-
ADVISORY BOARDM. CHEParis, France
A. CORMA CANSValencia, Spain
D.D. ELEYNottingham, England
G. ERTLBerlin/Dahlem, Germany
G. HUTCHINGSCardiff, UK
E. IGLESIABerkeley, California, USA
H. KNZINGERMunich, Germany
P.W.N.M. VAN LEEUWENTarragona, Spain
J. ROSTRUP-NIELSENLyngby, Denmark
R.A. VAN SANTENEindhoven, The Netherlands
F. SCHTHMlheim, Germany
J.M. THOMASLondon/Cambridge, England
H. TOPSELyngby, Denmark
-
Academic Press is an imprint of Elsevier
225 Wyman Street, Waltham, MA 02451, USA
525 B Street, Suite 1800, San Diego, CA 92101-4495, USA
Radarweg 29, PO Box 211, 1000 AE Amsterdam, The Netherlands
32 Jamestown Road, London NW1 7BY, UK
The Boulevard, Langford Lane, Kidlington, Oxford, OX5 1GB,
UK
First edition 2013
Copyright 2013 Elsevier Inc. All rights reserved.
No part of this publication may be reproduced, stored in a
retrieval system or transmitted in
any form or by any means electronic, mechanical, photocopying,
recording or otherwise
without the prior written permission of the publisher.
Permissions may be sought directly from Elseviers Science &
Technology Rights
Department in Oxford, UK: phone (+44) (0) 1865 843830; fax (+44)
(0) 1865 853333;
email: [email protected]. Alternatively you can submit
your request online by
visiting the Elsevier web site at
http://elsevier.com/locate/permissions, and selecting
Obtaining permission to use Elsevier material
Notice
No responsibility is assumed by the publisher for any injury
and/or damage to persons
or property as a matter of products liability, negligence or
otherwise, or from any use
or operation of any methods, products, instructions or ideas
contained in the material
herein. Because of rapid advances in the medical sciences, in
particular, independent
verification of diagnoses and drug dosages should be made
Library of Congress Cataloging-in-Publication Data
A catalog record for this book is available from the Library of
Congress
British Library Cataloguing in Publication Data
A catalogue record for this book is available from the British
Library
ISBN: 978-0-12-420173-6
ISSN: 0360-0564
For information on all Academic Press publications
visit our website at store.elsevier.com
Printed and bound in USA
13 14 15 16 11 10 9 8 7 6 5 4 3 2 1
-
CONTRIBUTORS
Tracy J. Benson
Center for Chemical Energy Engineering, Dan F. Smith Department
of Chemical
Engineering, Lamar University, Beaumont, TX, USA
Wm. Curtis Conner Jr.
Department of Chemical Engineering, University of Massachusetts,
Amherst, Massachusetts,
USA
Prashant R. Daggolu
Dave C. Swalm School of Chemical Engineering, Mississippi State
University, Mississippi
State, Mississippi, USA
Karl D. Hammond
Department of Nuclear Engineering, University of Tennessee,
Knoxville, Tennessee, USA
Rafael A. Hernandez
Department of Chemical Engineering, University of Louisiana at
Lafayette, Lafayette,
Louisiana, USA
Shetian Liu
Coal Chemical Catalysis Center, National Institute of
Clean-and-Low-Carbon Energy,
Beijing, China
Robert Schlogl
Department of Inorganic Chemistry, Fritz Haber Institute of the
Max Planck Society,
Berlin, Germany
Mark G. White
Dave C. Swalm School of Chemical Engineering, Mississippi State
University, Mississippi
State, Mississippi, USA
vii
-
PREFACE
The obituaries of two giants in catalysis anchor this
volume.
Paul Weisz created the field of shape-selective catalysis and
was at the
core of the research that gave birth to the science and
technology of
zeolite-catalyzed hydrocarbon conversion. He graced these
volumes as
the author of three chapters, as Associate Editor from 1956 to
1959, and
as Editor from 1959 to 1993. No one did more for Advances in
Catalysis.
Haldor Topse created one of the most influential and
consistently inno-
vative industrial organizations in our field, leaving a legacy
of robust new
technology, a commitment to basic research as the wellspring of
industrial
innovation, and a dedication to serving mankind. Researchers in
the com-
pany that bears his name have made numerous lasting
contributions to these
Advances as authors and editorial board members.
Our first chapter, by Hammond and Conner, will serve both as a
tutorial
and as an insightful review of the characterization of porous
materials by
physical sorption techniques. This field has been spurred not
only by the
widespread availability of automated equipment to facilitate the
measure-
ments but also by advances in data interpretation including
those guided
by density functional theory. As interest in porous materials
grows, the chal-
lenges of using surface areapore volumemeasurements continue to
expand.
The field is being stimulated with continuing discoveries in
zeolites and
related materials; in materials with ordered mesopores; and in
materials
with extremely high internal areas (such as metalorganic
frameworks)
there is a flood of innovations in the design of complex pore
networks
and tailored pore hierarchies. As the authors point out, good
practice in
the use of physical sorption techniques requires an
understanding of the
fundamentalswhich they set outand attention to details that are
often
overlookedwhich they highlight. This chapter is a readable,
practical
how-to guide that we believe may become required reading for
many
who work with porous materials.
Carbon is widely used as a (rather inert) support in catalysis,
but its own
catalytic properties have gone largely overlooked. Schlogl
explains the
fascinating chemistries of inorganic carbons, including
nanocarbons such
as fullerenes, nanotubes, nanofibers, and graphenes. He provides
deep
insight into the electronic properties of carbon, including the
effects of var-
ious structural elements on local, bulk, and topological levels.
For example,
ix
-
on an atomistic level, carbon hybridization, defects in graphene
sheets, and
surface functional groups are all significant. The author
compares the prop-
erties of basal and prismatic carbon faces and explains the
influences of
topologyand hence of curvature. This discussion demonstrates
that car-
bon has it allmetal-like properties; olefinic properties
associated with
localized double bonds; and acidbase properties stemming from
the surface
functional groups. This combination offers rich opportunities
for catalysis
but also presents a daunting challenge for characterization and
synthesis. The
analysis of the electronic properties of a particular type of
sitespresent
among a multiplicity of other sitesturns out to be quite
involved. Schlogl
assesses the characterization techniques and provides guidance
for controlled
syntheses of nanocarbons. He also reports on the inadvertent
formation of
carbons as surface deposits during catalysis and their
detrimental effects
and contrasts them with the perhaps surprising beneficial
catalytic effects
of some such deposits. Case studies (oxidative dehydrogenation
of ethylben-
zene and of alkanes) highlight the catalytic properties of
nanocarbons and
demonstrate the success of the authors rigorous approach to
synthesis and
characterization.
Addressing the rapid growth in research motivated by the need
for sus-
tainable fuels and chemicals, and the corresponding focus on
biomass con-
version, Benson, Daggolu, Hernandez, Liu, and White have taken
on the
substantial task of organizing the literature of the surface and
catalytic reac-
tions of oxygen-containing organic compounds. The authors start
with a
foundation in reaction thermodynamics and present an extensive
summary
of the adsorption chemistry of oxygenates on a wide range of
surfaces. Most
of the knowledge in this area has emerged from surface science
and compu-
tations, and the presentation illustrates the difficulty of
analyzing the behav-
ior of oxygenates on more complex surfaces. The authors address
various
classes of catalytic reactions that lead to removal of oxygen
from organic
compoundsas products including water, CO2, CO, and others.
A reaction class considered in detail is hydrodeoxygenation,
seemingly of
great potential for production of fuels from bio-oils but
hindered by a lack
of fundamental understanding and a clear path forward for
choosing viable
catalysts and processes. After presenting candidate processing
steps for bio-
mass upgrading, the authors wrap up with some thoughtful
concepts for
integrated process designs. This chapter will be a stimulus and
a valuable
resource for researchers working on biomass conversion.
BRUCE C. GATES
FRIEDERIKE C. JENTOFT
x Preface
-
PAUL B. WEISZ19192012
Paul B. Weisz, 93, an internationally recognized expert in the
areas of
zeolite catalysts and diffusion died on Tuesday, September 25,
2012, in State
College, PA.
In the 1960s and 1970s, working with several collaborators
atMobil Oils
New Jersey research laboratories, Weisz pioneered the use of
natural and
synthetic zeolites as catalysts for petroleum refining and
petrochemical man-
ufacture. These zeolite catalysts eventually revolutionized many
refining and
petrochemical conversion processes because they facilitated only
certain
reactions between molecules having specific dimensions. To
describe this
class of chemical reaction, Weisz coined the phrase
shape-selective cataly-
sis. In 1960, he first published his concept of acid-catalyzed
zeolitic reac-
tions and shape selectivity in a seminal article entitled
Intracrystalline and
Molecular-Shape Selective Catalysis by Zeolite Salts ( J. Phys.
Chem. 64,
382 (1960)) and in a 1962 article recognized the importance of
acid catalysis
in zeolites ( J. Catal. 1, 301 (1962)).
Weiszs vision of shape selectivity, initially probed using
natural zeolites
including erionite and chabazite, ultimately led to the
discovery of the zeo-
lite ZSM-5 and at least 15 significant and distinct
petrochemical and
xi
-
petroleum refining processes. His early work in multifunctional
heteroge-
neous catalysis (Adv. Catal. 13, 137 (1962)) was fundamental to
the
understanding and progress of multifunctional aromatic reforming
catalysts.
