544.4 LJ85 2_ N.

Autor: Laidler, Keith 1

Ttulo: Chemical kinetics.
546575
10 111
N" Pat.:0040 13
 CONTENTS

PREFACE vn
1. The Analysis of Kinetic Results 1
2. The M easuremeni of Reaction Rates 31
3. The M echanisms of Elementary Processes 49
4. Elementaru Gas-phase Reactions 115
6. Elementary Reactions in Solution 198
6. Reaclions on Surfaces and in lhe Solid State 2fi(J
7. Complex Reactions 32/
8. Some Complex Reactions in the Gas Phase 8firi
9. Homo eneous Catalysis 4.'14
10. Some Reaction M echanisms in Solution 488
A ppendix A Determinaiioti of Rate Constants fi, 8
by N umerical M ethods
A ppendix B N umerical Values
IJIBLIOGRAPHY 542
AME INDEX 549
UBJECT INDEX 561

ix
SUB]ECT INDEX 563

Hydration shell, 224 Initial rates, 15 16

Hydride ion transfer, 194-195 Initiation, of oxidation, 415 424
Hydrocarbon oxidation, 421-424 of polymerization, 427-429, 523-527
Hydrogen atoms, production of, 352-353 Integration of rate cquations, 5-15
reaction, with chlorine, 135-136,367 Interatomic distance in surra c catalysis,
with ethane, 129 302-304
with hydrogen molecules, 55-60, 93, Interconversion of energy, 135-137, 170-
129, 130, 352-353 172, 186-187
with methane, 129 Internal pressure, 207-209
Hydrogen-bromine reaction, 325-326, Internal rotation, 73
328-329, 356-360 Iodides, reaction with amines, 33, 65, 203,
Hydrogen-chlorine reaction, 360-370 209,242
Hydrogen-ethylene reaction, 274, 276- Iodine-acetone reaction, 455-456, 514-
281 517
Hydrogen-iodide decomposition, 116-119 Iodine-atom combination, 182-184
on gold, 293, 297 Iodine-hydrogen reaction, 119-124
on platinum, 292-297 Iodine-monochloride-hydrogen reaction,
Hydrogen-iodine-monochloride reaction, 322
322 Ion decompositions, 173
Hydrogen-iodine reaction, 119-124 Ion-dipole reactions, 225-230
Hydrogen-ion catalysis, specific, 461 Ion pairs, 222
Hydrogen molecules, activation of, 449 Ion yields, 348, 352
reaction, with deuterium molecules, 315 Ionic reactions, 192-195, 210-230
with hydrogen atoms, 55-60, 93, 129, Ionic recombinations, 535-537
130, 352-353 Ionic-strength effect,@-222, 229-230
with methyl radicaIs, 93, 129, 131 Isolation method, 17-18
Hydrogen-nitric-oxide reaction, 138, Isomerization, cis-trans, 164-165
321-322, 446-447 of cyclopropane, 143, 159-163, 171
Hydrogen-nitrous-oxide reaction, 277 Isotope effects, 90-98, 130, 131, 494-495
Hydrogen-oxygen reaction, 417, 419-42l
on surfaees, 276, 278
Hydrogen peroxide, decomposition of, 1, Kinematics, 102-107
163-164, 435, 436 Kinetic isotope effects, 90-98
Hydrolysis, of acetals, 504-506 Kinetic resulta, analysis of, 1-30
of amides, 233, 506-509 Kinetic spectroscopy, 47
of epoxides, 509-510 Kinetic theory of collisions, 63-68
of esters, 233, 243, 493-503
of ethers, 510-511
of halides, 245, 492-493 Lactone hydrolysis, 511-512
of lactones, 511-512 Langmuir-Hinshelwood mechanism, 266,
')'-Hydroxybutyric acid, conversion into 274-276, 280-283
lactone, 20-21 Langmuir isotherm, 260-262
Hydroxyl radicaIs in solution, 355 Langm uir- Rideal mechanism, 26l, 20U,
276-280
Life, mean, of radicals, 178-l 6
Induction period, 324 Lindcrnann theory, 144-147
Inductive cIT ct, 239 Linear rl'c - ncrgy relationshipa, 2~(i tU I

