for slurry hydrogenation
reactions with Pd-containing alumina powder catalysts at room temperature and
atmospheric pressure.

113

In hydrogenation of allyl alcohol, B-methallyl alcohol, and crotyl alcohol, we
observed mixtures of the isomerization product and the reduced product
(Scheme 3.1, page 119). Response factors were determined from the peak
areas obtained in the analysis of a 1:1:1 mixture of the starting material,
isomerization product, and hydrogenation product and were used to calculate the
turnover frequency (TOF) of the catalyst. TOF’s were calculated by plotting
percent hydrogenation vs. time and determining the slope at conversions less
than 50 percent. Each experiment was repeated twice and the :I: values

represent the difference between the average and the data points

3.2.6 Hydrogenation Reactions with Pd-Containing Membranes
Hydrogenation reactions performed with alumina membranes were carried
out by bubbling H2 into a substrate solution for a set time and then (while still
bubbling the solution above the membrane) passing the solution through a Pd
nanoparticle-modiﬁed membrane at a given ﬂux (Figure 3.4). For hydrogenation
of allylic compounds, samples of the membrane permeate were analyzed by GC
(Shimadzu GC-17A) without any sample pretreatment. For hydrogenation of
nitroaromatic compounds, samples of the permeate were analyzed by UV-Vis
spectrophotometry or GC-MS (Hewlett-Packard 618003 GCD system with a HP-
5MS column) to determine the amount of starting material that was
hydrogenated. Prior to GC-MS analysis, 3 mL of sample was extracted three
times into 1 mL portions of diethyl ether, which were then combined. For some

experiments, the permeate solution was collected, resaturated with H2, and

114

 

passed through the Pd nanoparticle-containing membrane multiple times to

achieve higher conversion.

      

; Collection
\_2 new: Beaker

Figure 3.4 Schematic diagram of the apparatus used for ﬂow-through
membrane hydrogenation reactions with Pd-containing alumina membranes.

 
   

 

 

3.3 Results and Discussion
3.3.1 Characterization of Polyelectrolyte/PdNP Films

Examination of the effect of nanoparticle size on catalytic selectivity
required the ability to prepare a series of catalysts with different, well-deﬁned
particle sizes. Kidambi et al. prepared [PAA/PEI-Pd(ll)]3PAA ﬁlms using 2 mM
Pd(ll) in the PEI-Pd(ll) deposition solution, and subsequently reduced the Pd(ll)
in the ﬁlm to obtain [PAA/PEI-Pd(0)]3PAA coatings containing Pd nanoparticles
with diameters of 1-3 nm.3 To vary the nanoparticle size, we increased the
concentration of Pd(ll) in the PEI deposition solution, but with Pd(ll)
concentrations >10 mM, the PEI solution became cloudy, perhaps because the
Pd(ll) cross-linked PEI chains. Thus we chose to add Pd(ll) to the PAA, rather
than the PEI, deposition solution. The 20 mM PAA deposition solution remained

clear at Pd(ll) concentrations as high as 15 mM. Previous studies by Rubner and

115

coworkers demonstrated adsorption of metal ions into preformed PAA/PAH ﬁlms
and subsequent formation of nanoparticles,”18 but in our work, we incorporated
the metal ions during the polyelectrolyte LbL assembly.

The sizes of Pd nanoparticles in [PAA-Pd(0)/PEI]3 ﬁlms increased with
increasing concentrations of Pd(ll) in PAA deposition solutions. Previous work
by Bhattacharjee showed that with Pd(ll) concentrations of 1, 4, or 8 mM in the
PAA deposition solutions, the average particle diameters in [PAA-Pd(0)/PEI]3
ﬁlms on carbon-coated copper TEM grids were 2.2 :I: 0.6, 2.9 :I: 0.5, and 3.2 i 0.6
nm, respectively.19 Due to the difﬁculty of imaging Pd nanoparticles on alumina
powder, the diameters obtained from ﬁlms on TEM grids were used as estimates
of the diameters of Pd nanoparticles in ﬁlms on alumina powder.

Alumina powder catalysts containing ﬁlms with preformed Pd
nanoparticles were also studied to provide a slightly larger particle size for
additional comparison. To investigate the effect of diffusion through the
polyelectrolyte multilayer ﬁlm on activity/selectivity, we modiﬁed alumina powder
with PAA/[PAH/PdNij ﬁlms containing 1, 2, or 3 PAH/PdNP bilayers. The
citrate-stabilized Pd nanoparticles exhibited a brown color with a peak
absorbance at ~250 nm (Figure 3.5), which is consistent with the literature for
particles with diameters of ~7 nm.20 Further analysis by TEM showed a
monodisperse size distribution with an average nanoparticle diameter of 7.2 i 1.2

nm (Figure 3.6).

116

§ 0.8 -

e“ 0.6 -

0

g 0.4 -
0.2 -

 

 

0 . . .
200 300 400 500
Wavelength (nm)

Figure 3.5 UV-Vis spectrum of a 0.10 mM Pd nanoparticle solution showing a
peak absorbance at 250 nm. Molarity is given with respect to Pd atoms.

b)

Number

 

  

456789101112
Nanoparticle Diameter (nm)

a

100 nm
Figure 3.6 a) TEM image of citrate-stabilized Pd nanoparticles on a carbon-

coated copper grid. b) Histogram of Pd nanoparticle diameters taken from
several TEM images. The average particle diameter was 7.2 t 1.2 nm.

The Pd content of each catalyst was determined by FAES or ICP-OES,
and Table 3.1 presents these results. Catalysts modiﬁed with [PAA-Pd(0)/PEI]3
ﬁlms exhibited increased Pd loading with increased Pd(ll) concentration during
deposition. The catalyst prepared with a Pd(ll) concentration of 8 mM exhibited
unusually high loading, presumably due to enhanced adsorption of the Pd(ll) ions

to the amine groups in the already-deposited polyelectrolytes. The Pd loading in

117

catalysts modiﬁed by PAA/[PAH/PdNP]x ﬁlms increased linearly with increasing

the number of PAH/PdNP bilayers (Figure 3.7).

0.60 1
0.50 1
0.40 1
0.30 '
0.20 "
0.10 '

Pd Loading (mg/g N203)

 

 

0.00 I I I I
0 1 2 3 4
Number of Bilayers

Figure 3.7 Plot of Pd loading vs. number of PAH/PdNP bilayers in catalysts
containing PAN[PAH/PdNij ﬁlms.

Table 3.1 Pd content of several different Pd catalysts used for hydrogenation
reactions.

 

 

Pd loading Mass of catalyst used in
Catalyst (mg Pd lg N203) hydrogenation (mg)
[PAA-Pd(0)/PEI]3 on N203,

Pd(lI) = 1 mMa 0.28 500
[PAA'Pd(0)/PEI]3 on N203,

Pd(ll) = 4 mMa 0.97 250
[PAA-Pd(0)/PEI]3 on N203,

Pd(ll) = 8 mMa 6.29 125
PAA/[PAH/PdNP], on AI203 0.18 250
PAA/[PAH/PdNP]; on Al203 0.39 125
PAA/[PAH/PdNP13 on Al203 0.56 100

 

3The initial concentration of KdeCl4 in the PAA deposition solutions.

118

 

3.3.2 Hydrogenation of Allylic Compounds with Pd-Containing Alumina
Powder Catalysts

To investigate catalytic selectivity as a function of nanoparticle diameter,
we determined TOFs (moles hydrogenated/moi Pd/h) for hydrogenation of allyl
alcohol, B-methallyl alcohol, and crotyl alcohol over several different catalysts.
Unfortunately, isomerization of the unsaturated alcohols to the corresponding
aldehydes (Scheme 3.1) complicated the interpretation of the TOFs because the
isomerization product does not readily undergo hydrogenation. Thus, TOFs were
calculated both for the initial hydrogenation rate and the sum of the initial

hydrogenation and isomerization rates.

\/\OH Pd catalyst _\/“:H+ +\/\0H

H2

Pd catalyst

/\/\OH —_.H2/\/u\ OH

Scheme 3.1 Pd-catalyzed hydrogenation of allylH alcohol, B-methallyl alcohol,
and crotyl alcohol.

Table 3.2 shows the hydrogenation TOF’s and corresponding selectivities
(T OF ratios) for the three [PAA-Pd(0)lPEl]3 catalysts as well as a commercially
available 5% Pd on alumina catalyst. Although the commercially available
catalyst showed approximately the same hydrogenation TOF for all three

compounds, the [PAA-Pd(0)/PEI]3 catalysts had signiﬁcantly different TOF’s for

119

Table 3.2. TOFs and TOF ratios (selectivities) for unsaturated-alcohol
hydrogenation catalyzed by Pd catalysts that contained Pd nanoparticles with
different diameters. Values in parentheses are the sum of TOFs for
hydrogenation and isomerization.

 

Average TOF (moles hydrogenatedlmol Poi/hf“b Ratio of TOFs
nanoparticle

 

 

Catalyst diamm, Allyl B-methallyl Crotyl Allyl I Allyl I
(nm)" alcohol alcohol alcohol B-methallyl Crotyl
[PAA-Pd(0)/PElla 2 2 2100 1 280 5 1 1 56 1 8 420 38
[Pd(ll)]=1 mMc ' (2300 1 300) (5 1 1) (65 1 9) (460) (35)
[PAA-Pd(0)/PE|Ia 2.9 1500 1 80 18 1 3 74 1 8 83 20
[Pd(ll)]=4 mM" (1800 1 80) (18 1 3) (90 1 8) (100) (20)
[PAA-Pd(0)/PE|]3 3 2 1700 1 300 29 1 9 300 1 60 59 5.7
[Pd(ll)]=8 mMc ' (2000 1 400) (46 1 10) (370 1 100) (43) (5.4)
Commercial _ 1200 1 300 1100 1 50 1400 1 50 1.1 0.9
5% Pd on Al203 (1600 1 300) (1700 1 100) (1600 1 50) (0.9) (1.0)

 

a50 mL of 25 mM substrate was hydrogenated in water. TOFs were initial values
determined at conversions less than 50%.

bThe catalysts used in this study came from different batches than the catalysts
used in reference 21 and have somewhat different TOFs and selectivities than
previously published data. Trends are similar in the two cases.21

c‘l’he initial concentration of K2PdCl4 in the PAA deposition solutions.

dAverage Pd nanoparticle diameter determined by Bhattacharjee.19

each compound. In general, the TOF was higher for allyl alcohol than for [3-
methallyl alcohol and crotyl alcohol for each [PAA-Pd(0)/PEI]3 catalyst. More
importantly, the selectivity (ratio of TOF’s) changed dramatically with the particle
size. The trend in Table 3.2 is that the TOF for hydrogenation of allyl alcohol
decreases or remains the same with increasing nanoparticle size, while the
corresponding TOFs for B-methallyl alcohol and crotyl alcohol increase with
increasing nanoparticle size. This occurs even though the surface area to
volume ratio for the catalyst is smaller for larger particles. As a result of this

trend, selectivities decreased by a factor of ~7 upon going from nanoparticle

120

sizes of 2.2 nm to 3.2 nm. Similar trends occurred when analyzing the sum of
hydrogenation and isomerization.

The high selectivities of nanoparticles in polyelectrolyte ﬁlms may occur in
part because the surrounding polyelectrolyte decreases the rate at which bulkier
molecules reach catalytic sites. Several studies showed that polyelectrolyte
multilayers provide a signiﬁcant barrier to the diffusion of small compounds, and
even molecules as small as glycerol and methanol show much slower transport
than water in polyelectrolyte ﬁlms.22 To examine whether simple diffusion
through polyelectrolyte ﬁlms enhances selectivity, we prepared catalysts with
PAA/[PAH/PdNP]x ﬁlms containing 1, 2, and 3 PAH/PdNP bilayers. By using
preformed PdNPs in the deposition process, we were able to ensure that the size
of the PdNPs was the same for each catalyst. Thus any differences in reactivity
or selectivity should be due strictly to diffusion through the polyelectrolyte and not
due to differences in particle size.

Table 3.3 shows that catalysts prepared by LbL deposition with preformed
Pd nanoparticles exhibit substantially higher TOFs (at least 5 times greater for
each substrate) than catalysts containing in-situ prepared Pd nanoparticles, even
though the preformed nanoparticles have a larger diameter and thus a smaller
surface area to volume ratio. This enhanced activity is most likely due to the high
accessibility of the reactants to the catalytic nanoparticles contained in the
exterior nanoparticle layer of the PAA/[PAH/PdNij ﬁlms. Conversely, in-situ
prepared nanoparticles may be completely embedded within the polyelectrolyte

ﬁlm leading to much less accessible catalyst sites.

121

 

Table 3.3. TOFs and TOF ratios (selectivities) for unsaturated-alcohol
hydrogenation catalyzed by Pd catalysts that contained PAA/[PAH/PdNP]x films
with 1, 2, and 3 PAH/PdNP bilayers. Values in parentheses are the sum of TOFs
for hydrogenation and isomerization.

 

 

 

TOFs (moles hydrogenated! mol Pdl l1)a Ratio of TOFs
Catalyst Allyl B-methallyl Crotyl Allyl I Allyl I
alcohol alcohol alcohol B-methallyl Crotyl
11400 1 700 270 1 50 1500 1 150 42 7.6
PANIPAH’PdNPI‘ (13100 1 1100) (350 1 80) (1800 1 150) (37) (7.3)
5600 1 100 90 1 50 600 1 50 62 9.3
PAA/[PAH’PdNPh (6400 1 80) (110 1 50) (760 1 100) (58) (8.4)
5100 1 50 170 1 60 640 1 60 30 8.0
PANIPAH’PdNPh (5600 1 90) (220 1 80) (780 1 120) (25) (7.2)

 

350 mL of 25 mM substrate was hydrogenated in water. TOFs were initial values
determined at conversions less than 50%.

More importantly, we see that while the number of layers does affect
catalytic activity, it has little effect on selectivity in the hydrogenation of allyl
alcohol, B-methallyl alcohol, and crotyl alcohol. When going from one bilayer to
two bilayers, the TOF decreases signiﬁcantly for each substrate. This decrease
in activity suggests that the addition of a second polyelectrolyte/PdNP bilayer
over the top of the initial nanoparticle layer inhibits diffusion of the reactants to
the interior nanoparticles. However, the extent of decreased activity is similar for
each substrate, as seen in the similar selectivities .for each catalyst.
Furthermore, the TOFs for the three-layer catalyst remained approximately the
same as the two layer catalyst. The 7.2 nm diameter nanoparticles exhibited a
much higher selectivity than expected based on the trend of decreased selectivity
with increased nanoparticle size shown in Table 3.2. However, the preformed Pd

nanoparticles had citrate stabilizers and thus a different composition than in-situ

122

prepared nanoparticles. Hence, direct comparison between the two nanoparticle
systems should be viewed with caution. Further studies with different sizes of
citrate-stabilized Pd nanoparticles could provide additional insight into selectivity
as a function of nanoparticle size.

Previous studies of nanoparticle reactivity suggest that the size-based
selectivity trends we observed may arise from geometrical effects.23 Doyle, et al.
proposed that the sensitivity of catalytic hydrogenation rates to nanoparticle size
occurs because the reaction takes place on specific types of atoms that are more
(or less) prevalent on small particles.5 Nanoparticles with large diameters mainly
contain terraces with atoms having high coordination numbers, whereas smaller
nanoparticles have more edge and corner atoms with low coordination numbers.
A recent study based on STM images showed that Pd nanoparticles with
diameters < 4 nm contain many defects, while nanoparticles with diameters > 4
nm have more large and well-deﬁnedfacets.12'24 Additionally, some studies have
shown that the dependence of TOF on particle size disappears when TOF is
calculated with respect to a deﬁned type of atom (edge atoms, for example),
which would make the reaction depend on nanoparticle structure rather than
8126.12’25'26

In general, we see that the TOFs for hydrogenation of allyl alcohol
decrease slightly with increasing nanoparticle diameter, while the TOFs for B-
methallyl alcohol and crotyl alcohol increase as nanoparticle size increases. One

explanation for this observation is that in the hydrogenation of monosubstituted

double bonds suchas allyl alcohol, defect atoms are the primary active centers

123

 

for catalysis, whereas hydrogenation of multisubstituted double bonds occurs
more readily on terraces. Hence, TOFs for the monosubstituted compounds
increase as nanoparticle size decreases because of a larger fraction of defect
atoms in the small nanoparticles. The higher TOF of allyl alcohol with the
smallest nanoparticle size could be attributed to the increased surface area of the
smaller particles. It is also possible that hydrogenation of multisubstituted double
bonds is more restricted on the terraces of the smallest nanoparticles because
the active site is smaller. Conversely, the terraces of larger particles have a
larger binding area that may allow hydrogenation of multisubstituted double
bonds to take place more readily.

