Carbonaceous adsorbents as heterogeneous catalyst supports

MISSING IMAGE

Material Information

Title:
Carbonaceous adsorbents as heterogeneous catalyst supports applications and characterization
Physical Description:
x, 130 leaves : ill. ; 29 cm.
Language:
English
Creator:
Lafrenz, Todd James, 1966-
Publication Date:

Subjects

Subjects / Keywords:
Carbon compounds   ( lcsh )
Adsorption   ( lcsh )
Heterogeneous catalysis   ( lcsh )
Genre:
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1995.
Bibliography:
Includes bibliographical references (leaves 124-129).
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by Todd James Lafrenz.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 002066462
notis - AKQ4661
oclc - 34360603
System ID:
AA00003199:00001

Full Text









CARBONACEOUS ADSORBENTS AS HETEROGENEOUS CATALYST
SUPPORTS: APPLICATIONS AND CHARACTERIZATION















By


TODD JAMES LAFRENZ


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY


UNIVERSITY OF FLORIDA

1 nc












ACKNOWLEDGMENTS


My journey through graduate school has been long and arduous,


oft inclined to take the more difficult path to the mountain top.


as I am


However, I am


also leaving a much wealthier person in terms of knowledge and fond


memories.


During this time I have learned much about myself, my friends,


human kind, and in no way the least--chemistry.


I have realized who and what


family and friends truly are, and it is to them I express my sincerest thanks


and gratitude.


When one looks around and tries to understand the diversity


and complexity of the world in which we live, it is only through the love and

support of family and friends that we can continue to strive for a better world,


and cope with our failures and disappointments along the way.

these same people that we share our joys and accomplishments.


It is also with

There are too


many to list by name, and I do not wish to leave anyone out, but I would like to

acknowledge one person specifically who has had the greatest impact on my


This recognition goes to one of the most important women in my life--my


mother, Joan Barnett.


It has been through her constant support and by her


own example that I have been successful in what I have undertaken.


I feel very


fortunate to have the large number of wonderful friends and family that I do, so

let me just say thank you once again to everyone.

Then where would my chemistry career be without the instructors and


advisors who had their impact? It is t

how's and why's of scientific research.


hese people who have shared with me the

First there was Dr. Sandra Etheridge at








R.K. Birdwhistle at the University of West Florida, who never allowed me to


submit or accept a "bullshit" solution to any chemical problem.


there is Professor Russell Drago, my research advisor.


And finally,


Through his patience


and his never-ending excitement for chemistry, he has taught me much about


how to be a chemist.


I can only hope that when I am his age, I will still have


that same intensity and interest for research and living (I also wouldn't mind

having his hookshot and sharp elbows for 2-on-2 basketball).

Finally, let me acknowledge a few of the people who made my stay in


Gainesville so enjoyable.

Chemistry Department.


First I must thank the University of Florida and its

Beyond their financial support, the opportunities and


treatment I received while there could not be matched anywhere.


I must also


thank the faculty, especially my committee members, for their support and


input.


It was through many conversations with them that my education and


research progressed.


I cannot forget the support, advice and friendship I


received from the Chemistry Departments support staff, especially the guys in


the electronics, glass blowing and machine shops. Individual thanks are

extended here to Vernon Cook and Rudy Strohschein. Also included in this list


are the people at Micromeritics, especially Tommy Hill, J.R. Berrett and Tony

Thorton--it was through many visits and discussions with them that I was


introduced to much of what I know of gas adsorption theory.


Lastly, I would


like to acknowledge the Rohm and Haas Company for their financial support

during a large portion of my research.

I cannot forget the daily "behind the scenes" help of our secretaries,


Maribel Lisk and Diana Williamson.


And a sincere "Thank you!" is extended to


Ruth Drago for welcoming me as a part of the ever-growing "extended" Drano







this has been the most diverse group of people, as far as interests and beliefs,


that Ive had the pleasure to work with.


I truly believe that every one of them


helped teach me something while I was at U.F. (although I still haven't figured


out what it was for a couple of individuals).


I feel especially lucky to have had


the opportunity to meet and work with the many postdoctoral researchers that


came through our lab.

Kaufman, David Singh,


People like Drs. Robert Beer, Krzysztof Jurczyk, Phil

Garth Dahlen and Doug Bums greatly aided my


research, as well as became friends.


Finally, I must thank the Florida Men's


Rugby Club, for the great times, sportsmanship and camaraderie.


It was with


this group, and the friends I found within, that some of my fondest memories

from graduate school lie.

It is not the critic who counts; not the man who points out how the


strong man stumbles


. The credit belongs to the man who is actually


in the arena, whose face is marred by dust and sweat and blood; who


strives valiantly


. who knows the great enthusiasm, the great


devotions; who spends himself in a worthy cause; who at the best knows
in the end of triumph of high achievement, and who at the worst, if he
fails, at least fails while daring greatly, so that his place shall never be
with those cold and timid souls who know neither victory nor defeat.
-Theodore Roosevelt














TABLE OF CONTENTS

page

ACKNOWLEDGMENTS ...........................................................................................ii

LIST OF TABLES ...............................................................................vii

LIST OF FIGURES ..........................................................................


ABSTRACT


CHAPTERS


AN INTRODUCTION TO HETEROGENEOUS CATALYSIS .................. 1

CHARACTERIZATION OF CARBONACEOUS MATERIALS .................7


Carbons Under Study ..................... ........... .......................................9
Experimental ................... .......... ........ ................... ............... ...... 1
Results and Discussion ............... ............................................. ......19
Conclusion .......................................................... ..................40

III GAS-SOLID ADSORPTION EQUILIBRIA ......................... ........ .......... 42


Introduction ......... .........................
Experim ental .................. ..................
Results and Discussion .....................
Conclusion ........................................


IV HETEROGENEOUS WACKER CATALYSIS ......................................66


Background ........... ......
Experimental ................
Results and Discussion
Conclusion ...................


V CONCLUSION .................. ....................................................... ....... 90

APPENDICES


IILIIIII~IIIIIIILII LLIIIIIIIIIIIIIIII) 1I111III11II11II1)1 11)(111111)111


........................................


Background .........,,


................., ................... .........,......., .,......,.87








B NUMERICAL ADSORPTION ISOTHERM DATA ................... ........... 1 04

C FORTRAN PROGRAM FOR SIMPLEX DATA ANALYSIS ................... 114


D SAMPLE OF SIMPLEX DATA FITTING ROUTINE FOR


A-572 ..........


LIST OF REFERENCE


S ................... ................... ................... .........,


BIOGRAPHICAI, SI(E=TCH ................... ................... ............,., ..........














LIST OF TABLES


Table


page


Elemental Analyses for Carbons Under Study


BET Surface Areas and Pore Volume Data from BJH Adsorption


Curve ..........


Pore Volume Data from BJH Desorption Curve


Degas Temperature Effects on Surface Area and Pore Volumes for


PPAN


a sascasa0


Elemental Analyses for A-POLY and A-SO4 Polymers ....


.......................3


Pyrolysis Temperature Effects on Surface Area and Pore Volumes
for A -SO 4 .............. ................... ........ ........... ................................ 35


Elemental Analyses for Pyrolyzed A-SO4


"II-i


Summary of Adsorbate Gases and Associated Physical Properties .......46


III-2


Gas-Solid Adsorption Equilibria Parameters for PPAN and A-57


2 .........54


111-3


Summary of


-AHads (kcal/mol) from InKads


1/T Plots ........................56


IV-1


Summary of Heterogeneous Wacker Catalyst Conditions and


Results ...


................... ...81


IV-2

IV-3


Surface Area and Pore Volume Data for Wacker Catalysts ...................84

Mass Susceptibilities of Ambersorb 563 and 572 With Dopants ........86


................... ......20


.................... .......124


............29


.. ........36













LIST OF FIGURES


Figure


page


Schematic of Glass Manifold for Adsorption Isotherm


Measurements .....................


*msm..m.eemmemm..m .mmsmmm.m..ms.mmm...e..e......mm.m.......... i...l1


Exploded View of Specific Volumes Vk,
M anifold .............................................


Vm and V,


from Glass


TGA Results for A-POLY


II-4

II-5

II-6

II-7

III-1


................. m ......... C **CC***CCCCCCC*Ce....... .. m. m. .a.. a. mm.m .......3


TGA Results for A-S04 .33.........................................

Pyrolysis Temperature Affects on Pore Distributions for A-SO04 ............36

Room Temperature N2 Adsorption Isotherms for Various Carbons .......38

Room Temperature CH4 Adsorption Isotherms for Various Carbons .....39

N2 Adsorption Isotherms at Various Temperatures for (a) PPAN;


(b) A-572 .......


III-2


CO Adsorption Isotherms at Various Temperatures for (a) PPAN;


(b) A-572 ........


.....m.C...C.m. ... CCCCSCCC.mmmtmsC ....... ........*...*5


111-3


CO2 Adsorption Isotherms at Various Temperatures for A-572 .............53


III-4


InKads vs. 1/T Plots for CO2 Adsorbed on


4d-5?'2 .....,,..........5


IV-1

IV-2


Gas Flow Apparatus for Catalyst Screening ...............


H20 Effects on CuPd-572 Catalyst and Production of CH3CHO ............83


................... ......16


.....................5 1













Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy


CARBONACEOUS ADSORBENTS AS HETEROGENEOUS CATALYST


SUPPORTS:


APPLICATIONS


AND CHARACTERIZATION


Todd James Lafrenz

December, 1995


Chairman:


Dr. Russell


Drago


Major Department:


Chemistry


Porous materials have long been the subject for chemical research in


applications as adsorbents and as heterogeneous catalyst supports.


Many of


the traditional instrumental methods for analyzing the chemical behavior and

functionality of compounds do not readily lend themselves to the analysis of


carbonaceous adsorbents.

general amorphous nature.


This is due in part to carbon's opaqueness and

Methods do exist to help probe these materials;


the most common and widely used of these is the determination and


interpretation of gas adsorption isotherms

interpret standard adsorption isotherms.


Several theories exist to help


From these analyses, surface areas


and porosity characteristics can be calculated.

Nine commercially available carbonaceous adsorbents were analyzed

using a variety of physical methods in order to further characterize these types


'rn -nZ, nl 4 nr nl r + arnl nltnhr-oal ll annlxro-So fth rrno1 he hcvuinvr card







isotherms.


A greater understanding of the physical properties that affect the


adsorption behavior of these types of materials was gained.

A detailed and critical analysis of the application and interpretation of


the standard BET equation was undertaken.


complementary to the existing theories was proposed.


From this analysis, a model


Derived from a multiple-


site Langmuir process, this proposed equation and its solution allow for a more

direct comparison of various adsorbents with a variety of gaseous probes.

Values for both the capacities and adsorption equilibrium constants for three


distinct adsorption processes were derived using this model.


The measurement


of adsorption isotherms at a variety of temperatures allowed for the calculation

of the thermodynamic parameters associated with these adsorption processes.

The understanding gained from the analysis of these carbonaceous

adsorbents has assisted in the characterization of a heterogeneous Wacker-


type oxidation catalyst.


Salts of Pd(II) and Cu(II) were pore-filled onto a porous


carbon, Ambersorb


572,


and an active heterogeneous


gas flow reactor system


for the oxidation of ethylene to acetaldehyde was developed.


The employment


of the porous carbon as a support was found to provide synergistic benefits for

this type of catalysis, arising from the oxidative capability of the carbon itself

and its high adsorption capacity and affinity for various substrates.













CHAPTER I
AN INTRODUCTION TO HETEROGENEOUS CATALYSIS


From the days of the alchemists,


ideas of catalysis permeated scientific


thinking and research.


The alchemists' greatest endeavor was the


pursuit of


the "philosopher's


stone,


" a material which would be able to transform base


metals into gold and promote good health and long life.


Its conception was


catalytic in nature, as it was considered that one needed only a small amount


of this material to effect large changes.


Ever since these ancient times, the


roles and understanding of catalysis in chemical research and industry, as well

as the tremendous biological importance of catalysis, have continued to be


studied.


A greater understanding has been gained, but the concept of the


philosopher's stone has changed many times since.

The traditional definition of a catalyst is a compound that facilitates the

attainment of equilibrium in any chemical process much more rapidly than in


its absence.


Most often this is accomplished by the lowering of the activation


energy for a reaction.


The amount of catalyst required to do this


is usually very


small in comparison to the molar amounts of reagents and products present, so


the catalytically active species itself is used over and over again.


A catalyst


cannot affect the position of the thermodynamic equilibrium; however, it may


change the mechanism of a reaction significantly.


Due to the inherent


reactivity of many catalyst systems, one must be careful in investigating new


T, -- +. -- i.^il: n* n1* niT n 1-. a a +11 Al e -a 4 n n stirl n^ n+ Itl n1t nt- I an s-r: 11 ^


nnCnlr*C:~




2



reactivity problem in catalysis exemplifies that, in addition to present efforts in

tailoring the reactivity of systems for specific applications, there already exists

a large amount of research in the catalyst systems presently being used.

There are numerous examples in which catalytic processes have affected


the consumer market.


Included in these are hydrocarbon cracking and


reformulation catalysts used in petroleum refining, polymerization catalysts in

plastic production, hydrogenation of fats and edible oils for food-stuffs,

catalytic converters in automobiles, industrial catalysts for NH3 and SO3


syntheses,


and the medicinal value of understanding enzymatic catalysis


occurring in both plant and animal tissues.


Of the many types


of catalytic


processes that exist, all fall under the general classification of either


homogeneous or heterogeneous catalysis.


Homogeneous catalyst systems are


characterized by the existence of reactants, products and catalyst all in the


same phase.


Heterogeneous catalyst systems exist in two or more phases.


most common examples of heterogeneous systems are gas phase reactions


occurring over the surface of a solid catalyst material.


Other types include two-


phase liquid mixtures (i.e., organic and aqueous phases), and liquid-solid batch

reactor systems.

Heightened interest has been given to the heterogenizing of typical


homogeneous catalyst systems since the early 1970's.


1 Advantages of


employing heterogeneous catalysts include ease of separation of products from

the catalyst, recovery of used catalyst, and the synergistic benefits in cases


where supports are used.


The immobilization of a catalyst also inhibits its


incorporation into the product composition.


Synergistic benefits may include


r r r 'r' 1 1 1 1 1 1





3



acid sites (e.g., silicas, heteropolyacids), and concentration effects due to

adsorbent-adsorbate interactions and capillary condensation (e.g.,


carbonaceous adsorbents).


It has also been observed that the electro-chemical


behavior of metal species is affected by doping the metal on solid supports,

which may have advantages in sensors and electrocatalytic processes.2,3,4

There exist some inherent disadvantages to employing heterogeneous


catalyst systems.


In a homogenous environment almost every active catalyst


species is capable of coming in contact with the reactant and, therefore, likely


to participate in the reaction process.


In a heterogeneous system the only


catalytic sites available are those on the surface of the solid or at the interface


between two layers.


