SOLID CATALYSTS

and
REACTION RATES General Principles

Based on the theory of activated adsorption, general rate equations are de-
veloped for chemical reactions ~ataly5edby solids. Cbnsideration is given
to the differences in equation forms resulting from control of the rate by
different types of activated s t e p . Quantitative expressions are proposed for
the effects of adsorption, catulyst activity, particle &e, porosity, j b w condi-
tions,poisons, anddiluents as well as temperature, pressure, and concentration.
Experimental procedures for the evaluation of the constants of the equa-
tions are discussed. A general method is outlined for integration of rate
equations in the design of commercial reactors.

0. A. Hougen and K . M . Watson

UNNERSITY OF VISCONSLN, MADISON, VIS.

5 A BASIS for the development of rata equations, it is 3. The d v a t e d adsorption of reactants and activated d e

A poetulated that when catalyzed by a solid, a liquid-

or gasphase chemical reaction actnaUy occurs on the
surface of the catalyst and involves the reaction of moleculea
wrption of pmducta at tb catslyti0 interface.
4. The nvface reaotion of adso-
vation-adsorbed products.
resotants to form acti-

or atoms wbich are adsorbed by activation on the active cen-
tern of the surface. From this viewpoint the catalyst in-
c- the rate of reaction through ita ability to adsorb the
reedsnts in such a form that the activation energy necesssry
for reaction is reduced far below ita value in the uncatalyd
reaction. Theee principles are extensively diseuaeed in the
periodical literature and in theoretical treatises on catalytic
rate phenomena (4, 7). The purpose of this paper is to d e
velop from theae principles equations which may be applied
to the interpretation and extension of rate data and to point
out expebental procedures best adapted to the complete
evaluation of such equations from a minimum of data.
Methods am slso developed for the u8e of such equations in
the d& of industrial reactors.
In order that a reactant in the main fluid phase may be
converted catalytically to 8 prcduct in the main fluid phase,
it is nBee8BBIy that the reactant be trsn%ferreerredfmm its psi-
tion in the fluid to the catalytic interface, be activatdy
adsorbed on the surface, and undergo reaction to form the ad-
sorbed prcduct. The prcduct must then be desorbed and
transferred from the interface to a psition in the fluid phase.
The rate at wbich each of these Bteps occm Muencea the
&bution of concentrations in the system and plays a part
in determining the over-all rate. Because of the differencea
in the mechanisms involved, it is convenient to claasify these
steps 88 follows when dealing with cstalynb in the form of
prow particlea:
1. "he mssa transfsr of reactsnts and products to and from
the gross m o r d a c e of the catalyst particle and the main
hodv of the ttuid.
2: The di5u8iond and tiOw transfer of rsaotants and, pmducta
in and out of the pore structnre of the aataljat partide when
reaction tat- place at interior interfaces.
MSY. 1943 INDUSTRIAL A N D ENQINIIERINQ CHEMISTRY
It is evident that the rate of these four types of operations
are d-dent on widely different factors in addition to the
concentrations or concentration gradients involved. Type 1
is determined by the flow characteristics of the system, such
w the m m velocity of the fluid stream, the size of the par-
ticles, and the diiTwiona1 charactmktics of the fluid., Type
2 is determined by degree of porcaity of the catalyst, the di-
mensions of the pores, the extent to which they are inter-
connected, the size of the particlea, the diffusional chsrsc-
t8riatics of the system, and the rate at which the reaction 00.
curs at the interface. Type 3 is detmnined by the charscter
and extent of the catalytic swfaca, and by the p i 6 c acti-
vation energiea required for the adsorption and dmrption of
each of the components of the fluid. Tspe 4 is determind
by the nature and extent of the catalytic surface and by the
activation energies required for the resotion on the surface.
The relative importance cf these four operations in deter-
mining the over-all rate variea widely. Type 1 is important
only when dealing with rapid reactions or where flow condi-
tions are unfavorable. Since the rate of this operation is
little dected by temperature, its relative imprtance tends
to vary for a particular system; it is frequently negligible at
low and highly important at bigb temperatures. Type 2 is
frequently negligible for catsl@a of low activity in small
particles with large interconnected pores, and ceaeea to be a
faator for nonpmus catalysts having no internal surface.
However, in the general ease of an active catalyst in mcder-
ably large particles having large internal surfaces with re-
stricted capillarity it may be of major importance.
Operations of types 3 and 4 are chemical phenomena gen-
eraUy involving relatively large activation energies and a~
therefom highly d t i v e to tamperatnm. The actual
chemical transformations frequently p m d by m r a l mi-
&ve stspas, each with ita own characteristic rate. This is
particularly true where aeveral molecules are involved. Gince
chemical rates vary over wide ranges,it is improbable that
the rates of any two step of typea 3 and 4 w i l l be of equal
order in any given system. For this m m n in msny CBBBB
it is permissible to consider only the Slowest single Btep of
types 3 and 4 and wume that equilibrium is maintained in
dl other steps of these types. Under such conditions the
eloweat activated step may be termed the “rat4eontmllinp
step”, and the o d rata is determined by consideration of
it in combination witb the physical steps of types 1and 2.
ACTIVATED A D S O P m O N
Activated adsorption, 88 differentiated from ordinary van
der Wade adsorption and capillary condensation, ia a highly
speoisc reaction betwean the adsorbate and the surface, and
the aharacteristica of a reversible chemical reaction.
Bince this concept wan introduced by Taylor in 1930, much
attention has been devoted to it in the Literature, recently
BummsriBed by Taylor (71, Emmatt (71, and Ghatone, Laid-
k,and Eyring (4).
It may be m e d that a unit area of catalytic surface
contains L‘ active centers on which activated adsorption can
take place and that all of these centera behave similsrly.
Frequently the active centers are not uniform, and 88 adsorp
tion proceeds, the remaining centem are pmgreaaivfy
active; the rmulta rn an increase in energy of &vatton
+ SUIIPACR REACTIONS
and a decrease in heat of adsorption. In the p m n t dip
d o n the energy of activated adsorption will be aesumed
the m e for all spate,or an average value will be aesumed to
represent the entire d a c e .
The rate of h r p t i o n of a component, A, from a fluid in
contact with the surface is then proportional to ita activiityc.g
in the fluid at the inkdace and to the conmtration 4 of
vacant active centera per unit area of surface. A surface
concentration of e: adsorbed A molecules per unit area will
result.
In the development of rate equationsfor application to i n d w
trial pmcesees, it ie convenient to expm rates r in moles per
unit time per unit msaa of catalyst at uniform conditions.
The total volume occupied by the catalyst particles, including
intemticeu between particles; is dated to the maen of the
catalyst by the bulk density, f i , 88 contrasted to the deasity
of the individual peUeta, p,, and the true density of the cata-
lytic =Lid pc. Similarly, it is convenient to expm surface
concentrations in moles per unit mass of aatalpt. Thus, if
A is the catalytic area per unit m m ,
L = AL’IN. - marimum molal sdso on capedts per
nnun of c a ~ v s tsiZ?om m o M e r
m
Expmsing the rate of adsorptton, r, in moles per wit time
per unit msaa of catalyat:
I - kAclArCl . (1)
Since activated adsorption is a reversible phenomenon,
Oomponent A is slso deeorbed from the Burface at a rate pro-
portionst to the concentration of adsorted moleculea,on the
surface. Thus the rate of desorption is e x p d by:
r - kioA
When adsorption equilibrium is reached, the rates of ad-
Sorption and deeorption become equal. Equating (1) and
(21,
kAaA<CZ - k h
sr) I N D 0S T R I A L A N D E N G I N E E R I N G C H E M I S T R Y
(3)
If component A is in admixture with other components B,
R,S, and I which are aka adsorbed on active centera of the
name type, rate and equilibrium equations similar to 1,2,and
3 may he mitten for each component. Then,
01 +
L - (CA + CB + CB + C8 G I . . .) (4)
At equilibrium conditions each of the adsorbate concentra-
tion terms in Ekption 4 may be replaced by an e x p e o n
CI -
similar to that obtained by solving Equation 3 for or:
L - Or(aA& + O d B -!-amKz + as& + a1iKn.. .) (5)
An expresion for the equilibrium surface concentration of
A in terms of interfacial fluid activities is obtained by wm-
beEquations 3 and 6:
Similar equations may be written for the equilibrium surf-
concentratiom of the other componenta of the mixture.
Surfacemctions may be assumed to tske place either be-
tween an a h r b e d reactant molecule and a molecule in the
fluid phaee or between adsorbed molecules on adjacently
eituated active centem. With the latter mechanism a rem-
tion proceeds at rates proportional to the concentrations of
adjacently adsorbed &ta. Thus, if adsorbed molecule
A reacts with adsorbed molecule B,the rate of the reaction k
pmportional to the nnpber of pairs of adjacently adsorbed
A and B molecules per &it area of d a c e . Similarly,if in a
monomolecular reaction, adsorbed molecule A is reacting
with a vacant active center to form a complex which subas
quenty breaks down to form two adsorbed pmduct molecules,
the rate of the reaction is proportional to the number per
unit area of A molecules adsorb%d adjscent to YBCant active
Centers.
It i s d t h a t thp active cyntersof the surface are dis-
tributed in a regular +metrical pattern detarmined by the
lattice atruoture such that each individual center is sur-
roqnded by a other caters equidistant from it. Thus, ifthe-
pattern is mtangdar with the active centem forming the
corners of squares, 8 -.4. If the pattern ia such that the
centers form the corners of equilateral trbgka, 8 = 6.
An average adsorbed molecule of A has adjacent to it 4
vacant centers, where 0, is the fraction of the total centera
which is vacant. S i b , 80. molecules of B m adjacent
tp each adsorbed A mpleoula, where 0. is the fraction of the
total centers which is occupied by adsorbed B molecules.
Then the d m conc8ntration of A moleculea and vacant
active centers adjacent to each other b m e s ‘/rsc;B, and
the concentration of A moleculea and B moWcs adsorbed
in adjacent poeitiom becomes I/lsc;e.. The factor re-
d@ from the fact that, in the summation r8p&ted by
the product of the concentration and frsotional adsgrpton
terms,each pair of spjgoent moleculea and centers is counted
(a)
twice.
Since 0, equals 4/L’and 0. equals c;/L’,
cii - 2L
8
7 oh’ (s)
(9)
Vol. 35, No. 5
where CAI = surface concentration of psirs of adsorbed A mole- the over-all rate of reaction. Thia net sdsorption rate is the
cnles and vacant centers in adjacent pitions difference between the rates of adsorption and desorption, or
c i s = surface concentration of pairs of adsorbed A and B
rnolecnlas in adjacent p i t i o n s
Equations 8 and 9 may be similarly written with concentra-
tions in moles per unit mass.
The rate of a monomolecular surface reaction involving in- Similar expressions may be written for the other reactant and
teraction of an adsorbed A molecule and a vacant active cen- product adsorptions, and substituted in Equation 12:
ter then becomes:
1 - --
kcai k
S
2L C I C Z (10)