Tnduetom ri ofl' ('~, 240 London oquaton, 55, 61

ln I'HZ;IUI otT( ot, 170 172, :14:1
J nhihitinn, 17
h nit.l'il\ oxido, :1\1:1 :mu, 404 I\n, 1\ 10, M/lIIIIIIII'I do ltl('thodH, :1:1 :J.I
41' 1/1 '1'1'I'II'IIIII(ltl', :If., I ,:1:11
!lf 11111fllllll 1'1'111\1111I1 , "UM "tlU 11'11 1111'1111 '11 I IIf 1'(1/11'1,11111 1111, " It 1
 chapter
Reactions on Surfaces and
6
in the Solid State
THE RATESOF a great many cheITcalreactions are influenced by solid
surfaces. This fact is frequently a serious complication to the study of
homogeneous gas reactions, in that an appreciable arnount of reaction
may occur on the surface of the reaction vessel and must be allowed for if
the homogeneous reaction is the one of special interest. In experiments
of this kind the amount of surface reaction can be determined by measur-
ing reaction rates under conditions in which the area of the surface is
varied over as wide a range as possible.
In addition to investigations in which there is incidental catalysis by
the walls of the vessel, a great many experiments have been made on the
kinetics of rcactona occurring on surfaces that are deliberately intro-
duced into tho systom. As a result of such work, it has been found that
many prOCeAA('1l oall ho made to occur much more rapidly by the introdu c-
tion of a suitablp /illl'fneo upon which the reaction proceeds; the solid
niuterial is tholl /inid 1.0 c/tl,l.\lyzethe reaction. In particular, a numbor of
processes thal, do 1101, ()('OUI' hornogeneously aI, an appl'eciablo rato can
111\ mado to oceur 1111 xu I fnc(A. The subicct of CU.tU.IYHil'lby solid su I'fl.\('('1l
1\
REACTIONS ON SURFACES AND IN THE SOLlD STATE 267
iA therefore of very great technical importance. The majority o investi-
I/;utionsof surface reactions have, in fact, been carried out with tho object
of finding a suitable substance to catalyze a technical gas reaction, and
have not been carried out under conditions suitable for an elucidation of
Lhefundamental processes concerned. In this chapter primary attention
will be devoted to a consideration of investigations of a more fundamental
nature, the essential criteria of such being that they include specification
of the surface used, in particular its smoothness and area, and that
absolute rates and activation energies of the reactions have been measured
accura tely.
The most important result that has emerged from such investigations
is that catalysis by surfaces involves specific chemical interaction between
Lhe surface and the reacting gas molecules, which must become adsorbed
on the surface before reaction can occur. This view was first proposed
in 1825 by Faraday, who studied the nature of adsorption on surfaces,
and considered that reaction occurred in adsorbed films. The original
idea was that the main effect of the catalyst was to cause the molecules to
be present in much higher concentrations than in the main body of the
gas. This concept is, however, shown to be false by the fact that in
ertain cases different surfaces give rise to different products of reaction;
e.g., ethyl alcohol decomposes primarily into ethylene and water on an
alumina catalyst, and mainly into acetaldehyde and hydrogen on copper.
This result, along with many others, clearly indicates that specific chem-
ical forces are involved at surfaces. In view of the importance of adsorp-
tion forces in connection with surface catalysis, some discussion will now
be devoted to modern views on this subject.
ADSORPTIONl
It is now recognized that two main types of adsorption may be clearly
distinguished. In the first type the forces are of a physical nature, and
the adsorption is relatively weak; the forces correspond to those assumed
in the van der Waals equation of state of gases and are consequently
known as van der Waals forces. These play only a very unimportant
part in connection with surface reactions (except for certain atomic
processes), since they are not sufficiently strong to influence appreciably
the reactivity of the molecules adsorbed. The second type of adsorption
1 For reviews see J. K. Roberts, "Some Aspects of Adsorption," Cambridge Uni.
vorsity Presa, Now York, 1939; S. Brunauer, "Physical Adsorption," Princeton II11 i
versity Press, Princeton, N.J., '1943; A. R. Miller, "The Adsorption of GaRPH1111
Solida," Cnmbridgo University Press, New York, 1949; K. J. Laidler, Chemisorpt luu,
in P. 11. EllllllotL (etl.), "Catalysis," vol. 1, Reinhold Publishing Corporatiou, NIV
York, IOIi~; li. M. W. '1'1'11])11011, in W. E. Garner (ed.), "Chemisorption," 11011"1'
wurth & ('o. (1'IIhliHhl'l'lI), 1.ld., London, 1955.
 (;lIEMICAL KINETI" ACT\ONS ON SURFACES AND lN TIlE SOUD STAT 9
is considerably stronger; according to Langmuir's original concept of il.,'
tiVi! HiLpH,which Taylor rcerrcd to as actiue centers. H wa ()f'iI-\ll/dly
the adsorbed molecules are held to the surface by valence forces of LII"
111(III~hL t.hat. atoms present on surface peaks would be tho most uetivo,
sarne type as those occurring between bound atoms in molecules. '1'11"
IlIll, d, lias later bccome clear that the matter is more complica teu than
heat evolved in this type of adsorption, known as chemisorption, should
t 1111'1, nctive centers may actually correspond to certain types of lattice
be of the same order as that found with chemical reactions and, in agro dtril('L, This question of the nature of active centers is considered in
ment with this idea, values of 10 to 100 kcal have been observed. 13y
fllom dotail late r in this chapter.
contrast, the heat changes found with van der Waals adsorption are usu-
alIy less than 5 kcal per mole. Another complication that exists in connection with chemisorption
11. thuL there are interactions, usually of a repulsive nature, between
An important consequence of this concept of chemisorption, regardod 1\ tOlllS or molecules adsorbed side by side on a surface. The first evidence
by Langmuir as an essential part of his theory, is that after a surface haa
1!lI' this was obtained by Roberts,' who measured heats of adsorption of
become covered with a single layer of adsorbed molecules it is essentially
hydrogen on a tungsten surface that was considercd to be quite smooth.
saturated; additional adsorption can occur only on the layer already
'l'ho heat was found to fali from about 45 kcal per mole for a bare surface
formed, and this usualIy takes place only with some difficulty. This idea
I() 15 kcal for what was believed to be a fully coverod surface; later evi-
of the unimolecular layer is capable of direct experimental test, and a
di'nce,2 however, indicated that the surface was only 70 percent covered,
number of investigations have shown that a definite adsorption limit
und that the heats of adsorption become close to zero when the surface is
occurs, beyond which further molecules cannot be adsorbed readily.
nctually fully covered. If the surface is really homogeneous, this change
When the surface area is known, this limit is found to correspond to the
in heat of adsorption must be due to repulsive interactions between the
existence of a unimolecular layer.2 This result is good evidence for the
ndsorbed molecules; when the surface is sparsely covorcd, the molecules
theory that chemical forces are involved in chemisorption.
are sufficiently far apart so as not to interact with each other, but the
It was suggested by Taylor! that chemisorption is frequently assoc-
repulsivo interactions beco me more important as the coverage increases.
ated with an appreciable activation energy and may therefore be a rela-
Additional evidence for repulsive interactions has been obtained by
tively slow process; for this reason it is now frequently referred to as
Emmett and Kummer" using an isotope method, and by Weber and
actiuated adsorption. The energies of activation are often of the order of
Laidler! on the basis of measurements of rates of desorption.
20 kcal, and consequently the adsorption is extremely slow at low tem-
peratures; under these conditions, van der Waals adsorption, which
requires no activation energy, will predominate. ADSORPTION ISOTHERMS
It must not be assumed that ali surfaces are smooth and that
adsorbed molecules are arranged on them in simple layers. In reality Work on chemisorption has been concerned with equilibria and with
ratos. The amount of gas adsorbed after equilibrium is established
surfaces are never smooth from the molecular standpoint, and the kind of
depends on various faetors, including the nature of the surface and the
behavior to be expected on rough surfaces was discussed by Taylor,4
Constable,6 and oLhers.6 Some surface sites will be more active than absorbate, the temperature, and the pressure. If for a given system one
others, and chemical processes wiU occur predominantly on the most koeps the temperature constant and studies the amount of adsorption as a
function of pressure, the resulting relationship is known as an adsorption
1 r. Langmuir, J. Am. Citem. Soe., 38, 2221 (1916).
isotherm. A number of such isotherms have been suggested, some being
2 r. Langmuir, ibid., 40, 1361 (1918); F. A. Paneth and W. Vorwerk, Z. Physik.
Chem., 101, 445, 480 (1022); J. K. Roberts, Proc. Roy. Soe. (London), A152, 445 (1935); empirical and others obtained theoretically. Of the theoreticaJ equations
cf. Miller, op. cit., PI>. 16 18. the simplest is that of Langrnuir," whose isotherm has the special sig-
H. S. Taylor, J. Am. Chem, Soe., 53, 578 (1931); Chem, Rev., 9, 1 (1931).
3 nificance of being the one that applies to the ideal case of chemisorption
H. S. Taylor, Proc. Roy. Soe. (London), AI08 (1925); J. Phys. Chem, 30, 145
4
(1926); cf. N. P. Kpior and S. Z. Roginsky, Dokl. Akad. Nauk., SSSR, 57,151 (1947); 1 J. K Roberts, Proc. Roy. Soe. (London), A152, 445 (1935); cf. W. G. Franken-
J. Phys. Chem, ussn, 28, 897 (1949); J. T. Kummer and P. H. Emmett, J. Am. burg, J. Am. Chem, Soe., 66, 1827 (1944); O. Bccck, Rev. Mod. Phys., 17,61 (1945).
Chem, Soe., 78, 28HO(1\)5[); P. H. Emmett and J. T. Kwnmer, J. Chim, Phys., 47. 'E. K. Rideal and B. M. W. Trapnell, J. Chim, Phys., 47, 126 (1950); B. M. W.
67 (1950); H. S. Taylor und S. C. Liang, J. Am. Citem. Soe., 69, 1306 (1947); H. Sadek Trapnell, Proc. Roy. Soe. (London), A206, 39 (1051).
and H. S. Taylor, ibid., 72, 1168 (1950).
3 P. I-I. Emmett and J. T. Kummer, J. Chim. Phys., 47, 67 (1950); J T. Kumnu-r
sF. H. Constablo, Proc. Roy. Soe. (London), AI08, 355 (1925)
and P. H. Emmett, J. Am. Chem, Soe., 73, 2886 (1951).
6E. Cremer and CL M. Schwab, Z. Physik. Chem., A144, 243 (1929); G. M. 4 J. Weber and K J. Laidler, J. Chem. Phll~., 18, 1418 (1950); 19, 1089 (1951).
Schwab, ibi , B5, 406 (J 029).
5 Langmuir, J. Am. Chem, Soe., 38, 2221 (1916); 40, 1361 (1918).
 1
N SURFACES AND lN TlIE SOL1D STAT
o CllEMICAL KINETICS

n a pcrfcctly smooth surface with no interactions bctween adsorbcd l-'~ip\

molccules. The Langmuir isothcrm therefore has an importancc in
adsorption theory which is equivalent to that of the idcal-gas laws, and it -------------------- ---
is convenient to speak of adsorption that obeys Langmuir's isotherm as
ideal adsorption. The equations for ideal adsorption play an important
role in surface kinetics, and they will now be derived from different points
of view and for several differcnt situations.
The Langmuir Adsorption Isotherm. The simplest situation exists Kp
--- /1=t+K;
when the gas atoms or molecules occupy single sites on the surface, and
are not dissociated; the adsorption and desorption processes may then
be represented as
I G
G+-S-~ I
-S- /1zKp
Langmuir's well-known kinetic derivation I of the isotherm is briefly as
follows. Let O be the fraction of surface that is covered and 1 - O the
/
fraction that is bare. The rate of adsorption is then k1P(1 - O), where p p
is the gas prcssure and Tel a constant; the rate of desorption is Te_IO. At
equilibrium the rates are cqual, so that Fig. 68. The relationsbip between tbe fraction of surface
covered o and tbe pressure, for the case of absorption without
O Te dissociation in which the Langmuir equation (2) applies. The
-- = -pl = J(p (1) rute varies with pressure in exactly the same manner for both
1- O Te-I
1\ unimolecular reaction [Eq. (29)1 and a bimolecular reaction

where J(, equal to TeI/Te_I, is a constant. This equation can be written as proceeding by a Langmuir-Rideal mechanism [Eq. (54)1.

J(p
(2)
O = 1 + Kp thercfore be written as
(3)
VI = TeIP(1 - 0)2
Figure 58 shows the type of behavior that corresponds to this equation,
and givcs the behavior at the limits of low and high pressures. The desorption process involves reaction between two adsorbed atoms,
Adsorption with Dissociation. ln certain cases there is evidence and the rate is therefore proportional to the square of the fraction of
that thc proccss of adsorption is accompanied by the dissociation of the surface covered,
moleculc on the surfacc. It is found, for example, that hydrogen is LI = Te_102 (4)
adscrbcd on the surf'acos of many metals in the form of atoms, each of
which occupies a surf'ace site. Similarly methanc adsorbcd on metals is At equilibrium the rates are equal, whence
usually diasociatcd into CITa, CH2, and hydrogen atoms. ln the case of O (TeI)\6 (5)
--- -p
the dissociat ion of a molccule into two species (e.g., of H2 into 2H) the 1- O - LI
= J(~"p'/J (G)
process may 1>0rcprcscntcd as