In a heterogeneous system, therefore, it is most desirable


to suspend the catalyst in a thin film, as in a membrane type system, or


disperse the active catalyst species on a high surface area material.

this, the chemical behavior of the support must now be considered.


In doing

This


chemical behavior would ideally be a synergistic benefit by design, but this is


not always the case.


When employing a solid in a catalytic process, one must


also consider its mechanical strength.


This may be a solid support, as when


employing zeolites, silicas or carbons, or a solid constructed of the catalytic

species itself, as in the case of various metal oxides or nickel/platinum


hydrogenation catalysts.


There are obvious concerns if the solid decomposes


mechanically or chemically under the desired reaction conditions.

Despite these apparent disadvantages, heterogeneous catalyst systems

have proven to be very useful and applicable to many commercialized


processes.


A large amount of research continues in all areas of heterogeneous








gas phase chemistry.


Analyses of the chemical structures of carbonaceous


solids and macroreticular polymers are further complicated by their

opaqueness, and their highly cross-linked network structures which render


them insoluble in most traditional solvent systems.


Methods do exist that help


characterize solid materials, including solid state nuclear magnetic resonance

(NMR), X-ray powder diffraction, X-ray photoelectron spectroscopy (XPS), magic

angle spinning NMR, diffuse reflectance infra-red fourier transform (DRIFT),

photoacoustic infra-red spectroscopy, electron paramagnetic resonance (EPR),

low energy electron diffraction spectroscopy (LEEDS), high resolution electron

energy loss spectroscopy (HREELS), scanning electron microscopy (SEM),


scanning tunneling microscopy (STM)


atomic force microscopy (AFM), and


extended X-ray absorption fine structure (EXAFS).


However, many of these


techniques only provide qualitative information, or are limited to probing the


surface or only a few molecular diameters beyond the surface.


introduction of '29Xe-NMR experiments6,7 and the combination of calorimetric

and spectroscopic data8S9 on solids have been helpful in elucidating internal


structural and surface information


their applications.


, but these techniques are also limited in


It remains difficult to characterize and understand the


intricate chemistry that occurs within the internal porosity of solids employed

in catalysis.

The ease with which gaseous or liquid reactant molecules can achieve

access to the internal surfaces and pores of a solid is often of crucial

importance in the development of a catalyst suitable for industrial applications.

If one is to be able to understand, and then further predict catalytic behavior, a
- -- .5 5 4 C .4<* i I




5



isotherms continues to be an elucidating method to probe the internal porosity


of these materials.


Various theories have been developed to help interpret gas


adsorption isotherms to yield surface area and pore size information of porous


solids. 10o,1


It is well accepted in the literature, however, that many of these


interpretations are derived while making several assumptions, some of which


are considered to not be truly representative of the surface being probed.


qualitative discussions and direct comparisons of data under the same

conditions are considered valid.

Carbon containing polymeric adsorbents and their derivatives are

quickly expanding as the newest areas of interest for heterogeneous catalysis.

However, because most carbonaceous materials are opaque and amorphous,

they do not lend themselves readily to many of the characterization methods


previously mentioned.


Because of growing interest in the application and


characterization of carbonaceous solids as adsorbents and in catalytic

applications, it is of great interest to further quantify their chemical and


physical behavior.


Therefore, an understanding of the interactions involved in


the physical and chemical adsorption of gases by solids is of fundamental

importance to many of the applications of these types of materials.

In this series of studies, numerous high surface area carbonaceous


adsorbents have been characterized.


Comparisons of their surface areas and


pore distributions, as well as their physical composition, have been made.


model is presented to help interpret gas adsorption isotherm data in attempts

of obtaining more quantitative measurements for comparing these types of


materials for adsorption and catalytic applications.


One of the goals of this


-,,,, I,. L -- a,.,,:, -t, lr- C .. ---- .41. .-...a .-




6



to assist in the choice of an appropriate solid for whatever application is

envisioned.

Finally, the understanding gained in characterizing carbonaceous solids

has been applied in the development of a viable, heterogeneous catalyst


system.


This system involves a heterogeneous Wacker process, specifically


looking at the oxidation of ethylene to acetaldehyde.


Many factors have been


found to affect the activity of these types of catalysts, including the surface and

pore characteristics of the support, dopant levels, temperature and flow rates of


gaseous reactants, and the components of the active catalyst species itself.


of these parameters were considered in detail in the development of this

system.













CHAPTER II
CHARACTERIZATION OF CARBONACEOUS MATERIALS


Background


There exist a large variety of carbonaceous materials derived from both


natural and synthetic polymeric materials.


Natural sources include wood fiber,


coconut husks, peat moss, petroleum pitch and coal tar residues.


Synthetic


precursors include polymers of styrene, acrylonitrile, vinyl and furfuryl


alcohols, and phenolic resins.


Applications of the carbonized derivatives of


these materials include adsorbents for purification of gaseous and liquid

streams, pressure-swing applications for gas separations, chromatographic

stationary phases, heterogeneous catalyst supports, prosthetic devices in

medicine, and composite and electrode materials.12

Pyrolysis of many of these polymers produces carbons that can be


tailored for specific physical characteristics.


These characteristics are


determined by the original polymer structure, as well as the pyrolysis


conditions and pretreatments.


Typical carbonization conditions involve heating


the materials under an inert atmosphere (e.g.,


3000C to 30000C.


N2, Ar) from temperatures of


The porosity and extent to which these materials graphitise


(at temperatures >20000C) is largely dependent on the properties of the original


polymer and the carbonization conditions.13


Elimination of various volatile by-


products during pyrolysis produces additional porosity in the remaining solid.14








Various treatments of these polymers are employed prior to pyrolysis to


increase their thermal stability, therefore increasing carbon yields.


naturally occurring polymers, this may include chemical treatments with

agents such as ZnCla, carbonates, alkali, sulfates, and sulfuric or phosphoric


acids.


This results in dehydration and further cross-linking of the polymers,


which produce higher carbon yields unaccompanied by tarry residues from


pyrolysis.16


Synthetic polymers are made as macroreticular polymers to impart


high thermal stability.


Macroreticular polymers are characterized as being


highly cross-linked, exhibiting little or no swelling upon wetting with a solvent,


and have an open, rigid macroporous structure.


Advantages of synthetic


polymer precursors include greater control over the homogeneity of the starting


material


, and greater reproducibility from batch to batch.


Synthetic polymers


may also be imbibed or treated with chemical agents as in the case of natural


polymers.


The resultant benefits of a highly cross-linked polymer system, as


well as the carbon-fixing agents previously mentioned, are to allow the carbon


to char without fusing, therefore retaining its original pore structure.


process referred to as activation may also be employed during or after pyrolysis


to impart various functionalities to the surface or increase porosity.


includes treatment with gases such as steam, air, NH3,


This


HC1, C12, or CO2, at


temperatures that typically range from 3000C to 10000C.

The resultant materials are generally amorphous in nature and best

described as consisting of twisted networks of defective carbon layers bridged


together with other similar layers.


Although they do not possess uniform pore


sizes as in the case of zeolites, they are found to contain unique pore shapes


a~~~~~~~I 5 0 -. W-5


1








structure and surface functionality.17


With this capability, carbon has


advantages as an adsorbent or support over other traditional materials like


zeolites or silica.


Advantages may include the absence of strong Lewis acid


sites as seen with aluminum in zeolites.


Additional advantages may include


solution and mechanical stability, carbon's general stability in many corrosive

environments, and high adsorption capacities for various organic substrates.

The very nature of carbon will lead to supported catalyst systems that behave

differently than other types of supports currently used.


Carbons Under Study


Ambersorb


carbonaceous adsorbents, made by the Rohm and Haas


Company, are among the materials that have been studied.


Derived from


macroreticular, sulfonated polystyrene, the pyrolyzed product has shown


interesting behavior as both an adsorbent and catalytic support.


The original


polymer is made through a suspension polymerization method, with


divinylbenzene as a cross-linking agent.


The resultant polymer is in a bead


form, which is retained during pyrolysis and mechanically very strong.


Two


types of Ambersorb carbons were examined Ambersorb 563, lot# 91/3298

(A-563) and Ambersorb 572, lot# 2125 (A-572). Together with differences in


pore distribution and surface area, A-563 is reported to be more hydrophobic


than A-572.


More information on the physical properties, adsorption


characteristics and syntheses of these materials can be found in the text by

Neely and Isacoff,is and in the patent literature.16'19


Pyrolyzed polyacrylonitrile (PPAN) was also investigated.


This material, a


- -r .---








the past several years because of its unique mechanical strength, catalytic


behavior and electronic properties.20,21,22


However, there is a concern in


developing any large scale synthesis because of HCN production during

pyrolysis.

Two different carbons were obtained from the Calgon Carbon


Corporation.


Calgon BPL


granular carbons are derived from bituminous coal.


Their production involves a steam activation treatment, resulting in a high

surface area, highly porous material, suitable for vapor phase adsorption and


as a heterogeneous catalyst support.


For example, these carbons are used in


the acetylene process for the production of vinyl chloride and vinyl acetate.

Filtrasorb 300, also made from coal, has been tailored to be an active


adsorbent in water purification.


Calgon BPL 4x10 (BPL) and Filtrasorb 300


8x30 (F-300) were the specific materials examined in these experiments.

A powdered carbon, AX21, was obtained from the Anderson


Development Company.


This material is made from a patented process


involving direct chemical activation of petroleum coke with large amounts of


KOH at elevated temperatures.23


Further details on the preparation and


properties of this material are reported elsewhere.24,25


AX21 has also been


previously studied in this research group as a catalytic support.26

Another powdered material examined was NORITTM 211 from Kodak (N-


Labeled as a decolorizing carbon, it is derived from pyrolyzed peat moss.


This is further treated with a steam activation step to increase porosity,


followed by an acid wash to leach out impurities.


was Norit N.V


The original manufacturer


., and the NORITTM line is now being distributed by Fisher








spherical bead like the Ambersorb


and PPAN materials, XUS-43493.01 (DOW


493) is designed for the concentration of organic from both air and water.


reported to possess a unique pore structure and high surface area, with good


physical strength and a hydrophobic surface.


These materials are reported as


having very little catalytic activity themselves due to extremely low ash content.

This is in contrast to some activated carbons which contain residual metal


salts from their processing or original sources.


An advantage of low ash


content would be the minimization of adsorbate decomposition, which is not

desirable in some applications.

The final carbon considered in these experiments was Kureha BAC-SP.

This is a spherical, activated carbon made by Kureha Chemical Industry Co.,


Ltd. in Japan.


These "bead activated carbons" are spherical and reported as


having a fairly uniform diameter of 0.7 mm.


They are reported to have a low


ash content and very high resistance to attrition on repeated usage.

The nine carbonaceous solids described above were chosen in order to


examine a broad range of carbon materials.


Most are commercially available,


and are derived from a variety of both natural and synthetic precursors.


wide range of composition, pore size distributions, surface areas, and

functionalities made these ideal materials for the focus of these investigations.


Experimental


Characterization of Supports


All carbons were used as received without further purification or


- S t -r -r -


.. C








gases employed in adsorption studies were purchased from the Matheson Gas

Company with a minimum 99.99% purity, and were used without any further


purification.

Micromeritics


Surface area and pore information were obtained on a


ASAP 2000 instrument.


Elemental analyses (carbon, hydrogen


and nitrogen) were performed by the University of Florida elemental analysis

laboratory.

Surface areas were determined using a five point Brunauer-Emmett-


Teller (BET) type analysis27 of the N2 isotherm obtained at 77K.


For this entire


range of materials, the best fit line for the BET transformation plot was found


to be at 0.01< P/Po


< 0.07 (see Results and Discussion for further details).


Micropore volumes were calculated using the Harkins-Jura (HJ) model28 with a


t-plot statistical analysis.29 The statistical thickness parameters used for the

best fit line were from 5.5-9.0A. An algorithm employing the Barret-Joyner-


Halenda30 (BJH) model was used to determine mesopore and macropore


volumes from the adsorption portion of the isotherm.


All the above


calculations were carried out using the software supplied with the


Micromeritics


ASAP 2000 instrument, and further details can be obtained


from the instrument manual.31


The interpretation of pore volumes follows the


IUPAC recommended definitions with micropore diameters smaller than 2 nm,

mesopores between 2 and 50 nm, and macropores having pore diameters larger


than 50 nm.


32,33


Thermal gravimetric analyses (TGA) were obtained on a T.A.


Instruments, Inc., Thermal Analyst 2000 system under flowing nitrogen.


This


instrument was equipped with a Hi-Res TGA 2950 module, differential









ramping rate dependent on the rate of weight loss.


The maximum temperature


ramping rate was set at 100C/min, with a resolution setting of 5.00.


This


enhances the resolution of weight changes, while minimizing run time and

avoiding transition temperature overshoot.


Gas Adsorption Studies


Gas adsorption isotherms were obtained on the Micromeritics


2000 Chemi system, or on a glass manifold line (Figure II-1).


ASAP


This manifold


system consisted of a glass chamber of known volume; threaded Teflon high

vacuum stopcocks (Ace Glass, 8190 series); a Welch vacuum pump and

mercury diffusion pump; two pressure heads from MKS Baritron, type 390HA

(one capable of measurements from 1-1000 torr, the other from 10-5-1 torr);

MKS Baritron type 274 channel selector, 270B signal conditioning unit, and a


390A sensor head; and various condensers and connecting tubing.


The system


itself was not thermostatted, but temperatures were found not to vary more

than +1C over the course of a run.

The procedure to obtain isotherm data on the glass vacuum line, known


as the successive addition technique, is a static, gas volumetric technique.


employs a known volume and changes in pressure to determine the amount of


gas adsorbed.


Initially, after the sample tube has been connected to the


vacuum line via an o-ring seal, the entire system is evacuated and the sample


degassed.


After closing the valve to the sample, the manifold (Vm) and known


volume (Vk) are filled with helium, a gas that is assumed to be non-adsorbing


(see Figure II-2 for further detail of specific volumes).


The valve to Vk is closed,


































































































































* I i II l i Ii llll l l li I Ii
lLI Li ll l l lii i II II i i I

L i ii i i Ii i i ii ii ii i tiil


oaa









allowing it to equilibrate with Vm yields P2, the equilibrium gas pressure.


Assuming ideal gas behavior,


Vm was then determined by the following


equations:


= PiVk/RT


(11-1)


where ngas is the total number of moles of gas in the system, R the ideal gas


constant


, and T the temperature (in Kelvin).


Since R, T and ngas stay constant,


when Vk is opened up to Vm:


PiVk


= P2(Vk+Vm)


(11-2)


which rearranges to


PI
Vm = Vk --
P2


(11-3)


This series of steps can now be repeated, where P2 becomes Pi, the valve closed

and Vm again evacuated, and then reopening Vk to Vm gives a new value for P2.

This sequence is repeated a total of five times, with progressively lower gas

pressures, to give a determination of Vm over the entire range of experimental

pressures, as well as a measure of the accuracy and reproducibility of these

measurements.

Next, it is necessary to determine the free space volume (Vs), which

encompasses the volume above the sample in the sample tube and the volume

between and within sample particles (sometimes referred to as void volume).