Similarly, his work with Dwight Prater in the interpretation of
measure-
ments in experimental catalysis (Adv. Catal. 6, 143 (1954)),
which drew
heavily from the work of giants in diffusion with chemical
reaction,
including Thiele and Damkohler, significantly advanced the
understanding
of how practical catalysts function. His work in experimental
measurements
in catalysis became essential reading for all who entered the
field.
His 1962 article with J. S. Hicks, entitled The Behavior of
Porous
Catalyst Particles in View of Internal Mass and Heat Diffusion
Effects,
Chem. Eng. Sci. 17, 265 (1962) was selected as one of the 50
most influential
articles in Chemical Engineering Science in the publications
1995 Frontiers in
Chemical Engineering Science commemorative edition.
One of Pauls formidable strengths was his ability to communicate
com-
plex theories succinctly. He was a constant contributor to the
ACS publi-
cation ChemTech throughout the 1970s and 1980s, where he
continued
to enlighten and delight readers with his insightful
observations of how phe-
nomena like diffusion and kinetics applied to everyday life. His
views of
catalysis were succinctly stated in numerous prefaces of
Advances in Catalysis,
which he edited from 1956 until 1993.
Processes based on Weiszs concept of shape-selective catalysis
were
first commercialized in the early 1960s. In displacing the more
conven-
tional amorphous catalysts in gasoline production, the zeolite
catalysts
were found to greatly increase the amount of gasoline produced
from a
barrel of oil as well as the octane number of the gasoline at a
time of dra-
matically increased demand for more and higher-octane-number
gasoline.
Weiszs work and its significant impact inspired efforts to look
for new,
synthetic zeolites, which continue to this day. Several zeolites
developed
under his guidance proved applicable for many other petroleum
and pet-
rochemical processes, including those that produced precursors
for poly-
ester (p-xylene) and high-quality lubricants. One of these
zeolites (ZSM-5)
was the basis for Mobils methanol-to-gasoline process, the first
new syn-
fuels process since the World War I development of the
FischerTropsch
process. Throughout the 1970s and 1980s, Paul was intimately
associated
with Mobils discovery of new catalytic materials and the
processes that
were developed around them.
While working at Mobil, Paul Weisz took a sabbatical leave to
earn his
doctoral degree from the Eidgenossische Technische Hochschule
(ETH) in
xii Thomas F. Degnan
-
Zurich, Switzerland, in 1966. His thesis was based on an
analysis of the per-
meation of dyes into fibers. His analysis was the basis for some
of the fun-
damental laws associated with diffusing molecules in fibers
based on the
molecular dimensions and chemical properties of the
molecules.
Born in Pilsen, Czechoslovakia, Paul Weisz grew up with an
innate
desire to become a scientist. He published his first article in
a ham radio jour-
nal at the age of 16. Paul emigrated to the United States in
1939 from Berlin,
interrupting his graduate studies in Germany to attend Auburn
University,
where he completed his B.S. degree in less than one year.
Following his
graduation, he worked as a researcher at the Bartol Research
Foundation
of the Franklin Institute in Swarthmore, PA. He later moved to
the
Massachusetts Institute of Technology, where, as an electronics
engineer,
he participated in the development of LORAN, a long-range
radio
signal-based aid to navigation.
Weisz joined Mobil Research and Development Corporation in 1946
as
a Research Associate at Mobils Paulsboro, NJ, research
laboratory. He
progressed through a number of technical assignments, in 1961
reaching
the position of Senior Scientist, the highest technical position
in Mobil.
He managed Mobils Exploratory Process Research organization
from
1967 until 1969 and its Central Research Laboratory in
Princeton, NJ, from
1969 through 1982. He retired from Mobil in 1984.
He then began a third, highly productive career, applying
chemical and
physical principles to biomedical research, first at the
University of Pennsyl-
vania and then at the Pennsylvania State University. Working
with Made-
leine Jouille at the University of Pennsylvania, he synthesized
molecules that
mimic some of the healing properties of heparin while not
exhibiting hep-
arins potentially dangerous side effects.
For his numerous industrial research accomplishments and
contribu-
tions to the science of catalysis, Paul Weisz earned many
awards, including
the E. V. Murphree Award in Industrial Chemistry from the
American
Chemical Society (ACS) (1972), the Pioneer Award from the
American
Institute of Chemists (1974), the Leo Friend Award from the
ACS
(1977), the R. H. Wilhelm Award from the American Institute of
Chem-
ical Engineers (1978), the Lavoisier Medal from the Societe
Chemique de
France (1983), the Langmuir Distinguished Lecturer Award from
the ACS
(1983), the Perkin Medal, the ACSs highest award (1985), the
Carouthers
Award from the ACS (1987), and the National Medal of
Technology,
awarded by President George H.W. Bush in 1992. Paul Weisz was
elected
to the U.S. National Academy of Engineering in 1977 and received
an
xiiiObituary Paul B. Weisz
-
Honorary Doctorate (Sc.D. in technological science) from the
Swiss Fed-
eral Institute of Technology in 1980.
The catalysis community will miss one of its most original and
influential
members.
Thomas F. Degnan
xiv Thomas F. Degnan
-
HALDOR TOPSE19132013
Dr. Haldor Topse, founder and Chairman of the Board of
Haldor
Topse A/S, passed away on May 21, 2013, only 4 days before his
100th
birthday. His passing away was a great loss to his family, the
company,
the catalysis communityand the world at large.
Dr. Topse was a researcher, an entrepreneur, and a businessman.
He
was also an idealist and a humanist, deeply engaged in the
global community.
In his words: The corporate world in itself means nothing unless
it
improves the lives of people and the conditions in poor
countries. He cre-
ated a unique company that today is world-leading in catalysis.
His achieve-
ments have reached much further: he contributed significantly to
the global
community through perseverance and dedication, leading to
technological
and scientific contributions to solve key global challenges
related to energy,
food, and the environment. Dr. Topse set high standards within
the indus-
try; he never stopped pushing the technological boundaries. His
vision
inspired and enabled progress that made the chemical industry
more sustain-
able, profitable, and competitive.
xv
-
Haldor Topses leadership of the company led to
accomplishments
ranging from fundamental understanding to technological
advancements.
One of the first business areas of the companyand to this day
one of
the most significantwas ammonia synthesis, in which Topse
established
a dominant position based on innovations in the catalyst and the
synthesis
process. Other achievements include environmental technologies
and
ground-breaking catalytic process equipment. One of Haldor
Topses last
efforts was the development of solid oxide fuel cells, and his
persistence led
to the establishment of the subsidiary Topse Fuel Cell and a
dedicated effort
to develop and commercialize this energy-efficient
technology.
Haldor Topse was born in 1913, into a world recovering from a
world
war and descending into recession. As a young man studying in
Copenha-
gen, he saw the lines of people queuing for jobs, queuing for a
free meal
and he vowed to himself that he would make a difference, that he
would use
his knowledge and privileged position to contribute to making
the world a
better place.
Topse continued his studies with, among others, Niels Bohr,
the
Danish physicist and Nobel Laureate. By 1940, he had already
made a name
for himself, but on the 9th of April of that year, Hitlers
troops invaded Den-
mark. Topse had been offered a job in the United States. A plane
was ready
to take him and his family from Denmark, but his departure was
impossible,
because two of his children were hospitalized. His wife, Inger
Topse,
turned to him and said, Now that we cannot leave, you must make
some-
thing which will be useful after the war. And early on, Topse
saw the
potential of catalysis, which today is the basis for most of the
worlds chem-
ical production.
Topses base was the natural sciences; his years with Bohr were
forma-
tive. He based his company on the strong belief that fundamental
under-
standing of the chemistry and engineering of the processes in
the
companys portfolio would provide a competitive edge. His policy
was
always to invest significantly in researchin manpower and
research instru-
ments, bench-scale test units, and pilot-scale units. This
visionary approach
serves as a model for industrial as well as academic research
institutions
worldwide.
Topses conviction was that research and new ideas are vital to
staying
competitiveeven research that may not contribute directly to the
bottom
line. Ideas that seemed unrealistic 20 or 30 years ago have
developed into
xvi Bjerne S. Clausen and James Dumesic
-
technologies that contribute significantly today. Topse
encouraged these
ideas to grow. Through his own scientific understanding and
visionary
insights, he inspired the people around him, acting as a
catalyst for the com-
pany he created and for the catalysis community as a whole.
Haldor Topses vision and his dedication to fundamental science
fos-
tered the development of new catalysts and processes, allowing
the company
to grow into new business areas: early on, he identified the
business potential
of the refining industry, and thanks to his commitment,
significant efforts
were made to develop hydroprocessing catalysts based on
cutting-edge
research, resulting in the BRIM catalysts and technology. The
develop-
ment of Topses WSA technology (wet gas sulfuric acid) started as
the
side-project of one of Topses brightest researchers, and
although Haldor
Topse doubted the potential of this technology, he encouraged
the
research to continue. Today, Topses WSA technology
contributes
significantly to a more sustainable world.
Dr. Topse was driven by a determination to make a difference.
The
roots of this determination lie in the depression of the
1930sand his
opportunity to contribute to change. Today, seemingly everyone
who
has known Haldor Topse is familiar with his stance that if you
have
the ability and the means to contribute to the world, you also
have the obli-
gation. Thus, beyond his contributions to science and
engineering, Topse
contributed funds for schools, programs for street children, and
other wor-
thy causes.
Topse held many influential posts. His knowledge and insight
about
sociopolitical megatrends served the globe, exemplified by his
work in
the Population Council in the 1970s. He had a knack for
networking that
brought him into contact with people of influence: politicians,
world
leaders, kings and queensfrom Nelson Mandela to Mahatma
Gandhi.