I I where J( is equal to TeI/Te-1 This equation can be written as

G2 + -S-S -~ GI G
I
-S-8- J('/Jp\6 (7)
/1 = 1 + J('/Jp~'J
The process of adsorption must be considered to be a reaction between
the gas moleculc and two surface sites, and the rate of adsorption may When the pressure is sufficiently small so that K~"p~" is much amnllcr
than unity, the fraction covered is proportional to p~". Another case o
1 Langmuir, loco cito
 02
CHEMICAL KINETICH 1(J~At:TlONS ON SURFAt:JtH AND IN TlIE SOUD STATE 3
special interest occurs when the pressure is high; Eq. (7) may be WI"tt,('1I
as III'OOlllOadsorbcd 01\ two diff'crcnt sets of surface sites, in which caso L!wro
1Mno compotition. The isotherms would then be the same as if two com-
1 - 8 = . -. - -
1..-." .~ (8)
plc\Ldyseparate surfaces were involved; i.e., Eq. (2) will apply to each gas .
Statistical Treatment of Ideal Adsorpton.! By the methods of sta-
so that when KYJp~~ 1 the fraction of the surface that is bare, 1 - 8, ia l,iHLicalmechanisms it is possible to derive equations that are equivalent
inversely proportional to the square root of the pressure. \'0 Lheisotherms given above, but which express the constants in terms of
Competitive Adsorption. The isotherm for two gases adsorbed on pnrtition functions. The case of adsorption without dissociation will be
the same surface is of considerable importance in connection with tho considered first. Let the volume of the gas be V cc, and the area of the
kinetics of surface reactions involving two substances. Let the fraction surface be S sq cm. At equilibrium the number of bare sites may be
of the surface covered by molecules of type A be 8, and the fraction written as N., and the number of adsorbed molecules as Na; the total
covered by B be 8'; the fraction bare is 1 - 8 - 8'. The substances A number of molecules in the gas phase is No. It is now convenient to
and B are assumed to be adsorbed without dissociation; the modification define concentrations as follows:
for dissociation can easily be made. The rate of adsorption of A is given
by loncentratioti in gas phase: Cg = No/V molecules per cc
oncentration of bare sites: c. = N./ S sites per sq cm
VI = lclP(1 - 8- 8') (9) oncentration of adsorbed molecules: c, = Na/S molecules per sq cm
where p is the partial pressure of A. The rate of desorption of A is
The equilibrium constant for the adsorption process is
lc_18
V_I

At equilibrium these rates are equal, whence

=
(10)
s, = c:~. = (No~V)N. (15)

8 By the use of the procedures explained in Chap. 3, the equilibrium con-

. _, = Kp (11) stant can be written as
where K is equal to lcI/lc_l. ln the same way it can be shown that for the K F
= _G_ef/kT (16)
equilibrium of the gas B, c FoF.
8' where the F's are the partition functions per unit volume, and E is the
1 - 8 - 8' = K'p' (12) energy released, per molecule, at OOK. This can more conveniently be
expressed as
where p' is the partial pressure of B, and K' is the equilibrium constant
for the adsorption of B. Equations (11) and (12) are two simultaneous ~ = L: e,/kT (17)
equations, the solution of which gives, for the fractions covered by A coe. Fof.
and B, where fa and f. are the total partition functions (rather than those per
8 _ Kp
unit volume or unit area). If 8 is the fraction covered,
- 1 + Kp + K'p' (13) ca 8
(18)
K'p'
~=1-8
and 8' = 1 + Kp + K'p' (14)
so that _8_ = Co ~ e,/kT (19)
Equation (13) reduces to (2) in the event that p' = O (gas B is not 1- 8 Fof.
present) or thnt K' = O (gas B is not adsorbed). It is seen from Eqs. The concentration Co is equal to p/kT, and Fo may be written as
(13) and (l4) Lhat the fraction of the surface covered by one gas is
F = (27rmkT)~ b (20)
reduced if the prcssure of the other gas is increased; this is due to the fact o h3 o
that the molecules are competing with one another for a limited number
of surface sites, and one may speak in this case of competitive chemisorp- 1 R. I-I. Fowler, Proc. Cambridge Phil. Soc., 31, 260 (1935); R. I-I. Fowlor untl

tion. There is an indication that in some special cases two gases may E. A. Guggenheim, "Statistical Thermodynamics," p. 426, Cambridge Univeraity
Press, New York, 1939.
 CHEMICAL KINETICS RltACTIONS ON SURFACES AND IN TIIE SOLID STATE 266
whcro bg representa the rotational and vibrational factors in the partition 'I'ho thcorctical significanco of this equation has been considered by a
function. The adsorption sites have very little freedom of motion so that numbor of authors, and it has been shown- that an equation of the same
the partition function f. may be taken as unity. The partition function Iorm as Eq. (25) is obtained if there is adsorption on a variety of sites
for the adsorbed molecules als o involves only internal factors, which may thaL have a certain type of statistical distribution. The isotherm may
be written as ba The adsorption isotherm thus becomes
ultcrnatively be explained in terms of repulsive forces between adsorbed
(J _ h
3
b; ./kT (21) III olecules.
1- (J - P (271"m)~(kT)% li;, e Another isotherm that has frequently been applied to kinetie prob-
lcms is that of Slygin and Frumkin," which is
This equation has the same form as Eq. (2), but now the constant K is
given in explicit formo 1
(J = lln ap (26)
In a similar way it can be shown that for adsorption with dissociation
the isotherm is
where f and a are constante. This isotherm can be interpreted theo-
_(J_ _ H h~ ~ ./kT (22) retically on the assumption that heat of adsorption is a function of the
1- (J - P (271"m)~~(kT)~'bgH e . fraction of surface covered.
where E is the energy liberated at OOK when one molecule is adsorbed.
For the' adsorption of two gases on the same surface Eqs. (13) and (14) CHEMICAL REACTIONS ON SURFACES
apply, with K given by
Mechanisms of Surface Reactions. A reaction occurring on a sur-
3
K - h b ./kT (23) face may usually be regarded as involving five consecutive steps, as
- (271"m)~(kT)H bg e follows:
A corresponding expression holds for !C. 1. Diffusion of the reacting molecules to the surface
The equations above apply only to the situation in which the 2. Adsorption of the gases on the surface
adsorbed molecules are localized on the surface ; that is to say, they have 3. Reaction on the surface
no translational freedom. At the other extreme is the situation in which 4. Desorption of the products
the adsorbed molecules can move perfectly freely in two dimensions. 5. Diffusion of the desorbed products into the main body of the gas
Ths isotherm that then applies is
At one time it was fairly generally believed that one of the diffusion
(} - h ba ./kT (24) processes, 1 or 5, was the slowest process, and, therefore, determincd the
- P L(271"m)~2(kT)~ bg e
overall rate. More detailed investigations of surface reactions showed,
where L is the total number of molecules that are adsorbed on 1 sq em however, that this could not be the case, except perhaps in certain tech-
when tho surface is fully covered. This relationship only applies for nical processes involving porous catalysts. This is evident from the
sparse covorago, since when there is a greater degree of coverage there can fact that heterogeneous processes nearly always involve appreciable
no longer bo a Iroo movcment of the molecules on the surface. activation energies, whereas diffusion in the gaseous state involves no
Deviations from Ideal Behavor.! It is beyond the scope of this activation energy j the diffusion process is, therefore, much more rapid
book to considor Lho treatments that apply to inhomogeneous surfaces, than the overall process and cannot constitute its slow step. Langmuir!
and to tho caso of intoractions between adsorbed molecules. Brief refer- presented the same type of argument in a somewhat different form when
ence should, howovor, be made to two isotherms that are applicable to he showed that, in order for diffusion to be the slow step, it would be
certain kinotie problorns. The first is that of Freundlich," which relates necessary to postulate such a thick diffusion layer on the surface that iL
the volumo v of gas adsorbed to the pressure p as follows: would be visible. Since diffusion is slower in solution than in the gas
v = kpl/n (n> 1) (25) 1 J. Zelelowitch, Acta Physicochim. URSS, 1,961 (1935); G. D. Halsey anel 11. H.
Taylor, J. Chem. Phys., 16,624 (1947); R. Sips, ibid., 18, 1024 (1950); F. C. Tompkina,
1 For a review RCO IC.J. Laidler, Chemisorption, in P. H. Emmett (ed.), "Cataly- Trans Faraday Soc., 46, 569 (195Q).
sis," vol. 1, pp. 98114, Itcinhold Publishing Corporation, New York, 1954. 2 A. Slygin anel A. Frumkin, Actu Phusicochiui. URSS, 3, 791 (1935).
2 H. Freundlich, "Kapillarchemie" Leipzig, 1909.
3 L Lsngmuir, J. Am. Chem. Soc., 38, 1145 (1916).
 66 CHEMICAL KINETICS
phase, it may sometimes be the rate-determining step in solid-liquid
reactions.
The processes of adsorption or desorption are much more likely to b
the slow steps in heterogeneous reactions, since both may involve appreci-
able energies of activation. The activation energies for desorption are
particularly high, and it is likely that in very many reactions the desorp-
tion of the products is the rate-determining step. ln practice, however,
it is not always convenient to separate steps 3 and 4 because one usually
does not know the rate of desorption of the products; it is usual, therefore,
to regard the reaction on the surface, giving the gaseous products, as a
single step. This concept is in fact the basis of the usual modern treat-
ment of surface reactions, due to Langmuir ' and Hinshelwood," This
treatment involves, first, obtaining an expression for the concentrations
of reactant molecules on the surface, and then expressing the rate of
formation of gaseous products in terms of these surface concentrations:
the rate is then expressible in terms of the concentrations of the gaseous
reactants. If there is a single reactant, the surface process is a simple
unimolecular change; if there are two, A and B, reaction may take place
between two molecules adsorbed on neighboring surface sites, and the
probability of this happening is proportional to the individual concentra-
tions of adsorbed A and adsorbed B. The Langmuir-Hinshelwood
mechanism for a reaction between A and B may be formulated as follows.
I A B A-B I I
A + B + -s-~ I I ~ I I ~-S-s- + products
-S-S- -S-S-
(adsorption) (activated complex)
Another type of mechanism for surface reactions was also considered
by Langmuir. According to this mechanism the reaction occurs between
a gas molecule and an adsorbed moIecuIe, so that only one of the reactants
has to be adsorbed. This mechanism may be represented as
B Kp
A + I -t I + products
-S- -S-
It is not noccssary that A is not at a11adsorbed; it is simply postulated,
in this mechanism, that an adsorbed A molecule does not react. lnterest
in these mcchanisms was revived by RideaP and, as will be seen, they
probably do apply to certain atom and radical combinations. They may
also play a rolo in other reaction systems, but on the whole the Langmuir-
1I. Langmuir, 'I'rans, Faraday Soc., 17, 621 (1921).
2C. N. Hinshelwood, "Kinctics of Chemical Change," Clarendon Press, Oxford,
p. 145, 1926; p. 187, J!)40.
3 E. K. Rideal, Proc. Cambridqe Phil. Soc., 35, 130 (1939); Chem. Ind. (London),
62, 335 (1943).
IONS ON SURFACES AND IN THE SOLlD STATE 267
11icI<-ltlmoohanisms do not appear to be as common as the Langmuir-
1IIIIHholwood ones, in which reaction occurs between two adsorbed
IlIOI('culcs.
In connection with the mechanisms of surface reactions, an important
c'ollcept is the molecularity, which is the number of molecules that come
lo~cther during the course of reaction; it is convenient not to count the
Htlrface sites. The molecularity of a surface reaction is deduced from the
.lnetica on the basis of the experimental results and of theoretical con-
lIiderations. The actual relationships between molecularity and order
wiU be considered in the fo11owingsections, in which reactions are classi-
cd according to their molecularity. One or two examples may here be
IIlcntioned briefly. Reactions involving a single reacting substance are
usually, but not invariably, unimolecular. The mechanism of the surface-
('l\talyzed ammonia decomposition (cf. p. 271) is, for oxample, usual1y
unimolecular. On the other hand, the kinetics of the decomposition of
l\cctaldehyde on various surfaces! can only be interpreted on the hypoth-
esis that two acetaIdehyde molecules, adsorbed on neighboring surface
sites, undergo a bimoIecuIar reaction. Reactions involving two reacting
substances, such as the reaction between nitric oxide and oxygcn on
glass," are usually bimolecuIar. When reactant moIecules are dissociated
on the surface, the reaction may invoIve interaction betwcen an atom or
radical and a molecule; for example, the exchange rcaction bctwccn
ammonia and deuterium on iron" must be rcgardcd as a bimolccular intcr-
action between a deuterium atom and an ammonia molecule (cf, p. 283).
UNIMOLECULAR SURFACE REACTIONS
The simplest treatment of surface reactions involving one molecul.
of a reactant is in terms of the Langmuir adsorption isotherm [Eq. (2)],
according to which the fraction (J of the surface covered is related to thc
pressure p by
(27)
(J = 1 + Kp
The rate of reaction is proportional to (J and may therefore be written aA
v = k2(J (28)
_ k2Kp (2\))
- 1 +Kp
where k2 is the proportionality constant. This formulation is bascd on
the assumption that the adsorption equilibrium is not disturbed by Lho
occurrence of the reaction, a condition that is generally satisfied.
1 P. C. Allen and C. N. Hinshelwood, Proc. Roy. Soe. (London), A121, 141 (1\)21').
2 M. Temkin and V. Pyzhev, Acta Physicochim. URSS, 2, 473 (1935).
3 J. Wcber and K. J. Laidler, J. Chem. Phys., 19, 1089 (1951).
268 CHEMICAL KINETICS REACTIONS ON SURFACES AND IN TlIE SOLID STATE