Hydrogenation of mixtures of the different alcohols was performed to
determine if monosubstituted and multisubstituted compounds are hydrogenated
on the same sites or different sites. In these experiments, mixtures of allyl
alcohol with either B-methallyl alcohol or crotyl alcohol were hydrogenated with
the [PAA-Pd(O)/PEI]3 catalysts under the same conditions as in the single
substrate hydrogenations. These experiments were carried out with the catalysts
prepared with 1 mM K2PdCl4 and 8 mM K2PdCI4. The TOFs and selectivities
obtained with the 1 mM catalyst were similar to those determined from single
component systems. However, the large difference in hydrogenation rates
among the three substrates makes it difﬁcult to see if they were reacting at
different active sites. Conversely, the initial TOFs of B-methallyl alcohol and
crotyl alcohol were much lower in the mixed component systems using the 8 mM

catalyst, which resulted in much higher selectivities than those determined from

124

the single component systems. Furthermore, once the majority of allyl alcohol
reacted, the rate of hydrogenation of B-methallyl alcohol or crotyl alcohol
increased to values similar to those in the single component systems (Figures
3.8 and 3.9). Mixed component experiments were also performed with a
commercial 5% Pd on alumina catalyst, and again the activities of B-methallyl
alcohol and crotyl alcohol were lower than single component experiments until
most of the allyl alcohol reacted. This suggests that all 3 compounds react at the

same active sites, and apparently monosubstituted double bonds are able to bind

 

 

 

 

100.0 - a)
'O
.33; 80.0 - . . .
r: . .
81600 -
.g 40.0 ‘ I
>
I 20.0 1 I
°\ 0.0 III e . .
0 5 10 15
Time (min)
20.0 - b)
8
a; 15.0 1
C
$100-
9 0
">’.
I 5.0 ‘
o\°
0.0 644’" . . .
0 5 10 15
Time (min)

Figure 3.8 Percent hydrogenation of a) allyl alcohol and b) B-methallyl alcohol
during competitive hydrogenation of these substrates using a [PAA-Pd(0)/PEI]3
catalyst prepared with 8 mM Pd(ll). The initial concentration of each substrate
was 25 mM. After 6 min, allyl alcohol was consumed due to hydrogenation and
formation of ~20% propionaldehyde. In (b), lines represent slopes before and
after reaction of allyl alcohol.

125

 

more easily to the surface than multisubstituted double bonds. Therefore, the
enhanced selectivity observed with smaller particles is most likely due to the

ability of allyl alcohol to bind more easily to the active sites of smaller

 

 

 

 

nanoparticles.
amp m
'0 I I
9600l
2 - -
0)
840.0 I
I
£200 -
g I
0.0 I . . .
0 5 10 15
Time (min)
40.0 - b)
8
*5 30.0 -
C
g200-
E .
E.
I 10.0 "
o\°
0064*T‘. . .
0 5 10 15
Time (min)

Figure 3.9 Percent hydrogenation of a) allyl alcohol and b) crotyl alcohol during
competitive hydrogenation of these substrates using a [PAA-Pd(0)/PEI]3 catalyst
prepared with 8 mM Pd(ll). The initial concentration of each substrate was 25
mM. After 6 min, most of the allyl alcohol was consumed due to hydrogenation
and formation of ~20% propionaldehyde. In (D), lines represent slopes before
and after reaction of allyl alcohol.

126

3.3.3 Hydrogenation of Allylic Compounds with Pd-Contalning Membranes
The results in Table 3.3 demonstrate that catalysts containing citrate-
stabilized Pd nanoparticles are especially active in the hydrogenation of allylic
alcohols. While powder-supported catalysts are beneficial for understanding
reaction kinetics, membrane-supported catalysts may offer additional advantages
by providing enhanced contact between the reactants and the catalyst,
separating the catalyst from the reactant mixture, and allowing reactions to run
continuously. We modiﬁed alumina membranes with a single PAA/PEl/PdNP
layer and performed initial ﬂow-through hydrogenation studies with allyl alcohol.
In the hydrogenation experiments with membranes, solutions were
sparged with H2 before passing them through an alumina membrane coated with
a PAA/PEI/PdNP ﬁlm. To determine whether the sparging time was sufficient to
saturate the solution with H2, we performed allyl alcohol hydrogenation
experiments with all conditions remaining constant except for the initial H2
bubbling time (Figure 3.10). Upon increasing the initial H2 bubbling time from 2
minutes to 15 minutes, the concentration of allyl alcohol hydrogenated to 1-
propanol after passage through the membrane increased from 1.0 to 1.4 mM.
However, samples collected from the feed solution after bubbling with H2 for the
given amount of time showed similar increases in the concentration of 1-pr0pan0l
from 0 to 0.4 mM. Though the feed solution did not pass through the PdNP-
containing membrane, it was still in contact with the upper membrane surface
during the initial bubbling time, and some of the allyl alcohol was hydrogenated to

1-pr0panol via catalysis at the upper membrane surface. The plot in Figure

127

3.10b shows that the difference between the concentration of 1-propanol in the
permeate solution and in the feed solution is nearly the same for all initial
bubbling times. Thus, H2 saturation occurs fairly quickly, so subsequent

experiments utilized a 2 minute bubbling time before passing the feed solution

through a modiﬁed membrane.

 

 

 

 

73 2.00 1 OPermeate a) 73 2.00 1 b)
N A a A
E: 1.50 IFeed . . . €31.50 -
- 8 1.00 - O I 3 B 1 00 2 A
a E “a E ‘ ‘ ‘
g .2 0.50 - I 3'12 0.50 -
8 I I I I 8
0.00 . . . 0.00 . u .
0 5 10 15 0 5 10 15
Initial Sparging Time (min) Initial Sparging Time' (min)

Figure 3.10 a) Conc. of 1-propanol formed vs. initial H2 sparging time in
membrane-catalyzed hydrogenation of 5 mM allyl alcohol. The ﬂux through the
Pd nanoparticle-modiﬁed alumina membrane was 0.023 mL/cmz-sec. Circles
represent the concentration of 1-propanol in the permeate solution and squares
represent the concentration of 1-propan0l in the feed solution. The plot in b)

represents the difference in the 1-propanol concentration in the permeate and the
feed.

The results in Figure 3.10 show that nearly 20% of a 5 mM allyl alcohol
solution is hydrogenated to 1-propanol upon passing through a Pd-containing
alumina membrane. To determine if the amount hydrogenated is limited by
reaction kinetics or hydrogen solubility in water, we performed reactions with a
series of initial allyl alcohol concentrations. At initial concentrations < 1 mM,
nearly all of the allyl alcohol is hydrogenated to 1-pr0panol during passage
through the membrane because at these concentrations, allyl alcohol is the

limiting reactant. However, at allyl alcohol concentrations > 5 mM the amount

128

hydrogenated remains essentially constant because H2 is the limiting reagent.
The solubility of H2 in water at 1 atm of H2 is roughly 0.8 to 1.0 mM,27'28 and this
corresponds to the maximum concentration of allyl alcohol that could undergo

hydrogenation (Figure 3.1 1).

 

 

_ 11.
8 e 0
“A 1' .
g:
kg 0.8' .
3° 06-
to-0 '
‘35 04
3»? '
8 0.2f
OT I ﬂ l
0 5 10 15

Allyl Alcohol Starting Conc. (mM)

Figure 3.11 Concentration of 1-propanol formed in membrane-catalyzed
hydrogenation with several initial allyl alcohol concentrations. Solutions were
sparged with H2, and the ﬂux through the Pd nanoparticle-modiﬁed alumina
membrane was 0.023 mL/cmz-sec.

We performed allyl alcohol hydrogenation with multiple passes through the
membrane to see if additional conversion of allyl alcohol to 1-propanol could be
obtained. When collecting the permeate and simply passing it through the
membrane again, no additional conversion was achieved. As expected, to obtain
higher conversions, the solution needed to be resaturated with H2 before again
passing it through the membrane. In subsequent multiple pass experiments,
permeate solutions were collected, resparged with H2 for 2 min, and then passed
through the membrane again. The plot in Figure 3.12 shows the concentration of

1-propan0l after 5 passes of an allyl alcohol solution (initial concentration = 20

129

mM) through a PdNP-modiﬁed membrane. The concentration of 1-propanol that
formed in each pass through the PdNP-modiﬁed membrane remained essentially
constant at ~1.1 mM. Thus, essentially all of the H2 was consumed each time

the solution passed through the membrane.

 

 

- 6.0 a
2 e
N A 5.0 "
a: .
$31.0 -
z 3 3.0 - .
0 . O
g;- 2.0
o 1.0 1 0
o
0.0 I I j r I
0 1 2 3 4 5
#Passes

Figure 3.12 Cumulative concentration of 1-propanol formed in membrane-
catalyzed hydrogenation of a 20 mM allyl alcohol solution after multiple passes
through an alumina membrane modiﬁed with a PAA/PEI/PdNP ﬁlm. Solutions
were sparged with H2 for 2 min before each pass, and the ﬂux through the
membrane was 0.023 mUcmz-sec.

For hydrogen-sparged solutions containing 10 mM allyl alcohol, increasing
the ﬂux through the PdNP-modiﬁed membrane from 0.023 to 0.13 mUcmzsec did
not signiﬁcantly alter the amount of 1-propanol that was formed upon passage
through the membrane (Figure 3.13). This suggests that the reaction conversion
was still limited by the amount of H2 in the solution even at higher ﬂow rates.
The use of solvents with higher hydrogen solubility leads to increased conversion
of allyl alcohol to 1-propanol. For example, the solubility of H2 in ethanol is ~3
mM under atmospheric conditions.29 When performing reactions with a variety of

initial allyl alcohol concentrations in ethanol, nearly all of the allyl alcohol was

130

converted to 1-propanol or the isomer product at allyl alcohol concentrations < 2
mM (Figure 3.14). At allyl alcohol concentrations > 5 mM, the reaction again
began to be limited by hydrogen solubility as the concentration of 1-propanol in

the permeate was ~3.1 mM.

 

 

'5 1.2 -
:
§§ ‘-°' 0 e e e
“E, 0.8 -l
9'1:
.‘I a 0.6 -
0 E
d o 0.4 '
g..-
0 0.2 -
0.0 I I I
0 0.05 0.1 0.15

Flux (lecmzsec)

Figure 3.13 Concentration of 1-propanol formed during membrane-catalyzed
hydrogenation as a function of ﬂux through an alumina membrane modiﬁed with
a PAA/PEl/PdNP film. The initial allyl alcohol concentration was 10 mM, and
solutions were sparged with H2 for 2 min before a single pass through the
membrane.

 

 

'5 4.0

C

“A .

3230

5g 0 '

3820

°.E

ge1o«0

O
O

U 00 . . .
0 10 20 30

Initial Allyl Alcohol Conc. (mM)

Figure 3.14 Concentration of 1-propanol formed in membrane-catalyzed
hydrogenation with several initial allyl alcohol concentrations in ethanol.
Solutions were sparged with H2, and the ﬂux through the PAA/PEl/PdNP-
modiﬁed alumina membrane was 0.023 mL/cmz-sec.

131

3.3.4 Hydrogenation of Nitroaromatic Compounds with Pd-Containing
Membranes

To further demonstrate the use of Pd-modiﬁed membranes for
hydrogenation reactions, we utilized alumina membranes modiﬁed with a
PAA/PEIIPdNP ﬁlm for the hydrogenation of 4-nitrophenol to 4-aminophenol
(Scheme 3.2). As we discussed in chapter 2, the reduction of aromatic nitro
groups is important in synthetic chemistry, and the formation of aromatic amines
via hydrogenation is especially attractive because the reducing agent is readily
removed from the reaction mixture during the reaction. In initial studies of
membrane-based 4-nitrophen0l hydrogenation, aqueous solutions of 4-
nitrophenol were sparged with H2 and passed through an alumina membrane
coated with a PAA/PEl/PdNP ﬁlm. Similar to the previous experiments with allyl
alcohol, reactions were primarily limited by the amount of 4-nitrophenol when its
initial concentration was < 0.1 mM (Figure 3.15). At higher concentrations, the
conversion reached a plateau, presumably because of the limited amount of H2
in the solution. Roughly 7% of a 4 mM 4-nitrophenol solution was reduced upon
passing through a PAA/PEl/PdNP-modiﬁed alumina membrane. This
corresponds to a change in 4-nitrophenol concentration of ~ 0.3 mM, which

requires a H2 concentration of 0.9 mM in the solution.
N02 NH2

+3H2 ig—b +2H20

OH OH
Scheme 3.2 Pd-catalyzed hydrogenation of 4-nitrophenol to 4-aminophenol.

132

O
«19:91»?
NU!“
Ill
0
O
O

a:
.

 

Decrease in 4-
Nitrophenol Conc. (mM)
9 c

H

GUI-l
.—.l1

I I I I I

1 2 3 4 5
Initial Conc. (mM)

Figure 3.15 Decrease in 4-nitrophenol concentration in membrane-catalyzed
hydrogenation with several initial 4-nitrophenol concentrations. Solutions were
sparged with H2, and the ﬂux through the Pd nanoparticle-modiﬁed alumina
membrane was 0.023 mL/cmZ-sec.

G

We performed additional hydrogenation reactions with nitrobenzene and
4-nitrotoluene. For both compounds, ~7% reduction of the nitro compound
occurred upon passing a 4 mM aqueous solution through a Pd nanoparticle-
modiﬁed membrane at a ﬂux of 0.023 mL/cmz-sec. The fact that the reaction is
limited by H2 concentration should be especially useful for minimizing side
reactions. Since the hydrogen is used up in the reaction with the starting
material, there is less hydrogen available to react with the initial hydrogenation
products in further reactions. Thus, future studies of membrane-based
nitroaromatic compound hydrogenation could examine the chemoselective
hydrogenation of nitro groups in the presence of other reducible functional

groups.

133

3.4 Conclusions

This chapter showed that Pd nanoparticle size greatly affects selectivity in
the hydrogenation of monosubstituted over disubstituted double bonds with
polyelectrolyte/PdNP ﬁlms deposited on alumina powder. While a commercial
5% Pd on alumina catalyst showed no selectivity for the hydrogenation of allyl
alcohol over B-methallyl alcohol or crotyl alcohol, catalysts containing [PAA-
Pd(0)/PEI]3 ﬁlms showed selectivities as high as 420 for allyl alcohol over [3-
methallyl alcohol and 38 for allyl alcohol over crotyl alcohol. Moreover, as the
nanoparticle size increases, selectivities decrease by a factor of 7. In studies
utilizing PAA/[PEI/PdNij ﬁlms with 1, 2, and 3 PEI/PdNP bilayers, the selectivity
was not signiﬁcantly affected by the number of layers in the ﬁlm. This suggests
that the changes in selectivity observed with the [PAA-Pd(0)/PEI]3 catalysts are
due primarily to differences in particle size and not to diffusion through the
polyelectrolyte multilayer ﬁlms. Differences in selectivities might occur because
monosubstituted double bonds bind more easily to the active sites of smaller
nanoparticles than multisubstituted double bonds do. Thus, the selectivities are
lower with larger particles because it is easier for the multisubstituted double
bonds to adsorb and react at the less hindered active sites.

Additionally, we demonstrated that alumina membranes modiﬁed with Pd-
nanoparticle-containing ﬁlms effectively reduce allyl alcohol to 1-pr0panol in ﬂow-
through hydrogenation reactions. However, these reactions gave relatively low
conversions due to the low solubility of H2 in water. When using ethanol as the

solvent, conversion increased by a factor of three as a result of the increased

134

hydrogen solubility. In each case, the reaction conversion was primarily limited
by H2 solubility once the initial allyl alcohol concentration singiﬁcantly exceeded
the concentration of H2 in the solution. Similar results were obtained in the
hydrogenation of several nitroaromatic compounds. Future hydrogenation
studies with Pd-containing membranes should focus on methods to overcome

the H2 solubility limitations.

135

 

3.5 References

(1) Nishimura, S. Handbook of Heterogeneous Catalytic Hydrogenation for
Organic Synthesis; John Wiley & Sons, Inc., 2001.

(2) Kidambi, S.; Bruening, M. L. Chem. Mater. 2005, 17, 301-307.

(3) Kidambi, S.; Dai, J.; Li, J.; Bruening, M. L. J. Am. Chem. Soc. 2004,
126, 2658-2659.

(4) Bhattacharjee, S.; Bruening, M. L. Langmuir 2008, 24, 2916-2920.