Similarly, for a surface reaction between adsorbed A and B

molecules: The concentration of vacant centers dso may be expressed
ks in terms of fluid-phase activities at the interface:
I = kcrs = - CdCB (11)
2L

O V E R A L L SURFACE RATE EQUATIONS

-
c, = L - (c.4 + C B + CR + c,. . .)
+
L - c!(Knari Kmsi + K m , + K,a,i ...) +
r( 1z +-1 --)1
On the basis of the theory outlined above, a general equsr ka kz
tion may he developed for the followingsurface reaction:
A+BF?R
-
-

(1
L+ c i kh '1
ki
+ KnaA. + K m s . + Kmn. + K m l . )
(15)

In the geueral caw the system will include inert components

in the fluid phase represented by I. ...
The net rate of the forward reaction is the dserence b-
tween the rates of the forward and reverse reactions which
are represented by Equations 11and 10, respectively. Thus,
I = -
8
2L (kcics - k'cmcr) (12)
The surface concentrations may he expressed in terms of the
activities in the fluid at the interface from Equations 1 and 2.
Thus, the net rate of adsorption of component A is equal to
Combining Equations 15 and 14 yields a complete expression
for the rate of reaction in terms only of activities in the fluid
at the interface and the constants of the,system. If the con-
stants were all evaluated, the equation could be solved graphi-
cally for the rate at specified conditions. However, in gen-
eral this complete equation involving the rates of d adsorp-
tion steps as well as the rate of the surface reaction is so cum-
bersome as to be of little value.
As previously mentioned, in many cases it is Batisfactory
to assume that the rate is controlled by a single slow step and
that all other steps are so fast that equilibrium may be m