This procedure, continued from above, is done in a similar fashion as in

determining Vm, however, this time Vm is used as the known volume to find Vs.

An initial pressure exists in the manifold (P'2) and in the sample volume (P'i).


Obviously, for the first data series P'1


pressure is obtained


Vm is then opened to Vs, and a new


The initial moles of gas present can be expressed as


-~~n l 7 jlPr..'


I n










P' IVs


since n3 = ns+nm:


P'2Vm


P'a(Vs + Vm)


(11-7


which rearranges to


\13=


(11-8)


- P'2
- P'3


gas supply


vent/bubbler


/-^l (


manometer


menometer


sompie


I.


I


tt
)fS




~T3









This series of measurements is then repeated over the entire experimental


pressure range (0-760 torr), as was done in the determination of Vm,


by the


following method.


After closing the valve connecting Vm and V


dosed to a higher pressure yielding a new value for P'2.


the manifold is


The previous


value of


P'3 becomes the P'1 value for this data series, and a new value for P'3 is


obtained when Vm and Vs are allowed to equilibrate.


obtained for V


As before, the values


s over this entire pressure range are averaged, giving an estimate


of the precision of these measurements.

strong adsorption of helium by the samp]


A large deviation may be indicative of

le. This was not observed for any of


the samples investigated.


With Vm and Vs now known, gas adsorption isotherms can be obtained

using the particular adsorbate of interest. Data were obtained in a similar

fashion as to that in the determination of Vs. After the manifold, known volume


and sample volume were evacuated, the sample valve was closed and Vm again

determined from Vk (Eqns. II-1 through II-3 and procedure outlined therein).

This was repeated to help minimize any errors due to non-ideal behavior of the


adsorbate gas, as well as for verification of prior results.


Any large deviation


from the value of Vm determined using helium would be indicative of the probe

gas not behaving ideally, or it strongly adsorbing to the walls and seals of the


manifold itself.


This was not observed with any of the gases used in these


studies.

After determination of Vm with the adsorbate gas, pressures were

obtained in a fashion similar to that used for determination of Vs with helium.

The initial pressure contained in the manifold (P"2), which was filled to specific









equilibrium pressure measurement, P"3.


The number of moles of gas adsorbed


by the solid (nads) was found from the difference in the number of moles present

initially in the gas phase (nini) and that present after equilibrium was allowed to

be established (nequil):


nads


= nj~~-nii


(11-9)


where


P'2Vm P"'Vs
nini = -- + --
RT RT


nequil =


P' '3(Vm + Vs)


(II-10)


(11-11)


Substituting equations II-10 and II-11 into II-9 gives


nads =


(P' '2Vm + P' 'iVs)


- P' '3(Vm + Vs)


(II-12)


This value was then divided by the mass of solid sample to give an amount

adsorbed per gram of adsorbent.


Pressure readings were recorded in torr,

yield pressures in units of atmospheres (atm).


but were all divided by 760 to

Equilibrium was considered to


be established when the pressure reading did not change 2 units over a period

of 30 seconds (values were recorded to four significant figures).

The Micromeritics ASAP 2000 Chemi system performs its analyses and

calculations in a similar fashion. Further details on the free space


determination, adsorbed gas volumes and equilibration intervals can be found


in the ASAP 2000 Chemi reference manual.34


Equilibrium was considered


attained for runs on this instrument when the pressure change per ten second


time interval was lea.a


than 0 1/% nf the average nreasire d urinu that time





19



and 1 atm, as most interest in these materials involves applications under

standard atmospheric pressures.


Results and Discussion


Elemental Analyses


Because of the variety of sources and precursors for the various carbons

under study, an initial determination of their physical composition was


desirable.


Since all of these materials have been pyrolyzed under relatively


high temperatures, it can be assumed that their chemical composition is

mainly carbon, with small amounts of hydrogen and oxygen, as well as

nitrogen if it was present in the original polymeric precursor as in the case of


PPAN.


Mechanistic details on the decomposition under pyrolysis conditions for


polymeric precursors of these materials are well characterized and have been


reported. 13,35,36


Other chemical components are assumed to have been present


in the original material as impurities, or incorporated during their preparation.

All characterizations and isotherms were obtained on the same batch and lot

number of a specific material.

Elemental analyses were done on all carbons as received, without


further washing or treatment.


The rationale was that this would most likely be


the way these materials are used in the development of large scale production


for a commercial catalyst or adsorbent system.

exhibited less than 5% weight loss upon drying.


The majority of the carbons

Exception to this was


observed for N-211


,PPAN, and AX21.


However, CHN analyses were obtained







Table II-1.


Elemental Analyses for Carbons Under Study


PPAN

A-572

A-563

N-211

BPL

F-300

Kureha

DOW 493

AX21


AX21


dried)a


70.18

91.13

83.57

85.24

88.82

89.60

95.56


44.43

87.13


1.66

0.33

1.70

0.46

0.50

0.34

0.67

7.05

2.60

0.31


5.31

0.00

0.00

0.39

0.70

0.55

0.04

0.02

0.07

0.15


total %CHN


77.15

91.46

85.27

86.09

90.02

90.49

96.27

94.32

47.10

87.59


a Dried under standard conditions for gas adsorption experiments,


approx. %

weight loss

upon dryinga


which


involves degassing at 200C under a vacuum of <10-3 torr for a minimum of 8
hours.

the weight loss upon drying accounts for the majority of the remaining mass


balance.


Because of this, it was assumed that most of this weight loss occurs


simply due to loss of adsorbed water


atmospheric gases.


and, to a much lesser extent, adsorbed


Any residue left after CHN analysis was attributed to metal


salt impurities, as alluded to earlier, or other functionalities existing due to


activation treatments. These may include groups containing oxygen, sulfur,

phosphorous or chlorine. It is well accepted in the literature that these types of


functionalities can exist on carbon surfaces,12 but specific analyses for other


functional groupings or metal salts were not done on these materials.


The data







Surface Area and Pore Volumes


In employing highly porous solids as adsorbents and catalyst supports,

an accurate determination of the physical structure of the surface, internal


porosity and interconnecting channels is desirable.


One of the standard


techniques used for the determination of this information employs the N2


adsorption and desorption isotherms obtained at 77K.


Nitrogen is most often


chosen as the adsorbate in these measurements because its small molecular


size allows it access to small pores.


Furthermore, it is easily handled,


economical, approximately spherical, and inert, hence not likely to participate


in chemisorption.


Other gases can be used,


the most common of which are


krypton or argon.

Various theories exist to interpret isotherm data leading to the


determination of surface areas, pore sizes, and pore volume distributions.


37,38


Determination of surface area turns out to be very complex in the absolute


sense, however.


Factors that may affect the determination of surface areas


from standard methods that contribute to the inaccuracies can include

irregular surface features, the fact that most surfaces of interest are


energetically heterogeneous, resolution of the method,


molecule being used.39


and the size of the probe


Although certain assumptions have to be made in the


interpretation of gas isotherm data, the relative ease of data acquisition and the

non-destructive nature of the technique give N2 adsorption a large advantage

over other methods such as mercury porosimetry.

The BET equation (Eqn. II-13) was employed in determining surface

areas from gas adsorption data obtained on these carbons, employing N2 gas as









nads(l


1 (C-1)
C nm C m


(II-13)


The relative pressure, X, is P/Po, where P is the measured equilibrium pressure

(in torr) and Po is the adsorbate saturation pressure at the given temperature.


The value nads is the molar amount adsorbed at the relative pressure


values of nm, the monolayer capacity, and C, a constant dependent on isotherm


shape, are both determined from the best fit linear plot of


n(1-X)


vs. X


After


determination of nm and C


the surface area can then be calculated from


S.A.BET


where NA is Avagadro's number


nm-NA'-a


6.022x1023


(11-14)


and a is the cross-sectional area


of the probe molecule.


For N2, a value of a


= 0.162 nm2 was used.


This is yet


one more point of debate in the usage of gas adsorption data for the

determination of surface areas, and several publications have discussed how

the cross-sectional areas of adsorbed gases vary dependent on both adsorbent


and temperature.40,41


The appropriate conversion factors are combined with


Equation 11-14, and the final value divided by the mass of solid in the

experiment yielding a value for the BET surface area (S.A.BET) in units of m2/g.

The multi-point methods used in these studies for surface area


determination are derived strictly from equations II-13 and 11-14.


For this


multi-point method, analysis of the BET equation (Eqn.


II-13


finds that


plotting T versus P/Po yields a straight line, where Tis


nads _
P


(11-15)


where nads is the number of moles of gas adsorbed per gram of solid at the


mnaen1nred relative nreR'alir.


Thep lep t-snnrlrpe fit analvsis nf the line nassing








1/C'nm


(11-17)


Simultaneous solution of the two equations gives nm and C, where nm is then


used to calculate the S.A.BET.42


The Micromeritics software employs this


relationship to find surface areas, employing the volume of gas adsorbed


instead of the number of moles.31


the molar gas volume.


These two quantities are related simply by


Therefore, in the case of the isotherms measured on


this instrument, Equation II-15 becomes


Vads(I~


(I-18)


and the S.A.BET (in units of m2/g) is calculated from the slope and intercept of

this line by


S.A.BET


eoNA


22414


(S+Yint


( (1I-19)


Data points from the adsorption isotherm chosen for the determination of the


S.A.BET are typically chosen between relative pressures of 0.05 and 0.30.

this region it is found that BET theory predicts adsorption values below


experimental values.


Below


Micropore filling arising from high adsorption potentials


is the major cause of this inaccuracy.


Above P/Po


0.30, multilayer adsorption


occurs, and BET theory is found to predict an adsorption value that is too high.

In this research, five points were chosen from the region of the isotherm

that yielded the best fit line in the BET transformation plot (Eqn. 11-18).


Although the chosen data region, 0.01


< 0.07


for these carbons occurred


near and below the lower limit for BET analysis previously mentioned, these


limits are not exact and do vary from solid to solid. The presence of large

micropore volumes also affects the choice of this data set. In these analyses,


10'"








was found to be greater than 0.99999.


As previously mentioned, the


application of BET theory is based on several assumptions, and surface area


determination is very difficult in the absolute sense.


Therefore, for these


materials the choice of the best fit line and the use of the same data region for

each solid was assumed to be the most appropriate for making relative


comparisons.


The experimental data set used in applying BET and the pore-


volume theories (BJH and HJ) should always be kept in mind when making

quantitative comparisons based on these types of reported values.


Table II


BET Surface Areas and Pore Volume Data from BJH Adsorption
Curve


CARBON


PPAN

A-572

A-563

N-211

BPL

F-300

Kureha

DOW 493

AX21


Surface Area


m2/g)


880

1159

606

762

1075

1026

1382

1340

2745


Pore Volumes (cc/g)


micropores


0.334

0.428


0.22


0.263

0.432

0.390


0.498

1.160


mesopores


0.119

0.284

0.206

0.188

0.079

0.143

0.062

0.314

0.318


macropores


0.090

0.207

0.226

0.050

0.009

0.022

0.007

0.324

0.016


BET C

constant


1685

866

1850

888

313

416

287

328


The data presented in Table II-2 shows the determined surface areas and


pore volume distributions for the carbons analyzed.


As was mentioned earlier,








pore volume distribution curves in Appendix A is made here, as well as the


data presented in Table 11-2.


From this it can be seen that a variety of


carbonaceous materials with a large diversity of relative surface areas and pore

volumes have been considered.

The initial steep slope of the experimental isotherms (see Appendix A) is

attributed to a large enhancement to the energy of adsorption due to the

presence of pores with diameters of only a few molecular diameters (with

reference to the probe molecule), and narrow openings to larger pores in the


internal volume of the porous solid.


This large interaction potential arises from


the close proximity of opposing walls, resulting in a large volume of the

adsorbed gas being concentrated in the micropore volume at pressures below

its normal saturation pressure, to the point of condensation of the adsorbate.

This large potential, which might be accounted for by the occurrence of

chemisorption, was ruled out in the case of these carbons as all of the N2

isotherms were found to be completely reversible.

Following this steep slope in the adsorption curve attributed to


micropore filling, the isotherm levels off.


The knee in the curve in this region


occurs near the monolayer coverage, or filled micropores in the case of


microporous solids.


In this leveled region of the isotherm, monolayer and


multilayer formation occurs, which is followed by another steep increase in the


slope.


This second large increase in adsorption is attributed to mesopore filling


due to capillary condensation.


The final leveling off of the adsorption curve


represents attainment of the adsorption capacity of the porous solid.


multilayers thus formed more closely resemble the bulk liquid than the








packing of the molecules.43 For these reasons, the usage of BET t

surface area determination is not considered an absolute method.


theory for


For similar


samples run under identical conditions, however, relative comparisons are


considered valid.


It is on this assumption that direct comparisons were made


for the various carbons and the data presented in Table 11-2.

The overall shape of the N2 isotherms (Appendix A) can be classified as


Type II or Type IV according to the BDDT classification scheme.44


Exception to


this can be seen in the case of Kureha, BPL and AX21, which more closely


resemble the Type I isotherm.


Analysis of the data in Table II-2 and the pore


distribution curves in Appendix A shows that these three solids are almost


entirely microporous.


These classifications are indicative of porous solids, and


interpretation of the various regions of these isotherms follows the description


in the previous


paragraph.


In general, reversible isotherms exhibiting a steep


increase in adsorbed volumes at low relative pressures are indicative of highly


porous materials and thus, the presence of micropores.37S38


This steep increase


will yield a greatly enhanced value for the BET C value (> 300), and a surface

area that may be erroneously high.

The difference between these three isotherm types is that Type I


isotherms are generally observed for solids that are mainly microporous.


steep increase is observed in the initial low pressure region of the isotherm,


with a leveling off over the remainder of the pressure range.


The shape of the


Type II and IV isotherms are very similar, where Type IV is generally seen to


level off near P/Po =


1.0 and a hysteresis often observed, yielding further insight


into the pore size distributions and shapes.


More detail on the meaning and


5 A' t 5 4 *r 4 4 it A1 flf inr 'n f









curvature on the condensed surface, and the lower the vapor pressure of the


condensed phase.


Therefore, upon desorption of an adsorbed gas, a lower


pressure must be reached before the adsorbate desorbs and a hysteresis loop is


thus observed.


A similar but opposite effect would arise from narrow entrances


to larger pores (ink-bottle pores).


The restricted opening, and hence the small


diameter and greater curvature of the surface of the condensed phase at the

pore opening, would make it necessary for a greater pressure to be reached


before the pore would fill.


The rapid filling of this pore volume would result in


an erroneously high value for the volume of the smaller pore size, which was


derived from the pore entrance diameter.


This would also contribute to the


occurrence of a hysteresis in the adsorption/desorption isotherm.

The relationship between the vapor pressure and the condensed film

surface radius, in addition to the thickness of the adsorbed layers, provides the


basis of BJH theory.30o


Pore volume distribution plots of the BJH desorption


data can be found in Appendix A.