(5) Doyle, A. M.; Shaikhutdinov, S. K.; Freund, H.-J. Angew. Chem. Int.
Ed. 2005, 44, 629-631.

(6) Doyle, A. M.; Shaikhutdinov, S. K.; Jackson, S. D.; Freund, H.-J.
Angew. Chem. Int. Ed. 2003, 42, 5240-5243. 1

 

(7) Wilson, O. M.; Knecht, M. R.; Garcia-Martinez, J. C.; Crooks, R. M.
J. Am. Chem. Soc. 2006, 128, 4510-4511.

(8) Shaikhutdinov, 8.; Heemeier, M.; Baumer, M.; Lear, T.; Lennon, 0.;
Oldman, R. J.; Jackson, S. D.; Freund, H. J. J. Catal. 2001, 200, 330-339.

(9) Xu, 8.; Liew, K. Y.; Li, J. J. Am. Oil Chem. SOC. 2007, 84, 117-122.

(10) Ruta, M; Semagina, N.; Kiwi-Minsker, L. J. Phys. Chem. C 2008,
112, 13635-13641.

(11) Semagina, N.; Renken, A.; Kiwi-Minsker, L. J. Phys. Chem. C 2007,
111, 13933-13937.

(12) Silvestre-Albero, J.; Rupprechter, G.; Freund, H.-J. J. Catal. 2006,
240, 58-65.

(13) Vasylyev, M. V.; Maayan, G.; Hovav, Y.; Haimov, A.; Neumann, R.
Org. Lett. 2006, 8, 5445-5448.

(14) Niu, Y.; Yeung, L. K.; Crooks, R. M. J. Am. Chem. Soc. 2001, 123,
6840-6846.

(15) Oh, S.-K.; Niu, Y.; Crooks, R. M. Langmuir 2005, 21, 10209-10213.

136

(16) Dokoutchaev, A.; James, J. T.; Koene, S. C.; Pathak, S.; Prakash,
G. K. S.; Thompson, M. E. Chem. Mater. 1999, 11 , 2389-2399.

(17) Joly, 8.; Kane, R.; Radzilowski, L.; Wang, T.; Wu, A.; Cohen, R. E.;
Thomas, E. L.; Rubner, M. F. Langmuir 2000, 16,1354-1359.

(18) Wang, T. C.; Rubner, M. F.; Cohen, R. E. Langmuir 2002, 18, 3370-
3375.

(19) Bhattacharjee, S. PhD Dissertation, Michigan State University, 2008.
(20) Teranishi, T.; Miyake, M. Chem. Mater. 1998, 10, 594-600.

(21) Bhattacharjee, S.; Dotzauer, D. M.; Bruening, M. L. J. Am. Chem.
Soc. 2009, 131, 3601-3610.

(22) Liu, X.; Bruening, M. L. Chem. Mater. 2003, 16, 351-357.
(23) Bond, G. C. Surf. Sci. 1985, 156, 966-981.

(24) Freund, H. J.; Baumer, M.; Libuda, J.; Risse, T.; Rupprechter, G.;
Shaikhutdinov, S. J. Catal. 2003, 216, 223-235.

(25) Li, Y.; Boone, E.; El-Sayed, M. A. Langmuir 2002, 18, 4921-4925.

(26) Semagina, N.; Renken, A.; Laub, D.; Kiwi-Minsker, L. J. Catal. 2007,
246, 308-314.

(27) Jauregui-Haza, U. J.; PardilIo-Fontdevila, E. J.; Wilhelm, A. M.;
Delmas, H. Lat. Am. Appl. Res. 2004, 34, 71-74.

(28) Baranenko, V. l.; Kirov, V. S. At. Energ. 1989, 66, 24-8.

(29) Purwanto, P.; Deshpande, R. M.; Chaudhari, R. V.; Delmas, H. J.
Chem. Eng. Data 1996, 41, 1414-1417.

137

Some of the text and ﬁgures contained in chapter four appeared in the following
publication.

Dotzauer, D. M.; Abusaloua, A.; Miachon, S.; Dalmon, J.-A.; Bruening, M. L.
Appl. Catal., B, In Press, Corrected Proof.

138

Chapter Four: Catalytic Wet Air Oxidation with Pt-Containing Membranes

4.1. Introduction

The previous two chapters described the beneﬁts of using catalytic
membranes for chemical synthesis reactions and reactions in which the control
over product formation is important. However, the membranes described in
these chapters suffer from difﬁculty in scaling up the system for higher
throughput. This chapter focuses on the use of catalytic membranes for the
environmental application of catalytic wet air oxidation. Wet air oxidation is an
important wastewater treatment process in which hazardous organic pollutants
react with oxygen to give more benign compounds, ideally H20 and CO2.“3 This
technique is attractive for processing wastewater that is too dilute to be treated
by incineration‘“5 and too concentrated to be treated by biological methods.‘5‘9
Traditional wet air oxidation of organic and inorganic substrates often requires
high temperature and pressure (150-350 °C, 20-200 bar air),1°'11 but the use of
catalysts such as Pt, Ru, or other precious metals immobilized on inorganic
powders allows much milder reaction conditions (room temperature, 1 to 5 bar
air).‘2'14 However, implementation of catalytic wet air oxidation in conventional
stirred tank reactors requires a catalyst recovery step, and reaction rates are
often limited by diffusion of oxygen and/or the liquid phase compounds to the
catalyst surface.

Porous membranes are an attractive alternative to powders as catalyst
supports because the high internal surface area of the membrane affords a high

loading of the active catalyst material and there is no need to separate the

139

catalyst from the reaction mixture. Thus, reactions can be run continuously.
Furthermore, catalytic membranes operated as gas/liquid contactors enhance the
accessibility of the reactants to the metal catalyst.”16 As we showed in chapter
one, the two most common membrane conﬁgurations for gas/liquid reactions are
ﬂow-through and interfacial contactors. We discussed the use of ﬂow-through
contactors for hydrogenation reactions with Pd-containing membranes in chapter
three. Flow-through contactors, where all reactants ﬂow through the membrane
in a single solution, are advantageous because when the membrane pores are
sufﬁciently small, reactions will not be limited by the rate of mass transport to the

L17

catalys Also, by controlling the ﬂow rate and, hence, the residence time of a

substrate within the membrane, side reactions may be minimized to give high

selectivity for a particular product.""22

Unfortunately, as we demonstrated in
chapter three, the low solubility of the gaseous reactant in the liquid phase often
limits the extent of reaction in gas/liquid reactions with ﬂow-through contactors.17
A similar problem occurs in ﬁxed-bed reactors.

In interfacial contactors, the walls of a catalytic membrane serve as the
interface between the gas and liquid phases (Figure 4.1) to allow rapid transport
of gas to the solid-liquid—catalyst interface and provide a high catalytic activity.23
Recent work by Pera-Titus et al. also suggests that the enhanced catalytic
activity in interfacial contactors may be due to increased gas solubility in the

conﬁned pores of the membrane.24 However, if pore sizes are larger than 10 nm,

this effect is usually not observed.

140

Air or 02 H20 + 002

    

Catalytic
layer

Gas Phase Liquid Phase

Figure 4.1. Schematic diagram showing the interfacial contactor conﬁguration in
a tubular membrane and its application to wet air oxidation.

Recently, the “Watercatox” process was developed as a method to oxidize
organic content in industrial wastewater using catalytic membrane reactors while
allowing much lower temperatures and pressures than traditional wet air
oxidation or incineration. This is a joint project between groups in Lyon, France
and Oslo, NonIvay in which catalytically active tubular ceramic membranes are
used for wet air oxidation of model efﬂuents in the laboratory and then scaled-up
to pilot-size for oxidation of industrial efﬂuents.253° They showed that the
interfacial contactor conﬁguration leads to increased activity when compared to a
conventional stirred tank reactor,29 and that the high activity stems from the
ability to control the location of the gas/liquid interface within the membrane
(Figure 4.1).

The location of the gas/liquid interface is a crucial factor because mass
transfer of pollutants and O2 to the reaction zone can greatly affect the reaction
rate. Several studies by Vospernik et al. speciﬁcally examined the effect of

pressure difference on catalytic activity in wet air oxidation of formic acid.“33

141

These studies showed that the concentration of formic acid is depleted as a
function of distance travelled radially into the membrane. Likewise, the
concentration of O2 in the liquid phase decreases rapidly to zero as a function of
radial distance from the gas/liquid interface. As a result, the thickness of the
reaction zone is only on the order of 5 to 20 pm depending on reaction
conditions.33 They also found that as the transmembrane pressure increases,
the location of the gas/liquid interface moves from the support layer with large
pore sizes to the intermediate and skin layers with smaller pore sizes, and the
activity of the membrane for formic acid oxidation increases. This increased
activity is likely due to both the higher catalyst loading in the intermediate and
skin layers (higher surface area for catalyst deposition) and the shorter diffusion
distance of formic acid to the reaction zone.32

In the interfacial contactors thus far employed for wet air oxidation,
catalytic noble metal particles were formed in the membrane by
evaporation/recrystallization/reduction and anionic impregnation/reduction
methods,"“"35 but other methods of catalyst deposition may have better control
over catalyst deposition and provide even higher catalytic activities. Among the
many methods for incorporating precious metal catalysts in porous materials?“1
layer-by-layer (LbL) adsorption of polyelectrolyte/metal nanoparticle ﬁlms is
attractive because it offers ﬁne control over nanoparticle size and composition
and can be applied to a variety of membrane materials.42 Ideally, the catalyst

should be highly concentrated in the membrane region where the gas/liquid

interface occurs, and the interface should be maintained close to the inner layer

142

of the membrane to limit diffusion distances for the liquid phase reactants.
Because LbL nanoparticle adsorption within membranes is very rapid, the depth
to which deposited nanoparticles penetrate the membrane can be readily
controlled by limiting the amount of nanoparticle-containing solution passed
through the membrane. Hence it is a simple matter to localize catalyst deposition
in the inner layer of a tubular membrane.

This chapter discusses the wet air oxidation of formic acid, acetic acid,
and phenol using tubular ceramic membranes modiﬁed with Pt nanoparticles by
LbL deposition. Many of the results described in this chapter were obtained in
collaboration with Miachon and coworkers at the Institute of Researches on
Catalysis and Environment in Lyon, France (IRCELYON). Results from these
membranes are compared with results from membranes modiﬁed by
conventional impregnation techniques used in the Watercatox project. The LbL
modiﬁed membranes have especially high speciﬁc activities in the oxidation of

these model compounds.

4.2 Experimental Methods
4.2.1 Materials

Anodisc aluminum oxide membranes (25 mm disks with 0.1 pm diameter
pore sizes, Whatman), tubular ceramic membranes (Pall Exekia) and 100 mesh
aluminum oxide (Aldrich) were modiﬁed with catalytic nanoparticles using the LbL
technique. The tubular membranes (25 cm long, 7 mm inner diameter, 10 mm

outer diameter) consisted of three layers: a TiO2-covered alumina support layer

143

with 12 pm-diameter pores, a TiO2-covered alumina intermediate layer with 0.8
pm-diameter pores, and a ZrO2 inner layer with 50 nm-diameter pores.
Hexachloroplatinic acid, sodium citrate, sodium borohydride, poly(allylamine
hydrochloride) (PAH, Mw = 17,000), and poly(ethylenimine) (PEI, branched, Mw
= 25,000) were obtained from Aldrich. Poly(acrylic acid) (PAA, Mw = 5,000) was

obtained from Polysciences Inc.

4.2.2 Modification of Aluminum Oxide Powder

LbL modiﬁcation of the alumina powder involved: 1) stirring 2.5 g of
alumina powder in 20 mL of PAA solution (0.02 M PAA, 0.5 M NaCl, pH adjusted
to 4.5 with 1.0 M NaOH) for 10 minutes; 2) stirring the PAA-modiﬁed powder in
20 mL of PAH solution (0.02 M PAH, 0.5 M NaCI, pH adjusted to 5.0 with 0.1 M
HCI) for 10 minutes; and 3) stirring the PAA/PAH-coated powder in 20 mL of a Pt
nanoparticle solution for 10 minutes. (Polymer concentrations are given with
respect to the repeating unit.) After each of the above steps, the liquid was
decanted, and the alumina powder was washed three times with 20 mL of
deionized water. Pt nanoparticles were prepared with thiol or citrate stabilizing
agents. To synthesize the thiol-stabilized particles, under vigorous stirring 5 mL
of 0.0676 M NaBH4 was added to an aqueous solution containing 10 mL of 3.38
mM H2PtCle-6H2O and 1 mL of 0.0237 M mercaptosuccinic acid (MSA).43 The
resulting MSA-stabilized Pt nanoparticle solution was diluted by a factor of 4 prior
to use in LbL adsorption on alumina powder. To prepare the citrate-stabilized

particles, 30 mL of a heated 1 wt% aqueous sodium citrate solution was added to

144

255 mL of a reﬂuxing solution of 0.3 mM H2PtCl5-6H2O under vigorous stirring.
The solution was reﬂuxed for 4 hours to allow completion of the reaction.44 The
resulting Pt nanoparticle solution was used directly for deposition on alumina
powder. Alumina powder was also modiﬁed by the anionic impregnation/
reduction technique by stirring 2.5 g of the powder in a 0.1 g/L solution of
H2PtC|6~6H2O for 2 hours, washing three times with 20 mL of deionized water,
and reducing the Pt ions to nanoparticles by adding 20 mL of 0.1 M NaBH4 and

stirring for 10 minutes.

4.2.3 Modification of Disk-Shaped Alumina Membranes

LbL modiﬁcation of disk-shaped alumina membranes was described in
chapter two for membranes containing polyelectrolyte/AuNP ﬁlms. Brieﬂy,
deposition of each layer involved passing the polyelectrolyte or nanoparticle
solution through the membrane using a peristaltic pump located at the permeate
side of the membrane. During membrane modiﬁcation, the films were formed by
ﬂowing a PAA solution (0.02 M PAA, 0.5 M NaCl, pH adjusted to 4.5 with 1.0 M
NaOH), water, PAH solution (0.02 M PAH, 0.5 M NaCI, pH adjusted to 5.0 with
0.1 M HCI), water, citrate- or thiol-stabilized Pt-nanoparticle solution (as-prepared
nanoparticles diluted by a factor of 10), and water through the membranes.

Several alumina membranes were also modiﬁed by the LbL method with
Au, Pd, and Ru nanoparticles to compare with the activity of Pt nanoparticles.
The Au nanoparticles were citrate-stabilized with an expected diameter of ~2.6

rim and prepared according to a literature procedure45 in which 1 mL of 1%

145

 

HAuCl4 was added to 90 mL of H20 at room temperature and stirred for 1 min,
followed by addition of 2 mL of 38.8 mM sodium citrate and stirring for 1 min.
Reduction occurred by adding 1 mL of 0.075% NaBH4 in 38.8 mM sodium citrate
and stirring the colloidal solution for an additional 5 min. The Pd nanoparticles
were also citrate-stabilized and had an expected diameter of ~7 nm.46 In the
synthesis of these particles, 50 mL of an aqueous 0.393 mM solution of K2PdCl6
was heated to reﬂux, a heated 1 % solution of sodium citrate (6mL) was added,
and the reﬂuxing was allowed to continue for 4 h. The Ru nanoparticles were
prepared by adding 0.5 mL of a 0.1 M NaBH4 solution dropwise to 10 mL of 2
mM RuCl3 under vigorous stirring.47 The Ru nanoparticles had an expected
diameter of 1.8 nm and were positively charged due to the absence of any
additional stabilizing agent.47 Membrane modiﬁcation with the Ru nanoparticles
involved deposition of 1.5 PAA/PAH bilayers prior to nanoparticle deposition
because of the positive charge on these particles. Citrate-stabilized Au and Pd
nanoparticles were adsorbed on 1-bilayer PAA/PAH films in the same manner as
the Pt nanoparticles.

Anionic impregnation of alumina membranes was carried out by ﬂowing a
0.1 g/L solution of Pt (in the form of H2PtCl6) through a membrane at a ﬂow rate
of 0.0005 mL/cmz-sec for 4 hours followed by rinsing with water at a ﬂow rate of
0.01 mL/cmz-sec for 1 hour. Reduction of Pt was carried out by ﬂowing a 0.1 M
solution of NaBH4 through the membrane at 0.0005 mUcmz-sec for 30 min and

again passing water at 0.01 mL/cmz-sec for 10 min.

146

 

4.2.4 Modiﬁcation of Tubular Ceramic Membranes

Tubular ceramic membranes were modified by several variations of the
LbL method as well as by the evaporation/crystalIization/reduction and anionic
impregnation/reduction techniques. (See below.) During LbL modification,
polyelectrolyte and metal nanoparticle solutions were deposited by flowing from

the inside of the tubular membrane to the outside as shown in Figure 4.2.