Topse never sat back in awe; rather, he used his talent for
dialogue and per-
suasion to express his sociopolitical views, to the benefit of
mankind. These
views have been publicized in several books dealing with the
global social
and economic situation.
Haldor Topse was honored with numerous awards and medals,
among
them the Hoover Medal, in 1991. To Dr. Topse, one recognition
espe-
cially stood out: in 1999, the Association of Danish Engineers
named him
the Engineer of the century. The honor was given in recognition
of both
his technical achievements in catalysis and his commitment to
social issues.
xviiObituary Haldor Topse
-
Dr. Topses social engagement, passion for science, ambition,
and
determination created a legacy that continues to exert an
enormous world-
wide impact.We and our community will long remember him for his
strong
commitment to the world around himand his belief that we all
have an
obligation to contribute.
Bjerne S. Clausen*
James Dumesic*
*Parts of this text appeared in Angew. Chem. Int. Ed. 2013.
Angewandte Chemie by Gesellschaft
Deutscher Chemiker Reproduced with permission of WILEY - V C H
VERLAG GMBH & CO.
xviii Bjerne S. Clausen and James Dumesic
-
CHAPTER ONE
Analysis of Catalyst SurfaceStructure by Physical SorptionKarl
D. Hammond*, Wm. Curtis Conner Jr.*Department of Nuclear
Engineering, University of Tennessee, Knoxville, Tennessee,
USADepartment of Chemical Engineering, University of Massachusetts,
Amherst, Massachusetts, USA
Contents
1. Introduction 72. The Phenomenon of Physical Adsorption 8
2.1 History of adsorption in catalyst characterization 82.2
Choice of adsorbate 92.3 Presentation of adsorption data 122.4 The
Langmuir isotherm 142.5 Monolayers to multilayers 162.6 BET theory
18
3. A Tour of the Adsorption Isotherm: From Vacuum to Saturation
and Back 273.1 The micropore-filling region: 108
-
7.2 Sorption analysis techniques derived from simulations 698.
Details of Adsorption Apparatus 72
8.1 Volumetric adsorption systems 728.2 Start to finish:
Acquiring an adsorption isotherm 75
9. Common Pitfalls in Adsorption Experiments and Analyses 859.1
Definitions of standard temperature and pressure 869.2 Measurement
of the saturation pressure 869.3 Drift in bath
temperatures/compositions and equipment calibration 889.4 Reference
state for argon at 77 K 889.5 Misuse of the BET equation 899.6
Failure to degas properly 919.7 Inaccurate or nonequilibrium
pressure readings 919.8 Correcting (incorrectly) for thermal
transpiration 929.9 Interpreting hysteresis at low pressures to be
porosity 949.10 Reporting 1720 pores 94
10. Summary 95References 97
Abstract
Heterogeneous catalysis usually takes place by sequences of
reactions involving fluid-phase reagents and the exposed layer of
the solid catalyst surface. Estimation of the totalcatalyst surface
area, its potential accessibility to gas- or liquid-phase
reactants, and gen-eral catalytic activity are initially based on
themorphology of the catalyst. Universally, mea-surements of
adsorption and their interpretation are used to estimate the
surface area andporosity relevant to catalytic reactions. We
provide here a description of many traditionaland recent techniques
in adsorption-based catalyst characterization intended for
exper-imental practitioners of adsorption. Our chapter includes
descriptions of which regions ofthe isotherm correspond to
micropore filling, mesopore filling, surface coverage, and
sat-uration, supplemented by discussions of model isotherms, from
the Langmuir isothermand the BrunauerEmmettTeller theory to the
Halsey equation. Pore size distributionmethods include the
BarrettJoynerHalenda and relatedmethods for mesopores, empir-ical
methods developed for micropores, and simulation-based methods that
have finallyresolved the differences between adsorption (increasing
loading) and desorption(decreasing loading). This chapter also
includes a discussion of hysteresis and metastabil-ity, both of
which trip up experimentalists from time to time. We finish with a
descriptionof data acquisitionmethods and equipment, which are
often obscured behind the facadeof automation, and a discussion
ofwhat users should be aware of andwhat can gowrong.
NOMENCLATURE[A] activity of species A
as normalized adsorption isotherm on a standard, nonporous
material such as silica orcarbon; asVads/Vads(P/P 0.4)
g surface tension
2 Karl D. Hammond and Wm. Curtis Conner Jr.
-
u loading: yn/nm, interpreted as the fraction of unoccupied
sites in the first layerui loading in layer ik proportionality
constant between flux and pressure for an ideal gas; k (2pMakT)1/2m
chemical potential ((partial) molar Gibbs free energy)n vibrational
frequency along the normal mode that results in desorption from a
layerr density of adsorbates atomic/molecular diameterV grand
potential, OPVgAA surface area
AAA force constant between the adsorbate and itself in the HK
and ChengYang models
AAE force constant between the adsorbate and the adsorbent in
the HK and ChengYang
models
Adsorbate the atom/molecule that is adsorbing onto a surface
Adsorbent the solid species onto which the adsorbate adsorbs
Adsorptive a synonym for adsorbate
Am area of one molecule, typically reported in nm2 or A2 per
molecule or atom but can also
be in m2/mol or m2/cm3 STP (i.e., surface area per unit volume
adsorbed)
Ap surface area of a pore
ai sticking coefficient: probability of adsorbing given a
collision with the surface
b Langmuirs parameter, which has units of inverse pressure
AEI International Zeolite Association framework code for
materials such as AlPO-18,
SAPO-18, SIZ-8, and SSZ-39 (the code is derived from AlPO
EIghteen)
ATS International Zeolite Association framework code for
materials such as
AlPO-36, MAPO-36, FAPO-36, and SSZ-55 (the code comes from AlPO
Thirty-Six)
BET BrunauerEmmettTeller theory
BJH the BarrettJoynerHalenda method of pore size distribution
analysis for mesopores
CBET or C second BET fitting parameter (the other being the
monolayer capacity), a
positive number related to the difference in the heat of
adsorption between layers
Chemisorption chemical adsorption, the process of atoms
ormolecules chemically reacting
with a solid surface, in contrast to physisorption
CH empirical Halsey constant in the FrenkelHalseyHill
equation
CK constant of proportionality between inverse pore size and the
base-10 logarithm of
relative pressure: CK 2gVl = RT log10 CPDFT classical potential
density functional theory
D inner diameter of the glass tube used in physical adsorption
experiments
d0 arithmetic mean diameter of adsorbate and adsorbent
atoms/molecules in the HK model
Degassing the process of heating a sample to remove adsorbed
gases, such as water, from the
surface in preparation for an adsorption experiment; also called
outgassing
DFT density functional theory
DH the DollimoreHeal method of pore size distribution analysis
for mesopores
Dp pore diameter, Dp2rpf fugacity
F Helmholtz free energy
Fr density functional
FAU International Zeolite Association framework code for
materials such as faujasite,
zeolite X, zeolite Y, ECR-30, LZ-210, and SAPO-37 (the code
comes from FAUjasite)
G Gibbs free energy of adsorption
3Analysis of Catalyst Surface Structure by Physical Sorption
-
h height of the bath
H enthalpy of adsorption
H1 hysteresis hysteresis classification in which the loop has
similar shape on adsorption as it
does on desorption
H2 hysteresis hysteresis classification in which the loop has a
different shape on adsorption
than it does on desorption; the adsorption process is gradual,
whereas desorption is
sudden
HEU International Zeolite Association framework code for
materials such as heulandite,
clinoptilolite, and LZ-219 (the code is derived from
HEUlandite)
HK HorvathKawazoe model for adsorption in slit-like
micropores
HRADS high-resolution adsorption, a term applied to techniques
that measure adsorption
below about 1 Torr (103 atm)IBET intercept of the BET plot
Jm molecular flux hitting the surface
k Boltzmanns constant
ka rate coefficient for adsorption
kd rate coefficient for desorption
K equilibrium constant
length of the neck of the sample tubeLTA International Zeolite
Association framework code for materials such as zeolite A,
ITQ-29,
SZ-215,SAPO-42, andZK-21,ZK-22, andZK-4 (thecode isderived
fromLindeTypeA, the
original name for zeolite A as synthesized by the Linde group at
Union Carbide)
log natural logarithm (note that common (base 10) logarithms are
written explicitly (e.g.,
log10(x)))
m mass of the adsorbent
Macropore pore with a radius larger than 50 nm (such pores fill
near the saturation pressure
and are generally not resolvable on adsorption isotherms)
Manifold the chamber directly above a valve that connects it to
the sample tube in a
volumetric adsorption system
Ma molecular (or atomic) mass of the adsorbate
MEL International Zeolite Association framework code for
materials such as ZSM-11, SSZ-
46, silicalite-2, and TS-2 (the code is derived from Mobil
ELeven)
Mesopore pore with a radius between 2 and 50 nm
MFI International Zeolite Association framework code for
materials such as ZSM-5,
silicalite-1, EU-13, ISI-4, mutinaite, and KZ-1 (the code is
derived from
Mobil FIve)
Micropore pore with a radius of 2 nm or less (important: the
word micropore has
nothing to do with the word micrometer)
MTT International Zeolite Association framework code for
materials such as ZSM-23,
EU-13, ISI-4, and KZ-1 (the code is derived from Mobil
Twenty-Three)
MTW International Zeolite Association framework code for
materials such as ZSM-12,
CZH-5, NU-13, TPZ-12, and VS-12 (the code derives from Mobil
TWelve)
NA areal density of adsorbate atoms/molecules
Nanopore name sometimes used in place of the word pore,
especially by authors
attempting to draw analogies to nanotechnology (all micropores,
mesopores, and
macropores are nanopores, despite the illogic of the
prefixes)