MAY, 1943 INDUSTRIAL A N D ENGINEERING CHEMISTRY 531

 med. The slow step may be either the muface resotion or
the adsorption of any one reactant or the desorption of any
product. Simplified useful equations may be developed on
this basis.
SwpLlFlED EPUATIONS FOR MONOMOI.ECULAR SURFACE
RMCllONS
T w o mec6anisms may be p t u l a t e d for reversible surface
reactions which are monomolecular in both directions, such
as immeri5ation resctions. The reaation might proceed
through an sdsorbed molecule, squiring sufficient energy to
cam it to d and form a product molecule (dl the time
'
conlined to a &le active center). In the other mechanism
an adsorhed molecule b m e s sufficiently energized to form
a complex with an adjacent active center, which then d m m -
posaa to form an h r k d product molecule.
If the hut mechanism is followed, the rata of reaction is
proportional to the concentration of admrbed reactant mole
de8:
r - kcl (16)
If the second mechanism is c o m t , the rate of reaction is pro-
portiod to the concentration of psirS of sdsorbed molecules
and adjacent vacant active centers. This rate is e x p r e s s e d by
Equation 10.
ABBuming that only a single activated step is rate control-
Ling, aimplised over-all equations may be derived accordingto
either mechanism for the reaction:
A#R
Surface R a t i o n Controlling. If the surface reaction
is rate controlling, it is assumed that all adsorption steps
are in equilibrium, and the surface concentration terms in
Equations 10 and 18 may be replaced by Equations 6 and
7:
Mscanmem 1
r- kL ( a A 1 K Aa-nTa)n (In
+
- +
1 S A S A f amKB Or& ...
w h K' d a c e equilibrium conatant klk'
At eqnilibritunthe net rate becomea m a and
- A+B*R
Subtitutiug Equation 18 in 17,
The eeaential d8erence betwean Equations 19 and 20 is the
-
squared term in the denominator. According to them equa-
tionn the initial rate, when an, 0, of s reaction following
mecbqhn 1 is pmgreaSvely increased by i n c d activity
of tbe Wtap$@thefluid phsae. If mechanism 2 is followed,
the initial rat# ma*%- :jcrea&?d activity of A at
%
low activities, reach a maximum rrespondiegto an optimum
activity, md then diminish 88 the activity of A is further
inrreased. Etrecta of thin type have been experimentally o b
served in varying the pwaure on monomolecular catalytic
dions.
Adsorption of Reactant Controlling. If the adsorption
of the reactant is the rate-oontrolling etep, it is assumed
that the surface reaction and the other adsorption steps
are in equilibrium. The rate is expreased by Equation 1:
r = k*arrc! - k h (21)
532 INDUSTRIAL A N D ENGINEERING CHEMISTRY
In the application of Equation 21 all surface concentrations
may be expreaaed in terms of the equilibrium of Equation 3,
with the exception of cA. This surface concentration must
be arrived at from the condition of equilibrium in the surface
re&ion:
CI = L - ( C A + C B + C I . . .) = L - cr(aniKdK + as& + . .)
ar&.
L
=1 + aniKdK + amrKn + ariKr. . . (23)
Combining Equations 22 and 23 with 21,
Equation 24 is independent of the mechanism of the surface
reaction.
Comparing the form of Equation 24 with 19 and 20,it will
be noted that, when adsorption of the reactant is controlling,
the initial rate (where aB1= 0) is d-ctly proportional to the
activity of the reactant; if the surface resetion is controlling,
the initial rate will i n c w less than in proportion to in-
creased activity of the reactant.
Another ChSrsderiStic of reactions controlled by a d s o p
tion is evident by consideration of the net rate of the rp.
VBIIK) reaction of Equation 24. If the over-dl equilibrium
constant K is small, the net rate of the reverse reaction be-
comes independent of the activity of A. Thia behavior ia
characteristic of reactions in which the rate is controlled by
the rate of desorption of the product.
BWOL&CULAR-MONOMOLBLAR BEACCION
The c h of reaction which is bimolecular in one direction
and monomolecular in the other may be represented by the
equation:
If it is assumed that the resetion t&es place betweeq an-
cently sdmrkd molecules, it is necessary that the monc-
molecular revem regation follow the mechanism e x p r e s s e d
by Equation 10. By the me pmcedures followed in the
preoeding section, the foUowing rate equations are developed:
SUEFAm W C T I O N CoNTROLImQ
AOSORP~ONor B CONTROLLLVQ
knLK
I-
1 + aA& + as& + aAiaerKrK + anKr + ... X
In Equation 25 the initial rate of the reaction is increased
by increased activity of component B; in Equation 26 i n c m
ing the activity of B reduces the initial rate where the activity
of R is sem. Thus,' it is characteristic of a bimolecular reac-
tion in which the rate is controlled by the adsorption of a re-
Vol. 35, No. 5
sdent that the initial rate is reduced by i n d activity
of the other reactant.
In Equation 26 the o v e r 4 equilibrium constant of the re-
action appears in both the numerator and denominstor of the
multiplying W o n . Under conditions favorable for the for-
ward d o n where K ia large,the rate becomea independent
of the activity of the pmduct and may approach kmL/K.,
m d e p d e n t of the activities of both resatanta and pmduct.
If the reaction on the mrfsoe ommn betwem an &orbed
A molecule and an unabsorbed B molecule in the fluid p b ,
the rata is proportional to the pmduct of the surface concen-
tration of A and the intarfacial fluid activity of B. Equations
Bimilar to 25, 26, and 27 result, except that the ademption
equilibrium constant of B does not appear. If the surface
resotion is contmliii, the term a/Z doee not appear, and
the denominator group entera aa the first power instead of the
quare.
P&*CllONS BIMOLECULAR IN BoTa DIRECTIONS
Reactions of this type are repwanted by the foUowing
equation:
A+BFf.B+S
By extension of the methods demonstrated above, the follow-
ing rate equations may be derived, a~~uming the muface re-
action to take place between adjacently adsorbed molecules:
atom A reding with several other mohulea or atoms B, all
of the name qmciea. In ~ ~ accme h high rates are obtained
with a catalyut wbich drongly absorbs regctant B so that the
majority of the surface is covered with adearbed B molecules
or atoms, and moet of the A molecules or atom which are
adearbed are aurmunded by the requisite number of B unite
on adjacent centers.
ADSORPTION WITH DISSOCUTION
Although other mechanisms are poeaible, Glaastone, Laid-
ler, and Eyring (4) suggest that, where a molecule is dim
cisted during adsorption, it is commonly first &orbed on a
dual adsorption site consisting of two adjacent active centers.
This adsorbed molecule then decompoees, and the atoms
jump to two vacant active centers. Thus, if A, is a molecule
which dissociatesinto two A atom and b is a dual adsorption
site,
& + h FtaJt
Ash +
W 02Al+ h
It is believed that the first step is rate controlling and that
equilibrium is maintained in the seoond. Then if r is ex-
pressed in moles of &
- ~APACS -k h h (31
Smce the w n d step main& equilibrinm,

c4* - c
'
&

The conmtration of dual Saeorption etea is e x p d by

an equation Bimilar to 8 and 9:

Cl, -a 8
a
Combining 31,32,and 33,

At equilibrium,

Compsrisons between Equations 28,29, and 30 are similar

CA - -
CAI C l d 4 x (W
to those between 25,2#, and 27. A reaction whom rata ia If A, ia adsorbd with diesociation from a mixture slao con-
controlled by the adsorption of a reagent ia charaoteriaed by taining wmponenta B, R, and I, an expression Similar to
an adverse effect on initial rate resulting from an increw in Equstion 7 may be developed for sdsorption equilibrium con-
activitv of the otber reactant. A d o n wbose rate is oon- ditions:
boUedby adsorption of a product ia charadensed . bytheini- . m,
L
tid rata being independent of reactant activities under con- c. =
dit io^ of large over-all equilibrium comtmta and negligible 1i s +
f anRe a d m + a r K ~ . . .
(37)
+
r e m resation. meet of Dissociation on Reaction Rate Banatione
REACTIOIYS mVOLVING MOPE THAN TWO MOLECULES
From Equation 37 it is evident that, where amolecule
dimociatea 88 it ia adsorbed, the square root of the product
The adsorption theory of ostslysia doee not preclude r8ao- of ita activity timea ita adsorption equilibrium aonstsnt
tions involving the simultaneous interaction of several mole- appears in dl adsorption equilibrium expressions in place
cules, and no highly improbable mechsnism such an the m- of the lirat power of this product. Thus, Equations 19, 20,
multaneous collieion of aU mol& ia involved. The rates 25, and 28, in which d ademption steps are at equilibrium,
of such raectionS should be proportional to the conaentration may be m&ed to apply where a component is h c i a t e d
of group of the required number of moleaalea adsorbed on ad- and one half molecule participates in the reaction merely by
jacent active centers. Rate equation~might be built up by raising to tbe one hslf power the activity and adeorption
exheion of the procedures used in developing Equations 10 equilibrium conetsnt of the component which ia dissociated
and 11 and the foUowing relatiom. wherever either appears in the equation.
It would be expectd that the rates of such reaction^ would The etIeots of diesociation where adsorption rate is mu-
be low because of the low conaentrationn of propdy ad- tmlling may be demonstrated by coneiderstion of the &&,
sorbed groups except in the case of one reectant moleoule or on Equation 29 of the dissooiation of B 6rat and then A. If
May, 1943 I N D U S T R I A L A N D E N G I N E E R I N G CHEMISTRY 533
A is not dissociated the rate is e x p d by Equations 1 and ucb. It is Ba8umed that the ooncentration or activity of the
2: activated complex molecules msinteina thermodyosmio
I= k (ahCz - 2). (as)
equilibrium with the activities of the reaatSnta. The rate of
reaction is, then, the rata of decompwition of the activated
complex moleculea into products which in turn is pwportional
If B is dissociated and only */,E molecule entera into the reac- to the concentration of activated complex molecules accord-
tion, an expreeaion for cAis derived from the equilibrium of ing to a relation which wan derived (4) from the principles of
the surface reaction: Btatistical mechanics. Thus, for the muface phenomena
under discussion in which rates are e x p d in moles per
unit mBBB,

