analysis of a single-stage fluidized bed reactor should be useful B = bed voidage

in a rational design of a continuous multistage reactor of a I.( = viscosity, centipoise, g/cm sec
large capacity. p = gas density, g/lit, g/cm3

Acknowledgment Literature Cited

Thanks are due to the Director, University Department of Athavale, A. S., Altekar, V. A,, Indian J . Technol., 7 (2), 52-54
Chemical Technology, Bombay-19, for his keen encourage- (1969).
Danckwerts, P. V., Jenkins, J. W., Place, G., Chem. Eng. Sci., 3,
ment during this work. 26 (1954).
Dunn, W. E., Trans. 1Met. SOC. AZME, 218, 6 (1960).
Nomenclature Feinman, J., Anner, W. D., AZChE J., 10 ( 5 ) , 652 (1964).
Feinman, J., Dexter, T., zbid., 7 (4), 584 (1961).
a, = initial analysis of iron oxide, g mol/g inerts Hougen, 0. A., Watson, K., Znd. Eng. Chem., 53, 529 (1943;I.
do = initial density of solids, g/cm3, g mol/cm3 Kellogg, H. H., Trans. M e t . SOC. A I M E , 188, 863 (1950).
E‘ = feed rate of solids, g inerts/hr cmZ,kg/hr m2 Ketteridge, I. B., Brzt. Chem. Eng., 10 ( 7 ) , 452 (1965).
Krishnamurthi, RI. S . , CSIR Symposium, on “Fluidization and
G = gas feed velocity, g mol/hr cm2 Related Processes,J’ Kharagpur, India, 1964, p 248.
k,? = rate constant’, weight loss of iron oxide/cm2, sec a t m Lakshmanan, C. AI., Hoelscher, 11. E., Chennakesavan, B.1
PHCl = partial pressure of HC1, a t m Chem. Eng. Sci., 20, 1107 (1965).
n = number of the stage Leva, Max, “Fluidization,” McGraw-Hill, New York, N. y.,
19.59, PD 80-98.
Q = average kinetic and geometric constant, g mol/g in- Levenspid, O., “Chemical Reartion E n_
r i n --....
w.r i~
n_,
r ”~
2nd pd Wiley,
erts hr
~.
New York. N. ?t‘., 1962, pp 338.
LU, w., TriAs. M e t . SOC. A I M E ,
---
s = a,(l - X B ) , g mol of iron oxide/g of inerts
V = bed inventory, g inerts/cm2 McKewan, W. M., ibid., 212, 791 I l Y J o j .
bleissner, H. P., Schora, 1F. C., ibid., 218. 12 11960)
TI’= weight, grams Nordin, A. W.. Brit. Chem. Ena.. 8 (4). 2:i7 i1963).’
zo = initial dimensions of t’he particle, em bok. W. Or.’ ATPhE’ J . , 12
XB= percentage conversion of iron oxide
y = gas efficiency, g mol/g mol Ostberg, G., Jerhontorets Ann., 140 ( l ) , 46 (1960).
Overholt, J. L., Vaux, G., Rodda, J. L., Amer. Mineral., 35, 117
Ye = mole fraction of the gaseous component a t time 8 (1930).
I’, = mole fraction of the gaseous component a t the bottom Themelis, N. J., Gauvin, W. H., Trans. Met. Soc. AZME, 227, 290
of the reactor volume (1963).
YT = mole fraction of the gaseous component at the top of !enz. F. A,. Othmer. D. F.. “Fluidization and Fluid Particle
the reactor volume
0 = time, see, hr
@ = shape factor, area of same volume of sphere/act’ual RECEIVED
for review September 21, 1970
area of particle ACCEPTEDApril 12, 1971

Application of a Kinetic Model for Catalytic Cracking

Effects of Charge Stocks

Donald M. Nace, Sterling E. Voltz,l and Vern W. Weekman, Jr.

J . 08066
Research Department, Illobi1 Research and Development Corp., Paulsboro,

A kinetic model for catalytic cracking has been applied to gas oils of widely different molecular composi-
tions and properties. The model adequately describes conversion and gasoline selectivities over wide ranges
of reaction conditions. Charge stocks with high concentrations of paraffins and/or naphthenes have the
highest rate constants for gas oil cracking and gasoline formation. Recycle and coker feedstocks have lower
rate constants than virgin gas oils. Catalyst deactivation increases with increasing concentration of aro-
matics in the charge stock. Coke formation i s related to the concentrations of high-molecular-weight aro-
matics.

T h e cat.alyt’ic cracking of wide boiling gas oils over acidic decay. I n t,his paper, the effects of gas oil composition on the
catalysts produces a broad spectrum of products as well as rate constants of important cracking reactions as well as
carbon deposits which cause rapid decay of the catalytic cat,al\-st decay are presented.
activity. The chemical composition of the gas oil has a pro- Most of the previous studies on the kinetics of catalytic
found effect, on the rates of cracking and rat,e of catalyst cracking were concerned only with the overall conversion
of gas oil to products (Aindrews,1959; Blanding, 1953; Voor-
TGwhom correspondence should be addressed. hies, 1945; Waterinan et al., 1960). The coupling of catalyst