The relationship given in Eq. (29) between the rate and the pressur tions ariso when the inhibitor I is a product of thc rcaction: H()Il\(lI\XI~Il\I'I\
I
has exactly the same form as that in Fig. 58, which relates surface cover- of this are considered later.
age and pressure. At sufficiently high pressures the rate is independent Activation Energies. The rate constant k2 appearing in Il:q, ('lll
of the pressure, which means that the kinetics are zero order. Undcr will obey the Arrhenius law, which may be expressed as
these conditions Kp 1, so that Eq. (29) reduces to
d ln lc2 = E ('~(\)
v = k2 (30) dT RT2

At low pressures, when Kp o; 1, Eq. (29) reduces to The temperature variation of the equilibrium constant K follows til
v = k-Kp (31) unalogous van't Hoff relationship

and the kinetics are first order. Some examples of both cases, and of dlnK
(['f = - RT2 (37)
intermediate cases where the order is between unity and zero, are con-
sidered later. where is the heat evolved, per mole of reactant gas, in the adsorption
lnhibition. One complication that frequent1y exists in surface reac- processo The temperature dependence of the rate may now be con-
tions is that a substance other than the reactant is adsorbed on the sidered with reference to the two limiting cases.
surface; the result is that the effective surface area, and therefore the If the pressure is low, the rate is given by Eq. (31), so that the first-
rate, is reduced. Suppose that a substance A is undergoing a unimolecu-
order rate constant TcI is given by
lar reaction on a surface and that a nonreacting gaseous substance I
(referred to as an inhibitor or poison) is also adsorbed. If the fraction of TcI = k2K (38)
the surface covered by A is 0, and that covered by I is 8;, the fraction
covered by A can be expressed by [cf. Eq. (13)] Using Eqs. (36) and (37), it follows that
Kp d ln v = d ln kI = ~ + d ln K (39)
= 1+ Kp + K.Vi (32) dT dT dT
E-
dT
(40)
where Pi is the partial pressure of the inhibitor and K, its adsorption con- = RT2
stant. The rate of reaction, equal to k20, is therefore
The apparent activation energy E; is therefore given by E - ; it is tho
v __ ~k,::..~--,p,--==-- (33) "true" activation energy E reduced by ~eat of adsorption of tho
- 1 + Kp + Krp,
reactant.
In the absence of inhibitor this reduces to Eq. (29). If the pressure is sufficiently high, Eq. (30) applies, and the apparent
A case of spccial intcrest occurs when the pressure of the reactant is activation energy is now equal to the true activation energy E.
sufficiently low 80 Lhat tho availablo surface is only sparsely covered by These relationships may be considered with reference to the potcn-
the reactant. Tho term Kp is thcn negligible in comparison with tial-energy diagram shown in Fig. 59. Reaction first involves tho
1 + KiPi, and tho rato of the inhibited reaction is passage of the system over an initial energy barrier to give the adsorbcd
state, the energy of which is lower by than that of the initial stato.
lcd(p The system then passes over a second barrier of height E. If the prCSH\II'1
v = -=-l-+'=-';K~iPi (34)
is low, most of the reactant molecules are in the unadsorbed state, and 1.0
If the inhibitor iR very strongly adsorbed, Krp; is large compared with pass to the activated state they only have to acquire the energy 1e .

unity, so that At high pressures, however, the equilibrium favors the adsorbcd Ht,nt,\I,
and the system has to acquire the energy E in order to pass to tho 11111 i
k2Kp
(35) vated state.
v = K.Vi
When a reaction is inhibited, the activation energy is modificd \ly 11,,1
The reaction is thoref'oro first order in reactant, and the rate is inversely energy of adsorption of the inhibitor. There are several possibililll'll, 111\\
proportional to tho inhibitor pressure. Special cases of the above equa- simplest of which is that corresponding to Eq. (35), for which tho "'lilll.IIIIL
 270
CHEMICAL KINETICS R.EACTIONS ON SURFACES AND IN TIIE SOLID STATE t
A few examples of unimolecular surface reactions will IIIIW 1)(
discussed.
Decomposition of Ammonia. Ammonia decomposes into nit,I"OKoII
und hydrogen on a number of metal surfaces, and several kinetic invosga-
tions have been carried out. A study of the decomposition on tungston
und platinum surfaces was made by Hinshelwood and Burle 1 In their
work with tungsten they used a commercial tungsten-filament lamp as a
reaction vessel. The vessel was kept in melting ice and the filament
~ heated electrically to a constant temperature, which was measured by
G>
c:
G>
determining its resistance. The course of the reaction was followed
!.!? manometrically.

.f
ce 200
.........
180

160I
V
Adsorbed reaelant ,/ l{oo mm initial NH3 pressure
140I -I'
ou
Produet ~ 120
/
8. V
5 100
/
u
ouG>
!f. V
Reaetion eoordinate
E 80
/ ./

Fig. 69. Potential-energy diagram for a unimolecular sur-

E

60)
./ ~ 100 mm initial
, NH3 pressure
face reaction.
/~
40I
~ 50 mm initial t!m3 ~ressu~e I.
o No hydrogen added
is weakly and the inhibitor strongly adsorbed. The inhibition constant 20I ~