whem k = reaction velocity aonstsot, mdea per unit time per

AH%,A S S - unit mass of catsrlvst
molal standard enthalpy and e n v p y a
companying formation of aohvated eom Y%x"
Epuation 46 appliea both to activated adsorptons and surf&
MCtiOns.
Since in dl the rate equations the reaction velocity constsnt
irj multiplied by either L or 8L& it is convenient to group the
constank together:
Comparing Equations 41 and 29, if the component whose ad- --An%
+BT
T A8f
sorption rate is controlling is not diaaociated, the rate equa- kL or &
!2 = C'TE
tions are modihed .for application where a component is dis-
sociated and only one half molecule enters the reaction, by w h constant C' is a function of the nature and extent of
raising the activity and adsorption equilibrium constant of the catalytic surface.
the dissociated component to one half power wherever they To extend the catalytic rate equations to aU temperatures,
appear in the equation. the expreeaions of Equations 46 and 47 are mbstituted for all
If component A of Equation 29 is dissociated and only one equilibrium and velocity comtanta, respeotively. The re-
half molecule enters the reaction, it is neceassrg to expreen dtmt equations become cumbermme and it is prefmble
the rate of reaction by Equation 34. From the equilibrium to d d e r the temperature vsristion of the individual oon-
of the surface reaction, atants separately with the aid of simplifying m p t i o n s
which will be d i a c d later.
mieons
CI - + +
L The term "poison" has been low& used to denote any
+ +
+ '
amasra
asiK
an& amKm -tas& ariKi ... materid which i - e t d ~the rate of a catslytic &ion.
the producta of the reaction am sometima refend to as
(48) poisons because in-
Thus,
their concentration reduces the
Substituting Equations 42 and 43 in 34, rate. However, it is halieved more loaicsl to eliminate all
reactanta and products from cla&cati& 88 poiaons and re
aLk1
, w e the term for other components which retard the WBC-
+
o n S n a d s +as& + alrKl + .. .)
tion by reducingthe numbr of active centera available for the
reactanta. suchpieonsmaybe~edsseithert9mporsry
amsi
x [ah - (3)1 or permanent.
(4)
From inspection of the rate equations it is evident that my
This equation is quite ditiarent in form from Equation 29, and inert mponent, I, which is adsorbed on the activa centers
it is slso dependent on the meohaniam assumed for the ad- promoting the reaction, m es an a poison; the extent of the
sorption and diemciation of A. By aasUming a different retardation depends on the concentration of the inert and its
mhanim (41, an equation c811 be obtained in which the ademption equilibrium conatant. I f t h e h r p t i o n o f t h e i n -
initial rate is proportional to the one halfpower of the activity ert is revemible with a moderate equilibrium constant, the
of A iwtezul of the first power of Equation 44. poisoning &ect quickly reaches a conatant value which is
temporary and is eliminated by moval of the component
Effect of Temperature. Each of the equilibrium oon- from the nuid, mu, the , ,im in might
atanta appearing in the rate equations is expressed as a m e an a poison in a catslytic &&tion. he
funotion of temperature by the following equation a p atiects of such inh LvB in&,&J in the rate equations.
plicable to both chemical and adsorption equilibria: Another tspe of pmkming frequently encount3red resub
- -
-AO0 -AX' AB' from the permanent, substantially irreversible h r p t i o n of
K - S rn -e + T (45) componenh present in smsll quantities in the reactant Uuid.
The &eat of this type of poisoLling is promwively to reducs
Although in general AGO varies greatly with temperature, it L', the number of active centers per Unit area, an the catalsgt
is frequently sstisfactorg to m e A W and Meto be continuesinemvice. Thefmquentlyobserved1~ofBctiVity
iudepndent of tampmture at averags values. of a catalyat in &ce may result from poisoning of this type,
According to the Eprincl theory of absolute reaction ratea from structural change% afiecting the area and E, or from
(d), any reaction may be considered an proceeding in two physically coating the active surface with a solid ox semisolid
steps,first the formation of an activated complex and then the mahid either presomt in the reactantsor formedby BBcolldBsJ.
.decompoeition of this activated complex into the final prod- reactions. In high-temperature or&c d o n 6 c a w
534 INDUSTRIAL AND ENGINEERING CHEMISTRY VoL 35, N a S
 Below. commsrcfal G t d y t i c co-
dimerCamhaUnitotRi he; Thermal
Unit a t wt; \ L e in the
Csntsr
am frequently coated with high-molecular-weight compounds
approaching pure carbon in compmition. Such coatings
may be looked upon as reducing the dective concentration
of active c e n h L.
JJMITATIONS OF SUIIFACE RATE EQUATIONS
The rate equations derived above are ideal forms involving
~ v e r aasaumptiona
l which are not rigorom. The derivation
implies that all active centers on catalytic surface behave
eimilarly d e there ia evidmce that this is not the c~se.
Frequently the en- of activation gradually increaaas an
active centers are progressively oocupied. It is well known
that surfsoe activity depemds upon spatial amaement of
active centam which may not be uniform over the entire mr-
face. Ale0 it is as3umed that each eomponent is adsorbed
independent of intaraction between the adsorbed molecules
of like or unlike ap&es. Again them in evidence that mch ia
not the ~ 8 8 8 ,and correctionswere developad (4) for such in-
tersctton && in simple -. Intmsction between ad-
sorbed mole& temds to reduce the loss in enthalpy ordi-
narily accompanying activated adsorption.
Although numerous attampta have been made to establish
more rigoroun h r p t i o m relations, them is little evidence
that their use is justified in rate equationsfor complex reaction
systems. It ia belied more generally sstiefaatory to UBB the
May, 1913 I N D U S T R I A L A N D E N 0 I N E E R I N O .C H E M I S T R Y
equationsdeveloped above and evslu-
ate the c o n h t a from direct reaction
rate masgurements at conditions
representing the ranges of intereat
of the variables. Comtanta directly
evaluated from axperimentsl data in
this manner represent average ap-
parent or efiective values over the
range coneidered, and the reeultsnt
equations cannot be expected to be
entirelyrigorous.
As in d theoretical rate oalrmla,
tions, the form of the relation is d e
pendent on the mechanism. Thi%
situation is entirely diffenmt fmm
that encounted in thermodynamic
equilibrium d c u l a t i m in which the mechanism is imma-
terial. Furthermore, the correct me&a&m cannot be 88
lected by consideration of Stoiohiometric equations but must
be determined by comparison of the trends of exprimentd
data with the trends manifestedin the various rate equations.
EFFECT OF CATALYST muosm
To obtain large catalytic amaa and still have a c a t + &
form which may be easily handled and will produce g o d &JW
distribution, many industrially important catalyst6 are Usal
in relatively large partiales having a high degree of porc4ty.
In such particles the gmes e x t a d area of the partide is
negligible in comparison to the large internal area of the ?rea,
and the total catalytic area per unit mass is practically m d s
pendent of particle &e.
In the derivation of t&bssio rate equations it has been im-
plied that the activitim of the reactant6 and producta in the
fluid phase at the catalytic interface are constent throughout
the catalyst mass. However, where 6nite m t i o n rates are
produced, ooncentration gradients exist within a catalm
particle because of the Mueion and flow of mtgnta into the
inner porea and the diffusion and flow of prcducts out to the
main fluid stream at the partide d a c e . It in evident that
the regction rate per unit surface may vary widely from a
msximum at the exterior to a minimum at the centar of the
535