530 Ind. Eng. Chem. Process Des. Develop., Vol. 10, No. 4, 1971
 Table I. Properties of Charge Stocks
Conrodson
ASTM distillation Nitrogen, Hydrogen, carbon Bromine
Charge stock 5% ' F 95% F Av mol wt Sulfur,wt% wt % wt % no. no.
P1 451 650 206 0.09 0.01 13.85 0.01 0.8
P2 642 939 402 0.01 0.05 12.17 0.03 0.4
P3 467 901 276 0.10 0.03 13.48 0.02 0.6
N1 443 687 228 0.03 0.001 12.83 0.01 0.0
32 657 908 354 0.13 0.03 12,61 0.04 0.7
N3 457 877 27 2 0.10 0.02 12.65 0.04 0.4
PN33 46 1 836 27 1 0.11 0.02 13.28 0.03 0.4
PA31 483 913 250 1.37 0.02 13.28 0.02 1.9
PA32 482 882 28 1 0.23 0.03 12.67 0.09 1.3
PA33 473 908 320 0.28 0.06 12.65 0.10 1.3
PA331 477 886 332 0.42 0.09 11.44 0.22 3.0
PA34 48 1 911 295 0.80 0.06 12.60 0.15 1 .o
PA37 460 855 278 0.87 0.06 11.08 0.07 4.0
PA38 465 959 325 1.78 0.09 12,51 0.73 ...
*\A45 723 892 366 3.77 0.14 10.87 0.19 4.0
PC32 467 911 298 0.59 0.09 13.02 0.10 4 2

decay kinetics with the conversion kinetics was report'ed by Experimental Methods
Weekniaii (1968) using a time-depeiideiit decay term. One
of the earliest treatnieiits of decaying reactor systems was a Preparation a n d Characterization of Charge Stocks. A
t,lieoretical study by Frotnent aiid I3isclioff (1961, 1962). series of 16 catalytic cracking charge st'ocks was prepared
Subwqueiitly, AIasaniune and Sniit'li (1966) and Sagara which covered wide ranges of properties and molecular com-
et, al. (1967) studied tlie cotisequeiices of diffusion on cat,alyst positions. Particular emphasis was placed on boiling poilit
decay. Ozawa and 13ischoff (1968) employed a linear decay ranges and relative concentrations of paraffins, tiapht~lieiies,
law based 011 carlion formation in describing the kinet,ics of and aromatics. The desired range of charge stocks was ob-
crarkiiig 011 hilien-alniniiia catalyst's. 13y regression t~ec4iniques, t:iiiied by bletidiny together selected fract,ioiis of different
Kberly et al. (1966) had previously developed a relationship refinery gas oil rtocks. Some of their properties and niolecular
for cc~keformation with respect to catalyst residence time and compositioiis are given in Tables I aiid 11.
sl~nc'e velocity. Again, using time-dependent decay t,ernis P I and P2 were highly paraffinic fractions with boiling
Weeknian (1969) and Weekman and S a c e (1970) studied the point ranges representative of light and heavy fuel oil frac-
cotisequeiices of decay 011 the selectivity of fixed, moving, t'ioiis (430-650" F aiid 650" F+, respectively). 1'3 was a mis-
anti fluid t,ed reactors for the catalytic cracking of gas oils. ture of equal volumes of PI and P 2 and was equivalent to a
Szepe a n d Levenhpiel (1968) ill a general study of catalyst highly paraffinic charge stock of wide boiliiig point range.
de:icti\-at,ioti showed that, a simple n t h order decay law de- Corresponding charge stocks of highly naphthenic conceiitra-
of catalyst fouling. Kojcieclioivski (1968), tions (Le., S I , K2, 1 3 ) were also prepared. 5 3 was composed
Campbell and Wojciechorvski (1969), aiid X o j ciechoivski of equal volumes of ?;I and S 2 . T o cover a range of paraffin/'
et al. (1969) reported a kinet8icmodel with a tinie-depeiident iiaphthene coiicentratioiis;, P S 3 3 was prepared froni equal
decay rat'e which mas applied t,o catalytic tracking. volumes of P 3 and X3. These latter three charge stocks repre-
I n inost, of the earlier ~ ~ o dealing
rk ivit'li tlie kinetics of sented wide cut gas oils of relatively low aromat,ics coiiceiitra-
catalytic cracking, 110 attenipts were made t.0 show the effects tioiis and similar boiling point ranges, but with coiiceiitratiotis
of charge stock composition. The "n-d-W' method of ideiiti- of paraffiiis and naplithenes of 10-46 wt Yo and 35-64 wt o/G,
fyiiig c w b o i i atoms in l~araffiiiic,iiaphtheiiic, and aromatic respectively.
st'ructures was u ~ 11). d Rrif et' al. (1961) to develop poly- Refinery aromatic extracts were used to 1)rei)are most of
nomial correlatioiis of yields. X o r e rereiitly White (1968a,b) the blended charge stocks with high coiiceiitrations of aro-
correlated catalytic cracking yields wit'll various classes of mat,ics. The properties and niolecular cotiipositioiis of the
parafhns, iiaphthenes, aiid aroniat,ics based on a inass spec- extracts are given in Table 111. They all contained more than
t'rogral)hic tecliiiique of Fitzgerald et al. (1970). Shiiaider 70 wt 70aromatics as determined by silica gel separation and
(1969) and Slinaider et' al. (1969) developed a kinetic scheme m a s spect,rometry. Most of the aromatic blended stocks were
for catalytic cracking and studied t'he effect's of chemical irepared by co~nbinitigP3 with the extracts listed in Table
compositions of charge stocks on crackitig rates in fluidized 111. P.131, PA32, P*133, PX34, I'A37, PA38, and I T 3 2 con-
reactors. hkhmetshina et. al. (1969) studied the effects of sisted of mixtures of 30 vol 70of the respective aromatic ex-
pure hydroca~~bons and chemical composition of some petro- t'ract added to P3. PA331 contained 60 vol yo of - 113added to
leum fractions 011 coke formation on an aluminosilicate cata- 1'3. ;3.145 was a mixture of equal volumes of A4 and A5.
lyst. Hydrocarbon type analysis of the charge stocks involved
ITsiiig smreral techniques to ideiitify paraffins, naphtlieiiex, distillation to separate boiling poilit, ranges;, liquid chroma-
atid aromatics, the present paper discusses the effect of these t'opraphy over silica gel, and mass spertroinetric analysis on
classes of rompounds O I L the rate coiistaiits of a kitietic model each fractioii. The 1)erceiitagesof paraffiiis (c/c l'), tiaphtheiies
for rat'alytic cracking. -1series of charge st.ocks with widely (70I), and aromatics (70A) were obtained from these data.
differeiit molecular compositions and other properties has Additional information based 011 a correlative technique
been ciiaracterized, and detailed kinetic cracking data have (n-d-JI method) which uses refractive index, density, rnolec-
beeti obtained for each charge st,ock. ular weight', aiid sulfur content provided the percentages of