K, varies with temperature according to the equation

0 ,V
0
- --
1,000
~ mm hydrogen added
I
- - --
2,000
., , I
3,000
d ln K, i Time,sec
---cIT = - RT2 (41)
Fig. 60. Pressure-time curves obtained in the decomposition of ammonia on a
where . is thc hoat. of adsorption of the poison. It follows that heated tungsten surface. (Hinshelwood and Burk.) Added hydrogen is seen to
have no effect on the rate.
~l~ = d ln k2 + d ln K _ d ln Ki (42)
d'1.' dT dT dT
= E -
RT2
+ i (43) Some of their results, obtained at 856C, are plotted in Fig. 60. I I,
is to be noted that the initial slopes are all the same, in spite of differcneos
in the initial concentrations; this indicates that the reaction is of ZO'O
The apparont activation energy is therefore given by order under the conditions of the experiments. The same oonclusiou I
indicated by the fact that the pressure-time curves are linear OVOI' t.11iI
Ea = E - + i (44) initial range of pressures; deviations from linearity occur at late!" I,illltln,
since the kinetics approach first-order behavior at lower partial pl'IIM"" I 1I
The reason that Lheactivation energy is increased by i is that it is neces-
of ammonia. It is clear that under conditions of the expori 1111\111,
sary for a moleculo of the inhibitor to be desorbed in order for a molecule
of reactant to beco me adsorbed and undergo reaction. 1 C. N. Hinshelwood and R. E. Burk, J. Chem. Soc., 127, 1051, 1111 (IU~fI).
 72 CHEMICAL KINETICS ACTlONS ON SURFACES AND lN TlIE SOLlD STATE 273

of Ilinshelwood and Burk the tungsten surface is largely covered by
ammonia. Barrer! made a careful study of the zero-order reaction on
tungsten, and found the activation energy to be 42.4 kcal per mole.
Hinshelwood and Burk also made a study of the decomposition of
ammonia on platinum. They found that the reaction was inhibited by
the hydrogen produced in the reaction, and by added hydrogen, the rate
law being
v = k[[~~3]
This equation is of the same form as Eq. (35); evidently the ammonia is
weakly, and the hydrogen strongly, adsorbed under the conditions of
their experiments. There is no inhibition by nitrogen, which apparently
is only very weakly adsorbed.
Under other experimental conditions more complicated kinetic equa-
tions have sometimes been observed. Thus Winter" found that the
decomposition on iron follows the equation
= k[NH3]
V [H2P~
This and similar complex rate equations have been explained by Temkin
and Pyzhev" in terms of the Slygin-Frumkin isotherm [Eq. (26)], which
corresponds to variable heats of adsorption.
Decomposition of Phosphine. The phosphine decomposition was
first investigated by van't Hoff and Kooij! on a glass surface. They
found that the reaction was of the first order, so that their surface was
sparsely covered by phosphine. Barrer! made an investigation of the
reaction on tungsten over a wide range of pressures and observed the
expected transition from first-order kinetics at low pressures to zero-order
kinetics at higher pressures. He found the activation energy to vary
from 26.5 kcal at low pressures to 31.3 at high pressures; the difference,
4.8 kcal, is to be rcgardod, according to Eq. (40), as the heat of adsorption
of phosgono on the surfaco, A similar change from first-order to zero-
order kinetics with increasing pressure was observed on a molybdenum
surface by 1\1clvi Ile and Roxburgh.? The activation energy fell by
7.2 kcal OVOI' the prcssure range, and this may be approximately the heat
of adsorption.
1 R. M. Bnrror, Trens, Faraday Soc., 32, 490 (1936).
2 E. WinLcr, Z. Physik. Chem., B13, 401 (1931).
3 M. Temkin anel V. Pyzhev, Acta Physicochim. URSS, 12, 327 (1940); for a
review see K. J. Luidlor, Kinetic Laws in Surface Catalysis, in P. H. Emmett (ed.),
"Catalysis," vol, 1, Hrinhold Publishing Corporation, New York, 1954.
4 J. H. van't nerr anel D. M. Kooij, Z. Physik. Chem., 12, 155 (1893).
Barrer, loco cito
6 H. W. Melvillo and lI. L. Roxburgh, J. Chem. Soc., 586 (1933).
Decompositions of Other Hydrides. Arsine was found by van't
l loff" to decompose on glass according to a first-order law, and the same
t ypc of behavior has been found in the decomposition of selenium
hydride on selenium! and of methane on carbon.! For the decomposition
"r stibine on an antimony surface- the behavior is more complex, but is
r-onsistent with Eq. (29). Hinshelwood and Priohard! found the decom-
position of hydrogen iodide on gold to be zero order, while Hinshelwood
uud Burk" found the same reaction on platinum to be first order.
(45)
Decomposition of Nitrous Oxide. A number of investigations?
have been made of the decomposition of nitrous oxide into nitrogen and
oxygen, and several different surfaces have been used. Most of the rate
IILWS obtained are special cases of the generallaw
k[N20]
V = ::-1--'+-a~[N~20=]-=+~b[=0~2] (47)
This indicates that both nitrous oxide and the product oxygen are
adsorbed to an appreciable extent.
Decomposition of Formic Acid. Hinshelwood and coworkers"
(46)
studied the kinetics of the decomposition of formic acid on a number of
surfaces, including glass, platinum, silver, gold, and titanium dioxide.
The first-order law was obeyed under the conditions of all of their experi-
ments. On the glass the reaction proceeds partly as a dehydration,
(1) HCOOH ~ CO + H20
and partly as a dehydrogenation,
(2) HCOOH ~ CO~ H2
The activation energy for the dehydration was found to be 16.0 kcal, and
that for the dehydrogenation was 28.0 kcal. On metals the dehydrogena-
tion predominated, whereas on titanium dioxide the dehydration was
more important.
1 J. H. van't Hoff, "tudes de dynamique chimique," p. 83, F. Muller and
Company, Amsterdam, 1884.
2 M. Bodenstein, Z. Physik. Chem., 29, 429 (1899).
3 G. M. Schwab and E. Pietsch, ibid., 12, 155 (1893).
4 A. Stock and M. Bodenstein, Ber., 40, 570 (1907).
C. N. Hinshelwood and C. R. Prichard, J. Chem. Soc., 127, 1552 (1925).
6 C. N. Hinshelwood and R. E. Burk, J. Chem, Soc., 127, 2896 (1925).
7 E.g., C. N. Hinshelwood and C. R. Prichard, J. Chem. Soc., 127, 327 (1925);
Proc. Roy. Soe. (London), Al0S, 211 (1925); G. M. Schwab, R. Stager, and 11. H.
von Baumbach, Z. Physik. Chem., B21, 65 (1933); E. W. R. Steacie and H. O. FolkinH,
Cano J. Res., ais, 237 (1937).
8 C. N. Hinshelwood, H. Hartley, and B. Topley, Proc. Roy. Soe. (London), Al00,
575 (1922); H. C. Tingey and C. N. Hinshelwood, J. Chem, Soc., 121, 1668 (11122);
C. N. Hinshelwood and B. Topley, J. Chem, Soc., 123, 1014 (1923).
 lotACTIONS ON SURFACES AND lN TIIE SOLID STATE 275
274 CIIEMICAL KINETICS

Schwab and his coworkers! have made extensive studies of tho

decomposition of formic acid on a variety of alloy surfaces. Under tho
conditions of their experiments the kinetics were generally of zero order.
These investigations were mainly concerned with the relationship between
catalytic activity and the nature of the surface, and are considered again
later in this chapter.

BIMOLECULAR SURFACE REACTIONS Rot

Reaction between Two Adsorbed Molecules. It appears, as has

been mentioned, that most surface reactions between two substances
occur by reaction between two molecules that are adsorbed on neighbor-
ing surface sites; this is conveniently referred to as a Langmuir-Hinshel-
wood mechanism. The rate of such a reaction between A and B is pro- ~
portional to the probability that A and B are adsorbed on neighboring
sites, and this is proportional to the fractions of the surface, O and O',
Pressure of reactant
covered by A and B. These fractions were given by Eqs. (13) and (14),
and the rate of reaction is therefore Fig. 61. Schematic representation of the variation of rate with
Lhe pressure of one reactant (the other being held constant) for a
v = k200' (48)
himolecular reaction proceeding by a Langmuir-Hinshelwood
k2KK'pp' mechanism [Eq. (49)1.
(49)
(1 + Kp + K'p')2

If the pressure p' is kept constant, and p is varied, the rate varies in
accordance with Fig. 61; the rate first increases, passes through a maxi-
mum, and then decreases. There is a similar variation with p' if p is
held constant. The physical explanation of the falling off of the rate at
high pressures is that one reactant displaces the other as its pressure is
increased. The maximum rate corresponds to the existence of the maxi-
mum number of neighboring A-B pairs on the surface.
Two special cases of the general equation (49) may now be con-
sidered.
1. Sparsely Covered Surfaces. If the pressures p and p' are both
sufficiont1y 10w that Kp and Kp' may be neglected in comparison with
unity, tho rate cquation becomes
v = k2KK'pp'
The reaction is therefore second order, being first order in both A and B.
A numbcr of reactions obey this law: examples are the reaction between
nitric oxido and oxygen on glass and that between ethylene and hydrogen
under certain circumstances (cf. p. 278).
1 G. M. Schwah, Trans. Faraday Soc., 42, 689 (1946); G. M. Scbwab and S.
Permatjoglou, J. Pltys. Colloid Chem., 62, 1046 (1948); G. M. Schwab, Diseussions
Faraday Soc., 8, 166 (1950).
2. One Reactant Very W eakl~Adsorbed. If reactant A is very
weakly adsorbed, the term Kp in the denominator of Eq. (49) may be
neglected, and the rate equation becomes
k2KK'pp'
V = (1 + K'p')2 (51)
The rate is now proportional to the pressure of A, but as the pressure of B
increases, the rate first increases, passes through a maximum, and then
decreases (Fig. 61). Such a maximum in the rate has been observed in
the reaction between hydrogen and carbon dioxide on platinum;' in tho
exchange reaction between deuterium and ammonia on iron (p. 283), and
in the reaction between hydrogen and ethylene (p. 278).
If reactant B is sufficiently strongly adsorbed so that K'p' is much
(50)
greater than unity, the rate equation becomes
_ k2Kp (r:2)
v - K'p' ,)
The rate is now inversely proportional to the pressure of the stronaly
adsorbed reactant B, and the order of the rcaction with respect to B muy
1 c. R. Prichard and C. N. Hinsbelwood, J. Citem. Soc., 127,806 (1925).
276 CIJEMICAL KINETICS ltACTIONS ON SURFACES ANO IN THE SOLlD STATE 277