particle. In such a care the over-all rate of d o n is less
than would result if the i n t e r f a d fluid concentrations were
constant throughout at the exterior values.
The ratio of the actual d o n rate obtgined per unit
maea of catalyst to the rate which would reault if no conuen-
tration gradients existed within the particles may be termed
the “&eetiveness factof’ of the particle and dedgnated by
E..
It in evident that the &ectiveness factor of a catalyst in a
function of the particle &e, the pore dimeneiona and the ex-
tent to which the pores am intexmnneeted, the catalytic 80
tivity, and the di5uaional chsracteristiea of the reactants and
products. The relations between these fsetora were d e
veloped by Thiele (8) for e a 4 simple cases which, with the
aid of reaeonable eimplifying aesumptions, could be handled
mathematically. F&gomus equations of thia type involving
complex rate equations would be difficult to develop and cum-
bersome to use.
It in suggested that an empirical method for taking into 80
count variations in &wtiveneen faotor be b d on the curve
developed by Thiele for spherid particles catslyzing a fvst
order irreversible reaction. The eEdveneea factor in ex-
pm&d ae a function of a modulus D ; m by Figure
1, plotted from Thiele’s equation (8). The aective diam-
eter 0;is delined ne the diameter of a sphere having the -me
swfsce area per unit volume BB the particle. Thus D; =
B/A&. The resetion velocity constant in the modulus
w a ~defined as the proportionality Isctor relating reaction
rate and concentration or activity in an irreversible reaction.
However, the eane erpression results if k ia dehed as the
pmportiodty factor relating reaction rate and the differ-
ence between the reactant concentration and the equilibrium
value in a reversible reaction.
Application 01 Figure 1 to otber than hrseorder reactions
involves the assumption of a linear relation between reactant
concentration and reaction rate. Although by no meam
gemrally rigomorqus,e ia a fair appmximation in many meed.
Further wmpheatrorm rue introduced where B chmp of
volume occurs in the reaction, resulting in d a t e flow BB
well BB diRimion in the porea. However, Thiele’s studies in-
dicate that theee && do not greatly dud the over41
form of the relation, and it in believed that Figure 1 in a useful
appmximation for general application, particularly if the con-
536 INDUSTRIAL A N D ENGINEERING CHEMISTRY
stgnts o€ the modulus are empirically de-
termined at conditions in the range of opera-
I ting interest.
For use 88 an empirical approximation,
the squsre root terms of Thiele’s modulus
may be combined into a single factor which,
for a given catalyst and reaction system,
may be ae8umed to be a function only of
temperature, repmnted by the usual ex-
ponential relation. The modulus may then
b
-9
be written 88 D;a’a ,and
,
B A - ~ ~- If ) D f l 8
( ,-$)
. (4)
with the function representedby Figure 1.
ALL the p d i rate equations should
include the &e&venmfactor EAae a multi-
0 plier on the right-hand side, when applied to
porous catalynta for which the extemal ares
of the pellet in negligible aa compared to the
internal area.
MASSTRANS= TO p h n c u SUBFACE
Therateequationsintheheprecedingaactions
are all e x p d in termaof activitiesin the fluid p h at the
catalytic intarface, which, throughthe incorporation of the ef-
feetimess factorin the quation, are takenaa the m e BB the
activities in the flnid phaee at the grim exterior surface of the
catalpt particle. These interfacial activitiea are less or
-tar than the activitiea in the main stresm of fluid sur-
m d h g the particles by the amounts of the activity gradi-
mts to maintain difhvdon. Thus, if (I., is the 80
tivity of A m the main fluid stream,
SA< (IA - &A (49)
Similarly, for a pmduot, S, difiusing away from the surface,
(Is< - (I8 + &os (50)
The activity difiemcw AaA and Aa8 are functions of the
ratea of diRusion and the flow ChsraeteristiCS of the system,
which may be calculated with astistSctory accuracy for many
gaseous system and may be empiricaUy evaluated by similar
m e t h d for other 8ystem.
Chilton and Colburn (@, analyzing the m888 transfer and
difiusional problema in &as absorption and distillston, in&
duced the concept of the “transfer unit”; it may be de6ned as
the length or height of contacting zone required to produce
a unit concentration change in the mainbody of a fluid stream
in the dkations of flow 88 the result of a difiuaional driving
fome p d u o j n e the transfer of unit concentration diiference.
The contgoting- c ‘ ‘cs of a system am conveniently
e x p d in tern of the haght of such a transfer Unit,
H. T.U.,a factor having the dimension of length. Chilton
and Colburn pointed out that for a given packing material
the H. T. U. in a function of the modified Reynolds number,
D p / & and the klunidt number p/pD.. In terms of this
concept the dmp in partial preasura of gaseous component A
BB a d t of transfer may be repreanted by the follow-
ing equations:
For equimohl count8rdiEusion of the two components in a
binsryaysterm,
For di5nejon in a binary qystem resulting in rrrmwal of com-
ponent A, the other component remamng,
Vol. 35, No. S