Ind. Eng. Chem. Process Des. Develop., Vol. 10, No. 4, 1971 531
 Table II. Molecular Compositions of Charge Stocks

Mass spectroscopy n-d-M method ____

Charge stock Paraffins, wt % Naphthenes, wt % Aromatics, wt % cr, wt % c s , wt % CA Wt %
P1 51.9 33.7 14.4 66.5 24.7 8 8
P2 40.9 36.5 22.6 69.9 22.8 7 4
P3 46.2 35.1 18.6 66.7 25.0 8 2
N1 11.3 68.8 19.9 40.1 53.8 6 1
x2 8.6 59.4 32.4 48.4 40.7 10 9
N3 9.8 64.0 26.3 43.5 47.5 9 1
PN33 27.8 49.9 22.5 54.3 35.9 9 7
PA31 33.8 26.1 40.1 56.1 25.9 18 0
PA32 32.1 31.9 36.0 56.7 25.4 17 9
PA33 31.3 30.4 38.3 57.9 26.1 15 9
PA331 17.7 26.2 56.1 47.9 29 5 22 6
PA34 34.9 28,6 36 5 59.6 23.6 16 8
PA37 30.2 23.7 46.1 58.8 6.1 35 1
PA38 32.5 26.5 41.0 64.4 18.1 17 5
AA45 11.o 14.2 74.8 53.0 15.7 31 3
PC32a 37.3 29.5 30.0 60.6 24.4 15 0
a Contains 3.2YGolefins as determined by hydrogenation.

Table Ill. Properties a n d Molecular Compositions of Aromatic Gas Oils

Aromatics from virgin stocks
~ ___ Pre-processedu aromatics

Code ,41 A2 A3 A4 x5 h8 A7 c2
ASTM 570, O F 495 637 i54 756 740 868 7 14 626
Dist. 95%, O F 610 778 909 877 8 i7 994 832 924
Mol wt, esptl 202 285 363 364 359 406 257 330
RIolecular analysis, wt %
Aromatics 85.4 71.4 77.5 73 . O 76 7 85 6 95 6 53 8
Xaphthenes 7.1 25.5 21.1 15.2 13 2 9 4 3.2 27.4b
Paraffins 7.5 3.3 1.4 11.9 10 1 5 0 1 3 18 7
n-&M method, wt %
(2.4 37.8 31.3 31.4 32.7 30 7 34 6 88.0 30.6
20.9 36.4 33.5 16.6 13 7 11 2 ... 13.9
CP 41.3 32.2 35.1 50.7 55 6 54 2 ... 55.6
Analyses, wt %
H content 10 36 10 72 10 71 11 03 10 90 9 54 7 48 10 68
%S 3 74 0 51 0 53 1 89 4 34 4 94 2 25 1 78
%X 0 011 0 047 0 13 0 12 0 15 0 20 0 13 0 22
% C (Conradson) 0 02 0 22 0 29 0 21 0 08 1 6 0 14 0.22
Bromine no. 2 7 4 3 4 6 3 4 4 5 14 2 10.9

a From coking or recycle stocks. * Including 10.0% olefins as determined bv hydrogenation.

carbon atoms in the tot'al charge st'ock in ternis of paraffitiic Molecular coinpositio~i
carbons (% C,), napht'henic carbons (% CN),and aromatic Masq yiectrometiy
carbons (yoCA), By combining h with 70 C A , the per- Paraffiiif 8 6-51 9 n t 7 0
centage of aromatic molecules can be classified into aromatic Saphtheiies 14 2-68 8 l y t 7 0
ring carbon atoms and substituent carbon groups attached to lroniatics 14 4-74 8 lyt 70
aromatic rings (Le., yGA - 7, CA = aromatic substituent' n-d-N
groups). CP 40 1-69 9 Iyt 70
The ranges of t,he properties and molecular conipositioiis of CN 6 1-53 8 15-t 7
6
the 16 blended charge stocks are summarized as follo\ys: CA 6 1-35 1 Iyt 7 0
Average molecular weight 206-402
Sulfur 0.01-3.77 wt 7c Description of Fluidized Dense Bed Reactor and Ex-
Kitrogen 0.001-0.14 W t yo perimental Procedures. bench scale cracking unit
A\

Hydrogen 10.87-13.85 \\-t 7 0 with a stainlesh steel fluidized deiise bed (coiitained -110 re-
Conradson carbon 0.01-0.73 wt ye placement) react,or was used in this study. Gas oil wa.: vapor-
Bromine number 0.0-5.3 ized ill a separatc preheater; t,he oil vapor.: were bypassed

