480 SABRI

ERGUN Vol. 60
phase. The sole volume effect encountered was pheres, are distributed throughout the continuous
adsorption, which removes ions. Diffuse ionic elastomeric matrix. At high concentrations with
atmospheres around the particles of the black would coarse blacks and a t low concentrations with fine
tend to give a constant loss tangent. blacks, chain formation appears which eventually
The general picture suggested by our results is leads to short circuiting of the samples and trans-
the following: The particles of the black behave forms them into conductors. Before this happens,
as if each contained a perfectly conducting graphite a strong increase in capacitance is observed which
core inside a semi-conducting sheath of much can be quantitatively interpreted by Bruggeman's
lower density, surrounded by a diffuse ion atmos- theory provided the dispersion remains uni-
phere. These particles, together with their atmos- form.

KINETICS OF THE REACTION OF CARBON DIOXIDE WITH CARBON

BY SABRI
ERGUN
Contribution of the Branch of CoaGto-Oil Research, Bureau of Mines, Region V , Bruceton, Pa.
Received Oclober 7 , 1066
Kinetics of the reaction of carbon with C02 has been investigated in the temperature range 7W14OO0 at atmospheric
pressures employing three carbons of different origin having different physical properties. The following reaction mechanism
has been derived: Certain carbon atoms, reaction sites, can detach an oxygen atom from a gaseous COZmolecule, reducing
COZto CO and form an occupied site. Reversely, gaseous CO can remove the oxygen atom from occupied sites to form
ki
C02: CO, + +
Ct J _ CO Co. These interactions constitute dynamic two-phase (solid-gas) oxidation-reduction and are
ki '
termed "oxygen-exchange reactions." The carbon transfer from solid phase to gas phase originates froin occupied sites:
kr
CO--f CO. This is the slowest step of the reactions and can be considered as unidirectional. The sorption of coz is not
a prerequisite step of the reaction. The reaction rate is proportional to the concentration of occupied sites. GO retards the
rate by reducing the concentration of occupied sites. The rate constants kl, kl' and ka are functions of temperature only
and are common to all carbons having trigonal (coplanar) bonds, regardless of origin, particle sire, porosity and cr stal-
linity. The differences in the reaction rates of carbons are due to specific number of reaction sites. I n studying the vaidity
of the above theory, data obtained in the present study have primarily been utilized. The theory has been found to be
applicable t o published data as well. For the first time i t has been demonstrated that carbons of different origin, particle
size, porosity and crystallinity can be treated on a common basis.