where P A 106 mean of (r
~ = - pd and (r - PAC)
Rigorous treatment of the simultanems Musion of several
components, particularly when volume changes are involved
at the interface, leads to complicated and cumhemme expres-
mons (6,8). As an approximation it is Buggested that con-
centration Merences resulting from Mnsion in complex re-
acting sydems be evaluated by direct application of Equation
51 where no change in volume m l t a from the reaction.
For determination of the terms in the Schmidt group, each
component is considered aa diffusing in a b q system com-
posed of itself and of d the other compnenta of the mixtnre
in the average proprtions in which they appear in the dflu-
m o d film. On thie bssis Mwion coe0lcienta may he
evaluated from the correlation developed by Gilliland (6).
It ahould be noted, however, that the algebraic 8um of the
partial pressure differenen of all the compnenta involved
must equal m.
Where a volume change &ccompBnie8 the reaction, the
&e& of a m888 flow in the direction of the volume decrease is
superimposed on the M n s i o d &e& that would he calcu-
lated fromEquation 51. This effect reduces the concentm
tion Merenca necesaq to maintain Mnsion in the diredion
of the volume d m and i n c r a w the concentration dif-
ferencesfor Mnsion in the oppasite direction.
Aa an approximation, arrived at by comparison of Equa-
tiona 51 and 62, it is rnggeatd that where a change of volume
&T, 19(9 INDUSTRIAL A N D ENGINEERING CHEMISTRY
I Z C T I O N R A T E s'
1 N ID l l S T H I A L R E .
I
accompanies the reaction, Equation 51 be used as described
above and the reault corrected by the multiplying factorf,:
In
= "A->)!=?( (53)
where A?' = change in number of mol% per mole of reaction
Concentration ditierences of reactante are multiplied by fo
while those of p r c d u h are divided byf,.
Fortunately the corrections expressed by Equation 53 are
generally not large, and errors involved cancel to some extent
if a consistent method of calculating diffusionaleffects is ad-
hered to both in the experimental evaluation of the constants
in the rate equations and in their application. Greatly sim-
p u e d procedures are possible in particular caaea; in many
c a w where conversion ntea ara low and m a velocities high
the concentration differenen may be neglected entirely.
Camson, Thodos, and Hougen (3) evaluated the dimen-
sionlass group a(H. T. U.) in terms of the Reynolds and
Schmidtnumbers for the Bow of gam through granular solids,
and establiahd a unique relation independent of the &e,
density, porosity, and composition of the solid. This rela-
tion waa found to hold for M e m t shapes, provided the par-
ticle diameter D, wa8 baaed upon the diameter of a sphere
having the eame gross area as the particle. Thus for a cylin-
der,
4 - 4D.H +2
where D. =
H - diameter of cylindrid surface
height of cylinder
s3l
The resulting corralation is plotted in Figure 2. h m this 53, the interfacial activities are obtained, correapondii to the
diagram values of am. T.U.)are readily determined for any measured average rate. The use of log mean activities is
system of known dimensions and physical proprties. merely an improved approximation, which in many c&8esis not
The efiective diamekr D# is not the m e aa the DL wed eatistacbry except where concentration changes are small.
in the Thiele modulus. For example, for a right cylinder of
~tequaltodiameterD:=D,wherea%D,=~D.= l3Bectiveneas Factor. This factor is evaluated aa a
function of temperature by differential rate measure-
m2 0;. The values of a@. T. U.) obtained from ments with two widely different particle sises at each of
Figure 2 are restricted to conditions where the total ~ m s ex-
s two different temperatures, holding the activities constant
ternal aras of the particle is available for heat and m888 trans at each temperature. At each temperature the ratio of the
fer. rata, r l m = EAJEA,coResponding to the particle sizes D,,
EYALUA'l" OF RATE EQUATION -STANTS
and D,, is determind. The Thiele modulus may be read
directly from Figure 3, derived from Figure 1. From two
Esch of the equilibrium and velocity constants of the rate d u e a of the modulus at merent temperatures the con-
equations is a function of t e m p t u r e expressed by Equation stants a' and b of Equation 48 may he calculated, completely
46 or 48. The incorporation of 48 in a rate equation is sim- evaluating EAas a fnnction of particle t h e and temperature.
plised by comb+ aU the multiplying constants into an Where this factor proves to he of considerable importance,
overdl rate constant, C. Since the entropy changes may be it may be desirable to i n d i t e a series of several particle
taken as indewdent of temperature, they may lrlso be in- sises to verify the form of the relation expressed by Figure 1.
cluded in the o d constant. Thus for Equation%,
Adsorption Equilibrium Constants. These constants,
appearing in the rate equations, might be independently
determined by adsorption measurements on the individual
components. Such measurements may prove to be a
valuable adjunct to direct catalytic rate measurements but
and this equation, including the effectivenee8 factor, may be are open to question beoauae of the interaction effects and
written: variability of active centers previously mentioned. Until
theee relations are better u n d d , thesa constants in the
- -AIIO'
ET
rate equations are best determineed in the ~pecXcreaction
EACTe under study by making rate measmwnenta over a wide range
r-
(1 + a d * + am& + + as& + ar& + ...I X
aa& of concentrations.
(aam -7 alias6 )(W To minimim tedious algebraic solutions of simultaneous
equations, it is deiimble in complex systems to determine the
where A""' = AIfllt + Mi + AH; effectiveaverage value of the absorption equilibrium constant
of each component by a constant-temperature series of three
The over-all chemical equilibrinm donstsnt, K, can or- difIerentiS1 rate measurements in which the activity of that
dinarily be evaluated from thermcdynamic data. Experi- component is vsried over a wide range while holding the ac-
mental evaluation is required for the effec-
tiveness factor, the overdl rate constant,
the over-all standard enthalpy change, and
40
the parious apparent adsorption equilibrium
constants. 30
conatants in complex equations of this
type are heat determind by a diEer811tialre- 20
sctor containing a bed of catalpt so shal-
low that relatively small changes of com-
position are obtained. Such apparatusahould
be deaigned to produce the smallest com- 10
9
position changea that permit accnrate evalua- 8
tion of the rate of reaction with the available 7
analytical methcds. Because of the expo- 6
nentinl inauence of tempratun, it i% d e s
sirable that the dilierential bed be at aa uni- 4
form temperature as poesible, requiring a ml
3
ddiamekr reactor with good heat t r a w
fer provisions in the walk of the container.
2
In a difierentisl reador the average 80
tivity of each component may be taken as
the arithmetic average of the mlet and out-
let paluea for small changea or, better, de-
r i d from the logarithmic mean of the dif-
ferencea between the activities at the inlet .8
.7
and the outlet for the reactants. The mean .6
activity of a product, R,may be tsken as 0% .S
minua the log mean value of (a*, - as) at the A
t a m i d conditions where a=. is the activity 0 .I .2 .3 A .5 .6 .7 .8 9 I
of R when the reaction proceeds to equi- m2
librium. Then by calculatingthe diffusional K=T
activity ditierences from Equations 51 and Figure 3. Determinotion of Eflecticeneds Factor Modulus