532 Ind. Eng. Chem. Process Des. Develop., Vol. 10, No. 4, 1971
around the reactor while the catalyst was fluidized with The rate constants calculated from the model are:
nitrogen. T h e switch from nitrogen fluidization to gas oil
cracking in the reactor was practically instantaneous. Gas a1 72 = catalyst decay constants
oil vapors were diluted with 10 mole 70iiit'rogen in the pre-
K O= K 1 + K 3 = overall rate constant for gas oil cracking
heater. Temperat.ures in the fluidized dense bed reactor were
K1 = rate constant for gasoline formation
measured b y t,hermocouples a t t,hree positions in the reactor
Kt = rate constant for gasoline cracking
and automatically recorded every 25 sec. Stripping of catalysts K3 = rate constant for gas oil cracking to A,
was i n i h t e d immediately a t t'he end of the cracking run KiIKo = initial select,ivity ratio for gasoline forma-
t,ioii
with 900" F steam aiid cont,inued for 10 min. X commer-
cial FCC zeolite c a h l y s t mas used. KtIKo = ratio of gasoline cracking to gas oil cracking
The on-stream times (catalyst residence times) of oil Weekman (1968) and Weekmaii and S a c e (1970) derived
through the reactor were 1.25 aiid 5.0 min. The weight rat,io integrated rate equations for both catalyst tleca~-fuinctioiis in
of catalyst/total oil charged was varied from 0.5 to 7 by different types of reactors. In this work, the equations for a
changing catalyst volume in t,he dense bed andlor the feed fixed bed were used to calculate the rat'? const'aiits from the
rate of gas oil. Since 110 catalyst was added during a run, the cracking data of t'he contained fluidized dense bed reactor.
activity, aiid thus t,he cracking rates, decreased with on- The integrated rate equatioiir: which were derived for fiserl
stream time. beds, are equally applicable to this type of fluidized reactor
Gaseous products included a light gas which passed through assuming plug flow iii the gas phase. The catalyst, activity in
traps a t -40' and -110' F and a heavy gas produced by the latter reactor decays uniformly throughout the elitire
weathering all the condensed liquid products at 130" F. catalyst bed during a particular run hecause of the high rate
Distillations of the liquid products were conducted in a of catalyst mixing. The equations for conversion are suimma-
microstill with 10 in. of Cannon packing operated a t a con- rized below:
stant reflux ratio of 3 : I . T h e gasoline end point was 430" F. First-Order Decay
Cat8alystcoke was determined by burning a sample of 15-25
grams of coked cat'alyst in oxygen and analyzing the combus-
tion gas for carbon dioxide by Orsat gas analysis. T h e products
were collected over each entire experiment aiid represented
time-averaged products.

Kinetic Model

A kinetic model for the catalytic cracking of gas oils was (time-averaged)
developed b y Weekman (1968) and Weekman and Xace
(1970). This model can be represented schematically as:
KI t," Decay
-4,+ A t
Kz\ dKa
A3

iil is the weight fraction of unreacted gas oil, A z is the weight

fraction of gasoline, aiitl -42represents the weight fraction of
all other products (i.e., C1-C4, coke). Gas oil aiid gasoline
cracking are assumed to be second and first order, respec-
tively. The justifications for these reaction orders have been e
discussed previously b y Weekman (1968, 1969). The dif-
w
ferential rate equations can be w i t t e n as,
where t aiid :are, respectively, iiistantaneouq and time-aver-
aged conversions and S = space velocity. Equations 6 mid 7
give the conversioiis for @ = e and Equations 8 and 9 give
those for @ = t , - n . Equation 9 is normalized with respect to
catalyst residence time and t , represents t'he total on-stream
t,ime.
Instaiitaneous gasoline yields are calculated h y the follon--
-43 = 1 - A1 - A ? (3) ing equation which can be derived froin Equatioiis 1 and 2:
where to = oil resideiice time aiid @ iq a catalyst decay func-
tion. Various function. have been used to describe catalyst
deactivation during cracking reactions. Weekman has uqed
the following two:
@ = e-a''
(-2)

where

in which t , = catalyd residence time and cy a i d n are rate

constants for catalyst decay.

Ind. Eng. Chem. Process Des. Develop., Vol. 10, No. 4, 1971 533
 ~~

Table IV. Kinetic Rate Constants at 900" F

Charge stock cy KO Ki KP KdKo KdKa n KO"
P1 24.8 31.8 26.3 1.83 0.83 0.06 0.58 1.52
P2 29.9 32.7 26.2 1.09 0.80 0.03 0.71 0.82
P3 30.5 34.0 28.0 1.86 0.82 0.05 0.72 0.79
N1 18.5 39.2 33.5 1.54 0.85 0.04 0.42 0.42
N2 28.7 34.2 29.4 2.35 0.86 0.07 0.68 0.99
N3 25.5 36.2 31 . O 2.02 0.85 0.06 0.60 1.57
PN33 27.4 36.4 30.9 1.94 0.85 0.05 0.64 1.26
PA31 31.4 24.7 21.0 2.87 0.85 0.12 0.72 0.56
PA32 31.6 22.9 18.6 1.95 0.81 0.09 0.71 0.53
PA33 33.9 22.1 17.6 1.48 0.80 0.07 0.77 0.38
PA331 36.4 15.5 12.6 2.66 0.81 0.17 0.78 0.23
PA34 34.4 22.1 17.8 1.78 0.81 0.08 0.78 0.37
PA37 37.7 10.3 7.71 2.18 0.75 0.21 0.77 0.16
PA38 34.7 21.1 16.6 1.66 0 79 0.08 0.78 0.34
Ail45 40.1 12.3 9.30 2.28 0.76 0.19 0.83 0.14
PC32 31.6 19.3 15.0 1.15 0.78 0.06 0.71 0.44
a Rate constant for gas oil cracking associated with n decay constant.

1.0 Results and Discussion

EXPERIMENTAL CATALYST
RESIDENCE TIMES
Kinetic Rate Constants. R a t e constants were calcu-
0.8 0 - 1.25 MINUTES - lated for each charge stock from the cracking data obtained in
z *\* 'o-,
- 5.0 MINUTES

s 1 the fluidized dense bed reactor Equations 7 and 13 were used

Y 0.8 \ n
. ,
! to calculate a , KO, K1,K 2 .The values of n and KO (associated

1 *%*J
.I with n) were calculated from Equation 9. i i n improved version
: >Lo-\-- of the technique for least-squares estimation of nonlinear