4 Karl D. Hammond and Wm. Curtis Conner Jr.
-
NE areal density of adsorbent atoms/molecules
nads or n number of molecules (or moles) adsorbed, usually per
gram adsorbent
NamA number of moles in the adsorption manifold before the
sample valve is opened
NamB number of moles in the adsorption manifold after the sample
valve is opened
Ngas number of moles currently in the gas phase inside the
adsorption manifold and over the
sample
Nin total number of moles added to the adsorption manifold since
the start of the
experiment
NLDFT nonlocal density functional theory, a type of CPDFT in
which the functional
includes the contributions of density gradients to the free
energies
nm monolayer capacity: the number of molecules or moles of
adsorbate in one layer on the
surface, typically normalized per unit mass of adsorbent
NMR nuclear magnetic resonance spectroscopy
Outgassing synonym for degassing
P absolute pressure
PA pressure in the adsorption manifold before the sample valve
is opened
PB pressure in the adsorption manifold after the sample valve is
opened
P/P relative pressureP saturation pressurePg pressure on the
vapor side of a vaporliquid meniscus
Physisorption physical adsorption, the process of atoms or
molecules adhering to a surface
without forming chemical bonds, in contrast to chemisorption
P pressure on the liquid side of a vaporliquid meniscus
Ps standard pressure (usually 1 atm.101,325 Pa760 Torr)Q heat of
adsorption (negative enthalpy change of adsorption); subscript i
indicates
adsorption on the ith molecular layer
QL heat of condensation
QSDFT quenched solid density functional theory, a type of CPDFT
in which gradients in
both the adsorbate and adsorbent densities are factored into the
free energy calculation;
the net effect is more flexible pore walls than NLDFT
R universal gas constant (in J mol1 K1)RHO International Zeolite
Association framework code for materials such as zeolite Rho,
ECR-10, LZ-214, and pahaspaite
rc rate of condensation
re rate of evaporation
rK or hri mean radius of curvature of the meniscus inside a
filled or filling pore, often calledthe Kelvin radius
Re ratio of the volume of the empty tube to the manifold
volume
Rf ratio of the volume of the filled tube to the manifold
volume
Rn square of the ratio of the volume of a cylinder representing
the pore and another
cylinder representing the fluid added to/removed from the pore;
used in BJH
analysis
rp mean pore radius
S entropy
SBET slope of the BET plot
SABET surface area as extracted from the BET equation
5Analysis of Catalyst Surface Structure by Physical Sorption
-
SANS small-angle neutron scattering
SBA material code designating unique materials first made at the
University of California,
Santa Barbara, such as SBA-15
SEM scanning electron microscopy
SF SaitoFoley model of adsorption in cylindrical micropores
STP standard temperature and pressure, typically defined as
273.15 K and 1 atm
t statistical thickness of the adsorbed layer
T absolute temperature (in Kelvins)
Tb temperature of the bath
TEM transmission electron microscopy
TON International Zeolite Association framework code for
materials such as zeolite Theta-1,
ISI-1, KZ-2, NU-10, and ZSM-22 (the code comes from
Theta-ONe)
Tr room temperature, typically about 2025C
Ts standard temperature (usually 273.15 K0 C, but others are
also commonly used)Tt triple-point temperature (in Kelvins)
Type I isotherm an isotherm, such as the Langmuir isotherm, that
is concave and has no
apparent multilayer adsorption
Type II isotherm an isotherm, such as the BET isotherm, that has
an inflection point and
shows multilayer adsorption but no hysteresis
Type IV isotherm an isotherm that begins like a Type II isotherm
but exhibits hysteresis
Vads or V standard volume adsorbed, in cm3 STP/g adsorbent.
VadsnadsRTs/Ps
Vam volume of the adsorption manifold, typically in cm3
Vbulb volume of the bulb at the bottom of a typical glass
adsorption cell
VDS apparent volume of the chamber containing the sample from
the sample valve onward,
including effects due to temperature gradients, typically in
cm3
Ve volume of the empty calibration tube, typically in cm3
Vf volume of the calibration tube after the known-volume insert
is in place, typically
in cm3
Vg molecular or molar volume of the vapor phase
Vgas volume, at STP, of adsorbate currently in the gas phase
inside the sample and manifold
volumes
Vin total volume, at STP, of adsorbate added to the adsorption
manifold since the start of the
experiment
VET International Zeolite Association framework code for the
zeolite VPI-8 (the code is
derived from Virginia Polytechnic Institute EighT)
VFI International Zeolite Association framework code for
materials such as VPI-5,
AlPO-54, H1, and MCM-9 (the code is derived from Virginia
Polytechnic Institute FIve)
V liquid molar or molecular volume
Vm monolayer volume, or the volume all adsorbed molecules would
take up if desorbed and
returned to STP; VmnmRTs/PsVp pore volume
Vstd volume of a known standard, such as a cylindrical tube of
precision-bore glass
w width of slit-like pores in the HK model
W weighting function for each isotherm in the simulated kernel
of isotherms; this
becomes the pore size distribution function when fit to an
experimental isotherm
xm areal density of surface sites
6 Karl D. Hammond and Wm. Curtis Conner Jr.
-
1. INTRODUCTION
Physical adsorption results from interactions between
subcritical fluid
species and nearly any solid surface. The measurements are made
by a variety
of well-developed techniques and interpreted by using ever more
sophisti-
cated models. Physical adsorption experiments probe
thermodynamic phase
equilibria between bulk fluid phases and adsorbed phases, which
progress
from single, isolated molecules to a single layer of molecules
on the surface
(a monolayer) to multilayers to condensation (or sublimation).
Analyses of
equilibrium data characterizing the adsorption of physisorbing
gases are
commonly employed to estimate the morphology of the sample,
including
the total surface area, the distribution of the dimensions of
any pores (ranging
in diameter from about 0.1 to 50 nm), and the total pore
volume/void frac-
tion. These analyses are employed to guide understanding of the
influence of
morphology on sorption, separations, and catalysis.
Considerable progress has been made in the last several decades
in inves-
tigations of physical adsorption on high-surface-area solids,
both experimen-
tally and theoretically (18), such that we now understand the
phenomena
associated with sorption far better than we ever have.
Furthermore, materials
synthesis has developed to such an extent that we can now
produce materials
possessing very high surface areas (>1000 m2 g1 of solid) or
with uniformpores in the range 120 nm, or even solids with
multiscale porosity compris-
ing a network of pores of one size embedded within a network of
pores of
another dimension and/or connectivity. Physical adsorption
(physisorption)
is then employed to characterize, design, and optimize the
morphology of
the material for specific applications.
The purpose of this chapter is to provide an understanding of
what is
known about physisorption with respect to analyses and
interpretation as
these relate to the morphology of high-surface-area solids. To
these ends,
we begin by describing the sequence of phenomena associated with
phy-
sisorption, its history, and simple modes of adsorption on
relatively flat sur-
faces. We then begin our tour, considering the adsorption
isotherm, region
by region (Section 3). Adsorption onmaterials having pores less
than 2 nm in
fundamental dimensions (called micropores) exhibits unique
challenges,
as their pores fill at extremely low pressures, before the
surface is completely
covered; these challenges are discussed in Section 4. For larger
pores, called
mesopores and macropores, there is often hysteresis between the
adsorbing
7Analysis of Catalyst Surface Structure by Physical Sorption
-
and desorbing trajectories along an adsorption isotherm; the
phenomenon of
adsorption hysteresis is discussed in Section 5. We also discuss
methodolo-
gies involved in determining porosity and pore size
distributions and close
the chapter with discussions of the experimental aspects of
adsorption: how
adsorption apparatus works and how to avoid some common mistakes
in
measuring and interpreting adsorption data.
2. THE PHENOMENON OF PHYSICAL ADSORPTION
2.1. History of adsorption in catalyst characterizationIt was
recognized well over a century ago that solid surfaces could
enable
gases or liquids to react under conditions in which they would
not react
in the absence of the surface. These observations were quickly
understood
to occur when molecules or atoms stuck to the surface, changing
both
their relative reaction energies and their local concentrations.
This
processatoms and molecules adhering to a surfacewas termed
adsorp-
tion by Kayser in 1881 (9). Irving Langmuir (10,11) later
expressed the
kinetics of steps associated with the individual adsorption and
desorption
processes by assuming that each atom/molecule reacted with an
array of
sites on the surface,
AS !ka A S AS 1:1AS !kd AS AS, 1:2
where [A] represents the activity of an adsorbed molecule, the
adsorbate, and
[S] the activity of a site on the solid, the adsorbent. The
adsorbate can also be
called the adsorptive, especially in situations in which
adsorbent and
adsorbate may be confused. At equilibrium, an equilibrium
constant,
K, reflects the ratio of adsorption and desorption rate
coefficients, ka/kd.
Langmuir was attempting to provide a quantitative analytical
back-
ground for heterogeneous catalytic reactions, and the sorption
processes
to which he was referring were generally exothermic and
activatedwhat
we would now call chemical adsorption, or chemisorption.
However, it was soon
understood that less exothermic processes could occur under the
general
heading of adsorption, and indeed, adsorption could occur
without for-
ming chemical bonds with the surface. Such nonchemical
adsorption is
called physical adsorption, or physisorption.