   

Membrane

Figure 4.2 Schematic diagram of the apparatus used for depositing
polyelectrolyte/metal nanoparticle ﬁlms in tubular ceramic membranes. The
pressurized solution ﬂows though the membrane pores in an inside—out
configuration.

4.2.4.1 Method 1 - LbL Adsorption with Ex-Situ Nanoparticle Formation
[PAA/PAHIPtNPh

Modiﬁcation of tubular membranes with PAA/PAH/PtNP ﬁlms was similar
to modiﬁcation of disk-shaped membranes except the deposition solutions were
diluted more to limit pore blocking that could occur due to smaller pore sizes.
The modiﬁcation procedure included sequential ﬂow through the membrane (as
shown in Figure 4.2) of 250 mL of PAA solution (0.002 M PAA, 0.1 M NaCI, pH

adjusted to 4.5 with 1M NaOH), 500 mL of water, 250 mL of PAH solution (0.002

147

 

M PAH, 0.1 M NaCl, pH adjusted to 5 with 0.1 M HCI), 500 mL of water, and
1000 mL of a citrate-stabilized PtNP solution prepared by diluting 25 mL of the
as-prepared nanoparticle solution with 975 mL of water. The ﬂow rate of the
solutions through a given membrane was between 20 and 25 mL/min and was
maintained by applying a pressure between 0.2 and 0.5 bar. Figure 4.3 shows a
general scheme of this procedure. Similarly, tubular membranes were also
modiﬁed with PAA/PAH/AuNP ﬁlms. The deposition procedure was the same as
that described for Pt nanoparticles except 25 mL of the as-prepared Au

nanoparticle solution45 was used in place of 25 mL of the as-prepared Pt

nanoparticle solution.

 

 

   

PAA
deposition
-.* g; . c- . _
Bare support
~
, 2:le Pt-NP
‘5 2:10 deposition

 

 

Figure 4.3 Schematic diagram of modiﬁcation of membrane surfaces (including
pores) using LbL deposition of PAA/PAH/PtNP ﬁlms.

4.2.4.2 Method 2 - LbL with In-Situ Nanoparticle Formation [PAA/PEl-Pt(0)]1

In a slight modiﬁcation to previous procedures for modifying alumina
powder with polyelectrolyte/PdNP ﬁlms,48'49 method 2 incorporated a PEI-Pt(ll)
complex in the deposition procedure rather than preformed Pt nanoparticles.

Brieﬂy modiﬁcation included sequential ﬂow through the membrane of 250 mL of

148

 

PAA solution (0.002 M PAA, 0.1 M NaCl, pH adjusted to 4.5 with 1M NaOH), 500
mL of water, and 250 mL of PEI solution that contained Pt(ll) (0.002 M PEI,
0.0004 M K2PtCI4, pH adjusted to 9 with 1.0 M HCI), and 500 mL H2O. To form
Pt nanoparticles, 250 mL of 0.1 M NaBH4 solution was passed through the
membrane to reduce the Pt ions to PtNPs (Figure 4.4), and the membrane was

then rinsed by the passage of 500 mL of water.

 

 

 

PAA
deposition
Bare support
NaB H4
reduction

   

Figure 4.4 Schematic diagram of modiﬁcation of membrane surfaces (including
pores) using LbL deposition of PAA/PEl-Pt(ll) ﬁlms followed by reduction with
NaBH4.

4.2.4.3 Method 3 - LbL with ln-Situ Nanoparticle Formation [Pt(0)IPEl]2
Similar to a previous method for modifying alumina powder,50 the
membrane was ﬁrst immersed in a solution of hexachloroplatinic acid (0.1 g Pt/L)
for 20 hours. After mounting the membrane in the apparatus displayed in ﬁgure
4.2, 500 mL of water was passed through the membrane pores to remove excess
Pt solution. PEI was deposited by ﬂowing 250 mL of solution (0.002 M PEI, 0.1
M NaCI, pH adjusted to 9 with 1 M HCI) through the membrane, which was

subsequently rinsed by passage of 500 mL of water. A second PtClsz'lPEl

149

bilayer was deposited similarly before reducing the Pt with NaBH4 in the same

manner as in method 2 (Figure 4.5).

 

 

 

 

    
 

 
 

 

E , J‘ . ‘ i 1.. -._. -.. ' .- ' .,.;'-'._'__-- ' ﬂ "” V-f. 7.. '=-- .-" "I’m
Bare support 1 2 Repeat
1 ,2
W NaBH4 0.2!, 0-0 :0 0‘ 5!. 0
reduction “2.2.1.343;

 

Figure 4.5 Schematic diagram of modiﬁcation of membrane surfaces (including
pores) using LbL deposition of [Pt(lV)IPEI]2 ﬁlms followed by reduction with
NaBH4.

4.2.4.4 Method 4 - Evaporation/Recrystallizatioaneduction
The technique of evaporation/recrystallization/reduction was similar to a

previously reported procedure in the Watercatox project.35

Brieﬂy, the membrane
was immersed in a 0.1g/L chloroplatinic acid solution for 4 hours, removed from
the solution, and allowed to dry at room temperature. Evaporation of the solvent
led to concentration of the Pt precursor on the surface of the membrane with
more of the Pt located in the inner layer. Reduction of the Pt was performed by
placing the membrane under ﬂowing H2 at 200 °C for 12 hours (Figure 4.6).

H2PtCl6
4 hr

   
    

nitrogen stream

1 Evaporation under

 

200°C

Figure 4.6 Schematic diagram of modiﬁcation of membrane surfaces (including
pores) using evaporation/recrystallization of H2PtCI6 followed by reduction with
H2 at 200 °C.

150

4.2.4.5 Method 5 - Anionic lmpregnation/Reduction

The anionic impregnation technique was also performed in a manner
similar to previous Watercatox research.35 In this case, the support was
immersed in a 0.1 g/L chloroplatinic acid solution for 4 hours and then rinsed by
ﬂowing a 0.1 N nitric acid solution through the membrane pores for 1 hour to
remove any unbound Pt species from the membrane. After rinsing with water
and then drying under ﬂowing N2 at 100 °C for 1 hour, the Pt was reduced under

ﬂowing H2 at 200 °C for 12 hours (Figure 4.7).

 

   

 

 

Bare support Rinsed with 0.1 N
nitric acid
H Hz 2% O ‘LQ
t .

 

200 °C

Figure 4.7 Schematic diagram of modiﬁcation of membrane surfaces (including
pores) using anionic impregnation with H2PtCI6 followed by reduction with H2 at
200 °C.

4.2.5 Characterization

Nanoparticles were characterized by transmission electron microscopy
(JEM-2200FS microscope) to determine their approximate size and shape. TEM
Samples were prepared by diluting the as-prepared PtNP solutions by a factor of
5 with water and placing a 2 UL drop of this mixture on a carbon-coated copper
grid. TEM was also used to demonstrate the deposition of nanoparticle-

containing ﬁlms in disk-shaped porous alumina membranes. Prior to imaging,

151

the membrane was ground into a poWder with a mortar and pestle and dispersed
in water using a vortex mixer. A 2 uL drop of the resulting solution was then
placed onto a carbon-coated copper grid and dried before analysis.

The Pt content of the disk-shaped membranes and alumina powder was
determined by completely leaching the metal with aqua regia (3 parts HCI, 1 part
HNO3) and analyzing the leachate by ﬂame atomic absorption spectroscopy
(FAAS). For tubular membranes, the amount of deposited Pt was estimated by
chemical analysis of the deposition solutions before and after passing them
through the membrane. These values were veriﬁed by grinding the membranes
into powder with a mortar and pestle, dissolving the Pt in aqua regia, and

analyzing the solution by FAAS.

4.2.4 Catalytic Reactions

Formic acid, acetic acid, and phenol were employed as substrates for
oxidation reactions. Initial experiments were performed with formic acid and
powder catalysts to see if the nanoparticle stabilizer or deposition technique
affected the nanoparticle activity. In these reactions, oxygen was continuously
bubbled into 50 mL of a vigorously stirred solution containing catalyst and 5 g/L
of formic acid. Samples of the reaction mixture were collected after several time
intervals, ﬁltered to remove the catalyst, and analyzed by ion chromatography
(Dionex LC20, lonpac AS16 column) to determine the amount of formic acid that
remained in solution. Similar experiments were also performed with powders

prepared by grinding tubular membranes.

152

 

—>
liquid out b)

       

a)

Membrane

    

 

 

P

Figure 4.8 Schematic diagram of the wet air oxidation apparatuses that employ
a) disk-shaped membranes as ﬂow through contactors and b) tubular catalytic
membranes as interfacial contactors.

Pen'staltic Pump

Flow-through reactions performed with disk-shaped membranes were
carried out by bubbling 02 into a formic acid solution for 2 minutes and then
passing that solution during continuous bubbling through a nanoparticle-modiﬁed
membrane at a given ﬂux (Figure 4.8a). Samples of the membrane permeate
were analyzed by IC to determine the extent of formic acid oxidation. For some
experiments, the permeate solution was collected, resaturated with O2, and
passed through the nanoparticle—containing membrane multiple times to achieve
higher conversion.

For interfacial contactor reactions, the modiﬁed tubular membranes were
mounted in a gas tight module that allows the flow of liquid through the lumen of
the tube and countercurrent gas ﬂow on the shell side of the tube (Figure 4.8b).
The liquid ﬂow rate was typically between 7 and 10 mL/min, and the gas

overpressure was set to values between 0.2 and 4 bar. The gas ﬂow rate was

153

maintained at 50 mL/min with a mass ﬂow controller. The system initially
operated under N2 until the desired pressure was reached and remained
constant for 10 min. Then the gas was switched from N2 to air to begin the
oxidation reactions, which were carried out at 20, 60, or 80 °C by controlling the
temperature of the feed solution. The starting concentrations of formic acid,
acetic acid, and phenol were 5 g/L (0.108 M), 3.25 g/L (0.054 M), and 1.7 g/L
(0.018 M), respectively, which corresponded to carbon contents of approximately
1.3 g/L in each case. For these reactions, the conversion of each substrate was
monitored using total organic carbon (Shimadzu TOC 5050A) and/or HPLC
(Varian Prostar with UV-Vis detection) analysis. The uncertainty in the calculated
speciﬁc activities was < 10% for formic acid oxidation experiments and < 20% for

acetic acid and phenol oxidation experiments.

4.3 Results and Discussion

4.3.1 Characterization of Polyelectrolyte/PtNP Films

    

 

2 2.5 3 3.5 4 4.5
Particle Size nm

5nm ‘ .-'. -"- .. 5nm
— ~—

  

Figure 4.9 TEM images of a) mercaptosuccinic acid-stabilized Pt nanoparticles
and b) citrate-stabilized Pt nanoparticles on carbon-coated copper grids. The
inset in each image shows the nanoparticle size distribution determined from
several high resolution TEM images.

154

 

 

Figure 4.10 Low (a) and high (b) magnification TEM images of Pt nanoparticles
immobilized in the pores of an alumina membrane modiﬁed with a
PAA/PAH/PtNP ﬁlm. The Pt nanoparticles contained in the ﬁlm are citrate-
stabilized.

TEM images of the nanoparticles and membrane samples were collected
to determine the size and shape of the Pt nanoparticles and to see if these
particles were effectively deposited in the membranes. Figure 4.9 shows TEM
images and size distributions of MSA- and citrate-stabilized Pt nanoparticles.
The average particle diameters were 2.6 1: 0.3 and 3.2 1' 0.5 nm, respectively.
Figure 4.10 presents TEM images of the citrate-stabilized Pt nanoparticles
immobilized in a disk-shaped alumina membrane. These images show that the
LbL method yields a high density of nanoparticles within the pores of the
membrane and that there is minimal nanoparticle aggregation, which should lead
to accessible nanoparticles with a high catalytic surface area.

The amount of Pt in each of the types of tubular membranes was
determined by FAAS of precursor Pt solutions before and after passing them
through the membrane. These values were later confirmed by dissolving the Pt

in aqua regia and analyzing these solutions by FAAS. The two values were in

155

good agreement for each membrane type except those prepared by method 5
(Table 4.1). The difference in the two values for method 5 is likely due to some
Pt being washed away in the nitric acid rinsing step during the membrane
modiﬁcation. This Pt loss was not accounted for in the initial mass balance. The
Pt content was 200-1000 mg of Pt per m2 (1-5 mg of Pt per membrane), based
on the area calculated from the inner tube diameter. The relatively similar Pt
loading among the different membranes facilitates the comparison of the catalytic

activities for the different deposition methods.

Table 4.1 Pt contents in tubular membranes

 

 

 

Modiﬁcation Pt loading (mg Pt I m2)a
Method Mass Balance“ Membrane Powder“
1 220 :I: 40 220 1 20
2 220 1 20 200 1 20
3 910 1 90 830 1 80
4 470 1 60 480 1 40
5 690 1 90 400 1 40

 

8based on an internal membrane surface area of 0.00506 m2 calculated from an
internal tube diameter of 7 mm and an active length of 230 mm.

bdetermined by chemical analysis of precursor solutions before and after
deposition.

cdetermined by chemical analysis of solutions prepared by removing Pt from
ground membrane samples with aqua regia.

4.3.2 Wet Air Oxidation Catalyzed by Pt Nanoparticles on Alumina Powder
To examine the effect of the stabilizing agent on the activity of Pt

nanoparticles, alumina powder was modiﬁed using LbL deposition with either

citrate- or MSA-stabilized Pt nanoparticles to prepare heterogeneous catalysts

for conventional slurry reactions. Formic acid was chosen as a model compound

156

 

 

for this process because many organic molecules degrade to low molecular
weight compounds and eventually to formic acid, which is then oxidized to CO2

and H20 in the ﬁnal step of complete wet air oxidation (scheme 4.1).51

Catalyst
———>

HCOOH + V202 CO2 + H20

Scheme 4.1 Catalytic wet air oxidation of formic acid.

In wet air oxidation of formic acid with these materials, the catalyst
modiﬁed with citrate-stabilized Pt nanoparticles exhibited an activity of 1.3 1 0.2
mmol/(s*gPt) whereas the catalyst containing MSA-stabilized Pt nanoparticles
had an activity of 0.7 1 0.1 mmol/(s*gPt). The average nanoparticle size is
similar for both types of particles, so differences in surface area should not
account for the difference in activity. In fact, MSA-stabilized particles showed
slightly smaller diameters (higher surface area per mass) in TEM images. The
most likely explanation for the difference between the two types of nanoparticles
is that the thiol stabilizers bind more tightly than citrate to the surface of the
nanoparticle, and this stronger binding limits the number of active sites for
catalysis. Previous studies of catalysis by thiol-stabilized metal nanoparticles
also showed low reaction rates.52'53

Alumina powder modiﬁed by impregnation of PtClez' and subsequent
reduction of Pt(lV) to Pt nanoparticles had an activity of 1.0 1 0.1 mmol/s*gPt,
which is again lower than that of the catalyst containing citrate-stabilized

nanoparticles. Thus, the LbL deposition with citrate-stabilized nanoparticles

provides catalysts with comparable or better activities than traditional methods of

157

 

 

catalyst preparation. Furthermore, these data suggest that the polyelectrolyte
multilayer does not inhibit the activity of the Pt nanoparticles. To maximize
catalytic activity, the following studies with tubular membranes modiﬁed by

method 1 utilized citrate-stabilized PtNPs rather than MSA-stabilized PtNPs.

4.3.3 Oxidation of Fannie Acid with Disk-shaped Membranes

In initial studies of membrane-based oxidation, solutions sparged with 02 were
passed through an alumina membrane coated with a PAA/PAH/PtNP ﬁlm. At
initial formic acid concentrations < 2 mM, nearly all of the formic acid was
oxidized to CO2 and water during passage through the membrane because at
these concentrations, formic acid was the limiting reactant. However, at formic
acid concentrations > 10 mM the amount of oxidation was essentially constant

because 02 was the limiting reagent. The solubility of O2 in water at 1 atm of O2

 

3.0
U
E, I I I
0 A
g E 2.0
1% -
C 310
._ O .
33°
C
.2
O 0.0 1 . .
0 20 40 60

Initial Formic Acid Conc. (mM)

Figure 4.11 Change in formic acid concentration in membrane-catalyzed
oxidation with several initial formic acid concentrations. Solutions were sparged
with O2, and the ﬂux through the porous alumina membrane modiﬁed with citrate-
stabilized Pt nanoparticles (method 1) was 0.023 mL/cmz-sec.