1 p--,
0.4
s parameters, developed by Marquardt (1963), was used to
b
Y

5 0.2
I
,_L 1 obtain the best fit of the rate constants to the data.
The rate constants a t 900' F are given in Table IV and their
0
V ranges are shown below:
I
0.0 I I I I I I I I I
0 5 IO 15 20 25 30 35 40 Rate constant Range at 900' F
SPACE V E L O C I T Y , WT/(WT)(HR)
18.5-40.1
Figure 1. Conversion vs. space velocity for charge stock 10.3-39.2
P3 7,7-33.5
- Equation 7, rate constants Table IV 0.97-2.87
0.75-0,86
0.03-0.21
Equations 6 and 8 can be used with Equation 10 to obtaiii the
iiistmantaiieous
gasoline yields since, These ranges represent the kiiietic behavior of charge stocks of
widely different properties and molecular compositions. Some
E=l-Al (12) correlations between the rate constants and charge stock com-
Similarly the time-averaged gasoline yields are calculated with positions are immediately obvious. Catalyst deactivation (a)
Equations 6 anti 8 with Equation 10 as, is highest for charge stocks with the highest concentrations of
1
aromatics. The rate constants for gas oil cracking (KO) and
Xp = A2de (13) gasoline formation (Kl) increase with increasing concentra-
tions of paraffins and naphthenes. It is somewhat surprising
that the ratio of K1/Ko which is a measure of the initial
selectivity for gasoline formation varied only from 0.75-
E X P E R I M E N T A L CATALYST
RESIDENCE T I M E S 0.86 over the entire range of charge stocks. The ratio of gaso-
0 - 1.25 M I N U T E S line cracking to gas oil cracking, K 2 / K o ,changed by a factor
- 5.0 MINUTES
of seven over the same range of charge stocks.
L
The values of n. (from tCn decay law) and associated KO are
also given in Table IV. As expected, the correspondence be-
tween CY and n is reasonably good. The general trends of the
relative values of K Ofrom both decay functions are similar;
some discrepancies may exist for specific charge stocks.
Paraffinic Charge Stocks. P1 has a lower value of CY
than eit,her P2 or P 3 which is probably relat,ed to its lower
boiling range. P2 aiid P 3 have similar decay rates. The small
differences in K O and K l for these three charge stocks are
CONVERSION, WT. FRACTION
probably not significant.
Figure 2. Gasoline selectivity for charge stock P3 h plot of time-averaged conversioll vs. space velocity for
- Equations 10, 13, rote constants Table IV P 3 is shown in Figure 1. The time-averaged conversion a t the

534 Ind. Eng. Chem. Process Des. Develop., Vol. 10, No. 4, 1971
 1.0 ,_ I I 1 E X P E R I M E N T A L CATALYST
RESIDENCE T I M E S

J
EXP ER I M ENTAL CATALYST
-
RESIDENCE TIMES
'\a 1 1 0 1.25 M I N U T E S -
I 1 0
0
-- 1.25 M I N U T E S
5.0 MINUTES
?
U
4 0.6
\*
\..
I
1 0 .
0 -5.0

I
MINUTES

0-$ 0.4
I
~ X ? O - \

n
d
W
g
> 0.2 b-
U I

I
i I
0.0
0.0 1
0 5
I
IO
I
15
I ,
20 25
I I
30
I
35
I
40
SPACE VELOCITY, WT/(WT)(HR)

Figure 3. Conversion vs. space velocity for charge stock N3

- Equation 7, rate constants Table IV

cat,alyst residence time of 1.25 min is considerably higher than stant,s in Table IV, is quit,e good. Similar plots for S1 and 9 2
that, a t 5.0 min, as mould be expected. The points are the ex- show the same agreement between the esperimental and com-
perimental data, and the solid lines are the values computed puted results and verify that the coilversions are in the order
.with the kinetic model with first-order decay from the rate N 1 > N3 > N2.
constants in Table IV. The agreement between the experi- The select,ivity plot of time-averaged gasoline yield vs.
mental and computed results is very good. The corresponding time-averaged conversion is shown in Figure 4 for 5 3 . From
plots for P I and P2 are not shown, but they are quite similar similar plot,s for SI and N2, it can be observed that the m a l l
to Figure 1 except t h a t the conversion for P2 is slight.ly higher, differences in select'ivity are in the same order as coiiveriioii,
particularly a t 1.25 min. Le., X l > .V3 > 5 2 . The values of K 1are i i i the sanie ordcr and
The values of K ffor P1 and P3 are about the same. P2 has those of K 2are the reverse order.
a relat,ively low value which is somewhat unexpected. h plot iis mentioned previously, P S 3 3 was a ldeiid of equal
of time-averaged gasoline yield vs. time-averaged conversion volumes of P3 and K3. Plots of time-ayeraged conver.s1011 ' vs.
is shown in Figure 2 for P3. Again the points are the experi- space velocity and time-averaged gasoline yield \-s. time-
mental dat.a, and the solid lines are the computed values. The averaged conversion are shown in Figures 5 and 6. P3, 5 3 , and
time-averaged gasoline selectivity a t 1.25 min is higher than PN33 have similar boiling ranges and about the same concen-
that at, 5.0 min. The agreement between the experimental and trations of aromatics, nitrogen, and sulfur. The ratios of
computed result,s a t 5.0 min is quite good. The computer curve naphthenes/paraffins are significantly different in the t'hree
for 1.25 mill is slightly higher than the experimental points. charge stmocksComparison of t'he plots of time-averaged
The corresponding selectivity plots for P1 are about the same conversion vs. space velocit'y indicates slight differences in
as those in Figure 2. P2 has slightly higher gasoline selectivi- cracking rates (Le., 5 3 > P S 3 3 > P3) which is the same order
ties which are largely owing to the low rate of gasoline as the ratio of naphtheiies/paraffiiiF. These slight differences
cracking (low K 2 ) . are also reflected in the rate constants in Table ITT. P3 has the
Naphthenic Charge Stocks. T h e values of CY for t h e highest, catalyst decay constant ( a ) and t'he lowest rate con-
naphthenic charge stocks are in the order N1 < N3 < S 2 ; stant for gas oil cracking (&). 5 3 and P S 3 3 have similar
K O is the reverse order. These results suggest that the con- values of KO,but S 3 has a lower a. The gasoline selectivity
versions would be in the order of 1\;1 > K3 > X2. h plot of plots of K3 and PK33 are about the same which agrees with
time-averaged conversion vs. space velocity for N3 is showi in the closeness of their values of K l and K?.The gasoline selec-
Figure 3. The agreement between the esperimental points tivity of P 3 is slightly lower.
and the solid lines, which were computed from the rate con- Aromatic Charge Stocks. Efects of droniatic Concentra-