8 Karl D. Hammond and Wm. Curtis Conner Jr.
-
2.2. Choice of adsorbateThe choice of adsorbate (vapor to be
adsorbed) is largely dictated by the type
of information desired and the adsorbent that one wishes to
characterize. For
physical adsorption, the adsorbatemust be chemically inert with
respect to all
compounds present on the surface. Nitrogen and argon are common
choices
for this reasonand several other (very important)
considerations: they are
readily available, inexpensive, and relatively safe to handle.
Nitrogen and
argon also have another distinct advantage when it comes to
porous mate-
rials: they are very small molecules and can thus penetrate much
smaller
pores and cover smaller surface features than larger molecules
such as
cyclohexane.
Krypton has been employed as an adsorbate, particularly for
materials
with very low surface areas. The reason for this application is
that kryptons
vapor pressure at liquid nitrogen temperature is very low (2
Torr (12)),meaning that errors in the dead space (Section 8.2.1)
are less important.
Xenon is also an inert probe for low-surface-area materials;
like krypton, it is
typically used at liquid nitrogen temperatures. This choice is
one of practi-
cality: liquid nitrogen is much cheaper than liquid krypton or
xenon. It has
the drawback that 77 K is well below the triple points of both
krypton and
xenon. XenonNMR spectroscopy has also been employed to probe the
tex-
ture and chemistry of surfaces; see the many studies by
Fraissard et al.
(1318).
Water can be used as an adsorbate, but its highly polar nature
provides
some interesting analysis quirks. The most important of these is
that inter-
actions between water molecules are sometimes stronger than
interactions of
water with surface atoms/molecules, meaning that the isotherm
(see next
section) is convex to the pressure axis. This is particularly
true for graphite
and other carbonaceous (i.e., nonpolar) adsorbents. Water
adsorption there-
fore holds its own niche in the adsorption community and
requires
completely different interpretations than those characterizing
more typical
adsorbateadsorbent interactions.We do not discuss water
adsorption in this
chapter.
Carbon dioxide is an increasingly common choice as an adsorbate.
How-
ever, it poses challenges in the interpretation of adsorption
data: carbon
dioxide sublimes under the conditions of most adsorption
experiments,
which implies that any concept of monolayer adsorption is
complicated
by the differences between the surface phase, a supercooled
liquid reference
phase, and the solid phase that actually occurs at saturation.
It has also been
9Analysis of Catalyst Surface Structure by Physical Sorption
-
known to chemisorb on many materials (12), and it is also known
to inter-
calate between layers of carbonaceous materials (19,20), a
process that is
more akin to dissolution or absorption than adsorption.
Hydrocarbons such as methane or butane are commonly used as
adsor-
bates, but only for the purpose of estimating adsorption
capacities of that spe-
cific hydrocarbon, such as for interpretation of diffusion or
catalytic activity
data.We emphasize, however, that the surface area and porosity
accessible to
larger hydrocarbons such as benzene or nonane may be smaller
than those
available to small molecules such as nitrogen or argon. Indeed,
materials with
multiscale porosity may have many small pores connecting larger
ones; if the
adsorbing molecule is too large to diffuse through the small
pores at appre-
ciable rates, large fractions of the porosity may be
inaccessible. Furthermore,
if the surface is atomically rough, the adsorbate might not
adsorb on the
rougher areas, obscuring the surface area.
For the purpose of surface area determination and assessment of
porosity,
which are the primary topics of this chapter, the use of
nitrogen and argon is
nearly universal for the reasons discussed earlier in this
section. Other inert
gases, such as krypton or xenon, are also used in specialized
adsorption
experiments, such as determining occluded porosity (e.g., pores
accessible
to nitrogen but not xenon), assessing surface roughness (21), or
determining
surface areas of very low-surface-area materials.1 A list of
common adsor-
bates and their properties is included in Table 1.1.
It might be noted that nitrogen, oxygen, and carbon dioxide have
non-
zero quadrupole moments, whereas argon, krypton, and xenon do
not.
Quadrupolar interactions between adsorbing species and a surface
have been
invoked to explain several differences in adsorption phenomena
(22). These
include smaller molecular surface areas for nitrogen as a
consequence of
nitrogens ability to stand up on the surface because of
interactions
between its quadrupole moment and ions or polar functional
groups on
the surface, such as oxides or hydroxides. The coverage per
molecule is less
when the nitrogen molecule is perpendicular to the surface than
it is when
the molecule lies flat on the surface with its axis parallel to
the surface.
Quadrupolar interactions have also been suggested as the reason
why nitro-
gen adsorbs on zeolites and other microporous silicates at lower
pressures
than argon (which lacks a quadrupole moment). However, we note
that
silicalite-1 exhibits the same differences between nitrogen and
argon
1 Krypton and xenon are often used in low-surface-area analyses
because their vapor pressures at liquid
nitrogen temperatures are extremely low, minimizing errors in
the dead volume (Section 8.2.4).
10 Karl D. Hammond and Wm. Curtis Conner Jr.
-
Table 1.1 Common adsorbates and their properties under the
conditions of common adsorption experimentsVapor T (K) Am (
2) Am (m2/cm3 STP) g (N/m) V (cm
3/mol) P (Torr) CK () References
N2 77.36 16.2a 4.30 0.00888 34.7 760 4.16 (2325)
N2 87.30 0.00667 36.9 2130 2.95 (23,24)
N2 90.20 17.0 4.51 0.00606 37.7 2750 2.65 (2325)
Ar 77.36b 18.0c 4.78 27.5 230c (12,26)
Ar 87.30 16.6a 4.40 0.01242 28.6 760 4.25 (12,27)
Ar 90.20 14.4 3.82 0.01186 29.1 1020 4.00 (2426)
O2 90.20 14.1 3.74 0.0132 28.1 760 4.30 (25,28,29)
CO2 194.7b 14.1 3.74 28.1 1320d (28,30)
Kr 77.36b 20.8 5.52 1.78 (31,32)
Xe 77.36b 23.3 6.18 0.00187 (33,34)
aThere are two proposals (12): use 16.6 A2 for adsorption on
both carbons and oxides, meaning Am(N2, 77 K)16.2 A2 for adsorption
on carbons and 19.3 A2 foradsorption on oxides; or use 16.2 A2 for
adsorption on everything for nitrogen and use 13.8 A2 for argon
adsorption on carbons and 16.6 A2 for adsorption on oxides.Keeping
the value for nitrogen constant and changing the value for argon is
typical, although not necessarily motivated by any theoretical
considerations.bTemperature is below the triple-point
temperature.cThe value of 230 Torr refers to the supersaturated
liquid; the sublimation pressure at 77.36 K is 200 Torr (26). The
value Am18.0 A2 corresponds to BET plots usingthe sublimation
pressure rather than the vapor pressure (12).dThe value of 1320
Torr refers to the supersaturated liquid; the sublimation pressure
is of course 760 Torr (30).Symbols: T, temperature of bath; Am,
area of one monolayer for use with Equation (1.22); g, surface
tension; and V, molar volume for the Kelvin equation(Equation
1.23); CK 2gVl = RT log10 , the constant for the reduced Kelvin
equation (Equation 1.35); and P, saturation pressure. Rows in
boldface indicatethe most common and/or recommended bath
temperature for the adsorbate in question.
-
adsorption as ZSM-5, even though the former has no aluminum or
internal
hydroxyl groups. We are unaware of any independent spectroscopic
evi-
dence of significant quadrupolar interactions giving rise to
differences in ori-
entation or energetics under the conditions employed in
physisorption
measurements, but we do not discount the possibility.
2.3. Presentation of adsorption dataAdsorption data are largely
presented in two ways: isotherms, or plots of
quantity adsorbed against pressure or a similar abscissa at a
fixed temperature;
and isobars, or plots of quantity adsorbed against temperature
or inverse tem-
perature at constant pressure. Isotherms are muchmore common, as
it is typ-
ically much more difficult to control pressure at
non-atmospheric values than
it is to control temperatures. A list of common low-temperature
baths is
given in Table 1.2. The vast majority of adsorption isotherms
for the purpose
of catalyst characterization are nitrogen isotherms recorded at
77 K (liquid
nitrogen at its normal boiling point).
Adsorption isotherms are plots of quantity adsorbed against
pressure,
fugacity, activity, or chemical potential. Pressure is nearly
always the pre-
ferred abscissa in experiments: measuring the chemical potential
is (unfortu-
nately) rather difficult. Theoretical or simulated isotherms,
however, often
are based on chemical potential or activity instead. There is
little difference
Table 1.2 Common (and not-so-common) low-temperature baths used
in physical andchemical adsorptionBath Bath temperature Adsorbates
typically used Tt (K)
d
Nitrogen 77.36 K (boiling) N2, Ar, Kr, Xe 63.1526
Argon 87.30 K (boiling) Ar, N2, Kr, Xe 83.8058
Oxygen 90.20 K (boiling) Ar, N2a 54.36
CO2b 194.7 K (subliming) Hydrocarbons, CO2 216.55
Ammonia 240 K (boiling) Ammonia 195.40
Water/ice 273.15 K (freezing) Hydrocarbons, carbohydrates, CO2
273.16
Ambient air 292300 K H2,c COc N/A
Boiling water 373.15 K (boiling) 273.16
aLiquid oxygen baths are typically used only in specialized
experiments because of safety concerns.bCarbon dioxide baths are
usually dry ice in a low-freezing liquid such as acetone or
alcohol.cFor chemisorption.dThe symbol Tt indicates the
triple-point temperature.