The y-axis in these plots, dV/dlog(D),


generated from the following equation:


dlog(D)


log(DpI/DpI +


1) (11-20)
1)


where Vp, is the adsorbed gas volume (cm3/g @ STP) at data point I,


and Dp


and Dp+i are the pore diameter range over which the BJH calculation was

done. Graphed in this fashion, the change in pore volume per change in


magnitude (change per decade


of pore diameter is represented.


From this graphical data, it was observed that the majority of the

carbons analyzed generally have a bi-modal distribution of pore size ranges,


-t r.. ..~ a_~


n_ t.~_r_.. __ n r~n








around 320A.


In the case of Kureha, almost all of its pore volume is contained


in the micropore region (< 20A), with a small peak at -40A.


The BJH pore


volume distribution plots compliment the data reported in Table 11-2, and show


these materials have a limited


, well defined pore structure.


This information


can be used to further help in determining which adsorbent would be the most

appropriate choice for a specific application.

In general, it is proposed that the BJH desorption data is more


representative of the surface than the adsorption data.


However, both data


sets are subject to error from effects of narrow pore openings into larger pores


and failures in the assumptions made by the model.


The BJH model assumes


a cylindrical pore shape, which obviously cannot be rigidly correct as the

majority of these pores consist of irregular spacings between carbon particles.

Consequently, the reported pore sizes and volumes cannot be taken as

absolute, but may be interpreted as a statistical average of irregular pore


shapes.30


It is therefore more important in making comparisons of a variety of


solids to be consistent in data manipulation, and for these experiments the

BJH adsorption curve was used to calculate mesopore and macropore volumes.

For point of interest, the mesopore and macropore volumes determined from


the BJH desorption curve are reported in Table 11-3.

later discussion the values reported in Table 1I-2, bE


However, for purposes of


ised solely on the BJH


adsorption curve, will be the only values used.

The BET C constant is related to the increased effect on the heat of


adsorption for these systems.


For the determination of a reliable value of nm,


the monolayer coverage, it is beneficial to have a sharp knee in the initial








Table 11-3.


CARBON


PPAN


A-57


A-563

N-211

BPL

F-300

Kureha

DOW 493

AX21


Pore Volume Data from BJH Desorption Curve


Pore Volumes (cc/.)


mesopores


0.203

0.496

0.441

0.022

0.113

0.158

0.090

0.427

0.354


macropores


0.009

0.001

0.002

0.215

0.003

0.008

0.002

0.196

0.006


spherically symmetrical and tend not to localize on the surface.


Non-


localization demands a low value for C, as a large value is indicative of strong


interactions between the adsorbate and adsorbent.


The stronger the


interaction between the two, the more an effect the adsorbent lattice has on the


orientation and geometry of the adsorbed molecule.45


The BET C value has


also been related to the enthalpy of interaction by the following equation:46


In C


= (AHads- AHi)/RT


(11-21


where AHads is the adsorption energy and AHi the heat of liquefaction of the


adsorbate.


However, it is generally accepted that, although the value of C may


give an indication of the magnitude of the adsorbent-adsorbate interaction and

the isotherm shape, this value is not usually directly translatable to adsorption

enthalnies.32





30



Because of this, degassing at elevated temperatures under vacuum before


isotherms are obtained is necessary.


The data in Table II-4 illustrates the effect


of degas temperatures on surface area and pore volume for PPAN.


It has been


reported that for degassing of microporous, carbonaceous materials


temperatures of at least 3000C should be employed.


However, original


analyses included some of the polymeric precursors to the carbons reported

that were known to be thermally unstable at these higher temperatures (see


Figures 11-3 and 1I-4).


For this reason, a lower temperature of 2000C was


chosen for these gas isotherm studies.


From the data presented in the


following table, it was observed that increasing the degas temperature had its

greatest effect on increasing the surface area and micropore volume.


Table 11-4.


Degas Temp.


Degas Temperature Effects on Surface Area and Pore Volumes
for PPAN


Surface Area

(m2/g)


micropores


0.305

0.333

0.351


mesopores


0.120

0.119

0.118


macropores


0.086

0.090

0.094


BETC

constant


1549

1685

1714


TGA, Pyrolysis and Pore Volumes


Samples were obtained, courtesy of the Rohm and Haas Company, that

were similar in chemical composition to the precursor materials for the


syntheses of Ambersorb


adsorbents, for the purpose of studying the thermal,


surface area, and pore volume characteristics of the polymeric materials used


Pore Volumes Ice








considered representative of this class of compounds.


These analyses were


found to yield interesting information about the effects of functionality and

thermal treatments on the resultant pyrolyzed solids.

Two solids were analyzed with a variety of elemental analyses, surface


area and pore volume studies, and thermal gravimetric analyses (TGA).


One of


these materials, referred to as A-POLY, was a macroreticular copolymer of


styrene and divinylbenzene, with no additional functionality.


SO4, was a similar polysulfonated copolymer.


The other, A-


Both of these materials were in


bead form, made through suspension polymerization, and were used as


received without any


Table 11-5.


further treatment.


Elemental Analyses for A-POLY and A-SO4


total %CHN


approx.


% weight


loss upon drying


A-POLY

- pyrolyzeda

A-SO4

- pyrolyzeda


91.66

92.56

39.19

82.78


8.07

3.44

5.22

2.09


0.00

0.00

0.00

0.00


99.73

96.00

44.41

84.87


a The CHN results for the pyrolyzed samples were obtained from the chars
resulting from the TGA analyses (Figs. II1-3 and II-4) on the original polymeric
materials.


The two polymeric materials were both subjected to pyrolysis conditions,

and weight loss as a function of temperature was analyzed employing TGA.

Note that the major difference between these two materials is that A-SO4 has


been sulfonated, where A-POLY has not.


A... n) + l;n~:n


* 1 ** l***III UIli ~li ~ ll


It was observed that without


thiQ tvne rfnrrlU-rrn.r .qvten, rPedilv dea-rlde.s nt ternnern tliire










(UTw)


awrl'


C---]--


q4 wt


C (D CUl









[uT;W)


-- -2-


0 0 0 0
CD(0 C "U


awFI.




34



sulfonation of these polymers greatly increases their thermal stability, where


-45% of the original mass still remains at 6000C (Figure II-4).


a 94.2% carbon yield.


This represents


Elemental analyses for these two polymers and their


resulting chars are presented in Table 11-5.


One can infer from this information


that approximately 44% of the mass of A-SO04 was due to the presence of the


sulfonate groups.


The data from the last column in Table II-5 shows that


functionalization of this polymer greatly enhances its tendency to adsorb


atmospheric moisture.


The initial weight loss in Figure II-4 between room


temperature and 2000C is attributed to the loss of this moisture, and this value


agrees well with the observed weight loss reported in the table.


phase of weight loss, observed above


The second


2000C, is reported to be due to the loss of


the sulfonate functional groups,36 and this is in agreement with the results


observed in the TGA.


At 6000C, approximately 55% of the original mass has


volatilized, which is in agreement with the loss of sulfonate functionality and

adsorbed moisture inferred from the CHN data.

To observe how pyrolysis temperatures affect pore distributions and

surface areas, a series of experiments was performed with A-S04 varying the


maximum temperature by increments of 500C up to 6000C.


These runs were


done by placing -1.5g of A-SO4 in a porcelain boat, placing this inside a quartz


tube in a tube furnace


, and performing the pyrolysis under flowing N2 gas.


tube furnace (Hevi-Duti Electric Company, type M-3012) was controlled with


an Omega CN-2042 programmable temperature controller.

purged with nitrogen before elevating the temperature, anm


All samples were


I the ramp rate was


constant for all, at


2C/minute.


The sample was held at the maximum


F (r r r 11 lr 11 1 1 11 _r








However, the degas temperature for all these samples was lowered


due to the thermal instability of A-POLY and A-SO4.


For comparison purposes


this temperature was kept constant for all samples at 1100C.


Table 11-6.


Pyrolysis Temperature Effects on Surface Area and Pore Volumes
for A-SO4a


Pyrolysis

Temperature


A-POLYb

A-SO4b

300

350

400

450

500

550

600


Surface

Area

(m2/g)


Pore Volumes


micro-


0.006

0.005

0.152

0.194

0.228

0.240

0.255

0.262

0.252


meso-


0.265

0.130

0.207

0.227

0.239

0.246

0.260

0.246

0.229


cc/g)


macro-


0.191

0.137

0.236

0.255

0.265

0.257

0.250

0.249

0.238


a Degas temperature changed from normal for these samples.


kept at 1100C under


BET C

constant


42.8

88.6

599

729

849

1061

1355

1659

1987


Conditions were


<10-3 torr vacuum for a minimum of 8 hours.


b Unpyrolyzed polymers as received.


Table II-6 shows the changes in surface areas and pore features from the


original polymeric materials as a function of the pyrolysis temperature.


It can


readily be seen how increases in temperature increased all of these parameters


up to about 5500C.


After this point, increases were no longer observed,


except


for the BET C value


, and changes occurring above this temperature were much


outlined.




36



definite trend of rising %C and falling %H was seen as pyrolysis temperatures


increased.


Further illustration of these changes in pore volumes is given in


Figure II-5.


Table II-7


Elemental Analyses for Pyrolyzed A-S04


Pyrolysis

Temperature (oC)


300


350

400

450

500


-----


550

600


66.93


73

77


.53

.38


81.07

84.19

86.80

88.89


2.94

2.89

2.84

2.87


---


2.57

2.27


0.00
0.00

0.00

0.00

0.00

0.00
----


---

---

- -

---

---

- -


0.00


total %CHN


- --


69.8

76.4

80.2

83.9


86.9


89.3


'7

*2

:2

4

7

7


91.16


A-S04 300 350 400 450 500 550 600


Pu'rlu'iQ


Tamn (fl


Vr *ui -. V v v* ***







Gas Adsorption Isotherms


To gain further understanding of the adsorption characteristics of these

carbonaceous materials, a series of gas adsorption isotherms was obtained at


room temperature employing a variety of gases.


The tabular data for these


adsorption isotherms, as well as isotherms obtained at other temperatures, are


summarized in Appendix B.


Figure II-6 compares the N2 adsorption isotherms


obtained on all nine carbons at room temperature.


Relative comparisons of the


affinity and capacity of these materials can be made from this data.


Attempts


to correlate the adsorption capacities with any of the previously reported

surface area, pore volumes or the BET C constants yielded no apparent trends


for this data.


Similar data is reported in Figure II-7 for room temperature CH4


adsorption, except for data for F-300 and a limited data set for PPAN.


capacity of all of these solids was higher for CH4 than N2, yet again no direct

correlation for the relative ordering could be made with any of the physical

parameters previously found.

In comparing the N2 adsorption data with that of CH4, the relative

affinities and capacities of some of the solids change as a function of both the


adsorbate and equilibrium pressure.

adsorption data for BPL and N-211.


This is most obvious in the CH4

Since no physical parameter, such as


micropore volume or surface area, can be employed to adequately explain this

shift in adsorption behavior, the chemical functionality of the surface and the


adsorbate must have decided effects on the adsorption process.


It is of interest


in comparing these adsorbents to be able to better quantify the adsorption


affinity and capacity they would have under various conditions.


Therefore, it is


- 4 t 4. ,,,- 4 1. .-, -,,,,j, -- -- .Z -, t. -- A 1- S 4 4. ,-






























































flTTYrf(lE q I rirt rTflYTY


PC I
I~frE* 1* CI





39











IC











I
II I









I I I T I 2 3 I Q
o










000







I I

o o
T1 T







S1 1


I --
I-I
IIIIII
III1.I









nune 9 Ione corarry







Conclusion


Referring to Table II-1, despite the large variety of source materials and

the pyrolysis and treatment conditions the original polymers were exposed to,

the elemental compositions of most of the resultant carbonized materials were


very similar.


However, it can be seen from the data presented in Table 11-2, all


the variables in chemical treatments and pyrolysis conditions have a decided


impact on the porosity and surface features of these carbons.


From this it was


concluded, as stated in the introduction, that through varying a large number

of parameters a carbonaceous adsorbent can be tailored to have specific

capacities and affinities for a wide variety of chemical compounds, both


dependent on their shape and size, and chemical functionality


. Varying these


parameters would have a large impact on how these materials behave as

adsorbents, catalysts, and catalyst supports.

In addition to the surface area and porosity differences observed in these


materials


, their affinities and capacities for various adsorbates is also


dependent on the synthesis of these materials.


It was found that for two fairly


non-interactive probes, N2 and CH4, differences in adsorption characteristics

were observed that cannot be accounted for simply by the measured physical


parameters of these adsorbents.


It is therefore likely that the adsorption


character of these materials, although affected by parameters such as surface

area and pore volumes and distributions, is highly dependent on their chemical

properties such as surface functionality, elemental composition and acidity.

This observation has been observed in the development of heterogeneous


catalyst systems.47,48,49


It would therefore be desirable to develop another








and predictive ability of the adsorption characteristics of these types of solids.

Such a model is developed in the following chapter.


The composition of the remaining mass of these solids,


both as


additional functionality and residual metal salts (ash content), was not well


characterized in these studies.


From the CHN analyses of the


se materials


(Table II-1) it was found that several have significant amounts (>5% by mass) of


other elements incorporated in their structure or on their surface.


A more


detailed analysis for trace metals is warranted, as it has already been surmised

that the chemical functionality of these materials is very important in their


adsorption behavior.


The presence of trace metals would also have a large


impact on the catalytic behavior of these solids.

The shape of the isotherms and hysteresis loops can help in elucidating

information about the structure, shapes and sizes of the pores in these


materials.32375,s0


Although pore shapes and sizes have not been fully developed


here, a large amount of information has been gained about the relative


porosity,

studied.


surface area and adsorption characteristics of the nine carbons

The BET theory and interpretation of isotherms have been discussed


in great detail to understand the large amount of information gained from gas


adsorption data.


Further analysis of the BJH model and Harkins-Jura theory


employed to determine pore size distributions is beyond the scope of this work.

However, a good understanding of the BET equation (Equation II-13) is helpful

in understanding the informative value of the interpretation developed from the

isotherms presented in Figures II-6 and II-7 in Chapter III.













CHAPTER III
GAS-SOLID ADSORPTION EQUILIBRIA


Introduction


Porous supports have been examined in great detail for applications in


the area of heterogeneous catalysis.


Porosity, pore size, surface area, and the


accessibility to the internal surface area are all important factors facilitating


reaction pathways on solids. As was found in Chapter II

correlation of a solid's adsorption behavior cannot be del

physical structure alone. Certain pore structures may a

reactants near active catalyst sites. The same concentra

with reaction products, which could inhibit the reaction.

understanding and quantitative measure of a solid's cap


:, however, direct

termined from its

effect concentration of

Ition effect may occur


Therefore, an


city and affinity for


various adsorbates are essential in understanding and developing a rational

selection process for solid materials.

The general phenomenon of adsorption involves the enrichment of one or


more components at an interfacial layer.