1.0,
EX PER IM E NTA L CATALYST
RESIDENCE T I M E S
7
1.0,
1 E X P E R I M E N T A L CATALYST
RESIDENCE T I M E S , I
4
L
Y
f
; 0.8

L
LL
0 -
-
1.25 M I N U T E S
5.0 M I N U T E S

, I I

$
0.6
: o.6
I

n
i
n
_1

! 0.4 ! 0.4 I
**
W
f"
5
0 0.2
n
<

0.0 1 I 0 . o v 1 1 I
0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0
CONVERSION, WT. FRACTION CONVERSION, WT. F R A C T I O N

Figure 4. Gasoline selectivity for charge stock N3 Figure 6. Gasoline selectivity for charge stock PN33
- Equations 10, 13, rate constants Table IV - Equations 10, 13, rate constants Table IV

Ind. Eng. Chem. Process Des. Develop., Vol. 10, No. 4, 1 9 7 1 535
 I ,o tivities is P3 > PA33 > PA331 > AA45; the values of K 1 are
in the same order.
Effects of Boiling Point Range of Aromatics. series of
charge st'ocks was prepared by blending P 3 with five aromatic
extracts of boiling ranges as esemplified by the 5700--95%
0.8 ASTM boiling points (Al, 495-610" F; h 2 , 636-788" F ; -13,
754-909" F; -24, 756-877" F; d8, 868-994" F). PA31, P-132,
z Pr133, PA34, and PA38 coiit,aiiied approximately 30 vol yo of
2
c
V
the respective aromatic estract. As shown in Tables I and 11,
4 the five blended charge stocks have reasonably similar boiling
0.6
ranges and concentrations of paraffins, naphthenes, and
s aromatics. Their nitrogen coiicentrat,ions are quite low al-
though PA31 and PA38 have much higher sulfur coiicentra-
z
-0ln tioiis than the other blends.
," 0 . 4 Plots of time-averaged conversion vs. space velocity and
> time-averaged gasoline yield vs. time-averaged conversion
z
0 for PA31, PA32, PA33, PAA34,and PA38 are similar. It ap-
V
pears that the boiling ranges of the aromatic extracts have
no major effect 011 conversion or gasoline selectivity. The
0.2 actual differences in gasoline selectivities were also quite
small. Both K 1and KP decrease slightly with increasing boiling
range of t,he aromatic fractions. The additions of the aromatic
fractions to P3 significantly lower both the conversion and
gasoline selectivity. P3 has a higher K o and lower cr than the
0 five aromatic charge stocks.
0 IO 20 30
SPACE VELOCITY, WT/ ( WT)(HR) Recycle and Coker Stocks. T h e charge stocks described
previously were all formulated from virgin fractions. They
Figure 7. Comparison of conversions for different charge had iiot been subject'ed to any chemical reactions prior to
stocks
- being catalytically cracked in the fluidized dense bed reactor.
Equation 7, rate constants Table IV
Recycle stocks are significantly different than niost virgin
gas oils with respect to molecular compositions. This is to be
tions. The four charge stoc,ks, P3, Pr133, PA331, and AA45, espected, since the more readily cracked compounds such as
represented a wide range of aromatic concentrations of 19-75 paraffiiis, naphthenes, and alkyl- or naphthyl-aromat'ics have
wt yo based on analysis by mass spectrometry. The nitrogen undergone reaction during the first pass through the catalytic
cont,ent of hA45 was only slightly higher than the ot,hers; its cracking uiiit. Recycle stocks usually contain very high con-
sulfur cont,ent was considerably higher. A limited amount of ceiitrat,ioiisof refractory aromatic coinpounds.
cracking data from charge st'ocks of similar molecular coni- A blended recycle charge stock, PX37, was prepared by
positions and sulfur concentrations of 0.3-1.6 wt yo indicated adding 30 vol % of A7 to P3. Aromatic extract, '27, was a n
that, sulfur had no significant effects on t,he cracking rates. extract from recycle and coker stocks. P1ot.s of time-averaged
Plots of time-averaged conversion vs. space velocity for conversion vs. space velocity aiid time-averaged gasoline
these four charge stocks at a catalyst residelice time of 1.25 yield vs. time-averaged conversion are shown in Figures 9 aiid
mill are shown in Figure 7 . The differences in tile conversioiis 10. The agreement between computed and experimental re-
are quite large; the order of cracking is P 3 > PA33 > PA331 > sults is reasonably good in both plots. Thus, t,he kinetic model
AA45. This is also reflected in the rate constants; a increases is capable of fitting cracking data from both virgin and recycle
and K Odecreases in the same order. Figure 8 shows the time- charge stocks. The conversion of PA37 is considerably less
averaged gasoline selectivities a t 1.25 min. The order of selec- t.han that of eveii XAA45;the latter contains a much higher con-

1.0 1.0 I I
EXPERIMENTAL DATA
E X P E R I M E N T A L CATALYST
RESIDENCE TIMES

z 0.8 0 - l,25 MINUTES -

2 0 - 5.0 MINUTES

E
gO.6 1.25 M I N U T E S '
,I
~ A T A L Y S TR E S I D E N C E T I M E
U

L
5
0.0
I
4 z
0.4 2
z
W U
-J W
>
0 z
0
2 0.2
0

0 1 I I 1 I
0 0.2 0.4 0.6 0.8 1.0 0 5 10 I5 20 25 30 35 40
C O N V E R S I O N , WT FRACTION SPACE V E L O C I T Y , WT/(WT)(HR)