bc said to be -1. Under certain conditions the rate of the reaction t ion of Lhe rcaction on evaporatod films in fact suggested that it occurs by
between carbon monoxide and oxygen on the surfaces of quartz ' aud (:olllbination of the two mechanisms (p. 279).
11,

platinum 2 is directly proportional to the pressure of oxygen and inversely The surface combinations of atoms and free radicals appear definitely
proportional to that of carbon monoxide; the lattcr is therefore strongly to occur by Langmuir-Rideal mechanisms; these reactions are discussed
adsorbed, and as its pressure is increased, it displaces the oxygen from luter (p. 284).
the surface. Another example is the reaction between hydrogen and Adsorption of Two Gases without Mutual Displacement. A third
oxygen on platinum (p. 278); the rate under certain conditions is inversely possibility, which occurs in one or two cases, is that there is reaction
proportional to the oxygen pressure. hctween two molecules which are adsorbed on two different types of
Reaction between a Gas Molecule and an Adsorbed Molecule. surlace site, so that they do not displace each other from the surface.
There is also the possibility that the reaction is between an adsorbed 'I'hc isotherm for the adsorption of gas A on sites of type 1 is
molecule and a molecule from the gas phase; mechanisms of this type are __(} = Kp (55)
conveniently referred to as Langmuir-Rideal mechanisms. Suppose that 1 - (}
reaction occurs between an adsorbed A molecule and a gaseous B mole-
cule, The fraction of surface covered by A is given by Eq. (13), and the 'Similarly, for the adsorption of B on sites of type 2,
rate is now proportional to this fraction and to the pressure p' of B, viz.,
1 - (}'=
K"
p
(56)
v = k2(}p (53)
k2Kpp' The rate is proportional to (}O', so that
(54)
1+ Kp + K'p' k2KK'pp'
It is not assumed, in this mechanism, that B is not at all adsorbed, since
v = (1 + Kp)(l + K'p') (57)
the term K'p' in the denominator corresponds to the adsorption of B; This dependence of rate on p and p' is quite different from that of Eqs.
adsorbed B molecules do not enter directly into reaction, but they affect (49) and (54), and this case is readil~stinguished by measuring rates
the rate by occupying surface that might otherwise be occupied by A. over a wide range of pressures.
Equation (54) is to be contrasted with Eq. (49) for the Langmuir- Equation (57) has been found to be applicable to the reaction
Hinshelwood mechanism. There is now no maximum in the rate as the between hydrogen and nitrous oxide on gold.! and to the reaction between
pressure p (or p') increases; instead the rate varies with the pressure of hydrogen and carbon dioxide on tungsten."
either reactant in exactly the manner shown in Fig. 58. A decision Inhibition of Bimolecular Surface Reactions. Inhibition equations
between the two types of mechanism may therefore be made by seeing for reactions of this type may readily be derived, and only a few simple
whether the rate passes through a maximum with increase in the pressure cases need be considered. Consider a reaction between A and B occurring
of cither reactant; if it does, the Langmuir-Rideal mechanism can be by a Langmuir-Hinshelwood mechanism and inhibited by a gaseous sub-
excludcd. stance of partial pressure Pio The fractions covered by A and B are easily
Thcre are few well-defined examples of reactions occurring by a
shown to be
Langrnuir-Rideal mechanism. Allen and Hinshelwood" studied the Kp (58)
decomposition of acetaldehyde into methane and carbon monoxide on o = 1 + Kp + K'p' + K,p.
various surfaccs; at low pressures the kinetics are second order and at K'p'
o' = ----==-~,---,- (59)
high prcesuros first order. There seems to be no tendency for the reaction and 1 + Kp + K'p' + K;p.
to becomo zero order at high pressures, which would occur if there were
reaction betwccn two adsorbed molecules. The results therefore appear The rate is therefore
to favor a Langmuir-Rideal mechanism, but more work is desirable. The (iO)
v = k200'
reaction bctwccn ethylene and hydrogen has frequently been interpreted k2KK'pp' (<I I)
as occurring by a Langmuir-Rideal mechanism, and a careful investiga- - (1 + [(p + K'p' + !(;p.)2
M. Bodonstein und F. Ohlmer, Z. Physik. Chem., 53, 166 (1905).
J W. H. Hutchison and C. N. Hinshelwood, .r. Chem. Soe., 129, 1556 (102(1).
21. Langmuir, Trens. Faraday Soc., 17,621 (1922).
2 C. R. Prichard and C. N. Hinshelwood, J. Chem. Soc., 127, 1546 (1021).
8 P. C. Allen and C. N. Hinshelwood, Proe. Roy. Soe. (London), A121, 141 (1\128).
 278 CHEMICAL KINETICR REACTIONS ON SURFACES AND IN THE SOLlD STATE 279
If the inhibitor is a diatomic molecule that is adsorbed atomically, LIl( which is of the same form as Eq. (51). This result is consistent with a
corresponding equation is Lungmuir-Ifinshelwood mechanism, with the ethylene more strongly
k2KK'pp' ndsorbed than the hydrogen.
v = (1 + Kp + K'p' + K;'/2p;'2)2 (62) The facts regarding the reaction on a nickel surface have been a
matter of some controversy. Several workers! have obtained results
If there is sparse coverage by reactant molecules and large coverage by that are consistent with the rate equation,
inhibitor, the equation becomes
k[H2J[C2H4]
V = C""""2H~4]
::-l":"+-K-;;-;-:;[ (65)
k2KK'pp'
v = K;Pi (63)
This type of behavior is consistent with a Langmuir-Rideal mechanism,
A few examples of bimolecular surface reactions will now be con- and Beeok? and Jenkins and Rideal- have suggested that reaction occurs
sidered. between a pair of adsorbed hydrogen atoms and a gaseous ethylene
Hydrogen-Oxygen Reaction. The reaction between hydrogen and molecule:
oxygen has been investigated under a wide variety of conditions, and very H H I I
different types of behavior have been observed. On a porcelain surface I I + CH 2 4~ -8-8- + CH
2 6
-8-8-
Bodenstein I found that the rate was proportional to the pressures of both
hydrogen and oxygen, which means that the surface is sparsely covered Toyama.! on the other hand, obtained defini te evidence that the rate
by both reactants. On a variety of surfaces including silver, gold, copper, fell at high ethylene pressures, a fact that requircs that the Langrnuir-
and nickel the reaction has been found- to be first order in hydrogen and Hinshelwood mechanism play a role. A similar result was obtaincd by
zero order in oxygen, and therefore to be consistent with a Langmuir- Laidler and Townshend! in an investigation, using an evaporated nickcl
Rideal mechanism. On a platinum surface inhibition by oxygen has surface, that was primarily concerned with distinguishing bctwccn th
been detected," while on silver! there is inhibition by water, the product two mechanisms. They found that\ the kinetic behavior dependcd to a
of reaction. considerable extent upon which gas W1ts first admitted to the surfaco (01'
Ethylene-Hydrogen Reaction. A considerable amount of kinetic whether both were admitted together). Laidler and Townshend inter-
work has been done on the hydrogenation of ethylene, on a variety of preted their and other results on the basis of the hypothesis that two
surfaces, and there has been much uncertainty both as to the facts and to mechanisms are taking place simultaneously. One mechanism, favorcd
the mechanisms involved. Quite different behavior is observed under by prior addition of ethylene 01' by the presence of excess of ethylene, is
different conditions, and it appears that reaction may occur by two or the Langmuir-Hinshelwood one. The second, favored by prior addition
more mcchanisms at the same time. of hydrogen or excess of hydrogen, is the Langmuir-Rideal mechanism
An invcstigation by Pease! of the reaction on a copper surface indi- between adsorbed hydrogen and gaseous ethylene. These conclusions
cated that the rate is proportional to the hydrogen pressure and that it throw light on apparent discrepancies among the results of previous
passes through a maximum as the ethylene pressure is increased, the rate workers, who used a variety of experimental conditions and usually did
ultimatcly bccoming inversely proportional to the ethylene pressure. not specify how the gases were introduced to the reaction vessel.
The rato law is, to a good approximation, On the basis of much evidence? regarding the way in which hydrogcn
and ethylene are adsorbed, Laidler and Townshend concluded that 11
k[H2][C2H4] molecular rnechanisrn previously proposed by Markham, Wall, and
V = (1 + K[C2H4])2 (64)
I E. K. Rideal, J. Chem, Soe., 121,309 (1922); H. Zur Strassen, Z. Physik. eu;
I M. Bodonstoin, Z. Physik. Chem., 29, 665 (1899). A169,81 (1934); A. Farkas, L. Farkas, and E. K. Rideal, Proc. Roy. Soe. (LondQI/),
2 W. A. Bono and R. V. Wheelof, Phil. Trans, A206, 1 (1906); R. N. Pease and Al71, 55 (1939); O. Beeck, Rev. Mod. Pllys., 17,61 (1945).
H. S. Taylol", J. J1m. Chem. Soe., 43, 2179 (1921); 44, 1637 (1922); A. F. Benton and 2 Beeck, ibid., 20,127 (1948); Discussions Faraday Soe., 8, 118, 126, 193 (J!)!i(l).
P. H. Emmott, tu, 48, 632 (1926); cf. also R. P. Donnelly, J. Chem, Soc., 132, 2438 S G. I. Jenkins and E. K. Rideal, J. Chem, Soe., 2490, 2496 (1955).
(1929). 4 O. Toyama, Rev. Phys. Chem ..Japan, 11, 153 (1937).
sI. Langmuir, Trans. Faraday Soe., 17, 621 (1922). 6 K. J. Laidler and R. E. Townshend, 'l'rans. I?araday Soe., 57, 1590 (1961).
4 R. N. Peaso, J. Am. Chem, Soe., 45, 1196 (1923). S E.g., Jenkins and Rideal, loco cit.; C. Kemball, J. Chem. Soe., 735 (UI56).
280 CIIEMICAL KINETICS REACTIONS ON SURFACES AND IN THE SOLlP STATE 281