158

 

 

 

is roughly 1.25 mM,54 which would correspond to a concentration of 2.5 mM
formic acid that could be oxidized. This is similar to the maximum change in
formic acid concentration shown in Figure 4.11.

Alumina membranes modified with Pt, Au, Pd, and Ru nanoparticles were
all tested for the oxidation of formic acid using the ﬂow-through mode with
multiple passes through the membrane. In all experiments, the initial formic acid
concentration was 21.7 mM (1000 mg/L). Solutions were saturated with oxygen
by bubbling 02 into the solution for 2 min and then passing the solution through
the membrane at a ﬂow rate of 0.023 mL/cmZ-sec. The plot in Figure 4.12 shows
the change in formic acid concentration through 6 passes for each of the

nanoparticle-modiﬁed membranes.

 

 

 

:2 15-
2 :
.2A “‘4’
EE10‘ {’__
0v "'o‘
:2. 'A‘ a’.‘
I; O 5.l “ '. I
Go ’1'.
5 2 5"
5 0 " ~ — —— '
0 1 2 3 4 5 6
#ofpasses
+Pd --A--Au —I— Ru

—0— Pt-Citrate —A— Pt-MSA - -D- - Pt-Anionic

Figure 4.12 Cumulative change in formic acid concentration in membrane-
catalyzed oxidation after multiple passes through alumina membranes modiﬁed
with either Pt, Au, Pd, or Ru nanoparticles. Solutions were sparged with O2, and
the ﬂux through the metal nanoparticle-containing membranes was 0.023
mUcmz-sec.

159

The membranes containing Pt nanoparticles all exhibited changes in
formic acid concentration of 2 to 2.5 mM with each pass through the membrane,
which is again consistent with the expected solubility of O2 in aqueous solutions.
This suggests that essentially all of the O2 was consumed when passing the
solution through the Pt-containing membranes. Membranes containing metals
other than Pt did not effectively oxidize formic acid. In fact, the average change
in formic acid concentration for each pass was 0.20, 0.11, and 0.13 mM for Au,
Pd, and Ru nanoparticle-modiﬁed membranes respectively. The low activity of
these other metals suggests that Pt catalysts are ideal for wet air oxidation
reactions. In these experiments, membranes containing MSA-stabilized Pt
nanoparticles performed as well as those containing citrate-stabilized Pt
nanoparticles. The membranes modiﬁed by anionic impregnation with Pt also
had similar formic acid conversion. This shows that even though the three types
of Pt nanoparticles have different activities in slurry reactions, the difference in
activities is not observed in membrane reactions using the conditions described
above because the reaction is limited by the solubility of 02. In order to see
differences in the activity of the three Pt-containing membranes, the oxidation
reactions should be performed at higher ﬂow rates.

Nevertheless, for oxygen-sparged solutions containing 10.8 mM formic
acid, increasing the ﬂux through the citrate-stabilized Pt nanoparticle-modiﬁed
membrane from 0.023 to 0.12 mUcmzsec did not signiﬁcantly affect the reduction
in formic acid concentration that occurred upon passing the solution through the

membrane (Figure 4.13). This suggests that the reaction is still limited by the

160

amount of O2 in the solution at high ﬂow rates. If the reaction were kinetically
limited, we would expect less reduction of the formic acid concentration at higher
ﬂow rates due to lower residence times in the membrane. In most fast gas/liquid
reactions with ﬂow-through contactors, the solubility of the gas in solution will
limit the reaction rate unless high gas pressures are employed. For this reason,
tubular interfacial gas/liquid contactors are often more attractive than ﬂow-

through contactors for membrane-catalyzed gas/liquid reactions such as wet air

 

 

oxidation.
.12 3.0 .
8§20
E5 '
IL 6
.5 r:
3,3 1.0 -
c
II
5
0.0 . . .
0.000 0.050 0.100 0.150

Flux (mLIcmZ-sec)

Figure 4.13 Change in formic acid concentration during membrane-catalyzed
oxidation as a function of ﬂux through a Pt nanoparticle-modiﬁed disk-shaped
alumina membrane prepared by method 1. The initial formic acid concentration
was 10.8 mM, and solutions were sparged with O2 before passing through the
membrane.

4.3.4 Wet Air Oxidation with Tubular Membranes
This section compares the catalytic activities of ﬁve types of tubular
interfacial contactor membranes (Figures 4.3 - 4.7) in sequential studies of the

oxidation of formic acid, acetic acid, and phenol. Experiments were performed

161

‘ ““51

 

on at least two membranes modiﬁed by each method. Because catalyst

deactivation often occurs during phenol oxidation,”57

membranes were again
tested in the oxidation of formic acid after experiments with phenol to see if

catalyst deactivation occurred.

4.3.4.1 Wet Air Oxidation of Formic Acid

Initially, formic acid oxidation was examined at 0.2, 1, 2, 3, 4, and 4.3 bar
of air overpressure to determine the effect of overpressure on. reaction rates.
High air overpressures increase the solubility of O2 in the solution, but more
importantly, they shift the location of the gas/liquid interface so that it is closer to
the inner layer where there should be a larger amount of catalyst. The highest
activity in formic acid oxidation occurs at 4 or 4.3 bar of overpressure because
the gas/liquid interface is closest to the inner layer of the tube where most of the
catalyst is located. However, in some cases much of the air begins to come
through the defects in the membrane at an overpressure of 4.3 bar. As a result,
the catalytic activity sometimes starts to decrease at 4.3 bar because the
gas/liquid interface is no longer well-maintained in the catalytic layer of the
membrane. This is consistent with previous results.28 Figure 4.14 shows the
increase in activity as the air overpressure increases. Furthermore, when N2 is
used as the gas instead of air for oxidation reactions, no reaction occurs, which
shows that the change in formic acid concentration due to adsorption to the

polyelectrolyte/PtNP ﬁlm is negligible.

162

0.40 -

 

 

oAir O
5" INitro en
E 0.30- g
a
3 O
E 0.20-
3.7
(I
n: 0.10- . . o
0.00 ' 4 I I 1'
0 1 2 3 4

Gas Overpressure (bar)

Figure 4.14 Normalized rate of formic acid oxidation vs. air or N2 overpressure
for a tubular membrane modified with a PAA/PAH/PtNP ﬁlm. Normalization was
performed with respect to the area of the internal wall of the membrane.

The activities of membranes prepared by all 5 methods of modification
were compared with respect to membrane area and Pt loading. Each membrane
exhibits the expected trend of increasing activity with increasing air
overpressures. Membranes prepared by methods 1 and 2 exhibit similar
activities (normalized to membrane area) for formic acid oxidation and similar
specific activities (normalized to Pt content) as well (Figure 4.15). Membranes
prepared by method 3 show a similarly high reaction rate (Figure 4.15a), but
because the Pt content of these membranes is higher than that of all other
membranes (Table 4.1), their speciﬁc activity at overpressures >3 bar is lower
than for membranes prepared by methods 1 and 2 (Figure 4.15b). This is not
surprising because in method 3, the initial Pt deposition occurs throughout the
membrane, not just in the surface layer. Only the Pt that is near the gas/liquid

interface is efﬁciently used for formic acid oxidation. In Figure 4.153, the

163

 

 

relatively high activities at low overpressures for the membranes prepared by
method 3 likely occur because at low overpressures the gas/liquid interface is
deeper in the membrane, and these membranes still have signiﬁcant amounts of
Pt in these locations.

Membranes prepared by method 4 show a high rate of formic acid
oxidation (Figure 4.15a) but a lower activity per gram of Pt (Figure 4.15b) than

methods 1 and 2. This suggests that method 4 deposits the platinum deeper into

 

   

 

 

 

... 0.40 a)
.E
g 0-30 ‘ +Method 1
g -I-Method 2
E o 20 +Method 3
E +Method 4
. -O-Method 5
.g 0.10 . /
2 / e .
0.00 - . - . .
0 1 2 3 4 5
Air Overpressure (bars)
1.60 1 b)

11'

it: 1-20 ' +Method 1
=0 -I- Method 2

E “-30 ‘ +Method 3
E + Method 4

,5 “-40 ' -0- Method 5

E 0.00 - - . .

0 1 2 3 4 5
Air Overpressure (bars)

Figure 4.15 Normalized rate of formic acid oxidation vs. air overpressure for Pt-
containing membranes prepared by the methods shown in Figures 4.3-4.7.
Normalization was performed with respect to a) the area of the internal wall of the
membrane and b) the amount of Pt in the membrane.

164

the membrane than methods 1 and 2, and therefore the speciﬁc activity with
method 4 membranes is low at the higher air overpressures because all of the
platinum is not being used effectively. On the other hand, the membranes
prepared by method 5 exhibit a much lower activity than those prepared by the
other four methods. This is expected because Pt does not bind well to the
surface of ZrO2 and is easily removed during the nitric acid rinsing step.
Therefore, the Pt is mostly bound to regions that were not washed well with nitric
acid. Since these regions are most likely located away from the gas/liquid
interface, the method 5 membranes should have relatively low activity.

When comparing the LbL-modiﬁed membranes (methods 1-3) with
membranes described in the literature,26 the rate per membrane area is only 1I2,
to 1/2 as high as published values. On the other hand, the rate is 5 times higher
than published values when normalizing to Pt content. The rate is also ~50%
higher than previous values obtained with “low-loading” membranes, which had
Pt contents similar to the membranes prepared in this study.26 The speciﬁc
activities of membranes prepared by method 4 are similar to past results with
“low-loading” membranes. Additional formic acid oxidation experiments were
also performed with tubular membranes containing a PAA/PAH/AuNP ﬁlm. At
room temperature no formic acid oxidation was observed. However, Au is
typically an inactive oxidation catalyst at low temperatures. When performing
formic acid oxidation at higher temperatures, (feed temperature = 80 °C) some

oxidation occurs with the Au nanoparticles, but the activity at 4 bar air (0.07

165

mmol/s*m2) overpressure is still signiﬁcantly lower than with the Pt-containing

membranes.

4.3.4.2 Wet Air Oxidation of Acetic Acid

Acetic acid is a much more refractory compound to oxidize (scheme 2)
than formic acid and is often an intermediate produced in the oxidation of
oxygenated compounds like phenol or other carboxylic acids.1“'58'59 Supported
Ru or Pt catalysts are often used at high temperatures to oxidize acetic acid and
other short chain carboxylic acids completely to CO2 and H20.60432 However,
carbonate species are commonly formed on the catalyst surface, which results in
catalyst poisoning and reduced performance.”64

Though the exact mechanism of acetic acid oxidation is not known, it is
generally believed that short chain carboxylic acids undergo oxidation through a

free radical mechanism.”68

In general, for compounds containing 3 or more
carbon atoms, hydrogen abstraction occurs at a C-H bond to form a free radical
R-, which then reacts with oxygen to form an alkylperoxyl radical (ROO-). This
unstable radical then undergoes decarboxylation or other decomposition
reactions to form shorter chain intermediates. Propagation of the free radicals
continues rapidly until acetic acid is formed. Duprez et al. proposed that in the
presence of a catalyst, acetic acid is oxidized through formation of an acetate
radical followed by decarboxylation and formation of a methyl radical that forms a

peroxyl radical and is broken down into CO2 and water.66 Scheme 4.2 shows the

proposed pathway for this series of reactions.

166

02 02 /O 02
CH3COOH ——> CH3COO' ——> CH3' —> CH3OO' ——> H2C\| —-+ C02+ H20
' H02“ ' COZ ' H' 0

Scheme 4.2 Proposed mechanism for catalytic wet air oxidation of acetic acid.
Adapted from Duprez et al.66

In this work, we performed oxidation of acetic acid under conditions similar
to those used in formic acid oxidation. Analogous to results with formic acid, the
rate of acetic acid oxidation at room temperature is highest at 4 or 4.3 bar (Figure
4.16). The oxidation rate is determined by TOC analysis and thus tells how
much of the acetic acid is completely oxidized to CO2 and H20. It does not
account for any partial conversion to other intermediate species that may also
form.

Figure 4.16a shows that membranes prepared by methods 1 and 2 have a
much higher activity for room temperature acetic acid oxidation than membranes
prepared by the other 3 methods. The membranes prepared by method 3 show
a small speciﬁc activity, but membranes prepared by methods 4 and 5 exhibit
essentially no activity for acetic acid oxidation at room temperature. At 60 °C,
membranes prepared by methods 1-4 exhibit higher activity than at room
temperature, as expected. In the case of membranes prepared by methods 1
and 2, on increasing the temperature from 23 °C to 60 °C the activity increases
by about 60 percent and 90 percent, respectively. In contrast, the activities of
membranes prepared by methods 3 and 4 increase by factors of 9 and 10,
respectively. Even with this large increase in activity, the membranes prepared

by methods 3 and 4 still have a 3-fold lower activity at 60 °C and 4 bars air than

167

the membranes prepared by the ﬁrst two methods. The membranes prepared

by method 5 show low activities even at the higher temperature.

0.25 1

 

 

 

1:: a)
9") 0.20 - +Method 1
:3" -I- Method 2
g 0-15 ‘ + Method 3
g 0.10 , +Method4
g1 -0- Method 5
g 0.05 1
< 0.00 -——Q=-—-o==t“3——i

0 1 2 3 4 5

Air Overpressure (bars)

A 0.40 1
a; b) + Method 1
7:0 0'30 ‘ -I- Method 2
g 0 20 ‘ + Method 3
g ' + Method 4
g o_1o . / -0- Method 5
E 0.00 a .

0 1 2 3 4 5

Air Overpressure (bars)

Figure 4.16. Normalized rates of membrane-catalyzed acetic acid oxidation vs.
air overpressure for different membrane types at a) room temperature and b) 60
°C. Oxidation rates were normalized to the amount of Pt in the membrane.

4.3.4.3 Wet Air Oxidation of Phenol
Like acetic acid, phenol can be very difﬁcult to oxidize, but it is a common
compound for testing catalyst stability in wet air oxidation of organic

pollutants.“'57'69'71

We expect phenol oxidation to occur via a free radical
mechanism in which phenoxy and phenoxyperoxyl radicals are formed and

undergo further reactions to form smaller intermediates.72 Though the

168

mechanism of wet air oxidation of phenol is not necessarily the same for all
catalysts, previous studies showed similar intermediate and product formation
during oxidation with several catalysts.“'73'74 In general, phenol ﬁrst undergoes
oxidation to form intermediate ring compounds such as catechol and
hydroquinone. These ring compounds then undergo ring opening reactions to
form 6-carbon dicarboxylic acids, which are oxidized further to form short chain
carboxylic acids and eventually CO2. Scheme 4.3 shows a proposed oxidation

pathway.

OH

©
0 I+/ \‘é’
0 OH 1

0
COOH

l 00: if”

E... \ .... /

E... 7' (a... 7 31..." ff...
\. l / I

HOOC-COOH HaC—COOH

l l

HOOOH —' 002+H20 C°2+H20

Scheme 4.3 Proposed pathway for wet air oxidation of phenol, adapted from
Devlin et al.74

169

In phenol oxidation experiments, samples were collected at 1, 3, and 4 bar
overpressure and subsequently analyzed by both TOC and HPLC. In TOC
analysis, activities are determined from the difference in organic carbon content
between the inlet and outlet solutions. In HPLC analysis, activities are
determined from the decrease in phenol concentration. Similar to previous
experiments involving oxidation of formic acid and acetic acid, the highest activity
occurs at air overpressures of 4 bar.

Tables 4.2 and 4.3 show that, again, the highest activity occurs with
membranes prepared by methods 1 and 2, and that membranes prepared by
method 5 show little or no activity for the oxidation of phenol. Membranes
prepared by methods 3 and 4 exhibit more than 4-fold lower speciﬁc activities
than membranes prepared by the ﬁrst two methods, even at higher temperature.
Oxidation experiments performed at 60 °C result in higher activities (two times
higher or more) than experiments performed at room temperature. We would
expect to achieve even higher activities with temperatures in excess of 150
°C,”55 but these high temperatures are not compatible with the experimental
apparatus (Figure 4.8) used in this study. Furthermore, these high temperatures
may also lead to ﬁlm deformation and possible sintering of the catalyst. Future
studies need to explore the stability of polyelectrolyte/metal nanoparticle ﬁlms at

temperatures at or above 150 °C.

170

Table 4.2 Catalytic activities of several tubular membranes in phenol oxidation
at room temperature with 4 bar air overpressure.

 

 

 

Preparation Activi
method ammoll(s*gPt) mmoll(s:gPt)
1 0.064 0.13
2 0.050 0.1 1
3 0.006 0.015
4 0.000 0.000
5 0.003 0.012

 

3Determined by TOC analysis, bDetermined from HPLC analysis

Table 4.3 Catalytic activities of several tubular membranes in phenol oxidation
with a feed temperature of 60 °C with 4 bar air overpressure.