12 Karl D. Hammond and Wm. Curtis Conner Jr.
-
between these choices: the vapor phase obeys the ideal gas law
at the pres-
sures involved in most adsorption experiments, and thus the
pressure and
chemical potential are related via the following result, derived
from the
GibbsDuhem equation:
m m kT log a m kT log f =f m kT log P=P 1:3
where P is the saturation pressure, the pressure at which the
bulk liquid isin equilibrium with the vapor at this temperature,
and the quantity P/P isa dimensionless quantity called the relative
pressure. Strictly speaking,
the reference pressure is not necessarily equal to the
saturation pressure,
but the difference in chemical potential between two points on
the isotherm
will still be equivalent to a difference in the logarithm of the
relative
pressure.
The quantity adsorbed is usually expressed inmoles or the
equivalent (see
following paragraph) and normalized by the mass of the
adsorbent. This nor-
malization is dubious in some cases: the quantity adsorbed per
gram of a very
dense adsorbent may be quite high per unit volume of adsorbent,
for exam-
ple, despite low values of the number of moles per gram on the
isotherm.
Comparing the quantity adsorbed per gram of ceria (r7 g/cm3)
wouldnot be a good comparison to the quantity adsorbed per gram of
alumina
(r4 g/cm3), for example, because of the difference in density.
Whencomparing quantity adsorbed across different materials, one
should take care
to normalize the plots in such a way that the resulting
comparison makes
logical sense.
When measuring adsorption isotherms, it is common to substitute
the
number of molecules (or moles) adsorbed, n, for standard volumes
adsorbed,
V. The standard volume is simply the volume that the molecules
would take
up in an ideal gas at standard temperature and pressure
(STP).What precisely
it means to be at STP is somewhat varied; we discuss this in
Section 9.1.
Because these quantities differ by a constant, it makes very
little difference
which one is chosen. Because using standard volumes typically
requires one
fewer set of conversions (see Section 8), doing so is typical.
From this point
on, we make no distinction between the number of molecules or
moles and
standard volumes, and, in all instances when n appears in the
equations in the
preceding text, it can be replaced byVwithout changing the
meaning of the
equations.
The astute reader will note readily that it is the logarithm of
pressure that is
proportional to changes in chemical potential, the driving force
behind
13Analysis of Catalyst Surface Structure by Physical
Sorption
-
adsorption. Using logarithmic pressure axes, however, is
hindered by the
fact that taking the logarithm obscures most of the interesting
parts of
the isotherm related to porosity (Section 3), which occur in the
decade
between P/P101 and P/P1.Examples of adsorption isotherms can be
found in figures throughout this
document. We consider some model isotherms in the remainder of
this sec-
tion. We emphasize that none of these models predicts observed
adsorption
isotherms in perfect detail in all regions of the isotherms.
Indeed, they tend
to be good models of adsorption only for idealized systems or in
relatively
narrow regions of the adsorption isotherm.
2.4. The Langmuir isothermThe Langmuir adsorption isotherm (11)
is the simplest model of adsorption
that yields useful results. The Langmuir isotherm is based on
the following
assumptions:
1. The surface consists of a uniform two-dimensional array of
identical
adsorption sites.
2. The probability of adsorbing on or desorbing from a site is
independent
of the number of nearby molecules (the loading, y).3. The
activation energy for desorption is equal to the heat of
adsorption,Q.
4. The vapor phase obeys the ideal gas law.
5. A site may not adsorb more than one adsorbate species at a
time (no
layering).
With these assumptions, the number of molecules striking the
surface per
unit area (the flux, Jm) is the following (35):
JmP1
2p1
MakT
rPk 1:4
where P is the pressure of the gas,Ma the molecular mass of the
adsorbate, k
Boltzmanns constant, and T temperature. If we define the loading
(fraction
of occupied sites) as y, then the fraction of empty sites is 1y.
If we assumethat there is a probability, a, of a molecule sticking
to the surface, then the
rate of condensation is
rc aJm 1y aPk 1y : 1:5By the third assumption, the rate of
evaporation is
re xmneQ=kTy, 1:6
14 Karl D. Hammond and Wm. Curtis Conner Jr.
-
where xm is the number of surface sites per unit area and n is
the frequency ofvibration along the reaction coordinate that
results in desorption. At
equilibrium, rc re and, thus,aPk 1y xmneQ=kTy: 1:7
Solving for the loading, we obtain the Langmuir adsorption
isotherm,
y akPxmneQ=kT aLP
bP
1 bP 1:8
where bak/(xmneQ/kT) has units of inverse pressure and is often
a fittedparameter. If we define Ps as the standard-state pressure,
then the quantity
KbPs is the equilibrium constant of the reaction, that is,
K T eDGads=kT eDSads=keDHads=kT eDSads=keQ=kT , 1:9
where G is Gibbs free energy, S entropy, and H enthalpy.
These results are summarized in Figure 1.1. The primary
characteristic of
the Langmuir isotherm is the last assumption: only a single
layer forms,
A + ka
A
A kd
A +
KineticskaP(1 q) = kdq q
1 q=
kakd
P = bP
q =bP
1 + bP
At equilibrium
1 2
0.25
0.5
0.75
1b = 10 atm1
b = 1 atm1
b = 0.1 atm1
P (atm)
q
Figure 1.1 The Langmuir (Type I) adsorption isotherm. In this
model, a single layer, ormonolayer, forms when adsorbate particles
adhere to specific sites on the surface,resulting in a horizontal
asymptote at unit loading. The loading, y, is the quantityadsorbed
divided by the quantity adsorbed at saturation (infinite pressure).
Highervalues of b indicate stronger adsorbateadsorbent
interactions.
15Analysis of Catalyst Surface Structure by Physical
Sorption
-
meaning there is saturation (a horizontal asymptote) once
monolayer capac-
ity is reached. Equation (1.8) is the basis for most theories of
heterogeneous
catalysis and chemical adsorption.
2.5. Monolayers to multilayersThe Langmuir adsorption isotherm
(Equation 1.8) is based on the
assumption that adsorption proceeds from zero up to
saturation,
y!1. The adsorbate species in the gas or liquid phase are in
equilibriumat a specific temperature and pressure with the species
adsorbed on the
surface. All species adsorbed are presumed to have equal
chemical
potentials, which do not depend on the presence of other
adsorbed spe-
cies. Thus, it is assumed that the interactions of each
adsorbing species
with the surface are identical and that interactions between
adsorbate
atoms/molecules on the surface are much weaker than the
interactions
with the surface.
When applied to reversible chemical adsorption
(chemisorption),
the Langmuir adsorption isotherm applies to both uniformly
distributed sites,
in which case the loading is proportional to the fractional
coverage of
the surface, and sites distributed on only a (possibly small)
fraction of the sur-
face. In the latter case, saturation of the Langmuir adsorption
isotherm rep-
resents the covering of only a part of the surface.
In physical adsorption, in contrast to chemisorption, the entire
surface
accessible to the adsorbate is involved. Thus, to a good
approximation,
the surface is entirely composed of active sites, and saturation
would be
achieved when the adsorbing species completely covers the
surface. This
would represent a monolayer of adsorbed species. Intuitively,
the maximum
coverage would reflect the closest packing of adsorbing species
on the sur-
face. In physisorption, the forces between an adsorbed species
and a surface
site are relatively weak, and the adsorbed species are
relatively free to move
across the surface and to change the surface sites with which
they primarily
interact. This facile, two-dimensional mobility also
differentiates physical
from chemical adsorption.
Physisorption occurs as a consequence of the interactions
between any
surface and molecules as the temperature approaches the boiling
(or dew)
point of the molecules in the gas. It begins at a temperature or
pressure
substantially below the actual pressure or temperature at which
bulk
16 Karl D. Hammond and Wm. Curtis Conner Jr.
-
condensation would occur. As an analogy, we can feel the effects
of
humidity even though it is not raining, or even damp,
outside.
When an atom or molecule approaches any surface, it is
influenced by
forces of attraction (e.g., van derWaals forces). A molecule is
also influenced
by such forces when it approaches other adsorbed molecules. In
physical
adsorption, whereby adsorbateadsorbent interactions are
relatively weak,
a molecule that encounters a molecule already adsorbed will be
influenced
by similar (albeit weaker) forces of attraction than those
involving a surface
site. It will have a probability of adsorbing on top of a group
of already
adsorbed molecules, which will probably be less than the
probability of
adsorbing if it had encountered the uncovered surface. The
difference in
the probabilities between adsorbing on top of already adsorbed
species
and on exposed solid surface is directly related to the
difference between
the energies of attraction between the surface and an adsorbing
molecule
and between an adsorbed molecule and an adsorbing molecule. The
surface
also contributes to the forces of attraction for adsorption of
the molecules in
the second and higher layers, but the forces are reduced because
the mole-
cules are at a larger distance from the surface.
Physical adsorption will therefore involve the formation of more
than a
single layer of adsorbed molecules as the pressure increases.
Thus, multilayer
adsorption is primarily a property of physical adsorption. It
can, however, be
found for chemisorption if subsequent layers differ in
composition, as in
atomic layer deposition (36).
If we wish to interpret the relationship between the quantity
adsorbed
and pressure under isothermal conditions (or the quantity
adsorbed
and temperature under isobaric conditions), it is necessary to
understand
multilayer adsorptionspecifically, the relationship between
adsorption
of the first layer (the monolayer) and adsorption of subsequent
layers. Prob-
ability (and thus entropy) leads one to conclude that the second
layer should
start to fill before the first layer is completed, provided
there is not an
extremely large difference in the heat of adsorption between
these layers.
As the number of adsorbed layers increases, it is also
reasonable to assume
that the heat of adsorption will eventually approach the heat of
condensation
of the adsorbate. Several relationships have been proposed to
express the
changes in the amount adsorbed and the pressure and temperature
for
adsorption up to and in excess of a monolayer. We discuss
several of these
throughout this chapter.