Figure 8. Comparison of gasoline selectivities for different Figure 9. Conversion vs. space velocity for charge stock
charge stocks PA37
- Equations 10, 13, rate constants Table IV - Equation 7, rate constants Table IV

536 Ind. Eng. Chem. Process Des. Develop., Vol. 10, No. 4, 1 9 7 1
 1.01 I
Comparison of the data for P3, PX33, Pd331, and ,$A45
EXPER I M E N T A L CATALYST
RESIDENCE T I M E S shows that the constant a increases sharply with iucreasing
; 0.8 0 - 1.25 MINUTES concentration of aromatics in the charge stocks. Coilstant, b is
# - 5.0 MINUTES
relatively insensitive to differences iii aromatics conceiitra-
tions.
The value of a increases with increasing boiling range of
d l l l l l the aromatic extract for charge stocks PA31, l'A32, 1'2133,
PA34, and PA-i38.The value of b is about the same for these
five charge stocks.
T h e recycle and coker stocks, PA37 and PC32, have rela-
tively high values of a,as would be expected. Again b is rela-
tively unaffected by the cornposit,ion of the charge stocks.
0.0
0.0 0.2 0.4 0.6 0.8 1.0
C O N V E R S I O N , W T . FRACTION
Summary and Conclusions
Figure 10. Gasoline selectivity for charge stock PA37
A kiiietic model for catalytic cracking has heel1 applied
- Equations 10, 13, rate constants Table IV
to charge stocks of extremely widely different inolecular com-
positions and properties. The rnodel quantitatively descrilies
the conversions and gasoline selectivities. Charge st,oc
centrat'ion of aromatics. PA37 has a relatively high a aiid a
high concentrations of parafins and/or Iiaphthciies have the
very low KO.The gasoline selectivity of 1'-137 is also lower
largest rate constants for gas oil crackiiig a i d gasoline forrna-
than A.145. Gasoline cracking appears to be relatively high
tion. Catalyst deactivation increases with increasing con-
with PA37; the ratio of KliZ<Dis 0.21 which is the higliest
ceiitrat.ioii of aromatics in the charge st.ocks.
value of all the charge stocks studied.
PC32 was composed of 30 vol yoof C2 and 70 vol 76 of P3. The molecular compositions of recycle and coker st,ocks are
significantly different from those of virgin pas oils. T h e formcr
C2 was a heavy coker stock. Rate constants given iu Table
IV show that coiiversion arid gasoline selectivity of PC32 cotitailis higher conceiitrations of refractory aromatic com-
are also low compared to those of t,he virgin charge stocks. pouiids which have lower cracking rates. The kiiietic model
can also be used to describe the catalytic crackiiig of these
Coke Deposition. Coke deposit'ion on cracking catalyst's
charge stocks.
is priinarily dependent on catalyst residence time and is
Some correlations have been observed for thr ~ 0 0 1 . h i w
iiormally observctl to be practically iiidependeiit of space
cwist8antsfor coke tleposit,ion and charge stock cornpositiou.
velocity. The Voorhies equation, C = at: (where C = cataly$t
coke, t, = catalyst residence time, and a aiid b are constants),
can be used to quantit,atively describe coke formation. The Acknowledgment
values of a and b were estimated from p1ot.s of log C vs. log The authors are indebted t,o Benjamiii Gross, Solomon 11.
t , (min) and are given in Table V for each charge stock. Jacob, and ITooyouiig Lee for helpful comnients and suggcs-
T h e constant a is usually a function of charge stock propertics t,ions, ant1 to Xleksarider 1 2 . Topoliiicki for assirt.aiice in the
and is a measure of the cokiiig tendeiicy of a part,icular charge computer calculations.
stock. The order of constant a is P 2 > P 3 > PI and K2 >
N3 > N1 which is the order of increased concentrations of high-
Nomenclature
molecular-weight aromatics. The values of a are in the same
order as constaiit, a for these charge stocks. Interestingly, the wt fraction unreacted gas oil
=
constant b increases in the reverse order of constant a. wt fractiou gasoline
=
Aa = wt fraction C1-C4 and coke
~.
A 1 = time averaged yield of A
a = coking constant, (niin-b (wt fraction c)
Table V. Voorhies Constants for Charge Stocksa b = coking exponent
C = coke-on-catalyst, wt yo
c = at? Ki = gasoline formatioii rate constaiit, (hr-l)
Charge stack a b (wt fract'ion gas oil in charge)-'
K2 = gasoline cracking rate constant, (hr-l)
PI 0 22 0.30 (wt fract'ioii gas oil iii charge)-'
P2 0 46 0.12 K, = -1,formation from A1rate coilstant, (hr-l)
1'3 0 41 0.16 (wt fraction gas oil in charge) -l
N1 0 23 0.27 K O = K1 + KR= overall gas oil cracking rate constaiit,
x2 0 46 0.18 (hr-l) (wt fraction gas oil in charge)-'
?i3 0 36 0.23 n = esponent in t," decay, dimensioiiless
PN33 0 32 0.24 S = Weight hourly space velocity, (wt gas oil)/
PA31 0 42 0.21 (wt cat) (hr)
PA32 0 47 0.21 = catalyst residence t'ime, hr
PA33 0 54 0 24 = catalyst residence time a t end of fixed bed
PA34 0 51 0.24 run, hr
PA38 0 64 0.21
P.1331 0 69 0.17 GREEKLICTTIZRS
,u45 0 75 0.20 a = decay velocity constant, hr
PA37 0 78 0.21 E = instantaneous wt fraction converted
PC32 0 52 0.22 E = time averaged wt fraction converted
Note t , in mirnite5 for coking data 0 = normalized time-oil-stream, t , / t ,
@? = catalyst decay fuiiction