Introduction the over-all rates appreciably. An understanding

Carbon dioxide is both a reactant and a product of the possible influences of these phenomena on
of the gasification, combustion and various other gasification rates was considered essential. Studies
reactions of carbons. Its reactions with carbons on Aow conditions, diffusion and heat transfer
play important roles in numerous industrial opera- were undertaken; some of these problems have
tions, such as generation of water and producer been analyzed in previous ~ a p e r s . 1These
~ ~ studies
gases and smelting of ores. Carbon dioxide- were integral with the present analysis of the inter-
carbon reactions, therefore, have been the subject actions of carbon dioxide with carbon.
of many investigations. As a result, both empirical Theory
and theoretical treatments have been developed, Studies of sorption4 and surface reactions6-'0 with
some useful but all limited in applicability. carbon-oxygen systems have demonstrated the
Interest in the production of synthetic liquid ability of carbons to retain oxygen at certain sites
fuels has stimulated research on gasification of on their surfaces by chemical bonding.
coals. Synthesis gas constitutes a substantial On an initially cleaned carbon surface, carbon
fraction of the over-all cost of synthetic liquid fuels; dioxide is reduced t o carbon monoxide at tempera-
therefore, reduction in the cost of its production tures as low as 600". The oxygen lost by dioxide
is highly desirable. Work on improvement and remains on the carbon surface.6-'0-1* These studies
development of processes has been intensified.
(1) S. Ergun, Ind. Eng. Cham., 47, 2075 (1955).
However, basic solutions to the problem of gasi- (2) S. Ergun, Chsm. Eng. Progress. 48, 227 (1952).
fication can be given on fundamental grounds only, (3) 8. Ergun, Anal. Chem., 23, 151 (1951).
that is kinetics of the reaction, dynamics of flow, (4) H. H. Lowry and G. A. Hulett, J. A m . Chem. Soc., 42, 1408
diffusion and heat transfer. (1920).
( 5 ) T. F. E. Rhead and R. V. Wheeler, J. Chem. Soc., 101, 831
The reaction of carbon dioxide with carbon is (1912).
heterogeneous, and numerous problems are en- (6) I. Langmuir, J. A m , Chem. SOC.,37, 1139 (1915).
countered in attempting to explain its mechanism. (7) A. Eucken, Z . anusw. Chcm., 43, 986 (1930).
'The stoichiometric equation, COz +
C = 2C0,
does not reveal the mechanism of the reaction,
(8) L. Meyer, Z. phyeik. Chem., Bl7, 385 (1932).
(9) V. Sihvonen, Z . Elekfrochrm.,40, 456 (1934).
(10) W. E. J. Broom and M. W. Travem, Proc. Roy. Soc. (London),
which must be deduced from the over-all rates of AlS6, 512 (1932).
reaction under varying experimental conditions. (11) H. Martin and L. Meyer, Z . EEcklrochem., 41, 136 (1935).
The deduction is generally difficult and becomes (12) A. F. Semechkova and D. A. Frank-Kamenetsky, Acta Physio-
chim., U.R.S.S., 12, 899 (1940).
comdicated when ancillarv Dhenomena such as
Y . (13) J. Gadaby, F. J. Long, P. Sleightholm and K. W. Sykes,
flow conditions, heat transfer and diffusioninfluence proC.R O ~soc.. (London), A ~ W 357
, w~8).
April, 1956 KINETICSOF THE REACTION
OF CARBON DIOXIDE
WITH CARBON 481