998 INDUSTRIA L A N D ENGINEERING CHEMISTRY Vol. 3s. No. 5

tivities of all other components constant. A convenient
procedure is to make an initial run with the activity of eacb
component at approximately the middle of the range of in-
terest. Additional series of two ruua eacb are then made in
which the activity of each compnent is varied from a mini-
mum d u e in one to a &um in the other while holding
the activities of all other components constant. This proca-
dure may require two or more trial8 at eome of the conditions
with interpolation in order to obtain saWactory constsncy
of the grouped activities but is not diflicult with a flexible
diBwntial reactor appmtw which permits wide variation
in feed ratas and depth of catalyst bed.
Oncethe adsorption equilibriumconstantsare all eduated,
eacb single rate measurament permits calculation of the group
-&p
__
RT
Ce
Effect of Temperahue. Each adsorption equilibrium
constant is a function of temperature expressed by Equa-
tion 45. For a complete experimental evaluation of the
standard enthalpy and entropy changes in adsorption
under reaetion conditions, it is nwwary to repest at a
different temperature the evaluation of the individual aver-
age adsorption equilibrium constants as described above.
Such measuremente completely evalnate the equation with
determination of C and AHo’.
Where such complete data are not available, the standard
entropy chsnge accompanying adsorption at one atmoaphew
may be roughly approximated by amming a lom of 23.0 en-
tropy units. Such an m p t i o n is Bomewhat similar to
Trouton’s rule. Entropies of adsorption aka may be derived
fromactivated adsorption equilibrium messurememte for the
individual components on the cat+&. It is believed that V I
entropy changw determined in this maqner may be eatid@ %-
torily used in rate equations, whereas the oorrwponding en-
tbalpy changesmay be of little d u e .
If stsndard entropy ebangea in adprption are approxi-
mated or measured, a single deteumbtion of the equilibrium
con%tant permite evaluah of the stsndard enthalpy change
and complete expression of the c o h t as a function of tem-
perstwe. If these functions are evaluated for all compo-
nente, C and A””’may be evaluated from sng two rate meas- feed. Through the stoichiometry of the reaction and the
wements at di6mt tempersturea In thia procedure BIRII~ relatione between mole fractions and activitiw, the activity
in the temperature coedfcientaof the equilibrium constantsare of eacb component in a given resotorfeedmay be e v e . l d as
h r b d in calculation of AH*’and may have relatively little a function of zA and plotted in c u m mch 88 Figure 4.
efied on the overall accuracy of the equation. Where the resction is at constant pressure and a c t i v i t i ~are
Activated Msorption Meutrrsmenb. It 6as been as- equal to mole fractiom or p d d premuw, this calculation is
sumed that the correct meahanism and equation wan simple and yielda linear relationa if no volume cbange WOUI’S.
&bW. This may be done by qualitative anal* of From the relatiordip between and activities, the reac-
&a &&a of varying the activities of the componentsas pm- tion rate r for an isothermalor adiabatic reaction at oonstsnt
y i o d y discussed. However, it is believed that much ns3fuI preaaure may be calculated as a function of z4 by means of
Mormation dong theae linea may be obtained by meamre- the rate equation. When working with an adiabatia resdor,
menta of dvated sdsorption rata and equilibrium con- temparsture must be e x p d as a function of convemion
stsnts of ~e individual components on the catel&, even fromthe thermal propert& of the Bystem. Such a relation
lthough the resulting constan$a a n n o t be substituted di- is plotted in Figure 4 for a more complex osse; the upper rate
,rectlyinto the final rate equation. curve is based on the assumpton tbat intarfacial activitiea
Adsorption meaeurements will m e to establish the r e b are equal to theactivitieain the main stresm of fluid.
tive magnitudes of the equilibrium constank, detect dimcis, If activity di@mcea due to -on are.imporht, a
tion during adsorption,and permit calculation of the entropy trial-and-armr calculation ia required in all except the most
change in adsorption. Further invedigation is n~ce881vybe- simple anew. Tbe &e& of these diff- on the initial
fore the full value a l l b e estsblished of mlchmeaaurementain rate ia datarmined by Bs$uming an initial rata slightly lower
.
supplementingdirectrate determmtiow. thanresd from the upper curveof Figure 4 and calculatingthe
di6nsional conmtration diflmcea from Epuations 51 and
-BATION OF lUTB BQunnoNS 53. Unng these ditIereneea to obtain intarfacial activities, a
co17BctBd initial rate is detarmined. If this rate is mriously
The principal value of a rate equation is to determine the Merent fromthat originally B B ~ ~ ~ I I Lthe
B ~calculation
, should
volume of reactor required to produoe a specified catalytic herepeated. WherediBuaionisacontmllingfador,itmaybe
.convemion. The exteut of convernion may be designated n~ce881vyto repeat mch trial-and+mur calculatione through-
byz, the moles afraadsnt A ccmverted per mole of total out the renotar as the integration pmceeda nnwever in
Max 1943 IN D U S T R I A LZA N D E N (3 IN E E R IN 0 CHEY I STRY
m y OILBBB the complete rata curve h d on true interfacial
ectivitiea is Satisfactorily established by the foUowing a p
proximation:
where r, re - corrected rate, b a d on true interfaciqlactivities
r', r: - corresponding to convemion Z A and lnitial con-
ditions, v t i v e l y
conwpondmg ratea, neglecting diffusional decte
In this manner the lower rate curve of Figure 4 may be ap-
proximated from 4 single trial-and-emr calcnlation at the
initial conditions.
In an elementary cross section of catalyst bed, having a
volume dV, a conversion dz, - is -~mduced. From the de6ni-
tion of r,
F&A = rpBdV
Since r is a function of z4 expressed by the cnwe of Figure 4,
this equation may he written:
(57)
It is common practice to expm correlations of catalytic
conversion data in tern of the space velocity, S., which is
de6ned aa the volumes of feed, messured at standard condi-
tions, per unik time per volume of catalyst. Equation 57
may be rewritten to expm the relation between pace v e
loeity and conversion:
(58)
The integral is readily evaluated by plotting the reciprocal of
the corrected rate curve of Figure 4 and measuring the area
under it.
Although the above procedure is not particularly tedious,
considerable simpli6cation is poasible under ideal conditions
in~~lving simple forms of rata equations or where f(zA)eat&
factorily approximatea a stmight line.
In the general c w , frequently encount?red in proceas d e
sign problems, the integration of the rate equation may he
complicated by Significant preeaure drop in the reactor and
temparature conditions which are neither iaothermal nor Sdia-
batic, hut correspond to some speoified rate of heat transfer
to or from the reactor. Under these conditions the curyes of
Figure 4 may be developed by a progreeive seriee of trial-
and-error calculations stsrting from the reactor inlet. A
further complication is intrcducd by the transverse tempera-
ture gradients resulting from the heating or oooling of the
reactor walk and temperature difiereneas between the flnid
and the adjacent catalyst surfaces. The temperstwe cunrg)
of Figure 4 represent the dective average temperatures acmss
any trsnsvew section of the catalyst bed. Readors are
generally deeigned to keep these transverse temperature dif-
ferences ~ d l The . temperature d8erence between the
fluid and the catalyst is given by the following equation (8)
if no volume change aeoompnniea the reaction:
Where a volume change occm, AT of Equation 59 is divided
by fD of Equation 53.
A small increment of conversion, zA,is seladed in which
the avenge values of all variables may he matuned to be the
sllo I N D U S T R I A L A N D ENGINEERING CHEMISTRY
arithmetic mean of the terminal d u e s . An average rate r'
for this increment is aeaumed generally somewhat lower than
the initial rate r; calculated from the initial activitiw.
Based on this assumed value of r', the diffusion wmction~
are calcnlated and the corresponding value of the corrected
average rate, r.. ., evaluated. The volume of catalyst c o r n
sponding to the selected increment of conversion is then
FAzA,/psr.. .
Using this volume, the pressure drop (1, 8)
and temperature change in the increment are calculated from
the Bow and thermal characteristios of the system after ap-
proximation values of t, and p, have been established from
which the activitia corresponding to zll are calculated.
Using the average activities and average temperature of the
incremental section, a second value of the average rate is cal-
culated. If thia value differs greatly from that initially as-
sumed, the calculation is repeated. After satistactory values
for the average and terminal conditions of the 6rst increment
are establish& the pmedure is repeated for a second incre-
ment, AzA,, and these Btepwise calculations am continufd
una the desired conversion zAris reached. The total reactor
volume is then the sum of the incremental volumes, and the
'temperature and preeaure variations throughout the reactor
will have been evaluated. This procedure is not unduly te-
dious hecauee with experience the various curve8 esn he ex-
trapolated from increment to increment with nn5cient ac-
curacy to make second approximations rarely nacesssry.
CONCLUSIONS
The foregoing equations and methods have been developed
in the belief that a fundamentally sound approach will greatly
minimiaethe amount of pilot plant experimentation o M y
employed in evaluating the &e& of the operating variables
of catalytic 6yatems and arriving at optimum reactor designs.
In ideal a few dBerentia1 rate measuremente will ob-
viate the need of a pilot plant and permit immediate designof
large-scsle reactors for optimum resulte. However, it must
be recognised that the theoretical treatment is i d e a l i d and
in ite simple form doee not take into account semndary and
side reactions which may accompany the main d o n .
Furthermore, in the extenston of equations to conditions far
Werent from those at which the constante are evaluated,
mious errom may result from the assumptions of constancy
for auch variable fsctors aa enthalpy changea in activated
adsorption. Careful consideration of such limitations with
considerable discretion must he used in the application of
these methods until the basic principles have baen more
thoroughly explored and teated.
ACKNOWLEDGMENT
The work reportad here waa made possible by the h c i a l
~ ~ p p oofr tthe WiscOnSin Alumni Research Fo~ndationand
the sponsorship of the University Resesreh Committee. The
helpful suggastons and criticisms of J. 0. Hirschfelder are
gratafully acknowledged.
a - NOMENCLATURE
gross external aren of aatalyst particles/unit of total
a'
a., as,
am, am, etc.
etc. --
catalyst volume including interstices
= empirical constant
activities of components A, B, etc., in main fluid
phase
activities of components A, B, etc., at catalytic
interface
A =corn nentA
A = totaKnterfacial-/unit mass of catalyst
A, =w ee external area of catalyst particles/unit mass of
catalyst
b a- emnirical constant
B -componentB
C A , CB = surface concentrationsof adsorbed A and B,moleslunit
maas of catalyst
Vol. 3s. No. 5
 -
--aurf.y conceqtautiou of adearbed A and B~molecules/
umt cat.dytlc srsa
concentautim of o?LDILnt saaorption sites, equivalent
I
r'
2
--
-
rawtion rata, molea/unit
reaction rate,
-
of catslyat/unit time
exkmd difhmion @enta
average hyhulic radius of catalynt pores