Laidler! was applicable to the reactions. According to this achemo, d('uLcl'iUIn pressure. The mechanism proposed by them! involves the
ethylene may become "associatively" adsorbed on a pair of bare surfaco rormation of mixed ethyl radicals by reaction between adsorbed dente-
sites, rium atoms and ethylene molecules,

(1) C2H4 + -8 I-.-~8I C

I H -CHI
2
2
CH2D
-8 8- D /
(L) C2H4 + I ~ CH2
In addition, an ethylene molecule from the gas phase may add on to an
-8- I
-8-
adsorbed hydrogen atom to form an alkyl radical,
or
CHs CH2D
H / CH2-CH2 D /
(2) C2H4 + I ~ CH2 (2) I I I~ CH2
-8- I -8--8-8 I I I
-8- -8-8-8
'8urface exchange may occur by interaction between adsorbed alkyl
Adsorbed alkyl radicals may also be formed from adsorbed ethylene
molecules and neighboring adsorbed hydrogen atoms. Ethane is pro-
radicals and adsorbed ethylene molecules,
duced in the gas phase either by interaction between two neighboring CH2D CH3
adsorbed alkyl radicals, I I
(3) CH2 CH2-CH2 ~ CH2-CH2D CH2
CH3 CH3 I I I I I I
I I CH2-CH2 8 8--8 -8--8 8-
(3) CH2 CH2 ~ I I + CH3CH3
8pecies such as C2HD3 and C2D4 are obtained in the initial stages of the
I I -8-8-
-8-8- reaction,? and this is explained if re~ons (2) and (3) occur very rapidly;
the evidence, in fact, suggests that there is a complete mixing of the H
or by interaction between an adsorbed alkyl radical and an adsorbed and D atoms on the surface. Since the concentration of adsorbed D
hydrogen atom, atoms is proportional to the square root of the pressure, the exchange rate
will be proportional to the square root of the pressure, as is believed to
CH3
be the case.
I I I
(4) CH2 H ~ -8 -8- + CH3CH3 Methane-Deuterium Exchange Reactions. The exchange reac-
I I tions between methane and deuterium have been studied by Kemball,"
-8-8- who used evaporated nickel surfaces. He found that all the products,
CH3D, CH2D2, CHD3, and CD4, were produced in the initial stages of
This mechanism is of the Langmuir-Hinshelwood type; the Rideal
the reaction; this behavior is to be contrasted with that found with
mechanism is shown on page 279.
ammonia (p. 282), where the products are formed byconsecutive reac-
Ethylene-Deuterium Exchange Reaction. Some kinetic studies
tions. The rates of formation of all of the methanes are proportional to
have been rnade? of the exchange reaction between ethylene and deute-
the first power of the methane pressure. The rate of formation of CH3D
rium on a nickel surface. The results are not clear-cut, but Laidler,
is inversely proportional to the square root of the deuterium pressure,
Wall, and Markharn" have shown that the rate is probably proportional
whereas the rates of formation of CH2D2, CHD3, and CD4 are inversely
to the first power of the ethylene pressure and to the square root of thr-
proportional to the first power of the deuterium pressure.
These results suggest that two different exchange mechanisms are
1M. C. Markham, M. C. Wall, and K. J. Laidler, J. Chem. Phys., 20, 1331 (1952);
21, 949 (1953). involved, one leading to the production of CHaD, and the other to tho
2 G. H. Twigg and E. K. Rideal, Proc. Roy. Soe. (London), A171, 55 (1939);
G. K. T. Conn and G. H. Twigg, Proc. Roy. Soe. (London), A171, 70 (1939); C. Kem- I Cf. also Laidler and Townshend, loco cito
ball, J. Chem. Soc., 735 (1956). . 2 C. Kemball, J. Chem. Soe., 735 (1956).
3 K. J. LaidJcr, M. C. Wall, and M. C. Markham, J. Chem. Phys., 21, 949 (1953). 3 C. Kemball, Proc. Roy. Soe. (London), A207, 539 (1951).
 282
CIIEMICAL KINETICS RI~ACTIONS ON SURFACES AND IN TIIE SOLID STATE 283
production of CH2D2, CHD3, and CD4. KembalI suggested that CU D
is formed from adsorbed CHs radicaIs, and the other produets frorn 3 mil mcrcly add on a deuterium atom,
adsorbed CH2 radicaIs which rapidIy exchange with adsorbed deuterium NH2 D I I
atoms. The surface is IargeIy covered by deuterium atoms, and the I I ~ -8-8- + NH D 2

number of bare sites is inverseIy proportionaI to the square root of the -8-8-
I
deuterium pressure [cf. Eq. (8)J. The methyI radicaIs are adsorbed on Tho adsorbed radical NH2 apparently cannot undergo exchange with an
singIe sites, and their concentration on the surface is therefore propor- adsorbed deuterium atom.
tionaI to [CH4J/[D2JH. The formation of CH3D is expIained as due to A number of studies have been made of the dependence of the
the reaction
)xchange rate on the concentrations of ammonia and deuterium. Farkas!
H
-8-8-
3 ~ r ~ -J-J-
+ CH3D used apure iron surface and found that the rate was proportional to the
square root of the deuterium pressure and independent of the ammonia
pressure. Weber and Laidler" worked with an activated iron surface
8ince practicalIy every adsorbed methyI radical wilI have a deuterium and also observed the square-root dependence on deuterium pressure.
atom as a near neighbor, the rate of production of CH3D is proportional They found, however, that the rate passed through a maximum as the
to [CH4J/[Dd~.
ammonia pressure was increased, and could be fitted to the Langmuir-
The different rate Iaw for the othcr products indicates that they are llinshelwood equation,
not formed frorn CH3, which therefore cannot exchange readiIy with D
atoms. When methane is adsorbed it is beIieved that Some CH2 radicaIs k[D21~[NH3]
are formed on the surface, V = (1 + K[NH3])2 (66)

The conclusion is that the surface is rather heavily covered by NH2

I I I I H CH2 H
CH4 + -8-8-8-8- ~ I / "" I radicals, and only sparsely by D atoms, and that reaction occurs between
-8-8--8-8- these two adsorbed species.
Somewhat different results were obtained by 8ingleton, Roberts,
8ince these are adsorbed on a pair of sites, their concentration wilI be
and Winter, a who used evaporated films of iron, tungsten, and nickel.
proportionaI to the square of the number of bare sites, i.e., to 1/[D2J.
Their results are consistent with the hypothesis that the Langmuir-
The adsorbed CH2 radicaIs are assumed to exchange readiIy with neigh-
Hinshelwood mechanism applies at lower pressures of ammonia. At
boring D atoms, so that CH2, CHD, and CD2 radicaIs are present on the
higher pressures of ammonia the rate does not fall toward zero, but only
surface in amounts that depend on statisticaI factors. Their concentra-
to a constant leveI, and the rate is then proportional to the frst power of
tions wilI be proportionaI to [CH4J/[Dd. The rates of production of
CH2D2, CHDs, and CD4, by reactions such as the deuterium pressure. It appears that under these conditions a .Lang-
muir-Rideal mechanism predominates, reaction being between an adsorbed
D CUD D I I I I NH2 radical and a gaseous deuterium molecule:
I
-8-8-8-8_
/ "" I ~-8-8-8-8- + CHD3
NH2 D

wilI thel'cfol'e also be proportionaI to [CH4J/[D2J.