Gas-solid equilibria are typically


characterized by three types of adsorption interactions:


chemisorption and physisorption.


absorption,


Absorption occurs when the adsorptive


molecules penetrate the surface layer of the solid and enter the bulk structure.

Chemisorption involves irreversible bond formation to the solid, and is

generally associated with large enthalpies of adsorption (10-50 kcal/mol or


- .t- ~ _._--.l *. --~1-,- a- ...i -t: L..L U.....ts.....








may evolve chemically different than the original chemisorbed compound.

example of this is seen when 02 strongly binds to charcoal at room


temperature.


It desorbs


at elevated temperatures, but as CO and CO2.


Physisorption, which will include the process of reversible chemisorption for

this discussion, involves the reversible adsorption of molecules on a surface.

This adsorption process is dependent mainly on intermolecular interactions

such as induced dipoles, permanent dipoles, quadrupolar interactions, and


electrostatic forces.


Interactions between the interface and adsorbent are


usually found to be most significant for the first monomolecular layer


formed.


51,52


Therefore, it is found that most physisorption processes involve


adsorption enthalpies that approach the enthalpy of condensation for the

adsorptive.s3

As described earlier in Chapter II, up to three processes may be present


in the physisorption isotherm.


These are micropore filling, monolayer and


multilayer adsorption, and capillary condensation.


pores in the solid,


The size and shape of the


the adsorbate diameter, temperature, and the equilibrium


gas pressure of the adsorbate all have an effect on which process is occurring.

To examine the effect of pore sizes on adsorption, Everett and Powl54 calculated

interaction energies as a function of pore size based on the Lennard-Jones


potential model.


In cylindrical pores five adsorptive diameters or less in size,


the model predicts increasing adsorption energies with decreasing pore size.

Preferential micropore adsorption occurs because of these higher adsorption


energies.


Furthermore, if the temperature is below the critical temperature of


the adsorbate, capillary condensation is possible in pores larger than four








from the Langmuir and Kelvin equations, it better accommodates multilayer


formation and porous surfaces.


As discussed previously, this model does


have limitations on its applicability and the assumptions made in its


derivation.


In addition, the value of the BET C constant and its physical


meaning have often been debated.


This constant appears to be a complex


quantity related to the isotherm shape, surface coverage, the ratio of

equilibrium constants for the various adsorption sites and processes, and

contributions from the enthalpies and entropies of adsorption of the monolayer


and multilayer.


Literature attempts to interpret the BET C value have led to


various proposals,56 but none appear universal in obtaining thermodynamic

information about gas-solid interactions.

The focus of this research was to develop a method complementary to


existing theory for the analysis of adsorption data.


The goal of this analysis


was to obtain thermodynamic data for the interaction of the adsorptive


molecules with the strongest binding sites on the solid surface.


For these


studies, the adsorption isotherms of two carbonaceous adsorbents that had

been previously well characterized in this research group were focused upon.


Pyrolyzed polyacrylonitrile (PPAN) and Ambersorb 572, lot#


chosen.


2125 (A-572) were


More information on the physical characteristics and adsorption


behavior of these two solids can be found in the previous chapter.


Experimental


All gas adsorption isotherms were obtained in the exact manner as that

described in Chapter II, employing the glass vacuum line (Fig. II-1) or the


'C'


-. -. *5 -








obtained by immersing the sample in a Dewar filled with a slush made from


liquid N2 and an appropriate solvent. Toluene was the solvent for the -93C

isotherms, and acetonitrile for the -420C runs. An ice/water mix was used for


isotherms obtained at 0C. Variance of less than +1C was observed for these

systems during the course of a run. For isotherms obtained at room


temperature (25C) or higher, a heating mantle was employed to regulate the


temperature.


For these systems temperature control was not as precise, with


variations of +20C observed.


The numerical data for all the obtained isotherms


for the various gases and temperatures are tabulated in Appendix B.


Results and Discussion


Supports and Adsorptives


Two chemically different porous carbonaceous adsorbents and a series of

different gases were focused on to better characterize the processes involved in


gas-solid adsorption equilibria.

polystyrene, and PPAN, an ads<


where characterized in detail in Chapter II.


A-572, derived from sulfonated, macroreticular


orbent derived from pyrolyzed polyacrylonitrile,


It was observed that these two


materials differ in both their physical composition (Table II-1) and surface


areas/pore distributions (Table 11-2). 1

towards N2 and CH4 (Figs. II-6 and II-7


additionally, their adsorption behavior


at room temperature were observed to


differ in a fashion not solely accounted for by the physical characterizations

just mentioned.

The gases used in the adsorption measurements discussed in this








quadrupolar CO2 were chosen to investigate the influence of polarity on the

application of the developed adsorption model and to provide an indication of


the importance of these properties on adsorption by these two solids.


N2 and


CO are non-condensible gases in this study because their isotherms were


obtained above their critical temperatures.


These adsorptives will therefore be


subject only to gas-solid, micropore filling, and multilayer interactions, with no


possibility of capillary condensation occurring.


CO2 is, however, a condensible


adsorptive at room temperature, so there also exists the possibility of capillary

condensation with this adsorbate.


Table III-1.


He
N2
CO
CO2


Summary of Adsorbate Gases and Associated Physical Properties57


Mol.


(g/mol)


4.00
28.01
28.01
44.01


Polariz-
ability
(A3)


0.205
1.74
1.95
2.91


Dipole
Moment
(debye)


0
0
0.112
0


Molar
Volume
(mL/mol)a


32
35.4
34.9
40.0


7C)


Critical
Temp.
To (C)
-268.0
-146.9
-140.2
31.04


Boiling
Temp.
Tbp (C)
-268.9
-195.8
-191.5


8.44


AHv
(kcal/
mol)
0.0194
1.33
1.44
6.03


subl)


a Molar volumes of the liquids at their normal boiling point (unless otherwise
noted).


Equilibrium Analysis and Adsorption Model


The measured isotherms were analyzed using an equilibrium model


derived by analogy to a multiple site Langmuir adsorption model.


The resulting


equation is given below:


(111-1)


niKi,adsPatm
i (1 + Ki,adsatm )


where ntit is the total number of moles of gas adsorbed per gram of solid; ni the


available canacitv of process i


*


K1 ,. the eauilibrium constant for the adsorption









routine (Appendix C) was used to solve equation III-1 for the values of ni and

Ki,ads.

It has previously been found in using this data fitting routine that a very

shallow minimum in the solution series results from solving simultaneous


equations for data involving even two processes.58


In addition, it is an accepted


limitation of the simplex approach that misleading data may arise due to the


minimization of a complex equation in local or relative minima.59


Our


experience in solving Equation III-1 for room temperature gas adsorption data


found a shallow surface and uncertain quantities for ni's


three processes.


and Ki's for up to


Employing six parameters to fit these relatively simple curves


yields values for ni and Ki,ads, but one must critically evaluate the physical


meaning of the determined values.


In the Cal-Ad method,sso60 adsorption and


calorimetric data are solved simultaneously to better define the minimum.


adsorption enthalpies of physisorption processes for most carbon adsorbents


are so small that this approach cannot be employed.

isotherms were measured at various temperatures.


Instead, a series of

Assuming the ni values are


temperature independent, each isotherm at a different temperature introduces

only new Ki's as unknowns, leading to a better definition of the minimum in the

solution of the combined data sets.

The reproducibility and precision in the adsorption measurements were

determined from three successive experiments in which N2 adsorption by A-572

at 250C was followed by evacuating and repeating the N2 adsorption experiment


on the same sample.


reproducible.


The adsorption isotherms were found to be very


At higher equilibrium pressures (Patm > 0.05), the relative error


- --- ---





48



small volumes of gas adsorbed and the limits of sensitivity of the pressure


transducers.


Two successive adsorption isotherms were also obtained for CO


on A-572 at T = -930C.


At this low temperature, CO leads to much higher


volumes of gas adsorbed at lower pressures.


However, as observed in the N2


data, the greatest differences in successive runs were observed at the lower


equilibrium pressures.


For the two CO adsorption isotherms a 6


< 4% was


found for Patm < 0.05.


In this type of data analysis, it is essential to add


processes only until the adsorption data is fit to the precision and accuracy to


which it is known.


The relative errors discussed above were used as the


criteria for the minimum number of processes needed to fit the adsorption


data.


The procedure for obtaining ni's and Ki's for the gases N2 and CO in this


study is outlined as follows:


The isotherms for N2 and CO at -420C were fit to two processes using


the above criteria, leading to values for all four parameters (ni, n2, Ki,


At this temperature, the value for nl is best defined61 as the data


points range from zero to near full capacity for process 1.


Further


explanation of this assumption will follow, along with an interpretation


of the physical meaning of the different processes.


In addition


, the Ki's


sufficiently differ so that the two processes are felt to be well

distinguished.

(2) The -93C isotherms required three processes to fit the data to


within experimental error.

obtained in Step 1. This l


The value for ni was fixed to the value


power temperature data provides a better


definition of n2, so this value was not fixed, although the value of n2









data points over the majority of their capacity range.


Therefore, the


values of n2, Ki and K2 are felt to be well defined for this data set.

Process 3 is not as well characterized as there exist a limited number of


data points in the region where process 3 is occurring.

n3 and K3 are not believed to be as accurately known.


Because of this,

This is a limit


imposed by the apparatus and experimental data set, as it was not


possible to take measurements above Patm =


1.0, where process 3 is


occurring to the largest extent.


The above two isotherms (at -42 and -93C) were then reanalyzed,


employing a total of three processes.


Since ni, n2 and n3 have already


been determined, these values were not allowed to vary.


was also used on the room temperature data.


values for all Ki's


This procedure


This final step yields


at the three different isotherm temperatures.


In addition to finding the Ki,ads's at various temperatures, Step 3


constitutes a final check to determine if meaningful parameters for ni's


have


been obtained.


If the values found for ni, n2 and n3 from Steps 1 and


2 are well


defined, it should be possible to fit all the data sets to within the precision of


the experimental measurements.


A second check results from plotting the


natural logarithm of the Ki values versus the reciprocal temperature (in Kelvin).

This relationship, known as the van't Hoff equation, should yield a straight line


whose slope is proportional to the enthalpy of adsorption.


Although different


temperatures were used for the CO2 adsorption data, the same basic procedure


was followed for fitting this data.


In this case, however, a good data fit was


- Cr C~ *6 *1


1 '








Further detail and discussion of this method, and determination of what


is considered a good fit of experimental data, are found in Appendix D.


This is


exemplified and discussed with a sample set of data taken from Appendix B.

It is important to realize that the Ki,ads values found from this procedure

were derived from static adsorption data, where the established criteria for


equilibria (see Chapt. II, Experimental


- Gas Adsorption Studies section)


generally observed 4-5 minutes passing between each obtained data point. I

these materials are allowed to stay in contact with the adsorbates for several

hours, the reported adsorption capacities are found to increase, but by a


relatively small amount


10%


However, this rapid equilibrium process is of


greatest interest for these types of adsorbents when considering separation and

catalytic applications.


Adsorption Isotherms


Figures III-1 through III-3 show the adsorption isotherms at various


temperatures for PPAN and A-572 with N2, CO and CO2, respectively.


In all


cases the adsorption isotherms were observed to become more linear, and the


capacities of the adsorbents decrease, with increasing temperature.


numerical data in all the isotherms were fit to Equation III-1 as previously


outlined


, producing the best values for nl, n2, and n3, and the different values


of Ki, K2,


and K3 for all the measured temperatures.


The isotherms for all three


adsorbates, on both solids at all the temperatures measured, could be fit within


the accuracy of the experimental measurements using three processes.


previously mentioned, the numerical data for these, as well as several other
n ~a-I ,*- -- C-h l a nnjt: a .n jh A4 0A n fl-, on nr- t+ TT cr0 +0ht11i~i a + iA










5.0E-03


4.0E-03


bO 3.0E-03 -


2.0E-03 -


1.0E-03 -


O.OE+00
0.0


0.2 0.4 0.6 0.8 1.0


P (atm)



(a)


5.0E-03


4.0E-03

U)
C
3.0E-03


" 2.0E-03


" 1.0E-03



O.OE+00


0.2 0.4 0.6 0.8 1.0


P (atm)


/- \











6.0E-03 --


5.0E-03


- 4.0E-03


u 3.OE-03


, 2.OE-03
-4


0

1.OE-03 -


O.OE+00
0.0


0.2 0.4 0.6 0.8 1.0


P (atm)



(a)


6.0E-03


5.0E-03


4.0E-03


3.0E-03


c 2.0E-03 -


1.OE-03 -


O.OE+00


0.2 0.4 0.6 0.8 1.0


P (atm


I'-'

























































































OO C
o 0 C
I I
IIII I I I






























































































































































































I
I
4 I
4 -%
02 *
4- a
rM~Z 1) :LJ
I -'








Table III-2 summarizes the ni and Ki,ads values determined for the


isotherms shown in Figures III-1 through III-3.


In the table, ni is the capacity


of the solid for the specific adsorption process i, expressed in millimoles of gas


adsorbed per gram of solid


mmol/g).


The ni values are a measure of the


capacity of each process, and are functions of the pore size and pore


distribution of the solid, as well as the adsorbate size.


The reported Ki values


measure the affinity of a specific process on the solid for the gaseous probe.

is well known that for solid materials, surface interactions are not easily


characterized due to the solid being energetically heterogeneous.56,62 This is

especially true for highly porous materials such as those studied here. With


this in mind, it is important to recognize that the Ki's


found are likely to be


average values for adsorption processes at different solid sites whose Kads'S can


not be differentiated by this type of data analysis.


It is also likely that although


the Ki's are significantly different as to separate them into three different

processes, the second and third processes are occurring before adsorption

process 1 is complete.

The enthalpies of adsorption for the three processes on the individual

solids have been determined using the well known relationship between the


thermodynamic parameters of enthalpy (AH


and entropy (AS), and the


equilibrium constant K, which in this case will all be adsorption values:


-RTlnKads


= AHads -


TASads


(111-2)


where R is the gas constant and T the temperature in Kelvin.


Dividing both


sides of Equation III-2 by -RT yields


lflKads


-AHads Aads
RT R


(III-3)








and intercept is AS/R.


The results are summarized in Table III-3, and a


graphical plot of this analysis for CO2 on A-5


72 is shown in Figure III-4.


In this


graph,


the points shown are the experimental data, and the lines generated


from the least-squares fit analyses.


The least-squares fit of the resultant plots


are given as a footnote to Table III-3.


The proposed model (Eqn. III-1


is further


verified by the linearity of these data series, as shown in Figure III-4.

Additionally, the values calculated for AHads are consistent with those expected

for physisorption.


Table III-3


a Process 1


Summary of


= 2370


-AHads (kcal/mol) from InKads VS.


+ 66)/T


1/T Plots


- 8.64( 11


Process


=22


70 ( 158)/T


- 10.22(+


InK at 250C was omitted


because


Process 3:
b Process 1:
Process 2:
Process 3:
c Process 1:


it is poorly defined.
K = 1509 ( 27)/T 8.54(+
K = 2562 (+ .1)1/T 9.00(


= 2548 (


219)1/T


1492 ( 154)1/T


= 3739


323)1/T


.04); InK at 250C was omitted.
.03)


- 11.2(+
- 8.0(+ .
- 9.7( .