demonstrate the ability of certain carbon atoms to they became fixed, leaving oxygen atoms to remove
detach an oxygen atom from a carbon dioxide other carbon atoms;'* however, such sites are be-
molecule. In a reverse manner, oxygen retained lieved to constitute not more than l or possibly 2%
on the surface can be removed by carbon mon- of the total number of reaction sites.'* Sorption of
oxide.I2-l6 These reactions may be expressed as CO in the presence of C 0 2 and subsequent removal
ki of the oxygen from the sorbed CO by gaseous CO
+
con Cf 1_ co co
k:
+ (1) would result in carbon transfer from the gas phase
and in carbon deposition, which were not detectable
where Cf represents a free site capable of reaction in experiments with Cl4O2.l7 Temperature co-
and Co an occupied site, i.e., a site possessing an efficients for sorption of inert gases, e.g., He or N,,
oxygen atom. Equation 1 expresses the oxygen- on carbon rule out their sorption a t gasification
exchange phenomenon, not the carbon transfer temperatures. Therefore, reaction sites can be
from solid to gas phase. Rate constants k~ and k: regarded as either free or occupied by an atom
are general rate constants, i.e., probability factors, oxygen
functions of temperature only. These constants +
(C,) = (Cd ( 4 ) (5)
have not been determined on an absolute basis; where (Ct) is the total number of reaction sites per
experimental evidence indicates that oxygen- gram.
exchange reactions are relatively fast. Under steady-state conditions the rate of forma-
The transfer of carbon from solid phase to gas tion of occupied sites would be equal to the rate of
phase originates from the occupied sites and may be their disappearance
expressed as d(Co)/dt ki(CO,)(Ct) - k:(CO)(Co) - ka(Co) = 0 (6)
Co-
k:
+
co n c f (2) where (COz) and (CO) represent the concentrations
where n is an integer having a value of 0, 1 or 2, of COz and CO in the gas phase, molecules per cubic
when occupied sites are considered individually. centimeter. Solving for (eb)from equations 5
When the reaction of a macroscopic carbon sample and 6
is considered, n can have any statistical value (Co) ki(COn)(CJ/[ki(COz) + +
kXCO) k:)l (7)
between 0 and 2; but the most likely value appears Substitution of equation 7 in (4) yields
to be 1. The rate constant k3 is a general rate
constant function of temperature only. The C w o k~(COdka(Ct)dW/[ki(COz) WCO) W1 ( 8 )
E + +
reverse of the above reaction and the reactions Equation 8 represents the rate of reaction of carbon
shown in equation 1 would result in carbon transfer by carbon dioxide in a general differential form.
from gas to solid phase and to carbon deposition At lower temperatures, oxygen-exchange re-
when the concentration of CO exceeds that re- actions occur a t temperatures approximately 200"
quired by thermodynamic equilibrium between CO, lower than required for gasification a t comparable
and CO. Experiments with C14 in the gas phase rates; it follows that
showed no detectable carbon transfer from gas to kl(C0J >>kl and/or kXCO) >> k: (9)
solid.l7 Moreover, carbon deposition on pure Since activation energies of oxygen-exchange re-
carbons a t gasification temperatures has not been actions are reported to be 43 to 60 k ~ a l . , '(com- ~
reported and could not be detected in the present pared to 59 kcal. for ka as determined in the present
studies. Reaction (2) can therefore be regarded as study) the inequalities (9) can be expected to be
unidirectional. true over a wide range of gasification temperatures.
The instantaneous rate of gasification of a carbon The conditions of the inequalities (9) are identical
sample surrounded by a gas having uniform com- with the attainment of a solid-gas heterogeneous
position may be expressed as equilibrium by the oxygen-exchange reactions
dn/dt 3 NO = kr(Co)W (3) Ki = ki/G e (CO)(Co)/(COz)(Cf) (10)
where dnldt or Nc is the rate of carbon transfer Equilibrium constant K I , defined above, does not
from solid phase t o gas phase, atoms per second, involve any parameter that is a function of physical
(CO) is the number of occupied sites per gram and properties of carbon. K1 can also be expected to
W is the weight of the carbon, gram. In flow be independent of pressure; it should be a function
experiments the gas surrounding the solid generally of temperature only.
is not uniform in composition. Equation 3 would From inequalities (9) it follows that ka in the
therefore hold true for a differential amount of solid denominators of equations 7 and 8 can be neglected.
only T his simplification and substitution of KI in place
d N o = ka(Co) dW (4) of kl/ki lead t o
Experimental evidence indicates that adsorption
of C02 by carbon is negligible above 600°.12 +
(CO) = ( C M 1 (CO)/K*(COn)l (11)
Some experiments have been reported in which COz and
molecules were able to enter into the surface where +
W O= h ( C 4 dW/U (CO)/KI(CO~I (12)
(14) A. Key, "Gas Research Board Commun.," London. No. Equation 12 represents the differential reaction
GRB 40 (1948). rate of carbon with COz. It can readily be applied
(15) J. D. F. Marsh, "Inst. Gas Engrs. Commun.." London, No. to 00w and closed systems. According to the
393 (1951).
(16) A. E. Reif, THIBJOURNAL, 66, 785 (1952).
equation the reaction rate is not determined by
(17) F. Bonner and J. Turkevich. J . Am. Chsm. SOL, 78, 501 (18) F. Brown, Trans. Faraday SOC..1005 (1952).
(1951). (19) A. A. Orning and E. Sterling. TAIBJOURNAL, 68, 1044 (1954).
482 ERGUN
SABRI VOl. 60