---molea/unit maas of catslyst

ooncmtaution of vacant dnorptim sites/unit area
O V 0 F d Mk COUBtWlta
mean molal heat capaoity of fluid atrerun
-arslswconatant
actin

--
Thiele modulus 6/A+* -
dective average diametez of catalyst particles for 8
as' --
-componentS
standard entropy caange
dective average dinmeter of catalyst particles for
Reynolds number 8. --
as! entropy of actnration
qacevelocity
&ient
--of difhrsion ZA
-
moles of A aonverted/mole of reactor feed

--
dectiveneaa factor
d W o d correction factor (Equation 63)
reactor feed rata, moles/unit time
@A, 0s

e,
fraction of total sitea occupied by adsorbed A a n d B
-- molecules
fraction of total adsorption sites unoccupied

AHA
---
= maas velocity/unit tctsl mxa section
standard fm-dwqy change
Plan& constant
heat of reaction uw mole of A
P
p
PB -
--
~osity
-densit offluid
bulL
-trued&
Jkity
of catalyst
ofcatalyticnolid
--
po
AH' standard enthalpy ch-
AE" OW-all standard enthalpp change

- -
AH% atandard enthdm of s c t m t i r m
k, k' forward abd meme reactionvelocity constante
.
p~
p,
d&ty o f 7 4 at standard conditions
density of catdyst particle
-totalpreasura
IJTERA'NIIE CITED
k Boltamannconatant
h adsorption and desorption velocity constante of (1) Chilton. T. C.,and Colburn. A. P.. h.
--
eta. ENQ.C-., !23, 913
commnent A (leal).
4 therm& conductivitvof fluid
- -
(2) M.. 27. 266 (1935).
K overall fluid-phase k c t i o n equilibrium constant (a) G S ~ W .B. w.. TIWAS. G.. m d HOW, 0. A,. ~mtu.~ m .
R' surface &ion equilibrium oonstant m.et-. E W ~ . ,39, 14 (1~3).
KA.Ke,&etol+rption equilibrium conatants of eomponeots (4) Glsastons, 8.. Laidler, K.J., and Eyring, H., "Theory of Rate
L -- AandB
total molal h r p t i o n sitea/unit nuw of catalyst
Rocesaea". New York. McOnw-Hill Book Co., 1941.
(5) Oilliland. E. R..Im. ENQ.h. . (1984).
28.681
L'
0)

2 --totalactivecenterspunitsrsaofaurface
-Thielemodulua
a m molecular weight of reaatm feed
(6) Lewis. W.K.,and Chsng, K.C.. Tram. Ani. Inst. et-. Bnon..
21. 127 (1828).
(7) Natl. -rcb Couna 1% Rept. of Corn. on Catalmis. John

PA
PA'
Pd
- p""
-Avopdronnmber
psrtral ~RLWJE of component A in main fluid stream
p m of component A at
= OgmeaUOf(* - PA) and (w - PAC
wiley & eons, 1940.
(8) B h d . T. K.,"AbaomTon and Extrsotion". New York.
McQmw-Hill Book Co.,1937.
(8) Thiele, E. W.,h. EWQ. CHIOT., 31. 916 (1839).

(SOLID CATALYSTS and REACTION RATES)

Oxidation of Sulfur Dioxide 0. A. Uyehara and K . M . Watson

UNIVERSITY OF WWCDNSM. MADISON, WIS.

The A m of Lmia and Riu on the oridotion of Ideally, data should be obtained under flow conditions in a
~ I f u diocide
r in a fbu, system -r a plotin*ed di6ereuti.d reactor containing a bed of catalyst 80 shallow
SabeStOn Wm1r.t h o C e bssn O M h e d hO C e o d n C e that relatively d chsnges of composition occur. Since
with the principles outlined in the preceding h p e r a t u r e ha$ an important intluenca on reaction rate, it is
papsr SI#). The &m LVO -11 represented desirable that the di6erential section be kept at a temperature
by M equation Cmsd on the ornrmption thrt the ae uniform 88 possible. This rquirea a d l d i a m e t e r r e
rate arntrdling step is a mqfacs -tion between actor with good heat transfer at the walls.
.~tiartio&&ed nrlfw dbx*b ond atomic One of the most thomugbly invdgated catalytic mctions
ory(yen. he sxpsri-ml &tcr Shot0 m - t , h t is the oxidation of sulfur dioxide over a platinum catalyst.
bss decktion from this g-1 equation than The extensive experimental data ob- by Bodenatein
from the empiricnl ex#n
and Rim.
pmpowd by Let& and Ebk (9, Gets& (a,
and Taylor and Lenher (7)caunot
b e d for the development of general rate equations becaw
in each caee a static ayatem wae employed in which diffusional
ERIFICATION of the applicability of the rate eqw conmtmtion and temperature differenaas md&dt to pre-
tions developed in the preceding paper (pagx 529) is dict. Each of t h inmatigatom and Benton (1) proposed
'di9ioult becsuse of the b c s t complete lsak of suitable empirid exp&ona repreSemting their own data, but in dl
published data. Msny cataly%ic rate megwrements have eaea di&lsionsl&e& were nepjaded and the d h t qua-
been carried out in static s m in which large difiusionsl tiom m applicable only over limited rmgea of mn6itionn.
conmhtion di6ereum were inevitable and controlled by Lewisand Rim (6)carried out a aeriesof flow experimentsin
aonditioue of natural convection which m not subject to reactom reasonably approximating di6ereutid operation. A
~ ~ ~ e m l i anal*.
zed Of the data obtained in flow svstems. plstiniren aebeetos catalyst WBS used in &I! u#-tubelpsctors
k y m &table beoauae of the large mum-tratiod inuuetd in a molten metsl bath. Although dsts neeessry
c b a c e o m d n g ~~KIId- of convsrsion intended to for the evalwtionof diffusional concentration di6erenm are
not preamted, it secrms reasonable to ne&& then &e& be-
INDUSTRIAL AND ENGINEERING CHEMISTRT S4l