Ind. Eng. Chem. Process Des. Develop., Vol. 10, No. 4, 1971 537
 literature Cited Shnaider, G. S., Khim. Tekhnol. Topk. Masel, 1969 (5), 5.
Akhmetshina, 11. N., Levinter, hI. E., Tanatarov, M. A., Shnaider, G. S., Mukhin, I. I., Chueva, M. A., Kogan, Y. S.,
Nochalov, Y. D., Seftepererab. Neftekhim. (hIoscow), 1969 ibid., 1969 ( l ) ,10.
Szepe, S., Levenspiel, O., European Federation of Chemical
(81, 10. Engineers, Symposium on Reaction Engineering, Brussels,
Andrew, J. AI., Ind. Eng. Chem., 51, 507 (1959). Se tember 1968.
Blanding, F. H., ibid., 45, 1186 (1953). Vooriies, A., Ind. Eng. Chem., 37, 318 (1945).
Campbell, D. It., Wojciechowski, B. W., Can. J . Chem. Eng., 47, Waterman, H. I., Boelhowuer, C., Huibers, D. T. A., “Process
413 (1969). Characterization,” Elsevier, Amsterdam, Netherlands, 1960.
Eberly, P. E., Kimberlin, C. N., Miller, W.H., Drushel, H. V., Weekman, V. W.,Ind. Eng. Chem. Process Des. Develop., 7, 90
I n d . Eng. Cheni. Process Des. Develop., 5 , 193 (1966). (1968).
Fitzgerald, 31. E., hIoirano, J. L., Morgan, H., Cirillo, V. A,, 5’. W.. ibid.. 8- z. 38R
Appl. Spectrosc., 24, 106 (1970). Weekman.
.- .
...-
. I - - 11969).
\ - -
Froment, G. F., Bischoff, K. B., Chem. Eng. Sci., 16, 189 (1961). Weekman: 1’.W.: ru’ace, D . M., AIChE J., 16,397 (1970).
Froment, G. F., Bischoff, K. B., ibid., 17, 105 (1962). White, P. J., Hydrocarbon Process. Petrol. Refiner, 47,103 (1968a).
Narquardt, D. W., J . SOC.Appl. Math., 11,431 (1963). White. P. J.. PreDrint No. 24-68, API Division of Refining, - 33rd
Masamune, S., Smith, J. AI., AIChE J., 12, 384 (1966). Mid-Year Megting, Philadelphia, May 1968b.
Ozawa, Y., Bischoff, K. B., Ind. Eng. Chem. Process Des. Develop., Wojciechowski, B. W., Can. J . Chem. Eng., 46, 48 (1968).
7, 73 (1968).
Wojciechowski, B. W., Juusola, J. A,, Downie, J., ibid., 47, 338
( 1969).
Reif. H. E.. Kress. R. F.. Smith, J. S..Petrol. Refiner, 40 ( 5 ) 237
(1961). ’
Sagara, Y., LIasamune, S., Smith, J. 11.,AIChE J., 13, 1226 RECEIVED
for review October 29, 1970
(1967). ACCEPTEDMay 3, 1971

Application of a Kinetic Model for Catalytic Cracking

Some Correlations of Rate Constants

Sterling E. Voltz,l Donald M. Nace, and Vern W. Weekman, Jr.

Research Department, X o b i l Research and Development Corp., Paulsboro, N . J . 08066

The rate constants from a kinetic model for catalytic cracking have been correlated with the molecular com-
positions of a wide variety of gas oils. The catalyst decay constant and rate constants for gas oil cracking
and gasoline formation can b e quantitatively related to the ratio of aromatics/naphthenes in virgin gas oils.
The corresponding rate constants for recycle and coker feedstocks deviate from the correlations. Some re-
lationships between the rate constants and coke formation have also been observed.

N u m e r o u s studies have been made of the reaction mech- unsaturates. In the same year, Service (1960) discussed t h e
misins and kinetics of pure hydrocarbons over acidic cracking detailed charact,erization of feedstocks for catalytic cracking
RIost of the results are consistent wit,h the concept including the various correlation methods to determine carbon
that catalytic cracking occurs primarily through carbonjuni atom distributions. He also recognized the utility of mass
ioii ~nechaiiisnis.The catalyt,ic cracking of gas oils is compli- spect,ronietry to elucidate classes of hydrocarbons in cracking
cated by t’he lireseiice of thousands of different molecules; feedstocks. Reif et al. (1961) developed correlation equations
eac811 ttype has specific rates of adsorption and reaction. For- to predict, yields and properties based on the det.ailed charac-
t~unatrly,many of the catalytic reactions of hydrocarbons can terization of a wide variety of cracking feedstocks and cracking
lie described both mechanist’ically and kinetically in terms of data from a pilot plant unit. Their equations were based to a
the behavior of cert,ain classes of hydrocarbons. During the past large degree on the carbon atom distributions in t.he feed-
decade, there has been considerable progress reported 011 stocks.
analytical methods to determine various classes of hydro- One of the most comprehensive studies relating feed com-
carbons in gas oils and development of quantit#ativerelation- posit,ion with product yields in catalytic cracking was made
slii~isbetweell molecular composit,ion of charge stocks and by White (1968). He developed a regression correlation model
product yields in catalytic cracking. to predict product yields based on nine different hydrocarbon
(’roroll and Jaquay (1960) pointed out the necessity to con- types in the feedstocks. The detailed analytical methods for
sidrr the differences in the cracking behavior of major hydro- the gas oil charact~erizationwere recently reported by Fit’z-
r a r l m i tylies in charge stocks to accurately predict product’ gerald et al. (1970). Their analysis included distillation,
yields iii catalyt,ic cracking. They described the relat,ive separations by molecular sieves, mass spect,rometry, and
cracking rates of saturates (paraffins and naphthenes) and ultraviolet spectroscopy.
Shnaider et al. (1969) recently developed a kinetic model
To whom correspondence should be addressed. for cat,alytic cracking and utilized it to analyze the catalytic

538 Ind. Eng. Chern. Process Des. Develop., Vol. 10, No. 4, 1971