constant k3 of the slowest step alone, however small where N c o and Nco, represent the rates a t which
the constant may be, but also by the equilibrium CO and COa flow past the cross section considered,
constant K I of the oxygen-exchange reactions as N c the rate at which carbon is gasified between
well as by the composition of the gas phase. At the inlet and the section and (C,) the number of
low concentrations of CO and C02 the inequalities total reaction sites per gram. When no CO is
(9) cannot be expected to be valid. In these admitted a t the inlet, oxygen and carbon balances
instances equation 8 should be used in place of between the inlet and the differential section yields
(12). Both equations represent instantaneous
Nc = N;o2 - Ncot = (‘/z)Nco (14)
rates. Changes in concentrations of reaction sites,
if any, as carbon reacts are determined by n of where the superscript 0 denotes the conditions a t the
equation 2. Discussion of n will be made in a later inlet. Substitution of equation 14 in 13 and
study of reaction sites and their concentration as a dvvc = -dNcoZ gives
function of burn-off,
The validity of the above development can be
examined in several respects. According to equa-
tion 12 inert gas should have no influence on the
reaction rates. This fact has been established Direct integration of the second term of equation
experimentally in an earlier study.’ It is well 15 is permissible under two extreme flow conditions.
known that CO reduces the reaction rate of COz Gas flow and reaction along the path of the flow
with carbon. Some of the theories place the cause establishes a concentration gradient; diffusion on
of retardation on the accumulation of chemisorbed the other hand opposes the gradient and mixes the
02 on the surface,12 or on the adsorption of CO gases. The two extreme conditions in the gas
on reaction sites,18 while others explained it by phase are “no mixing” and “complete mixing.”l
equations similar to the reverse of equation l.l4-lS At fast flow rates back mixing is negligible; in-
The present study is in accord with the latter view; tegration of equation 15 without placing restriction
the equilibrium constant K1 furnishes a quantita- on the variation of NcoI1 substitution of X =
tive explanation of the role of CO. The validity N c o / N c o , = 2Nk3oJ”, - 2, and proper re-
of equation 12 can be tested experimentally with arrangement lead to
respect to the influence of CO. 2(1 + 2 / X ) In ( 1 + X / 2 ) - 2 =
Probably the severest test for the above develop-
ment lies in its treatment of all carbons with tri- -KI + Kiko(Cd(1 + 2 / X ) ( W / N ; o J (16)
gonal (coplanar) bonds on a common basis. For At low flow rates diffusion can mix the gases
example, K I should be independent of origin, poros- completely, ie., the composition gradient can dis-
ity, crystallinity, particle size, etc. Comparison of appear. Under these conditions N c o / N c o , would
the experimental values of K1 obtained with carbons be the same throughout the bed. Integration of
of different origin and physical properties should equation 15 then yields
serve as a severe test. Also, different carbons
should yield the same values of k3; however, the
X -Ki +
KMCt)(l +
2 / X ) ( W / N & ) (17)
experimental procedure adopted in the present When gases are not mixed, experimental da.ta
can be represented graphically by plotting the left
study permits determination of the product of k3
with (C,) rather than k3 alone. Nevertheless, side of equation 16 us. (1 +
2/X)(W/N&o,).
The plots should result in straight. lines as shown
activation energies of k3(Ct) can be determined and in Fig. la. The intercept on the abscissa corre-
the energy should be independent of the physical
properties of carbon. sponds to the value of l/k3(Ct) and the negative
According to the present theory, the differences intercept on the ordinate to the value of K1. For
in the rates a t which carbons react with COP are complete mixing of gases plots of the experimental
due to the differences in the specific number of data according to equation 17 should yield a straight
reaction sites (cj., equation 8 or 12). Specific line, the intercept of which on the abscissa should
number should, therefore, serve as a direct and correspond to l/k3(Ct) and on the ordinate to
absolute measure of reactivity. -KI. A typical set of experimental data for this
condition is shown in Fig. lb. Comparison of
Application of Theory equations 16 and 17 shows that their right sides are
identical. Because their left sides are not linear
The present experiments were conducted in iso- with each other over wide ranges of values of X ,
thermal fluidized beds. Integral equations there- except for values of X as X approaches zero,
fore have been derived for isothermal flow systems experimental data that yield a straight line on one
under steady-state conditions. The integral equa- type of plot should not yield a straight line on
tions can be used to test the validity of the above the other. Accordingly, such plots could be used
development and to determine the rate and equi- to ascertain which condition of flow prevailed.
librium constants. Equation 12 can be applied This test is illustrated in Fig. 2, where plot of
directly to a differential section perpendicular to equation 16 yielded a straight line and that of
flow of an isothermal reactor. Substituting the equation 17 a curve, indicating little back mixing.
ratio of the flow rate of CO to CO2 in place of their For partial mixing, plots of experimental data
respective concentrations and rearranging the according to both equations should yield curves,
equation as shown in Fig. 3. The values of k3(Ct)o1)tained
from Figs. 1-3 should be multiplied by 6.02 X
102s/2.24 x IO4 to convert, them to site/g. sen.
April, 1956 KINETICS
OF THE REACTION WITH CARBON
OF CARBON DIOXIDE 483