Process


Process 3:
d Process 1:


= 3196 ( 163)1/T- 9.8(+
= 2705 ( .59)1/T- 10.6(
= 2665 (+ 81)1/T 10.1


Process


= 1892 ( 1


23)1/T


-8.3


Process 3:
e Process 1:
Process 2:
Process 3:


= 1808 ( 119)1/T


= 2665


+ 39)1 /T


= 2695 ( 238)1/T
= 1512 ( 188)1/T


- 10.2


-9.3


- 11.9
-8.2


The determination of accurate and meaningful numbers for the three

processes is in part based on the ability of Equation III-1 to predict the


observed isotherms.


Additionally, measuring a large number of data points


A-572 PPAN
Process Na COb CO2c N2d COe
1 4.7 + 0.1 5.1 + 0.04 7.4 + 0.6 5.3 + 0.2 5.3 + 0.1
2 4.5 + 0.3 5.1 +0.4 6.3 + 0.3 3.8 + 0.2 5.3 + 0.5
3 3.0 + 0.1 3.0 + 0.3 5.4 0.3 3.6 + 0.2 3.0 + 0.4



















































































































































































































































































































V V








important to note that the final ni's

solely on the -93C isotherm. It ha


and Ki's were within 10% of those based


s already been noted that the ni values for


CO2 were derived exclusively from the 0C isotherm.


Both these results


indicate that if enough adsorbate is adsorbed by a solid in the region of the

isotherm where a specific process is dominant, reasonable estimates of ni are


possible from measurements at only one temperature.


If this is not the case,


as seen with the N2 and CO data, isotherms at various temperatures must be

obtained to better isolate the individual processes.


Interpretation of the Processes Involved


The equilibria, enthalpies, BET surface areas, and porosimetry values of

A-572 were utilized to provide an interpretation of the physical nature of the


three processes.


solids


Because it is the most thoroughly characterized of the two


, the interpretation of the processes involved is illustrated with A-572.


PPAN was studied with N2 and CO to afford a comparison of adsorbents.


different values of K1,ads, K2,ads, and K3,ads for A-572 are consistent with different


interaction potentials for different adsorption processes.


At low pressures,


adsorption of the gas into the small micropores, and the most energetic sites of


the solid, should be the dominant adsorption process.


As a measure of the


affinity of the adsorbate for these sites, Ki,


increases in the order N2


< CO


CO2 at any given temperature (Table III-2).


The adsorption enthalpies increase


in the same order (Table III-3) and are directly proportional to the polarizability


of the gas molecules (Table III-1).


This result suggests that the predominant


interaction of the adsorbates studied with this solid involves dipole and









The assignment of process 1 to adsorption in the smallest pores is

supported by the literature,54,63'64 and by pore resolved NMR porosimetry


results.


In the NMR spectra, it is observed that the smallest pores of A-572


fill first when CH3CN is adsorbed from a dilute CC14 solution.


This is a direct


observation that shows the equilibrium constant for this process is the largest.

For all of the systems in Table III-2, K1 is almost an order of magnitude larger

than K2.


Process


2 is attributed to adsorption in the larger micropores and


mesopores of the solid.


For N2 and CO on A-572, the enthalpies for processes 1


2 are the same, within experimental error.


The predominant contribution


to the enthalpy in process 2 is the solid-adsorbate dispersion interaction.


and CO are similar in size, and therefore it is reasonable to assume they have


access to the same internal pore volume.


This then indicates that the


increased interaction potentials arising from the close proximity of pore walls in

the microporous region have only a small enthalpic, but a large entropic


component.


With the larger molecule CO2, the enthalpic contribution to


process 1 is 1.1 kcal/mole larger than that for process 2.


The small pores


involved in the solids studied apparently have dimensions that lead to a larger


enthalpic interaction for the CO2 molecule.


molecules N2 and CO


For the smaller, less polarizable


, the enthalpic contribution is small and the process is


driven more by entropic contributions.


The predominant contribution to the enthalpy of process


2 for CO, N2,


and CO2 arises from the dispersion force interactions of these molecules with


the carbon surface.


The values of both ni and n2 are larger for CO than for N2.


r II* I








more strongly than N2.


If this consideration is limited to process 1 as


previously defined, this would result in larger values of ni and KI for CO over


N2, as observed.


The Ki and K2 values for CO2 are much larger than those


found for the other two gases, which is consistent with the above arguments.

However, the ni and n2 values for CO2 are smaller than those reported for both


CO and N2.


The smaller ni and n2 values are accounted for by CO2 being a


much larger molecule and,


therefore, not having access to the same micropore


volume as does N2 or CO.

Process 3 includes adsorption on the remaining surface and larger

pores, multilayer formation, and capillary condensation in cases where

isotherms are obtained below the critical temperature of the gaseous adsorbate.

As was previously mentioned, there exist a limited number of measured data

points under the temperature and pressure conditions where process 3 is


thought to be dominant for these solids.


Additionally, since the measured


isotherms were never observed to level off, the total capacity of the solid was


never reached.


Because of this


, the values for n3 and Ka are considered


approximations, and not as well defined as those for processes 1 and


A polarized surface molecule could lead to multilayer adsorption even

though isotherm temperatures are above the critical temperature of the


adsorbate.


Whether the adsorbed molecules are all on the surface or on


surface and bilayer sites is often of minor concern.


Therefore, in processes


where the solid-adsorbate interactions are comparable to the adsorbate-

adsorbate interactions, it is difficult to distinguish the contribution from


multilayer adsorption and surface bound interactions.


In this case, the two








Using an adsorbate cross-sectional area of 16.2 A2 for nitrogen, the

values of ni and n2 for A-572 correspond to a surface area of 45 and 155 m2/g,


respectively.


The value of na corresponds to


519 m2/g.


A similar analysis for


CO using an adsorbate area66 of 15.0 A2 gives 48,


values of nli, n2 and n3, respectively.


164, and 446 m2/g from the


In both these cases, process


attributed to adsorption by the larger pores and remaining surface.


Comparing


the results from this data analysis with those from BET theory suggests that


only a small portion of the total surface area for A-572 (S.A.BET =


occupied under these experimental conditions.


1159 m2/g) is


Knowing that a large amount of


the reported S.A.BET is because of the large micropore volume of this solid, it

appears that all of the adsorption occurring under these conditions is in the


microporous region.


The Micromeritics software reports a value of 1080 m2/g


for the micropore area of A-572.


This number is calculated based on the


difference between the reported total surface area (S.A.BET) and that calculated


for the external surface area.


This external area is defined as the surface area


of the solid excluding that of the micropores.


The value for the external area is


calculated from the slope of the Harkins-Jura t-plot analysis,


including what is


referred to as surface area and density correction factors.67

Comparing the ni values to the total amount of CO and N2 adsorbed for


the various isotherms (e.g.,


-1.7x10-3 moles/g at Patm=1.0 and -42C on A-572)


shows that up to an equilibrium pressure of 1.0 atm all of the adsorbate can be


accommodated by the ni and n2 values.


At the lowest temperature,


-930C, this


is not the case.


However, the ni and n2 processes still have the capacity for


over half of the amount of gas adsorbed.


These observations were found to be


-~~~ ~~ 1 -- U. -I ru.- -. 'V -C ,r1 .


i" 1


"




62



seen to be poorly defined at the temperature and pressure regions studied.

This is caused by only a fraction of na's capacity actually being occupied.

When similar calculations are done with the CO2 data, using a cross-


sectional area66 of 21.8A2

m2/g for processes 1 and


ns value.


, the corresponding surface areas are 24 and 181


An area of 1140 m2/g is calculated based on the


The total area thus calculated from ni, n2, and na for these three


processes exceeds the reported S.A.BET.


Thus, CO2 must involve both


adsorption by the solid surface and multilayer processes.


Both ni and n2 are


smaller for CO2 than for either N2 or CO, as expected from its larger size.

However, n3 is much larger, and provides the only clear manifestation of a

multilayer process in the systems reported in Table III-2.

It is of interest to note that the enthalpies of process 3 are larger than


the enthalpies of vaporization of the liquids.


From this observation, the


polarization of a surface bound molecule must increase the enthalpy of


interaction of a second layer with the surface layer. In general, if multilayer

formation occurs in each of the processes, the Ki,ads's are comparable to those


for surface adsorption of those processes.


The interaction potentials that give


rise to the different processes influence not only the surface interactions, but


also the multilayer interactions.


However, this multilayer interaction beyond


the monolayer coverage has been found to decrease rapidly with distance,51,52

and therefore it is not likely that many areas on the surface of these solids

contain more than a bilayer at the pressures studied.


Comparison of PPAN and A-572


-. -- ~ 4IF 44 4*tj U 4 .4 ..








A-572 in its adsorption characteristics towards the gases studied.


In general,


values for ni, n2, Ki, and K2 where found to be very similar for the two solids.

Thus, the nitrogen functionality of PPAN appears to do little to change the


interaction of the solid surface with the adsorbed N2 or CO.


As observed


before, the adsorption characteristics, and more specifically the trends

observed for these two solids ni and Ki,ads values, do not appear to have any

direct correlation to any of the surface area or porosity values previously

determined using standard gas adsorption theory.

The BET model was developed to account for multilayer adsorption by

assuming that the Langmuir equation, which involves only monolayer


adsorption, applies to each layer.


Following Adamson,


56 the BET C constant is


related to the ratio of equilibrium constants for both the monolayer and


multilayer adsorption.


It is also related exponentially to the difference in the


enthalpies of adsorption of the monolayer and the multilayer.


Therefore, C


encompasses several processes and thus is difficult to relate to the affinity or


strength of interaction of an adsorbate with the solid.


Although BET deals with


multilayer formation, it assumes the surface is energetically homogeneous.


previously discussed, this is not the case with most solids.

In the multiple process equilibrium model presented here (Eqn. III-1),

the BET C constant is essentially being separated into components for surface


adsorption in pores of different sizes and multilayer formation. By interpreting

the isotherms as a combination of energetically different processes, distinction


is also being made between energetically different sites on the surface of the


solid.


A significant advantage of this model is that Ki,ads for micropore filling is


a true thermondnamic roantitv and can be related to nhvsical nrnnerties. (e.g..








From this analysis, a thermodynamic measure of gas-solid equilibria and a

thermodynamic basis for understanding adsorption processes have been

developed.

It is of interest to note that both adsorbents were cycled through

adsorption experiments and degassing at 2000C up to fifteen times. No


changes were observed after repeated analysis on the same sample in the


overall surface area or pore size distribution from the original sample.


This


thermal stability and physical integrity is important when considering

separation and catalytic applications.


Conclusion


The affinities of three gases for two porous solids have been compared


using a multiple process equilibrium analysis.


For the non-condensible


adsorbates N2 and CO, a multiple process adsorption model is able to separate

the processes of adsorption in small micropores, adsorption by larger


micropores and mesopores, and adsorption by the remaining surface.


enthalpies of interaction for condensible adsorptives indicate that multilayer

condensation may accompany surface adsorption for each process and is


clearly involved for CO2 in process 3.


Comparison of the affinity of the solids


for large and small probe molecules is complicated,


however, by the fact that a


different distribution of pore sizes comprises processes 1 and 2 for different size

probes.

The Ki,ads and ni values for different adsorptives provide a quantitative


characterization of the forces involved in porous solid-gas equilibria.


For a


4 4 C zI. tr ,.. rr 11 ilL.. t








of the solid surface.


Comparison of the Ki,ads and the AHi,ads values provides


insight into the fundamental nature of the interaction of the adsorbent and

adsorptive molecules, and allows a quantitative means of selecting solids for


adsorption or catalytic applications. The gas-solid interactions studied here

were found to be predominantly physisorption. Accordingly, the carbonaceous


adsorbents studied here are expected to be effective in the separation of

molecules that differ in polarizability.

The excellent fit of the data for the non-condensible gases to the three

process model, using the same nl, n2, and ns values for isotherms at all


temperatures studied,


suggest that the ni and n2 values obtained are accurate


measures of the available capacity for these different processes.


It should also


be considered that the values found from this data analysis may prove to be


more useful in comparing the adsorption behavior of different solids.


applications as adsorbents, catalysts, or catalyst support materials, the

absolute surface area and pore structure of a solid are not necessarily the most


important parameters to be measured.


Instead, the effective or accessible


surface area and pore volumes for a specific adsorbate are of greater interest.

The ni values are a measure of the capacity of this effective pore volume. The

Ki,ads and subsequent AHi,ads are measures of the strength of these interactions.

Kinetic factors are also important in the practical applications of solids


in separation and catalytic processes.


However, this aspect of the problem was


not addressed in these studies.













CHAPTER IV
HETEROGENEOUS WACKER CATALYSIS


Background


Palladium catalyzed oxidation of olefins is a well known synthetic


approach in homogeneous catalysis.68,69


One of the better known applications


arose from the discovery by Smidt et al.70 of the homogeneous catalytic


oxidation of ethylene to acetaldehyde.


Most often referred to as the Wacker


process after the German company that first commercialized the process, it


involves an aqueous co-catalyst system of [PdCl4]-2 and CuCl2.


Pd(II


is reduced


to Pd(0) upon the oxidation of ethylene to acetaldehyde, and is then reoxidized


by Cu(II


The catalytic cycle is completed by the rapid oxidation of Cu(I) back


to Cu(II) by 02.71.72


The overall process is represented by the following


equations:


C2H4 + PdC12 + H20 -> CH3CHO + Pd(0) + 2HC1


(IV- 1


Pd(0) + 2CuC12


CuCI + 2HC1 + 02


-> PdC12 +


2CuC1


2CuCl2 + H20


(IV-3


C2H4 + O2


-* CH3CHO


(IV-4


The proposed mechanism involves the displacement of a Cl- from the


palladium complex followed by coordination of ethylene.


olefin for hydration by the solvent, water.


This activates the


Addition of another H20 molecule to


-. ~L ..Y a:.. '. gA l aIM -~n -rr thn.1 an t a nn nA in CE '1a*\f W rr'-


1r- 11 -1 -- 1 -


U`IYln ilr +Hcnrre


nCH n "








One of the major disadvantages to the industrial homogeneous system is


the production of chlorinated oxygenates as unwanted by-products.73


addition to the presence of chloride as the counterion, it has been suggested

that the presence of copper is a major contributor to the formation of these


undesired by-products.


A variety of research has been done investigating the


replacement of both copper and chloride ion in this catalyst system.


It has


been reported that this system is very susceptible to the counterion and

oxidant, and that different product distributions are obtained with changes in


these variables.74,75,76


As shown in Eqn. IV-2, the importance and function of


copper in this catalytic system is in the regeneration of the active palladium


species.


Halide free catalysis systems have been reported, and other species


have been found that can serve the function of Cu(II) in this system.