 

 

 

Preparation Activity
method ‘mmoll(s*gPt) l"mmol/(s*gPt)
1 0.15 0.21
2 0.1 1 0.21
3 0.026 0.035
4 0.027 0.025
5 0.003 0.000

 

a‘Determined by TOC analysis, bDetermined from HPLC analysis

The activity values determined from HPLC are generally higher than those
determined from TOC (Tables 4.2 and 4.3) because TOC analysis only shows
how much of the sample is transformed to C02 or insoluble species, whereas
HPLC shows how much phenol is oxidized to any product. The higher activities
seen with HPLC suggest that some phenol is oxidized to smaller organic
compounds, and not completely to C02. Because the conversion in phenol
oxidation is low (< 10%), the quantities of these other compounds in the analyzed
samples are below detectable levels in HPLC. As a result, the identity and
amount of each byproduct in the reaction were not determined. The activities

determined by TOC analysis and HPLC are generally in better agreement at 60

171

°C than at room temperature, suggesting that more of the phenol is converted to
C02 and H20 at higher temperatures.

After phenol oxidation, the various membranes were again used to
catalyze formic acid oxidation at room temperature to determine if the oxidation
of refractory compounds like phenol causes catalyst deactivation. While a
decrease in activity for formic acid oxidation could be due to poisoning or other
effects such as nanoparticle leaching or aggregation, a constant activity may
suggest that the extent of catalyst deactivation is minimal. However, if formic
acid oxidation is simply limited by 02 solubility, then catalyst deactivation would
not be observed by this method. In nearly every case, there was no decrease in
formic acid oxidation rates after using the membranes for phenol oxidation.
However, one of the membranes prepared by method 4 showed a 40% activity
decrease in formic acid oxidation after the membrane was used for phenol
oxidation. The decreased activity may have been due to leaching of Pt for that
speciﬁc membrane or to poisoning of the catalyst during phenol oxidation. With
that exception, each membrane maintained a constant activity for formic acid

oxidation.

4.3.5 Conventional Reactions with Pulverized Tubular Membranes

To show that the interfacial contactor conﬁguration is advantageous for
wet air oxidation reactions, each type of membrane was also ground into a
powder that was used as a heterogeneous catalyst in a conventional stirred tank

reaction. In these reactions, a solution containing 0.108 M formic acid was

172

 

continuously bubbled with oxygen while stirring rapidly. Pure oxygen was used
as the oxidant instead of air to provide as much oxygen to the reaction as
possible. The results in Table 4.4 show that all five types of membranes in the
powder form have similar activities when normalized to the amount of Pt in the
catalyst. However, the activities of different membranes operated as interfacial
contactors vary signiﬁcantly with the method of modiﬁcation. In the case of
methods 1 and 2, membranes operated as interfacial contactors at 4 bar
overpressure show activities that are ~2.5 times higher than those of membrane
powders used as heterogeneous catalysts. Conversely, membranes prepared by
methods 3 and 4 exhibit little difference in activity between interfacial contactors
and powder catalysts, and the membrane prepared by method 5 shows higher
activity in the conventional reaction. These results demonstrate that the
interfacial contactor conﬁguration can be quite valuable for gas/liquid reactions,
but to take full advantage of this conﬁguration, the catalyst must be localized in

the inner layer of the membrane.

Table 4.4 Catalytic activities in formic acid oxidation for tubular membranes
used as interfacial contactors and as powders in conventional stirred tank
reactors.

 

 

 

Modiﬁcation Activity (mmolls*gPtL
Method lnterfacial Conventional

Contactor‘I Reactor”

1 1.5 :l: 0.3 0.6 1: 0.05

2 1.3 1 0.2 0.5 i 0.03

3 0.5 :l: 0.1 0.7 d: 0.1

4 0.7 :l: 0.2 0.5 i: 0.03

5 0.1 :t 0.03 0.5 t 0.1

 

aactivity at air overpressure = 4 bar, bPure 02 was sparged into the reaction
mixture.

173

4.4 Conclusions

The overall objective of this study was to compare the catalytic activity of
membranes prepared using layer-by-layer deposition methods with the activity of
membranes prepared by the traditional methods of evaporation/recrystallization/
reduction and anionic impregnation/reduction. Although the rate of formic acid
oxidation with LbL-modiﬁed membranes was 50% lower than previous results
when normalized to membrane surface area, the rate when normalized to Pt
content was 5 times greater than previous results. In this study, the Pt content
for all membranes was less than 5 mg of Pt per membrane.

Membranes prepared by LbL methods 1 and 2 exhibited the highest
activity when normalized to the Pt content inside the membranes, most likely
because of strong localization of the Pt in the inner layer of the membrane.
Conversely, the other three methods deposit Pt on the entire surface of the
membrane, which means that any Pt that gets deposited on the support layer or
intermediate layer is most likely not being utilized when performing oxidation at
higher air overpressures.

The biggest limitation to the methods involving LbL deposition is the low
loading of Pt. Since the support is quite expensive, the cost of Pt is not as much
of a concern as in other systems; however, Pt cost cannot be disregarded. In the
future, low loading with the LbL method can be overcome by optimizing the LbL
deposition procedure or by depositing multiple layers. Further studies should
also include examination of catalytic activity in continuous experiments over

longer periods of time to learn more about the catalyst stability. Experiments in

174

 

 

 

this study were typically performed only for a few hours. Layer-by-layer
modiﬁcation is quite versatile and could also be applied to polymeric hollow ﬁber
supports, which are much less expensive than the traditional ceramic supports.
This should result in a more cost-effective system for wet air oxidation of

wastewater as long as the polymer membrane is sufﬁciently stable.

175

 

 

 

4.5 References
(1) Luck, F. Catal. Today 1999, 53, 81-91.

(2) Mishra, V. 8.; Mahajani, V. V.; Joshi, J. B. Ind. Eng. Chem. Res.
1995, 34, 2-48.

(3) Zimmerman, F. J. Chem. Eng. 1958, 65, 117-20.
(4) Guibelin, E. Water Sci. Technol. 2004, 49, 209-216.

(5) Kim, B. J.; 0i, 5.; Shanley, R. S. Water Environ. Res 1994, 66, 440-
55.

(6) Adriaens, P.; Vogel, T. M. In Microbial Transformation and Degradation of
Toxic Organic Chemicals; Young, L. Y., Cemiglia, C. E., Eds.; Wiley: New York,
1995, p 435-486.

(7) Fewson, C. A. Trends Biotechnol. 1988, 6, 148-53.

(8) Matatov-Meytal, Y. |.; Sheintuch, M. Ind. Eng. Chem. Res. 1998, 37,
309-326.

(9) Jeworski, M.; Heinzle, E. Biotechnology annual review 2000, 6, 163-
96. .

(10) Debellefontaine, H.; Foussard, J. N. Waste Manage. 2000, 20, 15-
25.

(11) Luck, F. Catal. Today 1996, 27, 195-202.
(12) Levec, J.; Pintar, A. Catal. Today 2007, 124, 172-184.

(13) Pham Minh, D.; Gallezot, P.; Azabou, 8.; Sayadi, S.; Besson, M.
Appl. Catal., B 2008, 84, 749-757.

(14) Pintar, A.; Batista, J.; Tisler, T. Appl. Catal., B 2008, 84, 30-41.

(15) Julbe, A.; Farrusseng, D.; Guizard, C. J. Membr. Sci. 2001, 181, 3-
20.

176

(16) Dalmon, J.-A.; Cruz-Lopez, A.; Farrusseng, D.; Guilhaume, N.;
lojoiu, E.; Jalibert, J.-C.; Miachon, S.; Mirodatos, C.; Pantazidis, A.;
Rebeilleau-Dassonneville, M.; Schuurman, Y.; Van Veen, A. C. Appl.
Catal., A 2007, 325, 198-204.

(17) Dittmeyer, R.; Svajda, K.; Reif, M. Top. Catal. 2004, 29, 3-27.

(18) Bengtson, G.; Fritsch, D. Desalination 2006, 200, 666-667.

(19) Lange, C.; Storck, S.; Tesche, B.; Maier, W. F. J. Catal. 1998, 175,
280-293.

(20) Purnama, H.; Kurr, P.; Schmidt, A.; Schomacker, R.; Voigt, l.; Wolf,
A.; Warsitz, R. AIChE J. 2006, 52, 2805-2811.

(21) Schmidt, A.; Haidar, R.; Schomacker, R. Catal. Today 2005, 104,
305-312.

(22) Schmidt, A.; Schomacker, R. J. Mol. Catal. A: Chem. 2007, 271,
192-199.

(23) Reif, M.; Dittmeyer, R. Catal. Today 2003, 82, 3-14.

(24) Pera-Titus, M.; Miachon, S.; Dalmon, J.-A. AIChE J. 2009, 55, 434-
441.

(25) lojoiu, E. E.; Landrivon, E.; Raeder, H.; Torp, E. G.; Miachon, S.;
Dalmon, J.-A. Catal. Today 2006, 118, 246-252.

(26) lojoiu, E. E.; Miachon, S.; Dalmon, J.-A. Top. Catal. 2005, 33, 135-
139.

(27) lojoiu, E. E.; Miachon, 8.; Landrivon, E.; Walmsley, J. C.; Raeder,
H.; Dalmon, J.-A. Appl. Catal., B 2007, 69, 196-206.

. (28) lojoiu, E. E.; Walmsley, J. C.; Raeder, H.; Miachon, S.; Dalmon, J.-A.
Catal. Today 2005, 104, 329-335.

(29) Miachon, S.; Perez, V.; Crehan, G.; Torp, E.; Raeder, H.; Bredesen,
R.; Dalmon, J. A. Catal. Today 2003, 82, 75-81.

(30) Raeder, H.; Bredesen, R.; Crehan, G.; Miachon, 8.; Dalmon, J.-A.;
Pintar, A.; Levec, J.; Torp, E. G. Sep. Purif. Technol. 2003, 32, 349-355.

177

(31) Vospernik, M.; Pintar, A.; Levec, J. Chem. Eng. Process. 2006, 45,
404-414.

(32) Bercic, G.; Pintar, A.; Levec, J. Catal. Today 2005, 105, 589-597.
(33) Vospernik, M.; Pintar, A.; Bercic, G.; Levec, J.; Walmsley, J.; Reder,
H.; Eduard Emil, l.; Sylvain, M.; Jean-Alain, D. Chem. Eng. Sci. 2004, 59,
5363-5372.

(34) Perez, V.; Miachon, S.; Dalmon, J.-A.; Bredesen, R.; Pettersen, G.;
Raeder, H.; Simon, C. Sep. Pun'f. Technol. 2001, 25, 33-38.

(35) Uzio, D.; Miachon, S.; Dalmon, J.-A. Catal. Today 2003, 82, 67-74.

(36) Braunstein, P.; Kormann, H.-P.; Meyer-Zaika, W.; Pugin, R.; Schmid,
G. Chem. Eur. J. 2000, 6, 4637-4646.

(37) Daub, K.; Wunder, V. K.; Dittmeyer, R. Catal. Today 2001, 67, 257-
272.

(38) Perego, 0; Villa, P. Catal. Today 1997, 34, 281-305.

(39) Centi, G.; Dittmeyer, R.; Perathoner, 8.; Reif, M. Catal. Today 2003,
79-80, 139-149.

(40) Grdschel, L.; Haidar, R.; Beyer, A.; Coelfen, H.; Frank, B.;
Schomacker, R. Ind. Eng. Chem. Res. 2005, 44, 9064-9070.

(41) Xu, J.; Dozier, A.; Bhattacharyya, D. J. Nanopart. Res. 2005, 7, 449-
467.

(42) Decher, G.; Schlenoff, J. B. Multilayer Thin Films: Sequential Assembly of
Nanocomposite Materials; Wiley-VCH: Weinheim, Germany, 2003.

(43) Chen, S.; Kimura, K. J. Phys. Chem. B 2001, 105, 5397-5403.

(44) Brugger, P. A.; Cuendet, P.; Graetzel, M. J. Am. Chem. Soc. 1981,
103, 2923-7.

(45) Brown, K. R.; Walter, D. G.; Natan, M. J. Chem. Mater. 2000, 12,
306-313.

(46) Dokoutchaev, A.; James, J. T.; Koene, S. C.; Pathak, S.; Prakash,
G. K. S.; Thompson, M. E. Chem. Mater. 1999, 11 , 2389-2399.

178

““u‘ “I
I

 

 

(47) Yang, J.; Lee, J. Y.; Deivaraj, T. 0; Too, H.-P. J. Colloid Interface
Sci. 2004, 271, 308-312.

(48) Bhattacharjee, S.; Bruening, M. L. Langmuir 2008, 24, 2916-2920.

(49) Kidambi, S.; Dai, J.; Li, J.; Bruening, M. L. J. Am. Chem. Soc. 2004,
126, 2658-2659.

(50) Kidambi, S.; Bruening, M. L. Chem. Mater. 2005, 17, 301-307.

(51) Harmsen, J. M. A.; Jelemensky, L.; Van Andel-Scheffer, P. J. M.;
Kuster, B. F. M.; Marin, G. B. Appl. Catal., A 1997, 165, 499-509.

(52) Alvarez, J.; Liu, J.; Roman, E.; Kaifer, A. E. Chem. Commun. 2000,
1151-1152.

(53) Eklund, S. E.; Cliffel, D. E. Langmuir 2004, 20, 6012-6018.
(54) Wilde, F. D. Field Measurements: U. 8. Geological Survey Techniques of

Water-Resources Investigations book 9, chap. A6,
h_ttp://pubs.water.usgs.gov/twri9A/ Accessed February 6, 2009.

(55) Nousir, S.; Keav, 8.; Jr., J. B.; Bensitel, M.; Brahmi, R.; Duprez, D.
Appl. Catal., B 2008, 84, 723-731.

(56) Taboada, C. D.; Batista, J.; Pintar, A.; Levec, J. Appl. Catal., B 2009,
89, 375-382.

(57) Liotta, L. F.; Gruttadauria, M.; Di Carlo, G.; Perrini, G.; Librando, V.
J. Hazard. Mater. 2009, 162, 588-606.

(58) Levec, J.; Pintar, A. Catal. Today 1995, 24, 51-58.

(59) 0|iviero, L.; Barbier, J.; Duprez, D.; Wahyu, H.; Ponton, J. W.;
Metcalfe, I. S.; Mantzavinos, D. Appl. Catal., B 2001, 35, 1-12.

(60) Diikkanci, M.; Giindtiz, G. Catal. Commun. 2009, 10, 913-919.

(61) Wang, J.; Zhu, W.; He, X.; Yang, S. Catal. Commun. 2008, 9, 2163-
2167.

179

 

 

(62) Mikulova, J.; Rossignol, S.; Barbier Jr, J.; Mesnard, D.; Kappenstein,
C.; Duprez, D. Appl. Catal., B 2007, 72, 1-10.

(63) Mikulova, J.; Barbier Jr, J.; Rossignol, S.; Mesnard, D.; Duprez, D.;
Kappenstein, C. J. Catal. 2007, 251, 172-181.

(64) Mikulové, J.; Rossignol, S.; Barbier Jr, J.; Duprez, D.; Kappenstein,
C. Catal. Today 2007, 124, 185-190.

(65) Barbier, J. J.; Delanoé, F.; Jabouille, F.; Duprez, D.; Blanchard, G.;
lsnard, P. J. Catal. 1998, 177, 378-385.

(66) Duprez, D.; Delanoé, F.; Barbier, J.; lsnard, P.; Blanchard, G. Catal.
Today 1996, 29, 317-322.

(67) lmamura, S.; Nakamura, M.; Kawabata, N.; Yoshida, J.; lshida, S.
Ind. Eng. Chem. Prod. Res. Dev. 1986, 25, 34-37.

(68) Nikolaou, N.; Abatzoglou, N.; Gasso, S.; Chomet, E. Can. J. Chem.
Eng. 1994, 72, 522-533.

(69) Cybulski, A.; Trawczynski, J. Appl. Catal., B 2004, 47, 1-13.

(70) Yang, S.; Zhu, W.; Li, X.; Wang, J.; Zhou, Y. Catal. Commun. 2007,
8, 2059-2063.

(71) Wang, J.; Zhu, W.; Yang, S.; Wang, W.; Zhou, Y. Appl. Catal., B
2008, 78, 30-37.

(72) Rivas, F. J.; Kolaczkowski, S. T.; Beltran, F. J.; Mclurgh, D. B.
Chem. Eng. Sci. 1998, 53, 2575-2586.