0.15-
/"
x Epuation 17(extensiva gas mixing1 /I -

: L
N .IO

0.05 lbl
I
' A
0
20 30 40 50
[ W I N E . ) ( I t Z / X ) , p.sec/cc.
,/' -KO
I I Fig. 3.-Reaction of COZwith an activated carbon a t 800"
-0.05cJ 50 100 I50 (partial gas mixing).

Fig. 1.-Reaction of COZwith an activated carbon: (a) at

equal to X / 2 . Therefore at low conversions plots
850O; (b) at 750". of data according to both equations should yield
straight lines, but the values of K 1 obtained would
differ by a factor of 2. The analysis of kinetics is
best accomplished from data obtained under one
of the extreme flow conditions. The flow condi-
tions prevailing a t the gas velocities employed can
be determined by preliminary experiments.
When CO is admitted at the inlet, equations 16
and 17 take the forms
(2 + X ) ( 2 + XO) l n -2- 2+= x
x - xo 2 + xo

The above equations can be used tc; account for

I5I- I Equotion 17 (extensive go5 mixing)

o Equation 16 ( negligible gon mixing ) /-

the influence of inlet CO.
Experimental Procedure
The experiments were conducted in small fluidized beds
under isothermal conditions to minimize the influence of
rates of thermal-energy transfer upon the over-all rates.
The reaction tube, 2.22 cm. inside diameter and 56 cm.
long was made of a high temperature porcelain. The tube
was held vertically in a furnace, the heated chamber of
which was a prismatic cavity of 11.5 X 11.5 X 23.0 cm.
The furnace was heated by four silicon carbide heating ele-
ments. Standard tapered Pyrex joints were fused to both
ends of the reaction tube to eliminate high temperature
seals and permit flexibility in operation. The temperature
was measured with a PtrPt +
10%Rh thermocouple placed
vertically in the tube from the to .
The position of the
thermocouple could be adjusted ra8ally and vertically. A
second thermocou le was inserted horizontally through a
hole in the side o?the furnace and pkced against the o u b
side wall of the reaction tube. The rates of flow of gaR
were measured by means of capillary flowmeters within a
standard error of =t0.2%.
0 0.4 OB I.2 Cl.osely screened and weighed amounts of carbons, rang-
(WINOm2)( It 2/)0,g.sec./cc. ing i n m e from 8-16 (1.8 mm.) to 170-200 mesh (0.081
mm.), IJ. S. Standard, were introduced from the top by re-
Fig. 2.-Renation of C 0 2 wit8h an activntetl cxrlmn at moving the up er half of the tapered joint. Product gaaes
1,025". 1)at.a are plot,t,ed according to 1iot.h cquxt.ion IG were passed tirough a capillary flowmeter and sampling
and 17. t.ubes. When the temperature of the bed and exit gas flow
484 SABRI
ERGUN Vol. 60
,
rate assumed constant levels, Le., steady state, gas Sam les employed: Ceylon graphite, an activated carbon and ac-
were taken for anal sis. The loss in weight of the carEon tivated graphite. The samples had little mineral content,
sample, Le., burn-og, usually was kept below 5%; the loss ranging from traces up to 0.5%. More than 360 individual
was calculated from the rate and composition of the effluent experiments were conducted at 50' intervals: At 700 to
gas and the time elapsed to reach steady state. Corrections 1,050' with activated carbon, at 950 to 1,150' with acti-