17Analysis of Catalyst Surface Structure by Physical
Sorption
-
2.6. BET theory2.6.1 The BET equationBy far, the best known
model of multilayer adsorption is that developed
by Brunauer, Emmett, and Teller (37), universally known in the
adsorp-
tion community as BET theory. This theory was developed to
describe
the initial adsorption of a monolayer and the simultaneous
adsorption of
multilayers. It starts with the premise that more than a single
layer can be
formed on a surface. It is further based on the assumption that
the energy
of interaction between the adsorbing species and the surface is
strongest in
the first layer and decreases for subsequent layers. To simplify
the ana-
lyses, Brunauer, Emmett, and Teller made a further assumption:
the
energy of interaction (heat of adsorption) between Nth and N1st
layersfor N2 and as N!1 is the same as the heat of condensation.
The BETtheory is also based on the assumption that the
corresponding sticking
coefficients and attempt frequencies for the second and higher
layers
are the same as for the second layer. Only the forces of
interaction
(and sticking coefficients and attempt frequencies) between the
surface
and the first layer are different in the BET theory.
Furthermore, it is
assumed that the volume of each adsorbed layer is identical.
This is equiv-
alent to assuming that the surface is flat (smooth on an atomic
scale).
Moreover, at PP, the saturation pressure, the number of layers
is infi-nite, and the adsorbate density becomes identical to that
of the
bulk liquid.
Just as in the Langmuir expression (Equation 1.8), it is
possible to express
the formation of a monolayer by considering the rate of
adsorption onto
empty sites and their rate of desorption. We express the
fraction of empty
sites as y0 and the concentration of those sites covered in the
first layer asy1 (and so on for higher layers). The heat of
adsorption in the first layeris Q1, and xm is the number density of
sites in the sample, as before.
By analogy to Equation (1.7), the rate of ad/desorption for each
layer i at
equilibrium is as follows:
aiPk 1yi xmnieQi=kTyi: 1:10
The attempt frequency ni is, at the microscopic level, the
vibrational fre-quency of the normal mode of the adsorbed complex
that, if sufficiently
excited, will result in desorption of a molecule. It is never
actually measured,
nor is it necessary to do so.
18 Karl D. Hammond and Wm. Curtis Conner Jr.
-
By definition,
X1i0
yi 1, 1:11
and the number of molecules adsorbed on the surface is nnmP
i01 iyi,
where nm is the monolayer capacity (total number of sites in the
sample). From
Equation (1.10) and the assumption that the second and higher
layers have
identical properties, we find the following set of
equations:
y1 a1xmn1
PeQ1=kT y0
y2 a2xmn2
PeQL=kTy1 a1xmn1
a2
xmn2P2e Q1QL =kTy0
y3 a2xmn2
PeQL=kTy2 a1xmn1
PeQ1=kTa2
xmn2PeQL=kT
2y0
. . .
yi a2xmn2
PeQL=kTyi1 a1xmn1
PeQ1=kTa2
xmn2PeQL=kT
i1y0 1:12
These equations can be written more concisely if we define
ay1/y0 andby2/y1. We define another constant, C, as their
ratio,
C ab a1a2
n2n1exp
Q1QLkT
1:13
and write the fractional coverage of each layer as follows:
yi abi1y0 biCy0: 1:14C is positive and dimensionless. From
Equation (1.11), we can write
y0 1X1i1
yi 1X1i1
bi" #
Cy0: 1:15
Because b
-
nnm Cb
1b 1bCb : 1:16
If we now factor in the assumption that n!1 as P!P, then we
know,from the definition of b, that
limP!Pb 1
a2
xmn2PeQL=kT , 1:17
which means that bP/P and thusn
nm CP=P
1P=P 1 C1 P=P : 1:18
Equation (1.18) is the BET adsorption isotherm. To find the
number of
molecules in one monolayer, which is proportional to the surface
area, it is
convenient to rearrange this equation into something easily
plotted, such as
P=P
n 1P=P 1
nmCC1
nmCP=P: 1:19
Equation (1.19) is called the BET equation. A plot of the
left-hand term, P/
n(PP), versus P/P yields (if the model assumptions are accurate,
at least) astraight line with slope (C1)/nmC and intercept 1/nmC.
The surface area,SABET, and the value ofC (often writtenCBET) are
therefore given in terms
of the slope, SBET, and the intercept, IBET, by the
following:
SABET AmSBET IBET and CBET 1
SBET
IBET1:20
where Am is the area one molecule occupies on the surface. The
general
approach employed in the BET theory is depicted in Figure
1.2.
2.6.2 The constant C in the BET equationThe value of C in
Equation (1.19) reflects the differences between the for-
mation of the first layer and the formation of subsequent layers
(i.e., a/b as inEquation (1.13)). In the BET formulation, this is
viewed as the difference
between the first and second layers, with all layers from 2 to 1
regardedas being similar. Differences between the reflection
coefficients and attempt
frequencies between the first and second layer are presumably
small (of order
unity), so that the value ofC is most sensitive to the
difference in interaction
energy (heat of adsorption) between the surface and the first
layer and
20 Karl D. Hammond and Wm. Curtis Conner Jr.
-
between the second and subsequent layers (these latter
differences tend to
the heat of condensation). That is,
C a1a2
n2n1exp
Q1Q2kT
exp Q1QL
kT
1SBET
IBET1:21
The value of C therefore reflects the difference in the heat of
adsorption
for the first layer compared with the heat of condensation. The
value of C is
thus sensitive to the enhancement resulting from adsorption in
comparison
with bulk condensation. Large values of C reflect high
adsorption energies
AdsorbentMonolayer
Liquid-like second and higher layers
kd,iqi = ka,iPqi1
Each layer is assumedLangmuir-like on top of prior
layers, but the first layer differsfrom all higher layers.
C = AeQ1Q2
kT > 0P/
Vads(1 P/P )=
1
CVm+
C 1
CVmP /P
BET Equation
P / P
P /P
Vads(1 P / P )
slope =C 1CVm
intercept =1
CVm
BET PlotVm =
1
slope + intercept
C = 1 +slope
intercept
P
Figure 1.2 Schematic representation of the BET adsorption
isotherm and its assump-tions. Themonolayer volume, Vm, from the
BET plot is often used to estimate the surfaceareas of catalysts,
provided that the value of C is reasonable and the assumptions of
themodel apply.
21Analysis of Catalyst Surface Structure by Physical
Sorption
-
for the first layer, whereas small values of C reflect small
differences in
adsorption compared with condensation.
The ability to calculate a monolayer volume from an adsorption
iso-
therm depends on the nature of the isotherm and, thus, on the
difference
in the energy for the interaction between the surface and the
first layer and
the energy of interaction between the first and subsequent
layers. Low
values of C calculated from the BET equation can mean that the
first layer
is not significantly enhanced in adsorption compared with
subsequent
layers, and it will therefore be difficult to determine a proper
value of
the monolayer volume. A rule of thumb is that C must be greater
than 50
for the BET theory to give rise to a reasonable calculation of
the monolayer
volume and, thus, the surface area (38). A C value of 20
corresponds to a
difference of greater than 1.92 kJ/mol in the heat of adsorption
of nitrogen
at 77 K, for example; the heat of condensation is 5.56 kJ/mol at
this
temperature.
At the other extreme, large values of C reflect (in the BET
theory) large
differences in the energy of interaction for the first layer
compared with sub-
sequent layers. The theory was developed to represent physical
adsorption
on a flat (or nearly flat) surface, not chemisorption or
adsorption in micro-
pores (i.e., pores less than approximately 2 nm in radius, for
which the
assumption of a flat surface is no longer valid). Thus, there is
an upper limit
to the value of C for which BET analysis is reasonable to
employ. Sing et al.
(12,38) suggested that values of C greater 200 found in BET
analyses would
make the analysis questionable, and therefore the surface areas
calculated
from such data should be used only with reservation. A C value
of 200 cor-
responds to a difference in the heat of adsorption of 19.2
kJ/mol for nitrogen
adsorption at 77 K. The shape of the BET isotherm as a function
of the C
parameter is shown in Figure 1.3.
It is apparent from Figure 1.3 that isotherms with low values of
C do not
exhibit a definite transition between the first and subsequent
layers (this tran-
sition occurs at a relative pressure of0.1), whereas isotherms
characterizedby higher values of C (>200) exhibit a transition
at much lower relativepressures. Intermediate values of C (and thus
the difference between the
adsorption in the first layer and that in subsequent layers)
give an easily dis-
tinguishable transition from monolayer to multilayer adsorption.
Higher
values of C imply strong adsorption (i.e., more than simple
physical adsorp-
tion onto a flat surface).
We emphasize that the calculation suggested earlier (C1
slope/inter-cept) can be extremely sensitive to the value of the
intercept. This point is
22 Karl D. Hammond and Wm. Curtis Conner Jr.
-
0.1 0.2 0.3 0.4 0.5106 105 104 103 102 101
Relative pressure (P/P) Relative pressure (P/P)
0
0.5
1
1.5
2
L
o
a
d
i
n
g
(
n
/
n
m
)
C = 1
10
50
100
200
1000
10,000
1
10
50
2001000
10,000
Figure 1.3 The shape of the BET isotherm varies significantly as
a function of the C parameter. Values of C between 50 and 200
(shadedregion) are generally considered reasonable; values outside
this range are found in situations for which the assumptions
underlying the BETtheory are likely invalid. The plot on the left
is represented with the pressure on a logarithmic scale; such plots
are typical for high-resolution(micropore) adsorption
isotherms.