I + D2 ~ I + NH D 2
-8- -8-
Ammonia-Deuterium Exchange Reaction. The exchange reaction
between ummonia and deuterium, with the formation of NH2D, NHD , SOME SPECIAL TYPES OF REACTIONS
2
and NDs, olTers an interesting contrast in that, on a number of surfaces
(incIuding nickel and platinum), the products are formed in consecutive Certain reactions are not conveniently classified as either unimolecu-
reactons.: The initiaI product is NH2D, and NHD2 is formed from it in lar bimolecular, and are now considered separately.
01'

a Subsequent independent reaction; finalIy ND3 is formed from NHD . The Parahydrogen Conversion. Parahydrogen becomes converted
2 into the equilibrium mixture of the ortho and para forms by being brought
These resulta are most simply explained on the assumption that amrnonia
on adsorption ia dissociated onIy into NH2 and H, and that the formar A. Farkas, Trans. Faraday Soe., 32, 416 (1936).
1
J. Weber and K. J. Laidler, J. Chem. Phys., 19, 1089 (1951).
2
1 C. Kemball, ibid., A217, 376 (1953). 3 J. H. Singleton, E. R. Roherts, and E. R. S. Winter, Trans. Faraday Soe., 47,

1318 (1951).
 N SURFACES AND lN TIIE SOLID STATE
286
284 ClIEMICAL KINETICH ItACTIONS

1'110 rato is proportional to 8 and to p, whence

into contact with metal surfaces, especially those of the transition motals
(68)
which are particularly effective in adsorbing hydrogen. The reaction V = kp8
probably occurs! by a mechanism originally proposed by Bonhoeffer and k[(p2 (69)
Farkas.? This mechanism may be represented as = 1 + Kp
which means
I I H H A L lower temperatures the surface may be fully covered,
p-H2 + -S-8-~ I I ~-S-S- + o-H 2
(,hat Kp 1; in this case
-S-S- (70)
V = kp
The hydrogen becomes adsorbed on the surface and is then desorbed; At higher temperatures, on the other
since adsorption involves dissociation, the desorption process leads to the und thc kinetics are first order.
equilibrium mixture. The order of the reaction is close to zero," a result hand, Kp 1, so that (71)
V = kKp2
that can be reconciled with the Bonhoeffer-Farkas mechanism.
Surface Combination of Atoms and Radicals, Atoms and free radi- nnd the kinetics are second order. It is to be noted that this mechanism
cals combine on most surfaces, and kinetic studies have been made of a is of the Langmuir-Rideal type.
number of systems. ln almost all the cases investigated the reactions Doubt as to whether this is the correct mechanism is, however, cast
are first-order processes' which become second order at higher tempera- by some work on the reverse reaction, as will now be considered.
tures. One mechanism that is consistent with this behavior is that The Production of Atoms at Hot Surfaces. It was first shown by
reaction occurs between a gaseous atom or radical and an adsorbed one; Langmuirl that hydrogen atoms are produced from hydrogen molecules
thus, for the recombination of hydrogen atoms this mechanism would be at a hot tungsten surface, and Bryce? showed that the rate of reaction is
approximately proportional to the square root of the hydrogen pressure.
H I
H + I --t -S- +H 2
Roberts and Bryce" proposed a mechanism in which a hydrogen molecule
-8- strikes a bare site, whereupon one atom is adsorbed and the other passes
into the gas phase,
If the surface is fully covered by atoms, the rate is proportional to the
I H
pressure of the atoms; i.e., the kinetics are first order. At higher tem- H-H + --8---t I +H
peratures the surface becomes sparsely covered, and the fraction covered -S-
is proportional to the atomic concentration; the kinetics are then second
This mechanism is the reverse of that considered above, for the combina-
order.
tion reactions. lt accounts for the pressure dependence of the rate,
The kinetic equations for this mechanism may be formulated as
provided that the surface is fully covered; the fraction of surface bare is
follows: The fraction covered depends on the atomic pressure p accord-
then inversely proportional to the square root of the prcssure, so that the
ing to the isotherm
rate, being proportional both to the pressure and the fraction bare, is
8 proportional to the square root of the pressure. Laidler' has discussed
1 - 8 = [(p (67)
the mechanism in some detail and concluded that this mechanism is the
1 For a dotnilcd discussion of the mcchanism, which involves some complex correct one.
features, sco K. J. Laidler, J. Phys. Chem., 67,320 (1953); also K. J. Laidler, in
A more recent experimental investigation by Brennan and Fletcher
P. 11. EmJ1lett (ed.), "Catalysis," voI. 1, pp. 178-180, Reinhold Publishing Corpo- has, however, led to results that were different from those previously
ration, New York, 1954. obtained, and has cast doubt on the Roberts-Bryce mechanism. As an
2 K. F. HonhoefTer and A. Farkas, Z. Physik. Chem., B12, 231 (1931); K. F. alternative they propose that the hydrogen becomes adsorbed in th
Bonhoeffer, A. Furkns, and K. W. Rummel, -u, B21, 225 (1933).
a D. D. 1';ley nnd E. IC Rideal, Proc. Roy. Soe. (London), A178, 429 (1941). 11. Langmuir, J. Am. Chem, Soc., 34,1310 (1912); 37, 417 (1915).
4 L. von Muling, "Handbuch der Katalyse," vol. VI, p. 94, J. Springer, Vienna, 2 G. Bryce, Proc. Cambridge Phil. Soc., 32, 648 (l936).
1943; W. V. Smith, J. Chem. Phue., 11, 110 (1941).
S J. K. Roberts and G. Bryee, .ibid., 653.
6 W. Buben und A. Schechter, Acta Physicochim. URSS, 10, 371 (1939); F. 4 K. J. Laidler, J. Phys. Colloid Chem., 66, 1067 (1951).
Paneth, W. Hofcditz, and A. Wunsch, J. Chem. Soc., 372 (1935); F. Paneth and 6 D. Brennan and P. C. Fletcher, Proc. Roy. Soco (London), A260, 389 (105\).
W. Lautsch, Ber., 64B, 2708 (1931).
 86
CIIEMICAL KINETICS J(I~ACTIONS ON SURFACES AND IN TllE SOLID STATE 87
atomio forrn,
urfuco ia rcasonably smooth; only if theso conditions are eatlscd cun
II H H 0110 mako a reliable estimate of the number of surface sites.
H2 + -S-s- ~ I I Absolute Rates of Adsorption. The process of adsorption is a
-8-8- himolecular reaction between surface sites and gas molecules; the case of
and that hydrogen atoms become desorbed, tho localized activated complex will be considered first. Let N., No, and
N I be the numbers of bare sites, gas molecules, and activated complexes,
H I respectively, and let the corresponding concentrations be c.(= N./S),
I -)-8-+H
-8- ('u(= Nu/V), and CI( = NtlS). The usual assumption of equilibrium
hetween the activated complexes and the reactants then gives rise to the
They further conclude that at the temperatures of their experiments the cquation
adsorbed atoms have fuU translational motion on the surface. It is !.. - p
likely that different mechanisms apply under different experimental CoCo - F of e-'
l/kT (72)
conditions.
As in Eq. (17), the partition functions j" and j, relate to unit surface area,
ABSOLUTE RATES OF SURFACE REACTIONS and F u relates to unit volume. The energy EI is the energy of the com-
plexes with reference to the reactants, at the absolute zero, and is there-
Attention has so far been confined, in this chapter, to the question of fore the activation energy per molecule at that temperature.
the influence of concentrations on rates, without regard to the absolute As pointed out in Chap. 3, the partition function Fu can be factorized
rates of the reactions. The treatment of this latter problem involves, as into translational, rotational, and vibrational factors. If the activated
with all reactions, two types of consideration. The first is a calculation complex is localized on the surface, it has no translational or rotational
by quantum-mechanical methods of the energy of activation; the second motion, but it undergoes various types of vibrational motion. As pre-
is the evaluation of the partition functions for the initial and activated viously (p. 76), one of the vibrational factors in the partition function
states. The first problem is a very difficult one; approximate methods corresponds to a very loose vibration, and may be expressed as kT /hl',
have to be used, and good agreement with experiment has never yet been where I' is the frequency of the vibration. Equation (72) may therefore
achieved. In addition to the usual complications of such calculations, be expressed as
there is the difficulty that relatively little is known about the exact ~ = ft(kT/hv) e-'l/kT (73)
nature of the surfaces that are involved in catalysis. CuCo Fuf.
The second type of procedure, the calculation of absolute rates
where the partition function ft lacks the contribution for the one degree of
using partition functions, use being made of experimental energies of
freedom. Equation (73) rearranges to
activation, has been carried out much more successfully. The applca-
tion of absolute-rate theory to surface reactions was first done in any vC. = c c kT A e-'dkT (74)
detail by Laidler, Glasstooe, and Eyring, I and a number of different types o h Fof.
of reactions havo beco treated, alI with comparative success, Since this
whole qucstion has been reviewed in considerable detail." only a very The frequency v is the frequency of vibration of the activated complexes
brief discussion ia given here. in the degree of freedom corresponding to their transformation into
It sbould be noted that the theory can only be profitably applied to adsorbed molecules. The expression on the left-hand side of Eq. (74) is
investigations in which the surface area is known, and in whieh the thus the product of the concentration CI of complexes and the frequency
of their transformation; it is therefore the rate of the adsorption process,
1 K. J. JJaidlor, S. Glasstone, and H. Eyring, J. Chem. Phys., 8, 667 (1940); which is thus given by the expression on the right-hand side of the equa-
cf. B. Toploy, Nature, 128,115 (1931); R. E. Burk, J. Am. Chem. Soc., 66, 1279 (1935). tion, viz.,
2 K. J. Laidlor, J. Phys. Chem., 67, 320 (1953); M. C. Markham, M. C. Wall, and
kT A e-'dkT (7rI)
K. J. Laidler, ibict., 321; K. J. Laidler, M. C. Wall, and M. C. Markham, J. Chem .. VI = Coc. 11Fof.
Phys., 21, 149 (1953).
3 K. J. Laidler, Absolute Rates of Surface Reactions, in P. H. Emmett (ed.), The partition functionsmay now be evaluated in the same muunor
"Catalysis," chap 5, Reinhold Publishing Corporation, New York, 1954.
as in the equilibrium equations. There are no translational and I'O(,IL-