(73) Quintanilla, A.; Casas, J. A.; Mohedano, A. F.; Rodriguez, J. J. Appl.
Catal., B 2006, 67, 206-216.

(74) Devlin, H. R.; Harris, |. J. Ind. Eng. Chem. Fundam. 1984, 23, 387-
392.

180

 

 

Chapter 5: Conclusions and Future Work
5.1 Conclusions

The ﬁrst goal of this dissertation is to show that polyelectrolyte multilayer
ﬁlms provide an excellent means for immobilizing metal nanoparticles on porous
supports while still allowing access to the catalytic surface of the metal
nanoparticles. Chapter 2 clearly shows that preformed citrate-stabilized Au
nanoparticles exhibit the same catalytic activity when immobilized on 100 mesh
alumina powder and when they are unsupported, semi-homogeneous catalysts.
These data demonstrate that the polyelectrolyte does not inhibit the nanoparticle
catalysis. Moreover, synthesizing the nanoparticles prior to immobilization on a
support allows control over important properties including nanoparticle size,
shape, composition, surface charge, and stabilizing ligand.

Chapter 3 presents in-situ synthesis of nanoparticles through layer-by-
layer (LbL) deposition with polyelectrolyte solutions containing sz+ ions and
subsequent reduction of these ions using NaBH4. This method is advantageous
in that it typically yields smaller particle sizes than formation of nanoparticles in
solution. Additionally, the in situ technique is applicable to a variety of metal ions
(eg. Pd”, Au3", Pt”, Ag“, etc), and affords control over particle size by simply
changing the concentration of the metal ion in the polyelectrolyte deposition
solution. Pd nanoparticles formed in this way show high catalytic activities in the
hydrogenation of allylic compounds with size-dependent selectivity in the

hydrogenation of mono vs. disubstituted double bonds.

181

 

 

 

In addition to alumina powder supports, LbL modiﬁcation readily allows
modiﬁcation of porous membrane supports. The catalytic activity of Au
nanoparticles in the reduction of 4-nitrophenol was the same for nanoparticles
immobilized on alumina, nylon, and polycarbonate membranes despite the
differences in material, pore size, and pore morphology. More importantly, the
normalized rate constants of Au nanoparticles immobilized in membranes were
the same as Au nanoparticles in solution. These results demonstrate the
versatility of LbL deposition and provide further conﬁrmation that LbL
immobilization of nanoparticles does not result in a decrease in catalytic activity.

The second goal of this dissertation is to show that porous membranes
are attractive supports for catalytic nanoparticles. Chapter 2 demonstrates that
membranes operated as ﬂow-through contactors facilitate contact between liquid
phase reactants and the solid catalyst and eliminate diffusion limitations. These
membranes achieved >99% conversion of 4-nitrophenol to 4-aminophenol at
ﬂuxes as high as 0.53 mL/cmz-sec, and conversion was controlled by adjusting
the ﬂow rate of solution through the membrane. Control over reaction conversion
was most apparent in the reduction of substituted nitroaromatics such as 2-
nitrotoluene, where the ratio of nitroso to amine products was greatly affected by
the residence time of the solution in the membrane. These nanoparticle-
containing membranes also selectively reduced nitro groups in nitroaromatic
compounds containing other reducible functionalities.

In an effort to ﬁnd greener applications and milder reaction conditions for

membrane catalysts, we examined hydrogenation reactions in chapter 3.

182

 

 

Membranes modiﬁed with citrate-stabilized Pd nanoparticles effectively catalyzed
the hydrogenation of allyl alcohol and several nitroaromatic compounds when
operated as ﬂow-through contactors. In these reactions, the extent of conversion
was primarily limited by the solubility of hydrogen rather than the reaction
kinetics, which was demonstrated by the negligible decrease in conversion when
increasing the ﬂow rate by a factor of ﬁve.

Chapter 4 shows that the interfacial contactor conﬁguration can overcome
this solubility limitation to some extent. Tubular ceramic membranes modified
with citrate-stabilized Pt nanoparticles showed high activities in the wet air
oxidation of several model pollutants. This high activity is due in part to the
ability to control the location of the gas/liquid interface within the membrane,
which leads to a high surface area for contact between the gaseous and liquid
reactants and the solid catalyst. The LbL technique provides more control over
Pt location, limiting deposition to the skin layer where there is a higher surface
area for catalyst loading and thus more area for contact between all three
phases.

This work has primarily applied LbL deposition to effectively immobilize
catalytic noble metal nanoparticles on a wide variety of support materials with a
major focus on membranes. Though the systems explored thus far are
attractive, there are many more potential applications for LbL deposition of
nanoparticle catalysts. Furthermore, many characteristics of these nanoparticle
systems are not completely understood and thus require further investigation into

reaction mechanisms and nanoparticle properties. The remainder of this chapter

183

’1
‘-
IA

discusses some promising directions for this research. Some potential new
directions include the development of novel membrane materials for large-scale
catalytic appliCations, further utilization of membrane properties or nanoparticle
size to achieve selective product formation in hydrogenation reactions, and in-
situ hydrogen generation with conductive nanoparticle-containing membranes to

overcome H2 solubility limitations in ﬂow-through reactions.

5.2. Future Work
5.2.1 Modification of Polymeric Hollow Fiber Membranes for Catalytic
Reactions

One of the primary limitations to working with disk-shaped membranes is
the difficulty of using the membranes in large scale applications. Because ﬂow
rate and the amount of material that can be processed are directly related to the
surface area of the membrane, it is important to use membrane geometries with
a high surface area. Thus, polymeric hollow ﬁber membranes are often utilized
for high-throughput applications. Hollow ﬁber membranes have a higher surface
area per module volume than ﬂat-sheet membranes and thus can allow
extremely high ﬂuxes in ﬂow-through applications. Modules consisting of
polymeric hollow ﬁbers also cost much less than tubular ceramic membranes and
typically have lower operating costs.

The advantages associated with polymeric hollow ﬁber membranes make
them attractive for a wide variety of membrane applications, and as a result they

are extensively used in industry for processes such as gas separation,

184

 

hydrocarbon separation, water puriﬁcation, and kidney dialysis. However,
relatively few studies exist in which hollow ﬁbers act as metal nanoparticle
supports for catalytic applications. The work of Vincent and Guibal focused on
the use of chitosan hollow ﬁber membranes loaded with Pd nanoparticles for
degradation of nitrophenol and nitroaniline by sodium formate.“2 These studies
utilized hollow ﬁber membranes in the liquid-liquid contactor configuration. The
relatively small amount of previous research in this area leaves much room to

develop novel hollow ﬁber materials for catalytic applications.

5.2.2 Catalytic Selectivity with Pd-Containing Membranes

Chapter 3 demonstrates that Pd containing membranes work well in
hydrogenation of a variety of substrates including allylic and nitroaromatic
compounds. However, we did not use these membranes for selective
hydrogenation of similar substrates. Even though hydrogen solubility limits the
conversion in these reactions, it could be beneﬁcial to perform hydrogenation
studies on substrates containing multiple reducible functionalities. In many
hydrogenation reactions, the desired product is an intermediate that we want to
produce with high selectivity while still having a high conversion of substrate. In
order to achieve the desired product, we want to control the reaction time and
use a catalyst that preferentially reduces one functional group over another.

The interfacial contactor conﬁguration is commonly used for
hydrogenation reactions because it is easy to ﬂow liquid through the inside of the

tube, while passing gas around the outside of the tube or vice versa. There are

185

also a number of commercial modules available that can hold multiple tubular
membranes for these reactions. Previous studies utilized interfacial contactors
for hydrogenation of nitrate and nitrite,3 a-methylstyrene,4 nitrobenzene,5
cinnamaldehyde,6 and methylenecyclohexane.7 lnterfacial contactors still suffer
from mass transport limitations due to diffusion into the pores, and the catalyst
needs to be at or close to the phase boundary in order to be used efﬁciently.
lnterfacial contactors also offer less control over contact time, making them less
effective in achieving selective product formation.

Operating membranes in ﬂow-through mode can overcome these
limitations, but it requires passing the solution through the membrane multiple
times and reincorporating hydrogen into the solution between each pass. So far,
only a limited number of publications describe this method for hydrogenation
reactions. Schomacker and coworkers have done the most work in this area
with previous studies including the ﬂow-through hydrogenation of q-
methylstyrene, 1,5-cyclooctadiene, 1-octyne, phenyl acetylene, and geraniol with
Pd-containing polymeric membranes.8'9 In each case, the membranes
demonstrated selectivity for the singly reduced products. They also studied the
ﬂow-through hydrogenation of sunﬂower oil10 and 1,5-cyclooctadiene11 with
tubular ceramic membranes. Bottino and coworkers studied
methylenecyclohexane under ﬂow-through conditions with ﬂat PVDF and tubular
ceramic membranes.7 In both cases, the isomerization product was selectively
produced over the hydrogenated species. The membranes used in these

previous studies still have some disadvantages including: plugging of the

186

membrane pores,12 low catalyst loading (~0.6 wt%),8 and low catalyst
accessibility (< 16% of nanoparticle surface area).13 The problem of pore
plugging is not common with the membrane pore sizes used in our studies. Our
method of membrane preparation also offers a simple strategy to have high
catalyst loading with uniform dispersity and easy access to the catalyst sites.

Due to the limited amount of work concerning hydrogenation with catalytic
membranes operated in ﬂow-through mode, there are a wide variety of potential
applications for these membranes. Chemoselective reduction of nitro groups in
the presence of other reducible functional groups. is an important reaction in
organic synthesis. Although the hydrogenation of simple aromatic nitro
compounds readily occurs with traditional commercial catalysts, the
hydrogenation of nitro groups in the presence of other reducible functional
groups is more challenging. Flow-through membrane hydrogenations could
provide the control over residence time necessary to facilitate reduction of the
nitro group to the amine while not allowing enough time for the reduction of other
functionalities. The primary limitation of these ﬂow-through reactions is the low
solubility of H2, but this can be overcome by ﬂowing solutions at much faster ﬂow
rates. Studying the effect of ﬂow rate on conversion and selectivity could provide
insight into the kinetics of these reactions. Likewise, nanoparticle composition,
size, and stabilizer may also play a role in selectivity. When studying compounds
that can be selectively reduced, comparing the activity/selectivity of membrane
reactors with that of slurry reactors would help show if the selectivity is only due

to the catalyst properties or if the support also plays a role. Selective

187

hydrogenation of oleﬁns is another important industrial process that should also

be investigated.

5.2.3 Catalytic Selectivity as a Function of Nanoparticle Size

The results in chapter 3 demonstrate the strong dependence of catalyst
selectivity on particle size. Future studies should focus on understanding the
origin of these size effects and their applicability to other reactions. A number of
literature studies suggest that selectivity may result from different substrates
reacting at different sites on the nanoparticle.1“'15 For example, compounds with
monosubstituted double bonds may react preferentially at edge sites while
compounds with disubstituted double bonds react at terrace sites. As the
nanoparticle diameter decreases, the fraction of edge sites increases relative to
the number of terrace sites, particularly if the particles become defective and are
not regular cubooctahedra.

The competitive hydrogenation studies in chapter 3 suggest that
compounds with mono and disubstituted double bonds react at the same sites on
the nanoparticles. Thus, there are two possible explanations for the enhanced
selectivities with the smaller nanoparticles. First, adsorption of the surrounding
polyelectrolyte to the nanoparticle may provide more steric hindrance to binding
of the double bond to the particle when the nanoparticles are small. Second,
changes in the electronic structure of the nanoparticles with decreasing diameter
may greatly favor the binding of the monosubstituted double bond.

Bhattacharjee and coworkers showed that the latter hypothesis is more likely

188

because the high size-dependent selectivity in hydrogenation occurs with a range
of different catalyst systems including: [PAA-Pd(0)/PE|]3 ﬁlms on alumina, a
solution containing a PAA-Pd(0)/PEI complex, and reduced PdCl42'/PEI ﬁlms.“5
There should be less steric constraints with the complex in solution because of
more conformational freedom. Similarly, reduced PdCl42'lPEl ﬁlms should be
less tightly packed and show less steric hindrance to binding.

Using nanoparticle systems with a range of sizes, hydrogenation of a
number of different compounds can help to elucidate the origins of size-
dependent selectivity. For example, adding additional substituent groups to the
double bond on allyl alcohol should offer insight into the effect of steric hindrance
on reactivity and selectivity with different particle sizes. Figure 5.1 shows several
compounds that could be used to probe steric effects of nanoparticles in
hydrogenation. Comparison of the cis and trans forms of crotyl alcohol could
offer insight into the orientation of the double bond as it interacts with the
nanoparticle surface. Furthermore, comparison of the hydrogenation TOFs of
disubstituted double bonds with methyl, ethyl, and isopropyl substituents should
provide the most information. Single substrate experiments would be useful for
determining the differences in reactivity/selectivity of different substrates;
however, mixed substrate experiments would be necessary to determine if

different substrates react on the same catalyst sites.

Figure 5.1 Compounds that could be used to probe steric effects in
hydrogenation as a function of nanoparticle size.

189

The method of in-situ nanoparticle formation in polyelectrolyte multilayer
ﬁlms provides some control over nanoparticle size, but the size distribution is
relatively wide compared to other methods. To more conclusively demonstrate
size-dependent selectivity, other nanoparticle systems should be investigated.
Unfortunately, the methods that produce reactive nanoparticles with diameters

less than 3 nm are rare. The citrate reduction method typical yields particles with

17-19 20,21

diameters > 5 nm, whereas the Brust method gives smaller nanoparticles,

but the Brust particles are capped with thiols that decrease catalytic activity."’2'23
Dendrimer-encapsulated particles provide several advantages over the citrate or
Brust methods?"25 the ﬁrst of which is that the particles are relatively

monodisperse in size, which allows for more effective evaluation of reactivity as a

function of particle diameter.

1) Pd2+ .
§———
2) NaBH4

   

l j 1 l
Figure 5.2 Synthesis of relatively large nanoparticles in dendrimers using
multiple cycles of metal-ion loading and reduction. Higher generation dendrimers
could be employed in the actual nanoparticle formation.

190

Diameters of nanoparticles in dendrimers are typically 1 to 3 nm, which is
in the region of interest for selective hydrogenation of monosubstituted double
bonds, but it would also be useful to produce larger particles to have a wider size
range. Synthesis of larger particles would likely require a technique in which the
nanoparticles are prepared in a multistep process as shown in Figure 5.2. The
metal ions can be loaded in the dendrimer with an approximate binding ratio of 1
metal ion per tertiary amine group and then reduced with NaBH4 to form
nanoparticles. After dialysis to remove the reducing agent, repetition of this
process several times should produce larger particles in a controlled fashion.
Dendrimer-containing nanoparticles can be deposited in polyelectrolyte ﬁlms or
used directly in solution. Previous work by Crooks and coworkers suggests that
TOFs may decline slightly for particles less than 2 nm in diameter,25'26 but
selectivities for hydrogenation of mono and multisubstituted double bonds have
not been investigated with these systems. Several studies suggest that there is
an optimum nanoparticle size to achieve high reactivity, but relatively few studies
have investigated selectivity.27'28

Regardless of whether steric or electronic effects are dominant in altering
reactivities, the selectivity of nanoparticles as a function of size should extend to
a number of reactions. Again, chemoselective hydrogenation of nitroaromatic
compounds offers an interesting reaction to study. It would be beneﬁcial to
determine if nanoparticle size plays a role in which groups are more easily

reduced. Initial results by Bhattacharjee suggest that even relatively large

nanoparticles show high selectivity in the reduction of nitrobenzaldehyde to

191

aminobenzaldehyde.29 With a [PAA-Pd(0)/PEI]3 ﬁlm containing 3.4 nm diameter
Pd nanoparticles, 94% of the nitrobenzaldehyde is converted to
aminobenzaldehyde. In contrast, with Pd on carbon, the doubly reduced species
accounts for 60% of the product. Future studies should investigate the selectivity
as a function of nanoparticle size for a number of different nitroaromatic

compounds.