-
particularly problematic because the BET surface area parameter
is often not
particularly sensitive to the intercept. The slope is always
positive (unless
C
-
difference is smaller than the uncertainty in the measurement of
Am. The
value of the area for a close-packed liquid is thus often used,
meaning
Am13.8 A2 per atom. A value of 14.2 A2 per atom is used in some
casesas well (39). In general, surface areas can be determined from
argon iso-
therms, but the value of the specific surface area assumed in
the calculation
should be specified along with the C constant in the BET
equation. Values
for other temperatures and other adsorbates are presented in
Table 1.1.
If the assumptions underlying the model hold, Vads(1P/P) is a
strictlyincreasing function of relative pressure in the range in
which the BET equa-
tion is applied, and if the value of C is reasonable, then
measured values of
the BET surface area are typically repeatable within 5%.
2.6.4 Rough estimates: The single-point BET surface areaThe
value of Am4.30 m2/cm3 STP for nitrogen at 77 K lends itself to
aneasy estimate of the BET surface area of a non-microporous
material. Pick a
point on the isotherm that is above monolayer coverage but below
any
mesopore filling; P/P0.2 is usually a good choice, although
values any-where in the range from 0.1 to 0.25 have been used by
various authors,
depending on their guess of Point B defined by Brunauer et al.
(37).
Now, multiply the volume adsorbed (in cm3/g) by 4.3 (or 4 for a
rougher
estimate). The result is a very rough estimate of the surface
area.
2.6.5 Weaknesses of the BET theoryThe BET theory was formulated
on the basis of a series of assumptions
(Section 2.6.1) that may or may not be too restrictive for a
particular system.
Fortunately, for a large fraction of solids, these assumptions
are appropriate at
relative pressures below P/P0.3 (i.e., an average of
approximately onemonolayer of adsorption on a smooth surface). One
problem with the for-
mulation of the BET theory is that each individual molecule
added on top of
another molecule in a partial layer is viewed as being adsorbed
with the same
energy as found for bulk condensation. The interactions between
molecules
in a given layer are also disregarded. Thus, the n1st and n2nd
layers maybegin to form before the nth layer is complete. This
picture also does not
fully account for the entropy of adsorption: it accounts for
changes in con-
figurational entropy (ways of arranging molecules on the
surface), but
neglects entropy arising from molecular mobility, as the
molecules are fixed
in position in the BET model (12). It is also difficult to
interpret rough sur-
faces in the context of BET theory, as such surfaces violate the
assumption of
an array of nearly identical adsorption sites.
25Analysis of Catalyst Surface Structure by Physical
Sorption
-
Halsey (40) observed that the BET theory includes the quite
untenable
hypothesis that an isolated adsorbed molecule can adsorb a
second mole-
cule on top, yielding the full energy of liquefaction, and that
in turn the
second molecule can adsorb a third.... One would expect that the
linear
picture of columns of molecules would not be formed, but instead
layers
would more closely approximate close-packed layers, in which
subsequent
adsorbing molecules can interact with more than one molecule in
the
layer(s) below. These effects, if accounted for, would all add
small correc-
tions to the BET equation, some of which become more important
for
specific systems.
The assumption that the second and subsequent layers all have
adsorption
energies that are equal to the energy of condensation neglects
the possibility
that the second layer may be influenced by the solid surface,
which is only a
few Angstroms distant. In many cases, the second layer will be
influenced by
the presence of the surface below and thus interact with the
surface. This net
interaction energy in the second layer will fall somewhere
between that of
the first layer and the energy of condensation. As layers above
the second
layer are formed, the differences between the first and the
second and the
second and the third layers become evident.
The BET theory therefore overestimates the rate at which
multilayers
form and does not account for the adsorption entropy. It also
simplifies
the energy of interaction between layers. However, these
problems occur
primarily at loadings above an average of one monolayer of
adsorption.
At loadings up to onemonolayer average coverage, the BET theory
has been
shown to provide the most consistent approach for the estimation
of the
exposed surface area for surfaces for which appropriate values
ofC are found
(12,38,41)that is, for C>50 (reasonably strong forces of
adsorption) andC
-
3. A TOUR OF THE ADSORPTION ISOTHERM: FROMVACUUM TO SATURATION
AND BACK
A physical adsorption isotherm can be analyzed to determine a
variety
of morphological characteristics of a solid. No single theory is
able to reflect
all physical interactions for sorption (adsorption and
desorption): from the
first few sorbing molecules, to a monolayer, to multilayers, to
condensation
of a liquid (or even a solid) throughout the system. Theories
have been
developed to represent each of the sequential processes
associated with
the measurement of sorption.
In this section, we offer a tour of the physical adsorption
isotherm,
starting at the lowest pressures that can be obtained by
conventional vacuum
equipment (typically P108 P or higher), progressing in order
throughthe following regions: micropore filling, surface coverage
(monolayer for-
mation), mesopore filling, macropore filling, saturation,
macropore empty-
ing, and mesopore emptying. The astute reader will recognize
that these
regions often overlapbecause the transition between them is
often
unclearand thus analysis is typically restricted to ensure
applicability of
the given model analyses. We use Figure 1.4 as a guide.
3.1. The micropore-filling region: 108
-
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 10
100
200
300
400
500
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 10
200
400
600
800
1000
Vol
ume
adso
rbed
(cm
3 S
TP
/g)
Vol
ume
adso
rbed
(cm
3 S
TP
/g)
Relative pressure (P/P)
Relative pressure (P/P)
Figure 1.4 Nitrogen isotherms at 77 K on SBA-15 silica samples
that incorporate bothmicropores andmesopores, suggesting different
regions of the isotherm as discussed inthe text. The top plot
indicates a solid containing larger mesopores than the lower
plot;the material represented in the lower plot has pores that lie
right on the transitionbetween micropores and mesopores.
28 Karl D. Hammond and Wm. Curtis Conner Jr.
-
adsorption, or HRADS (a term coined by Venero and Chiou (42)),
as dis-
cussed in Section 4.
A representation of the phenomena associated with this region of
the
adsorption spectrum is shown in Figure 1.5. The pores fill at
pressures well
below those required to give a monolayer on the exterior surface
because of
Figure 1.5 Schematic representation of pore filling in
micropores. The pores fill withadsorbate before the exterior
surface is covered. Reprinted with permission from Ref.
(43).Copyright 2005, American Chemical Society.
29Analysis of Catalyst Surface Structure by Physical
Sorption
-
the three-dimensional interactions between the sorbing molecule
and the
surface. It is not even clear what is the density of the sorbed
species when
the pores are filled, as this depends on the molecule-to-surface
and
molecule-to-molecule interactions, which can differ even for
physisorption.
3.2. Monolayer region: 0.05
-
three-dimensional network), these void spaces will gradually
fill with
condensing adsorbate as the pressure increases. The smaller
pores fill
and empty at lower relative pressures. At the same time, more
is
adsorbed as the exposed surface becomes thicker. The existence
of
hysteresisa difference in quantity adsorbed at the same relative
pres-
sure between the adsorption and desorption branches of the
isothermis discussed in Section 5. We stress that mesoporosity
often
is present when a sample consists of an agglomerate of
particles: such
porosity is created between the particles, and the voids created
by the
agglomeration are often similar in size to the dimensions of the
primary
particles (Figure 1.6).
Figure 1.6 Simplified representation of the process of
adsorption and desorption inmesopores, showing surface coverage,
pore filling, pore emptying (both by cavitationand otherwise), and
saturation. Note that the surface is covered before the pores
fill.Reprinted with permission from Ref. (43). Copyright 2005,
American Chemical Society.
31Analysis of Catalyst Surface Structure by Physical
Sorption
- 3.5. Adsorption on exterior surfaces: 0.45
-
range of pressures that can be measured with such transducers is
four
decades at best, with accuracy increasing near the top of the
range. Thus,
a 1000 Torr transducer (e.g., one that translates a pressure in
the range
01000 Torr into a voltage in the range 010 V) is most accurate
in
the range 1001000 Torr, is marginally accurate from 1 to 10
Torr,
and is not accurate at all below about 0.1 Torr (about 104 atm).
Con-sequently, at least two transducers are required to cover the
required
range of pressures, preferably a combination that includes a
transducer
that is accurate (1%) at approximately 105 Torr. Not all
adsorptioninstruments that claim to measure microporosity by HRADS
employ
pressure measurement systems (transducers) with this precision,
and in
the ones that do, the low-pressure transducer is often
optional
equipment.
In addition to the required pressure measurement accuracy, the
mea-
surements must be performed over a sufficiently long time that
adsorp-
tion equilibrium is achieved. This concern is often not readily
apparent:
if one watches the pressure drop, it may appear stable for
several minutes
before dropping as little as 104 Torr, but over a longer period
(say,30 min), the pressure will drop by more than two or three
times that,
and neglect of the continuing change can lead to significant
errors in
the determination of pore sizes. The experimental problem is
that the
heats of adsorption in microporous solids are unusually high,
often sig-
nificantly higher than the heat of vaporization. Furthermore,
the rates of
heat and mass transfer to and from the micropores are low
because of the
low pressures involved with samples that are essentially thermal
insulators
and held in glass (a good thermal insulator itself ). It
therefore takes a
considerable time for sorption equilibrium to be reached: as
much as
an hour or more between points may be necessary at the lowest
pressures
at which adsorption takes place. It is extremely important that
the mea-
surements be performed properlyconsequences of not doing so
range
from inaccurate determinations of pore sizes to nonphysical
results (such
as oscillating isotherms (44)).
The measurem