were made in the weight of the sample for the amount of vated graphite, and at 1,OOO to 1,400' with Ceylon graphite.
burn-off at the time of gas sampling. When a run was com- The precision of the experiments was estimated to be *6%.
pleted, the sample was removed by suction through a 6 mm. The data used in the calculations involve the weight of
porcelain tube inserted into the reaction tube. carbon, the flow rate of carbon dioxide at the inlet and the
The aerodynamic surface area and density of the carbon ratio of carbon monoxide to carbon dioxide at the exit,
samples were determined, as outlined previously,*." in order equations 16 and 17. Since the constants K,and ks( Ct) are
to estimate rates necessary for fluidization.' Approximate sufficient for a complete description of the results obtained,
rates of gasification were then obtained. Mixing in the the results of individual ex eriments are not included here.
gas phase was generally studied by conducting gasification The constants are listed in $able I.
experiments at successively increased rata of flow of carbon
dioxide and with increased amounts of carbon while a con- TABLE I
stant ratio of weight of carbon to rate of flow of carbon di- A N D RATECONSTANTS
EQUILIBRIUM OF THE REACTION
OF
oxide was maintained. The most precise experimental
values for the rate constants were obtained under conditions CARBON DIOXIDEWITH CARBON
of either negligible mixing or thorough mixing; however,
these conditions could not always be realized, owing to the Temp.,
Csrbon OC.
fact that experiments in fluidized beds can be carried out
only over a limited range of flow rates. This ran e lies Activated 700 0.025 0.13
between incidence of fluidization and bubble or slug forma- carbon 750 .050 0.41
tion and covers approximately a twofold flow rate. 800 .096 1.2
Final gasification runs were conducted at constant rates 850 .16 3.7
of flow of carbon dioxide with successively increased amounts
of carbon, for example, 1 to 12 g. Helium or nitrogen was 900 .27 11
used to flush the system and also to serve as a tracer in the 950 .40 34
experiments, that is, to check the material balance of the 1,000 .52 83
reaction. The rate of flow of inert gas was about one-tenth 1,025 .60 130
that of carbon dioxide.
T o determine the influence of inert gases, they were in- Activated 900 .25 0.72
troduced at increased rates while the flow rate of carbon graphite 950 .37 1.9
dioxide remained the same. As seen in Fig. 4 , helium had 1,000 * 54 5.1
no effect on the reaction rates when its addition did not alter
the flow pattern appreciably,' as expected from the present 1,050 .76 13
theory. 1,100 1.0 27
1,150 1.4 62
Ceylon 1,000 0.23 0.93
graphite 1,050 0.76 1.8
1,100 1.0 4.6
1,150 1.4 9.6
1,200 1.8 19
a
W
1,250 2.3 38
U 1,300 2.4 77
W
1,350 3.6 130
0"
U
1,400 4.5 230
0
IL

z
Discussion of Results
0 The experimentally determined values of equilib-
U
i
0
rium constant K 1of the oxygen-exchange reaction
IL
0
are plotted as a function of temperature in Fig. 5.
__ t I I Percent helium All three carbons are included in the plot. The
3 3 Percent helium equilibrium constant was found to be a function of
0