5.2.4 Electrocatalytic Hydrogenation with Pd-Containing Membranes

We are particularly interested in using Pd-containing membranes for
liquid-phase hydrogenation because these reactions have fast kinetics and are
used in a wide range of applications in the pharmaceutical, petrochemical, ﬁne

30'” Membrane-based

chemical, and wastewater treatment industries.
hydrogenation is attractive for minimizing mass-transport limitations and
performing continuous reactions, but the low solubility of H2 limits conversion.
Above, we mentioned the two most common methods for overcoming this
limitation: passing the solution through the membrane multiple times with
reincorporation of hydrogen into the solution between each pass and using an
interfacial contactor conﬁguration. The former case doesn’t really solve the
problem unless extremely high ﬂow rates are used. The interfacial contactor
conﬁguration is attractive, but reactions are still likely to be limited by mass
transport of substrate to the catalyst, and the catalyst loading may be small.

ln-situ hydrogen generation could offer a better solution for overcoming

this limitation as it would provide a constant supply of hydrogen inside the

192

membrane pores during the reaction. If the hydrogen generation occurs faster
than the reaction kinetics, then ﬂow-through hydrogenation reactions would no
longer have H2 solubility limitations. To generate hydrogen in-situ, Pd
nanoparticle-containing membranes must be conductive, and the hydrogen
should be produced at the surface of the nanoparticles where the hydrogenation
will take place.

Electrocatalytic hydrogenation occurs when hydrogen is produced at the
surface of an electrode by electrolysis of water or reduction of protons.33‘35 In
addition to avoiding the need for high hydrogen pressures, electrocatalytic
hydrogenation has the potential to be environmentally friendly because it uses
ambient temperature, aqueous solutions, and electrons rather than chemical
reducing agents. In conventional electrode systems, hydrogen production occurs
at the working electrode, but additional stirring and/or diffusion of the substrate to
the electrode surface is necessary to facilitate the reaction.”43 Pintauro and
coworkers examined the electrochemical hydrogenation of vegetable oil using
fuel cell membranes to supply electrochemically generated hydrogen.‘“’“7
Remarkably, when H2 was fed to the anode and hydrogen was subsequently
generated at the cathode, the fraction of H2 that hydrogenated fatty acid double
bonds ranged from 45 to 97%.46 Thus, electrochemically generated hydrogen
can, in principle, be used very efﬁciently for hydrogenation. Performing ﬂow-
through electrocatalytic hydrogenation with catalytic membranes (Figure 5.3) will

have the same advantages as traditional electrocatalytic hydrogenation plus the

added beneﬁts of using convective transport through membrane pores.

193

 

 

 

 

 

 

 

 

AglAgCl
reference
Solution
i Inlet
Pt
V
' outlet
—>
membra

 

Figure 5.3 Experimental setup for examining ﬂow-through electrocatalytic
hydrogenation in Pd nanoparticle-containing membranes. Hydrogen is
electrochemically generated in the conductive, nanoparticle-containing
membrane. The hydrogenation of allyl alcohol to 1-propanol is shown as an
example reaction.

The ﬁrst step toward performing electrocatalytic hydrogenation in
membranes is the development of a porous, conductive electrode capable of
electrochemically producing hydrogen in aqueous media. Most synthetic
membranes consist of polymeric or ceramic materials that have minimal electrical
conductivity and are thus not suitable for electrochemical hydrogenation. As a
test system for electrochemical hydrogenation, we investigated silver membranes
(obtained from SPI supplies), which were developed primarily for other
applications because of their chemical resistance and bactericidal properties.
The silver membranes are highly conductive, but clean silver does not catalyze
dissociation of H2, and the activity of Ag is orders of magnitude less than the

activity of Pt in the hydrolysis of water.“49 After modifying the surface of a silver

194

membrane with a monolayer of mercaptoundecanoic acid, we deposited a
uniform PEI/Pd nanoparticle ﬁlm inside the pores of the silver membrane. Figure
5.4 shows SEM images of silver membranes before and after coating with a Pd
nanoparticle-containing film. Due to the high chemical resistance of silver,
polyelectrolyte adsorption does not take place directly on the silver surface, thus
requiring the initial formation of an alkanethiol monolayer to facilitate
polyelectrolyte adsorption. Shorter alkanethiols may be necessary to improve

the electrical contact between the electrode (Ag membrane) surface and the Pd

nanoparticles.

    

500 nm ‘ 7' ‘ 500 nm
_’ _

 

\ . .
Figure 5.4 Cross-sectional FESEM images of a) a bare silver membrane and b)
a silver membrane coated with a mercapt0undecanoic acid/PEI/PdNP film.

As an initial experiment to test the capability of Pd-containing Ag
membranes for electrochemical hydrogen production, we performed
electrocatalytic hydrogenation of allyl alcohol. To facilitate hydrogen formation
and provide a conductive solution, 10.1 M HCI served as the supporting
electrolyte. However, electrolytes that contain CI' may lead to catalyst
poisoning,”51 so future experiments should utilize other supporting electrolytes

that do not contain Cl'. When passing a 5 mM solution of allyl alcohol through

195

the membrane at a ﬂux of 0.01 mL/cmZ-sec, more than 20% of the allyl alcohol
was electrocatalytically hydrogenated to 1-propanol. This is comparable to
results obtained in the traditional hydrogenation reaction (bubbling the solution
with H2) with the same membrane. We obtained similar conversions in the
electrocatalytic hydrodehalogenation of 4-bromophenol to phenol. Unmodiﬁed
silver membranes did not allow electrochemical or traditional hydrogenation to
take place.

Future studies should focus on understanding the mechanism of
electrocatalytic hydrogenation in membranes so that these systems can be
utilized for room-temperature hydrogenation at low pressures and higher
substrate concentrations (> 0.1 M). The ﬁrst step in this future work is to
investigate the electrochemical generation of H2 using disk electrodes that can
be readily characterized. An important question to answer is whether H2 is
generated primarily at the nanoparticles or the electrode surface. Ideally, H2
generation will occur exclusively at the nanoparticles because the efﬁciency of
hydrogen utilization should be much greater if electrochemically generated
hydrogen atoms could be used for hydrogenation prior to H2 formation. This
would not occur if H2 was generated primarily at the Ag surface, because Ag is
not an efﬁcient hydrogenation catalyst.52'53 Future studies could instead utilize
carbon membranes because carbon is well known to have a relatively high
overpotential for H2 generation.

Additional studies could look at the effect of multiple nanoparticle layers

on the rate of hydrogen generation. Deposition of additional layers is

196

straightforward on both membranes and electrodes, and chapter 2 showed that
additional layers of nanoparticles increase the rates of membrane catalysis.
Moreover, other studies showed that at least for the ﬁrst four or ﬁve layers, the
nanoparticles in polyelectrolyte ﬁlms are electrochemically active.”57 Coverage
of the Ag with multilayers of polyelectrolytes should also decrease the rate of H2
generation at the underlying electrode.58 Addition of a substrate such as allyl
alcohol to the solution may also affect the generation of hydrogen on
nanoparticle electrodes. The added substrate may shift the potential for
hydrogen generation, which could be helpful in developing more energy-efﬁcient
reactions.

Both the generation of hydrogen and the rate of hydrogenation likely
depend on the nanoparticle composition and size. Therefore, a study of different
nanoparticle materials and sizes could be beneﬁcial. Though Pd is most

common for hydrogenation reactions,“59

generation of hydrogen may occur
more readily at a Pt surface.60452 Once the ideal nanoparticle system is chosen,
catalytic studies should investigate conversion and current efﬁciency as a
function of overpotential, substrate concentration, and ﬂow rate through the
membrane. In aqueous ﬂow-through hydrogenation reactions, hydrogen is the
limiting reactant at substrate concentrations > 1 mM, but this will not likely be the
case with electrochemical generation of hydrogen and high mM substrate

concentrations. Higher substrate concentrations and ﬂow rates should increase

the current efficiency (utilization of adsorbed hydrogen).

197

5.3 Summary

In summary, polyelectrolyte multilayer ﬁlms provide a convenient platform
for immobilizing active and highly accessible metal nanoparticles for catalysis.
Porous membranes are especially attractive supports for polyelectrolyte-
stabilized metal nanoparticles as they demonstrate many advantages over
conventional catalyst supports. We’ve demonstrated that membranes containing
catalytically active metal nanoparticles effectively catalyze reduction,
hydrogenation, and oxidation reactions, but they also show great promise for a

number of other mtalytic applications.

198

 

 

5.4 References

(1)
(2)

(3)

(4)
(5)

(6)

(7)

(8)

(9)

(10)

(11)

(12)

(13)

(14)

(15)

Vincent, T.; Guibal, E. Environ. Sci. Technol. 2004, 38, 4233-4240.

Thierry, V.; Francisco, P.; Eric, G. J. Appl. Polym. Sci. 2004, 94,
1634-1642.

Centi, G.; Dittmeyer, R.; Perathoner, S.; Reif, M. Catal. Today 2003,
79-80, 139-149.

Cini, P.; Harold, M. P. AIChE J. 1991, 37, 997-1008.

Peureux, J.; Torres, M.; Mozzanega, H.; Giroir-Fendler, A.; Dalmon,
J. A. Catal. Today 1995, 25, 409-15.

Pan, X. L.; Liu, B. J.; Xiong, G. X.; Sheng, S. S.; Liu, J.; Yang, W. S.
Catal. Lett. 2000, 66, 125-128.

Bottino, A.; Capannelli, G.; Comite, A.; Di Felice, R. Desalination
2002, 144, 411-416.

Schmidt, A.; Haidar, R.; Schomacker, R. Catal. Today 2005, 104,
305-312.

Purwanto, P.; Deshpande, R. M.; Chaudhari, R. V.; Delmas, H. J.
Chem. Eng. Data 1996, 41, 1414-1417.

Schmidt, A.; Schomacker, R. J. Mol. Catal. A: Chem. 2007, 271,
192-199.

Schmidt, A.; Wolf, A.; Warsitz, R.; Dittmeyer, R.; Urbanczyk, D.;
Voigt, l.; Fischer, G.; Schomacker, R. AIChE J. 2008, 54, 258-268.

Reif, M.; Dittmeyer, R. Catal. Today 2003, 82, 3-14.

Grdschel, L.; Haidar, R.; Beyer, A.; Reichert, K.-H.; Schomacker, R.
Catal. Lett. 2004, 95, 67-75.

Freund, H. J.; Baumer, M.; Libuda, J.; Risse, T.; Rupprechter, G.;
Shaikhutdinov, S. J. Catal. 2003, 216, 223-235.

Silvestre-Albero, J.; Rupprechter, G.; Freund, H. J. J. Catal. 2006,
240, 58-65.

199

(15)

(17)

(18)

(19)

(20)

(21)

(22)

(23)

(24)

(25)

(26)

(27)

(23)
(29)
(30)

(31)

(32)

Bhattacharjee, S.; Dotzauer, D. M.; Bruening, M. L. J. Am. Chem.
Soc. 2009, 131, 3601-3610.

Turkevich, J.; Kim, G. Science 1970, 169, 873-9.

Turkevich, J.; Miner, R. 8., Jr.; Babenkova, L. J. Phys. Chem. 1986,
90,:4765-7.

Turkevich, J.; Stevenson, P. C.; Hillier, J. Discussions of the Faraday
Society 1951, No. 11 , 55-75.

Brust, M.; Walker, M.; Bethell, D.; Schiffrin, D. J.; Whyman, R. J.
Chem. Soc., Chem. Commun. 1994, 801-2.

Brust, M.; Fink, J.; Bethell, D.; Schiffrin, D. J.; Kiely, C. J. Chem.
Soc, Chem. Commun. 1995, 1655-1657.

Alvarez, J.; Liu, J.; Roman, E.; Kaifer, A. E. Chem. Commun. 2000,
1151-1152.

Eklund, S. E.; Cliffel, D. E. Langmuir 2004, 20, 6012-6018.

Kim, Y.-G.; Oh, S.-K.; Crooks, R. M. Chem. Mater. 2004, 16, 167-
172.

Niu, Y.; Yeung, L. K.; Crooks, R. M. J. Am. Chem. Soc. 2001, 123,
6840-6846.

Oh, S.-K.; Niu, Y.; Crooks, R. M. Langmuir 2005, 21 , 10209-10213.

Doyle, A. M.; Shaikhutdinov, S. K.; Freund, H.-J. Angew. Chem. Int.
Ed. 2005, 44, 629-631.

Valden, M.; Lai, X.; Goodman, D. W. Science 1998, 281, 1647-1650.
Bhattacharjee, S. PhD Dissertation, Michigan State University, 2008.
Bond, G. C. Chem. Ind. 1967, 2018-25.

Molnar, A.; Sarkany, A.; Varga, M. J. Mol. Catal. A: Chem. 2001,
173, 185-221.

Hoelderich, W. F. Catal. Today 2000, 62, 115-130.

200

 

(33)

(34)
(35)

(35)

(37)

(38)

(39)

(40)

(41)

(42)

(43)

(44)

(45)

(46)

(47)

(48)

Chapuzet, J. M.; Lasia, A.; Lessard, J. Electrocatalysis 1998, 155-
196.

Fuchigami, T.; Tajima, T. Electrochemistry 2006, 74, 585-589.
Lessard, J. Chemical Industries 2005, 104, 3-17.

Cirtiu, C. M.; Brisach-Wittmeyer, A.; Menard, H. J. Catal. 2006, 245,
191-197.

Cirtiu, C. M.; Brisach-Wittmeyer, A.; Menard, H. Catal. Commun.
2007, 8, 751-754.

Cirtiu, C. M.; Hassani, H. 0.; Bouchard, N.-A.; Rowntree, P. A.;
Menard, H. Langmuir 2006, 22, 6414-6421.

Dalavoy, T. S.; Jackson, J. E.; Swain, G. M.; Miller, D. J.; Li, J.;
Lipkowski, J. J. Catal. 2007, 246, 15-28.

Vago, M.; Williams, F. J.; Calvo, E. J. Electrochem. Commun. 2007,
9, 2725-2728.

Chen, G.; Wang, Z.; Yang, T.; Huang, D.; Xia, D. J. Phys. Chem. B
2006, 110, 4863-4868.

Do Amaral Ferraz Navarro, D. M.; Navarro, M. J. Chem. Educ. 2004,
81, 1350-1352.

Korotaeva, L. M.; Rubinskaya, T. Y. Electrochemistry 2005, 1-22.

An, W.; Hong, J. K.; Pintauro, P. N. J. Appl. Electrochem. 1998, 28,
947-954.

An, W.; Hong, J. K.; Pintauro, P. N.; Warner, K.; Neff, W. J. Am. Oil
Chem. Soc. 1999, 76, 215-222.

Pintauro, P. N.; Gil, M. P.; Warner, K.; List, 6.; Neff, W. Ind. Eng.
Chem. Res. 2005, 44, 6188-6195.

Warner, K.; Neff, W. E.; List, G. R.; Pintauro, P. J. Am. Oil Chem.
Soc. 2000, 77, 1113-1117.

Mohammad, A. B.; Yudanov, l. V.; Lim, K. H.; Neyman, K. M.;
Rosch, N. J. Phys. Chem. C 2008, 112, 1628-1635.

201

(49)

(50)

(51)

(52)
(53)

(54)

(55)

(56)

(57)

(53)

(59)

(60)

(61)

(62)

Uchida, T.; Mogami, H.; Yamakata, A.; Sasaki, Y.; Osawa, M. J. Am.
Chem. Soc. 2008, 130, 10862-+.

Mendyka, B.; Musialik-Piotrowska, A.; Syczewska, K. Catal. Today
1992, 11, 597-610.

Rasmussen, S. B.; Kustov, A.; Due-Hansen, J.; Siret, B.; Tabaries,
F.; Fehrmann, R. Appl. Catal., B 2006, 69, 10-16.

Emmett, P. H.; Skau, N. J. Am. Chem. Soc. 1943, 65, 1029-35.
Sarkany, A.; Revay, 2. Appl. Catal., A 2003, 243, 347-355.

Chirea, M.; Garcia-Morales, V.; Manzanares, J. A.; Pereira, C.;
Gulaboski, R.; Silva, F. J. Phys. Chem. B 2005, 109, 21808-21817.

Cho, J.; Caruso, F. Chem. Mater. 2005, 17, 4547-4553.

Zhao, J.; Bradbury, C. R.; Fermin, D. J. J. Phys. Chem. C 2008, 112,
6832-6841.

Zhao, J.; Wasem, M.; Bradbury, C. R.; Fermin, D. J. J. Phys. Chem.
C 2008, 112, 7284-7289.

Harris, J. J.; Bruening, M. L. Langmuir 2000, 16, 2006-2013.

Tungler, A.; Fogassy, G. J. Mol. Catal. A: Chem. 2001, 173, 231-
247.

Lofrano, R. C. 2.; Queiroz, J. V.; Romero, J. R. J. Mol. Catal. A:
Chem. 2001, 174, 231-238.

Machida, M.; Sato, K.; lshibashi, l.; Abul Hasnat, M.; Ikeue, K.
Chem. Commun. 2006, 732-734.

Pontolio, J. O. S.; Purgato, F. L. S.; Romero, J. R. Quim. Nova 2004,
27, 550-554.

202