These


include vanadium oxides,77


heteropolyacids,


78 peroxides, benzoquinone


derivatives,79 and activated charcoal.80


Considering that the function of copper


may be met by carbonaceous materials, it is feasible that doping a specific

adsorbent, like those discussed in Chapter II, with Pd(II) salts would yield a


functioning catalytic system for Wacker-type catalysis.


Due to carbon's unique


chemical and physical properties, it may also prove advantageous over more

traditional supports (e.g., zeolites or silica) to employ a high surface area

carbonaceous support as a heterogeneous oxidation catalyst and/or catalyst

support.


For this study Ambersorb


Ambersorb


carbonaceous adsorbents were chosen.


adsorbents have previously shown catalytic activity towards both


oxidation and reduction applications,8182,83 as well as resistance to self-





68



volumes that would be advantageous for catalytic applications (see Chapter II).

All the reactants and products for the Wacker oxidation can be easily handled

in the gas phase, and therefore lend themselves well to a gas flow system with


an immobile Pd(II) solid catalyst.


For these reasons it was felt that Ambersorb


adsorbents would be excellent candidates for the support of Pd(II) salts that

could lead to a heterogeneous Wacker catalyst system.


Experimental


Ambersorb 572 (lot# 2201) was employed as the carbonaceous support


for this series of experiments.


Metal salts were dispersed onto the support by


pore-filling, also known as the incipient wetness method.


This involves making


a concentrated solution of a soluble salt of the metal species in an appropriate


solvent, usually deionized water.


Aliquots of the solution are added to the solid


until its surface appears wetted, avoiding excess liquid so as to not suspend


the solid in the solution.


appears dry.


The wetted solid is then dried under vacuum until it


This procedure is repeated until the entire salt solution has been


added.


The final catalyst is then dried under vacuum at 100C.


This technique


is used in hopes of concentrating the active catalytic species in the internal


volume and/or microporosity of the support.


As was previously discussed, the


micropore regions in these types of solids are thought to be very important in

catalytic processes.

Catalyst systems for the oxidation of ethylene to acetaldehyde were


evaluated in a gas flow system (Figure IV-1).


Reactant gases, which included


moist air and an ethylene/nitrogen mix (ranging from 10-20% C2H4 in N2), were





69



placed on top of the catalyst bed to fill the remainder of the reactor tube.

Water was introduced into the system by passing the air stream through a


glass bubbler fitted with a coarse air diffuser.


The amount of moisture in the


air stream was regulated by either placing the bubbler in a Dewar containing


ice water, leaving it at room temperature (RT


~ 250C), or heating the bubbler


slightly with thermal tape to increase the vapor pressure of water in the


bubbler.


Catalyst temperature was regulated with an Omega solid state


temperature controller (model CN310) and variable transformer (Staco, Type

3PN1010) connected to an oven that surrounded the reactor tube.

Temperatures did not vary more than 2C from the set-point on the


temperature controller during the course of a run.


constructed entirely of glass.

a 10 mm I.D. glass tube. Co

ground glass joints, and a co


the catalyst bed was supported.


The reactor flow system was


The reactor tube was 40 cm long, constructed of


nnections were made at both ends via 10/30


,arse glass frit was mounted in its center on which


A thermocouple, surrounded by a glass jacket,


was mounted directly below the glass frit and used for oven temperature


control and monitoring reactor temperature.


The reactor gas flow was in a


downward direction, as diagrammed in Figure IV-1.


A Hewlett-Packard 5890 gas chromatograph (g.c


Alltech RSLTM


equipped with an


-160 capillary column and flame ionization detector were used in


the analyses and identification of organic reactants and products from the


catalyst flow system.


This non-polar, 100% dimethylsiloxane polymeric phase


separates organic analytes mainly on the basis of their boiling points.


product line no longer used by Alltech.


This is a


The same phase can now be found sold













AIR SOURCE


PRE-GAS


SAMPLE PORT










GLASS BEADS




CATALYST BED


POST-GAS


SAMPLE


PORT


ETHYLENE


SOURCE


" TO FLOWMETER


N- -


OVEN/TEMPERATURE
CONTROLLER








Gas samples were taken both before and after the catalyst bed, which


allowed for calculation of conversions and selectivities.


A Varian 3700


equipped with a CarboxenTM 30 foot, 60/80 mesh packed column and a thermal

conductivity detector was also employed for monitoring N2, 02, CO and CO2.

This was done for purposes of mass balance and to monitor any possible over

oxidation of ethylene or the carbon support under reaction conditions.


Preparation of H21PdC141 Catalyst:


(HPd-572)


To 0.31g of PdCl2 was added 1.0 mL concentrated HCl(aq and 4.0 mL


distilled H20.


572.


The resultant solution was pore-filled onto 5.0g of Ambersorb


The resultant solid was dried under vacuum at 100C for one hour before


use in any catalysis runs.


The doping level in the resultant catalyst was


3.5x10-4 moles Pd(II)/g of carbon (3.7 wt%).


Preparation of Na2 PdC141 Catalyst:


(NaPd-572)


To 0.26g of PdC12 was added 1.0 mL concentrated HCl(aq) and 1.0 mL


distilled H20.


To the resultant brown solution was added 0.18g NaC1 (-2 moles


Na(I) per mole Pd(II)), dissolved in


-2 mL H20.


This solution was evaporated to


dryness over a steam bath, the resultant solid (presumed to be Na2[PdCl4])

dissolved in a water/ethanol mix (1/1), and then this solution was pore-filled


onto 5.0g of Ambersorb 572.


The resultant solid was dried under vacuum at


100C for one hour before use in any catalysis runs.


The doping level in the


resultant catalyst was 3.0x10-4 moles Pd(II)/g of carbon (3.2 wt%).








Preparation of Na2 PdC14 + CuC12 Catalyst:


(NaPd-572+Cul


To 1.04g of the NaPd-572 catalyst was added 0.5 Ig of CuCl2-2H20,


dissolved in 3 mL of a water/ethanol mix (1/1), by the


pore-filling method.


The resultant solid was dried under vacuum at 1000C for one hour before use

in any catalysis runs.


Preparation of NHa/Na2aPdC14/ CuC12 Catalyst:


CuPd-572)


Prior to pore filling, the carbon was washed with 6M NH3(aq),


to stay in contact with this solution for -1 hour.


and allowed


This was followed by repeated


rinsing of the solid with distilled H20 until a pH-neutral effluent was observed,

dried under vacuum at 1000C, and pore-filled with a solution containing both


Pd(II) and Cu(II) salts.


The Naa[PdC14] was made in similar fashion as in the


preparation of NaPd-572.


To 0.2


3g of PdCl2 was added 1.0 mL concentrated


HCl(aq) and 1.0 mL distilled H20. To t

0.15g NaC1, dissolved in 2.0 mL H20.


the resultant brown solution was added

This solution was evaporated to dryness


over a steam bath, the resultant solid dissolved in water,


CuCl22H20 was added to this solution.


and 1.10g


This liquid mixture was then pore-


filled onto 5.0g of the NH3 washed Ambersorb 572.


The resultant solid was


dried under vacuum at 1000C for one hour before use in any catalysis runs.

The doping level in the resultant catalyst was 2.5x10-4 moles Pd(II)/g of carbon


(2.7 wt%),


and 1.3x10-3 moles Cu(II)/g of carbon (8.2 wt%).


Materials


The Cr2H4 gas employed (Matheson) was C.P. grade (99.5% pure).








and used as received.


All other reagents used were of common laboratory


grade, obtained from Fisher Scientific, and used without any further

purification unless otherwise noted.


Run Conditions


Typical catalyst run conditions used between 0


1.0g of catalyst in the


flow reactor system.


Reaction temperatures were varied between


250 and


150C.


It was found that above 1250C, significant amounts of CO2 were


produced, arising from either oxidation of the support, or over-oxidation of the


organic substrates.


The best reaction temperature for these catalysts was


found to be around 750C.


Total gas flows ranged between


2.0 and 10.0


mL/minute, with C2H4 concentrations typically ranging from 1-3% of the total

gas flow.


Results and Discussion


HPd-572 Catalyst Results


Initial catalysis runs employed the H2[PdCl4] doped Ambersorb 572


(HPd-572). Reaction conditions were varied in attempts to yield the most active

catalyst system. Using 1.0g of the HPd-572 catalyst with a bed height of 2.0


cm, flow rates were varied over ranges from 2-10 mL/minute.


Catalyst


temperatures were varied from 50-150C, and the moisture content from the


described method with the water bath temperature ranging from RT


-5%


humidity) to a


heated bubbler at 750C


-25% humidity).


Relative amounts of




74



C2H4 did not exceed 20%, and the selectivity to CHaCHO was less than 25%.


Several additional products were found by g.c. analysis.


However, since the


desired activity was not observed and the production of chlorinated alkanes

and ethers is known to occur from acid catalyzed reactions with olefins, no


further attempt was made at identification or quantification.


In the


homogeneous industrial process it is reported76 that small amounts of acetic

acid, 2-chloroethanol, chloroethane, chloroacetaldehydes and acetaldehyde


condensation products are the most often observed by-products.


It has been


assumed that the minor by-products formed in this reaction are also composed

of these compounds.

Appreciable activity was not observed in this system until temperatures

exceeded 1250C, and this also resulted in the production of large amounts of


CO2.


Analyses of g.c. data showed almost complete consumption of 02 at


temperatures above 1250C, unaccompanied by a proportional increase in


aldehyde or other oxygenated organic products.


Since no noticeable decrease


in the carbon mass of the support was noticed over the course of the run, the

production of CO2 at these elevated temperatures is attributed to complete


oxidation of the organic substrates.


It was also observed that turning off the


C2H4 in the gas feed caused an increase in the amount of CO2 observed in the


postgas stream.


Again, since no measurable decrease in the support mass was


observed over the course of a run, this was assumed to be due to the complete


oxidation of adsorbed organic on the support.


Returning the C2H4 feed to its


original levels resulted in an immediate decrease in CO2 production, back to its


original levels at a given temperature.


Based on calculated turnover numbers


i ... .. t








not found to be catalytic.


The low activity it did exhibit was found to quickly


taper off within three hours of the beginning of the reaction.


Although the


overall activity of HPd-572 was very low, the catalyst system was found to be

more active at the slower flow rates (2-3 mL/min).

Initial observations found the support to have a high affinity for

adsorbing organic substrates, as it was at least 1 hour before any organic

analytes were observed in the post-catalyst gas stream for all of the catalyst


systems analyzed.


Additionally, the adsorption capacities and affinities of


these supports84 indicate they will not only adsorb and concentrate organic

reactants, but will have even higher affinities for the products of this reaction.

Due to this observation, no quantitative data for reported conversions or

selectivities were obtained until adsorption equilibrium was considered to be


reached.


This was determined from an approximate mass balance of the


amount of organic reactants measured in the pre-catalyst gas sample,


compared to that measured after the catalyst bed.


During the initial stages of a


catalyst run, where adsorption was occurring to a large extent,


it was obvious


that the majority of the organic reactant was not accounted for in post-gas

analyses.


NaPd-572 Catalyst Results


Because of the low activity of HPd-572 and the apparent production of

acid-catalyzed products, the next catalyst screened was designed to eliminate


possible effects due to the presence of Brensted-Lowry acidity.


In this system


dividing this by the number of moles of active catalyst species.


A T.O.N.< 1


. I.- I- --:I j." C .. : u -------------- ----- ..... -----1 L -- ------ "




76



the sodium salt of tetrachloropalladate was employed as the pore-filling salt


(NaPd-572).


Reaction temperatures were generally kept below 1000C, since


results with HPd-572 had shown no benefit in going to higher temperatures,


and flow rates were also kept low (3.0 mL/min).


Using 0.5g of the catalyst,


with a bed height of 1.3 cm and the water bubbler kept at RT, this system was

run under the typical reaction conditions described earlier (see Run

Conditions).

Initially, the NaPd-572 catalyst showed activity towards acetaldehyde


production at 500C.


The activity of this system was found to maximize around


75C, with -70% conversion of C2H4 and >90% selectivity towards CHaCHO.

Other products included very small amounts of CO2, and trace organic


products.


Determination of original conversions and selectivities were initially


calculated as much lower, as large amounts of ethanol were observed in the


post-gas analyses.


It was later determined that the observed ethanol was not


the product of a reaction of C2H4, but residual adsorbed solvent from the pore-


filling preparation of the catalyst.


This was further verified by the drying of a


batch of this catalyst overnight at 2000C under flowing air, and then subjecting


to normal run conditions with wet air and C2H4/N2 mix.


At temperatures below


1000C, CH3CHO was the only observed organic product, and no ethanol was

observed in the post-gas of the air dried NaPd-572.

To further verify that all the acetaldehyde produced originated from the

ethylene, a mixture of air and ethanol vapors were passed over the catalyst


under standard run conditions.

ethanol was observed. This con


Under these conditions, no reaction with


firmed the observation that all the CH3CHO


.,','ri rnnA ar,-'.a f-rnrn,-, Alr-nn+ ar4Alntn ,f f-l--l. .m'tiltr-h fha ma4i A', l a h art nl fanm







The large amount of ethanol remaining in this catalyst once again points to the

strong adsorptive capabilities of this support.


When the reaction with NaPd-


was carried out above 1000C, a variety


of heavier organic components (determined from g.c.


elution times) were


observed, along with a decrease in conversions and selectivity to acetaldehyde.


No further analyses to identify these trace compounds were undertaken.


was previously observed in the HPd-572 catalyst, lower temperatures are more


desirable for minimizing CO2 production.


Lower reaction temperatures are also


advantageous when considering overall costs for industrial scale processes.

Although this system was found to be catalytic, exhibiting 3 T.O.N. in optimal

conditions, its activity was observed to decline rapidly after a period of 5 hours,


leading to conversions of C2H4 of


<5%.


A series of eight runs on the same sample of the NaPd-572 catalyst


(0.50g, under standard reaction conditions:


3.0 mL/min gas flow, 750C) were


evaluated in an attempt to regenerate the catalyst to initial reactivity levels.

These included a variety of overnight treatments with air at elevated

temperatures (2000C), in hopes of reoxidizing the palladium and/or carbon

surface, as well as extended exposure to a humid air stream at room


temperature. Results of the air treatments saw a partial return of catalyst

activity, with about 5-10% conversion of C2H4, 100% selective to CH3CHO, but


this activity steadily declined over a period of 4 hours.


Moist air treatments


were more successful, leading to a return of -40% conversion, again with 100%


selectivity towards CH3CHO.


As was observed previously, however, catalyst


activity declined steadily after the first few hours of the reaction, dropping to


<5%


conversions after -5 hours.








run conditions.


Heating the water bubbler to 750C to increase the overall


humidity in the system prolonged catalyst life to -5 hours, yet conversions still


dropped off to


<5% after this period of time.


From these results, it would


appear that the reoxidation of the active palladium species by the carbon


support is not totally efficient under reaction conditions.


Regardless, although


some measure of activity was observed to return after the various attempts at


regeneration, none restored the catalyst to its original conversion levels.