I 1 temperature aloneindependent of source, particle

size and other physical properties of the carbons.
This conclusion accords with the present theory.
( W / N ~ 0 2 ) g . sec / c c . (STP).
The heat of reaction is 23 kcal. per mole in the
temperature range of 800 to 1,400". Because of
Fig. 4.-Reaction of C02 with an activated carbon at its high temperature coefficient, the equilibrium
900'. Introduction of helium had no marked influence on has a pronounced effect on the rate of gasification.
the reaction rate.
For example, the values of the constants are 0.022,
The effect of carbon monoxide was investigated also by 0.52 and 4.4 a t 700, 1,000 and 1,400°, respectively.
introducing mixtures of carbon monoxide and carbon dioxide If, in the gas phase, the ratio of carbon monoxide
into the reactor. Data obtained were analyzed by equa-
tions 18 or 19. The same values of K1 and ka(Ct) were ob- to carbon dioxide is maintained a t 1, the fraction
tained as in the absence of carbon monoxide at the inlet. of the reaction sites oxidized would be 0.0215,
The result was not surprising, because carbon monoxide is a 0.34 and 0.81 a t the respective temperatures;
product of the reaction and any theoretical equation that that is, when the reaction temperature is raised
predicts the rates of gasification over wide ranges of carbon
monoxide formation must account for the influence of carbon from 700 to 1,000", the number of occupied sites
monoxide, whether i t is introduced at the inlet or is formed would increase by a factor of 16 and when raised to
as a reaction product. 1,400",by 38. If, on the other hand, the ratio of
Three carbons of different porosity and crystallinity were carbon monoxide to carbon dioxide is maintained a t
(20) S. Ergiin. Anal. Chem., 24, 388 (1952). 0.1 the corresponding increases would be by factors
April, 1956 KINETICSOF THE REACTION
OF CARBON
WITH CARBON
DIOXIDE 485

Fig. 6.-Constants of the gasification step as a function of

temperature. (Energy of activation is 59 kcal./mole.)
greatly different carbons range from 1 to 100,
1,0001T. they account for the differences in rates under
Fig. 5.-Equilibrium constant of the oxygen-exchange conditions of temperature (700 to 1,400’) and
reaction as a function of temperature. (Average heat of pressure (atrliospheric) that corresponds to actual
reaction between 800 and 1,400° is 23 kcal./mole.) gasification conditions. For charcoal and acti-
of 22 and 140. The gasification rate is proportional vated graphite, both of which are porous, the
to the number of occupied sites; thus, the effect relative number of reaction sites per gram was
of the equilibrium constant on the rates is through found to be independent of particle size in the size
its influence on its concentration of occupied sites. range of 0.08 to 1.8 mm. With the non-porous
The experimentally determined values of the Ceylon graphite, however, the number increased
product ka(Ct) are plotted in Fig. 6 as a function of with decrease in size.
temperature. The energy of activation is 59 kcal. In the above analysis the reaction was studied
per mole for all three carbons, which is possible in a statistical manner, that is, in terms of average
only if one of the following conditions is satisfied: properties during reaction and the constancy of k8
(1) The number of reaction sites is independent of and K1 for a wide range of pure carbons has been
temperature; (2) the number of sites changes with demonstrated experimentally. The treatment does
temperature exponentially, and all carbons have not preclude the possibility of the existence of more
the same temperature coefficient, regardless of the than one type of reaction site, for example, two,
range of temperature in which they react. The in which case K1and k3,as determined, would be
first condition most likely is the case. Like K1,k3 composite constants. The fact that these con-
therefore is a general rate constant for all carbons. stants were the same for carbons of different
The differences in the intercepts of the three lines crystallinity, porosity and particle size signifies
on the ordinate correspond to the relative numbers that the distribution of the types of the sites was
of reaction sites per gram. If (Ct) is taken as approximately the same. It signifies further an
unity for Ceylon graphite (80-100 mesh, U. S. identity of the energy and chemical nature of the
Standard), it is 6 for the activated graphite and reaction sites.
100 for the charcoal. As the same values of K1 Acknowledgment.-The author takes great pleas-
and ks apply to all carbons, the only individual ure in acknowledging the advice and constant
property affecting the rate of the reaction is the encouragement given by H. H. Lowry, former
number of reaction sites per unit weight of carbon. director of the Coal Research Laboratory, Car-
The concentration of reaction sites is therefore a negie Institute of Technology, in the course of this
direct and absolute measure of reactivity. Al- study. This paper is based on a Doctor’s Thesis
though the relative numbers of sites for these submitted to Technische Hochschuie, Vienna.