Once


this catalyst system was started, even when conversions were very low (<5%),


CHaCHO was always observed in the post-gas stream.


However, instead of


continued catalytic activity, this could simply be from the desorption of

adsorbed product, which is feasible considering the previously observed

adsorptive capacity of the support.


NaPd-572+Cu Catalyst Results


Unsuccessful attempts to regenerate the NaPd-572 system and its


limited lifetime suggested the addition of a better oxidant for Pd(II)

Cu(II) salt, to this catalyst system may improve its overall activity.


such as a

The


resultant catalyst, NaPd-572+Cu, was screened in the same fashion as the


other catalysts.


A total gas flow of 3.0 mL/min, and a reaction temperature of


75C were used on 0.50g of catalyst with a bed height of 1.1 cm.

showed almost no activity initially (<5% conversion to CHaCHO).


This catalyst

However,


after an overnight exposure to moist air at 2000C activity was observed.

Acetaldehyde production occurred at temperatures as low as 500C.


Selectivities greater than 98% were observed up to 1250C.


The conversion of


-








the activity of this system was found to rapidly decline after -4hrs.


This


system was found to be catalytic, producing 17 T.O.N. of acetaldehyde.


NaPd-572 Without Ethanol Catalyst Results


To further characterize and determine the effects ethanol may have on

the reactivity of this catalyst system, another batch of the NaPd-572 catalyst

was made using methanol and water in place of the ethanol/water mix used as


the pore-filling solvent (see Preparation of Na2[PdCl4] Catalyst).


Additionally,


the catalyst was dried in a vacuum oven overnight at 1000C prior to use, to


ensure removal of the pore-filling solvents. The standard reaction conditions

and setup were used, as previously described. Two different runs were done


using this prepared catalyst.

The first experiment on this catalyst involved heating the water reservoir

to 500C, as it was previously noted that not enough water was being retained


by the system to remain active if the water bubbler was not heated.

activity of this system was quite high, with conversions of -80% and


selectivity to CH3CHO at 750C.


The

>95%


As before, however, the activity of this catalyst


was observed to rapidly decline after -6 hours. This system was found to be

more active than the catalyst made with ethanol, exhibiting -10 T.O.N.


The second experiment with this batch of catalyst was meant to


determine ethanol's effect on the reactivity of the system.


In this case, a 1/1


ethanol/water mix was placed in the bubbler through which air passed,


instead of distilled water only.


This introduced both water and ethanol vapors


to the catalyst bed during the course of the reaction.


Using the standard run








appearance of other organic products was also observed.


This is not completely


surprising, as the existence of another organic substrate would obviously allow


for the production of other compounds in this catalyst system.


Additionally,


was observed from pregas and postgas analyses that a large amount of the


ethanol in the feed gas stream was simply being adsorbed by the support.


significant evolution of ethanol was noticed until the catalyst temperature was


taken above 1000C.


Both of these experiments point to the presence of ethanol


being undesirable in this catalyst system.


CuPd-572 Catalyst Results


In an effort to combine all of the observations previously seen with the

various Wacker-type catalysts supported on Ambersorb 572, a base washed

catalyst was made containing both copper and palladium salts (CuPd-572).

Base washing with aqueous ammonia was done in hopes of removing any

residual acidity from the solid that might exist, considering its precursor is a


sulfonated ion-exchange resin.


And as observed in the case of the NaPd-


572+Cu catalyst, although this system was observed to be catalytic without

copper present, addition of this ion greatly increased its activity.

Under standard reaction conditions, which included total flow rates of

3.0 mL/min, temperatures ranging from room temperature to 1000C, C2H4

concentrations of -2%, and employing a heated water reservoir at 50C

(resulting in -5-10% humidity after considering dilution of air stream with

C2H4/N2 gas mix), experiments were done with 0.50g of catalyst, and a bed


height of 0.95 cm.


This catalyst system was found to be the most active of all




81



10 days, and during this time activity was never observed to decline.

Conversion of C2H4 was found to be >90%, with greater than 98% selectivity


towards CH3CHO production.


During this time period, it was found that the


catalyst produced more than 120 T.O.N.

Small amounts of CO2 were observed, but no other organic products


were indicated by g.c. analysis.

was lost during the reaction. T


However, further tests indicated that chlorine


'his is surmised from elemental analyses of the


CuPd-572


catalyst.


Results for the catalyst before use found it to contain


9.83% Cl- by mass (theoretical value is 10.2%).


Analysis of the catalyst after


the 10 day experimental run found it to contain 6.24% Cl-


. This is a 37%


decrease in the amount of chloride ion, and is not accounted for in the detected


products


This is assumed to occur either because the chlorine-containing by-


products are in such low concentrations that they are below the detection


limits of the g.c.


Table IV-1


detectors, or that these products are adsorbed and retained on


Summary of Heterogeneous Wacker Catalyst Conditions and Results


Catalyst HPd-572 NaPd-572 NaPd-572+Cu CuPd-572

Dopant Level 0.0037g 0.0032g (see prep. of 0.0027g Pd(II)

_(g metal9_carbon Pd II) ___Pd(II ___catalyst __ 0.0082g CueII)
Total Gas Flow 2-10 3.0 3.0 3.0

mLminLml ______n ________

Catalyst Temp. (0C) 75-150 50-125 50-125 25-75

C2H4 Conversion 15% max. 70% max. 80% max >90%

Selectivity to <20% >90% >98% >98%

CH3CHO








the carbon surface under reaction conditions.


Regardless, only small amounts


of materials were made, and this loss of C1- did not appear to affect catalyst


activity.


More long-term studies of this catalyst are warranted to further


understand this observation.

Table IV-1 summarizes the results for the various catalyst systems

studied, excluding the observations from the second sequence of experiments

done on the NaPd-572 catalyst made without ethanol.

In an effort to further qualify the effect of moisture content on this

system, and gain further evidence that a traditional Wacker-type mechanism

was being followed, the water feed to the system was turned off for a brief


period of time. This was done by simply routing the air flow around the

aqueous bubbler. The effects of this, which are further detailed in Figure IV-2,


show a decrease in the conversion of C2H4 concurrent with a drop in the


humidity of the system.


As soon as the water feed was reintroduced to the


system, the catalyst returned to its original activity level.


Porosimetry and Surface Area Studies


The presence of micropores in these types of carbonaceous supports is


thought to be advantageous for heterogeneous catalysis applications.


discussed in Chapter II, these microporous regions allow for areas of


concentration of reactants.


Ideally, these are localized around active catalyst


sites.


Accessibility to these areas is also of great importance.


Therefore, the


size of the substrates must be kept in mind,


as well as the presence of


mesoporous and macroporous regions on the solid that would facilitate

















































0 x


"l)' TTfT 7rT^l T n*"




84



Ambersorb 572 as a support for this heterogeneous catalyst system facilitates


these important processes.


This was determined from the large adsorption of


ethylene that was initially seen in the beginning of a reaction, as well as the

retention of ethanol by this support that was observed in the case of the NaPd-


572 catalyst system.


The development of an active catalyst system (CuPd-572)


also indicates these processes are occurring with some degree of efficiency.

In an effort to further determine where the pore-filled salts reside on this

support, a series of surface area and porosity studies were done on these


catalysts.


This data is summarized in Table IV-2.


These values were obtained


in the same manner as those reported in Chapter II, with the only difference

being the degas temperature for all the samples listed was chosen to be 1100C.


Table IV-


CARBON


A-572a


NaPd-572

% Difference

From A-572)__


CuPd-572


% Difference

(from A-572)


Surface Area and Pore Volume Data for Wacker Catalysts


Surface Area

(m2/g)


1236

857

30.7


a This is a different lot number


micropor


0.457

0.314

31.3


0.205


55.1


#2201


Pore Volumes (ce/g)

.es mesopore ma

S


0.324

0.262

19.1



0.232

28.4


cropores


0.242

0.175


27.7


0.162

33.1


and degas temperature


BET C

constant


995

970


1077
- -


110C) than








As can be seen from the data, a decrease in all the physical parameters is


observed upon doping A-572 with the various salts.


not specific to just the micropores, it is

undergoes the largest change. This is (


Although this decrease is


this region of the pore structure that


evidence that the majority of the pore-


filled salts are residing in, or blocking the entrances to the micropore volume of

the support.


Further Characterization of Ambersorb Carbonaceous Adsorbents


Magnetic susceptibilities and electron paramagnetic resonance (EPR)


spectra were obtained on two different Ambersorb


carbons in efforts to further


understand the activity of these materials as catalysts and catalyst supports,

and gain further insight into the interaction of the metal dopants with the


carbons.


Magnetic susceptibilities were obtained using a Johnson Matthey


magnetic susceptibility balance (model MSB-1).


This employs a method where


the magnetic susceptibility of a sample is determined by the force the sample

exerts on a suspended permanent magnet, which is the equal but opposite

force determined by the traditional Guoy balance method where the force


exerted on the sample by a fixed permanent magnet is measured.


Table IV-3


summarizes the results of the measured mass susceptibilities (Xg) of the two

carbons with and without metal dopants.

No corrections were made to the measured values for the volume

susceptibility of air or diamagnetic contributions from chloride or water present


in the samples.


Close analysis of the data in Table IV-3 does yield some


interesting observations.


A negative value for Xg is found for A-563/2144,


- -r ,


,,,







Mass Susceptibilities of Ambersorb 563 and 572 With Dopants


Sample


A-563 (lot #2144)

A-563/2144 with 5% Cu(II)b

A-572 (lot #2201)


A-572/2201 with 5% Cu(II)b

CuPd-572 catalyst


CuC12-2H20 solid


4X (c.g.s.)


.08x10-7


2.62x10-7

4.82x10-6


3.88x10-6


7.89x10-6


a All samples were ground to a fine powder using a Wig-L-Bug (Crescent), to
minimize errors from inhomogeneous packing and trapped air volume between
sample particles.
b The 5% doping level of Cu(II) was a weight percent pore-filling value, using
CuCl2-2H20 as the dopant.

reported Xg is so small, it is most likely that the measured diamagnetic

character of A-563/2144 is a bulk property arising from either diamagnetic

contributions from the matrix, magnetic anisotropy of the sample, or perhaps

anti-ferromagnetic coupling of the carbon free radicals to some trace metal in


the solid.


The existence of a definitive EPR signal shows that this solid is


paramagnetic, however, and the g value is typical of carbon free radicals.

Three EPR spectra were obtained at 190K for pure CuC12l2H20, a 5%

Cu(II) glass made from a 2:1 ethylene glycol:water mix, and for A-563/2144


with 5% Cu(II).


As would be expected, a very broad, unrefined peak is observed


for the pure copper salt, and the hyperfine splitting typical of Cu(II)-d9 ion is

not seen until the sample is dispersed in the glass and is magnetically dilute.86

The hyperfine structure of Cu(II) is observed in the case of the doped A-

563/2144, along with a single sharp peak associated with the pure carbon.


Table IV-3.








between the metal ion and the carbon support, the appearance of the hyperfine

splitting does indicate that Cu(II) is well dispersed on the solid.

The A-572/2201 sample has a fairly large and positive Xg, typical of


paramagnetic compounds.


Its value of 4.82x10-6 c.g.s. is comparable to that of


pure CuCl22H20


7.89x10-6 c.g.s.).


Of further interest here is that upon doping


A-572/2201 with 5% Cu(II) ion, the magnitude of Xg decreased.


for this occurrence is not obvious.


also reported here.


An explanation


The Xg value for the CuPd-572 catalyst is


No EPR was obtained on the A-572 carbons as it was not


possible to obtain a lock on the instrument for this sample.



Conclusion


Surface area and micropore volume are considered to be very important


factors in heterogeneous catalysis.


It stands to reason that the greater the


exposed surface area for the solid catalyst, the more likely the possibility of

reactants coming in contact with an active site on the catalyst surface.

Maximizing exposed catalyst surface area is the major limitation in developing


a heterogeneous catalyst from a homogenous system.


Employing a support


that has synergistic effects on the desired reaction is of even greater benefit,


and this was observed in the catalyst systems described.


Included in these


observed synergistic effects are the ability of the support to adsorb and

concentrate organic reactants in the internal pore structure of the carbon,


where most of the catalyst is believed to be present, and the support's


ability to


serve as an oxidative compound for the catalytic cycle.
'ri-. *: +. 1^ 4f C-.n trfltnfl ,n; d1,.n nrtn, fi n r Et n I4tn 117rn +trr1n nfI* 4+rnt flltVC 4%








system.


This in turn yields countless possibilities and parameters to be


modified in tailoring a heterogeneous catalyst system for specific reactivities


and reaction conditions.


Included in these are the possibilities of replacing the


chloride counterion with others, such as nitrate or acetate, which eliminates

the possibility of forming undesired chlorinated by-products, and application of

this catalyst to other systems of environmental and commercial interests.

Although no chlorinated by-products were observed in the most active catalyst

system, CuPd-572, evidence did exist that chloride was somehow being lost


during the reaction.


Changing the counter-ion in such a way would also have


some effect on the overall reaction and product distribution, which could also


prove desirable.


The preferential oxidation exhibited by these catalysts for an


organic substrate versus the support itself would also allow employing these

supports at higher temperatures, without concern for their combustion.

It is often observed in dispersed heterogeneous catalyst systems like

those reported here that the actual percentage of active catalyst species is


much less than 100%.


T.O.N.'s


reported here were based on the assumption


that all of the Pd(II) was active, which is probably not a true representation of


this catalyst system.


Further studies involving chemisorption and other


surface methods would help in gaining a more complete picture of the actual


catalyst surface and active sites.


The possible error arising from this


assumption is advantageous, however, as it shows that the reported catalytic

activity of these catalysts is a conservative estimate of their minimum activity

levels.

Results of the magnetic susceptibility measurements and EPR studies








and III, they provide another characterization of the behavior of these systems.

The curious results from the EPR and susceptibility measurements definitely

warrant further investigation into determining the effect these supports have

on metallic species doped onto their surface, and the nature of the chemical

interactions between the two.













CHAPTER V
CONCLUSION



The purpose of the research presented was to develop a greater

understanding of the behavior and properties of carbonaceous adsorbents.

Through the employment of a variety of physical methods such as elemental


analyses, gas adsorption studies, and the determination of surface areas and

porosities, a better understanding of the physical characteristics of porous


carbonaceous materials has been gained.


This included the determination of


what have proven to be the most important factors of the adsorption


characteristics of porous carbonaceous solids.


It was found that the


adsorption behavior of a solid is governed by a complex series of parameters.

Surface area and pore distributions of a porous solid are definitely important in

facilitating transport and adsorption behavior, but the chemical functionality of


the carbon supportlO0,1112,37,48 plays a much more significant role.


determination of what these various functionalities may be, where they exist on

the surface of a solid, and their relative concentrations have proven to be


difficult to determine by most modem instrumental methods.


However, rapid


advances are being made in methods to help characterize these types of

materials, driven by their importance and successes in adsorbent and catalytic

applications.

As observed in these studies, a large effort in the characterization of