Chemical r-6+7

CHAPTER 6
HOMOGENEOUS FLOW REACTORS

TUBULAR FLOW REACTORS!
For the design for a tubular flow reactor it is advantageous to have
the laboratory data also on a flow basis. In other words, the most
desirable approach is to determine the rate equation for the process
in
a small-scale flow reactor operated at as close to constant-
temperature
and -pressure conditions as possible. Then this information along
with
some pilot-plant data may be applied with reasonable confidence to
the
design of the commercial reactor. From the engineer’s standpoint
the
project is a two-step problem: (1) interpretation of the laboratory
and
pilot-plant data in terms of a rate equation; (2) the application of the
rate equation to the design of the large-scale reactor. Since one
step is
the reverse of the other, the calculational methods are, in general,
the
same. The chief difference arises because it is desirable, although
not
always possible, to obtain the laboratory data in a reactor operated
at
constant pressure and temperature in order that the interpretation
of the
results is not complicated by many variables. In the commercial-
scale
reactor it may not be feasible to operate with a uniform temperature
and
pressure throughout the reactor so that the design methods must
be
flexible enough to take these complications into account. For
example,
the evaluation of the temperature variation with position in the
commercial
reactor requires the application of heat-transfer calculations along
with the basic reactor-design equations presented in Chap. 4. It is
logical
to expect, then, that the second step, the design of the commercial
reactor, will be likely to involve rather tedious calculations,
frequently
of a trial-and-error nature. Only brief consideration was given in Sec.
3-5 to the analysis of kinetic data obtained on a flow basis.
Therefore
this first step will be treated again here before proceeding to the
design
problem.
6-1. Determination of Rate Equations from Laboratory Data. The

difficulty in interpreting laboratory kinetic data depends upon the
number
of variables involved. Operation at constant temperature and pres-
‘This is the second type in the classification of reactors discussed in
Chap. 4 and
illustrated in Fig. 4-1B,
130
HOMOGENEOUS FLOW REACTORS 131
sure is of considerable advantage. This advantage is augmented if
in
addition the measurements are made in a reactor small enough in
size
so that the change in composition is also small. Under these
restrictions
the rate is essentially constant throughout the reactor. The
calculation
of the rate where such differential reactor conditions exist is
illustrated in
the following example:
Example 6-1. The homogeneous reaction between sulfur vapor and
methane
has been studied in a small silica tube reactor of 35.2 ml volume.’ In
a particular
run at 600°C and 1 atm of pressure the measured quantity of
carbon disulfide
produced in a 10-min run was 0.10 g. Assume all the sulfur present
is the molecular
species So. The sulfur-vapor (considered as S») flow rate was 0.238
g moles/hr.
1. What was the rate of reaction expressed as gram moles of carbon
disulfide
produced per hour per milliliter of reactor volume?
2. The rate at 600°C may be expressed by the following second-
order equation:
r = kpcu;Ps,
where p = partial pressure, atm
Using the rate determined in (1) and this form of the rate equation,
calculate
the specific reaction rate in units of g mole/(ml)(atm?)(hr). The
methane flow
rate was 0.119 g moles/hr and the H.S and C8, concentrations in the
reactants
were zero.
3. Also compute the value of k without making the assumption that
the rate is
constant and that average values of the partial pressures may be
used. That is,
consider the equipment to operate as an integral, rather than a
differential,
reactor. By comparing the results, comment on the suitability of the
apparatus
as a differential reactor.
Solution
CHa + 28> Sn CS. 4- 2H>S
1. The carbon disulfide formation in gram moles per hour is
Then the rate of reaction per milliliter of reactor volume will be
i
r = 0.0079 35.2
2.2 < 10-4 g mole/(hr) (ml)
This rate is not a point value at a constant composition but
represents an integrated
value over the range of compositions between the entrance and exit
of the
reactor. However, the composition changes are small and the
average values in
the reactor may be used to approximate the specific reaction rate.
1R. A. Fisher and J. M. Smith, Ind. Eng. Chem., 42 :704 (1950).
loz CHEMICAL ENGINEERING KINETICS
2. Assuming that at 600°C and 1 atm pressure the components
behave as perfect
gases, the partial pressure is related to the mole fraction by the
expression
Pou, — Pic, = (1) you,
where you, represents the mole fraction methane in the reaction
mixture. The
average composition in the reactor will be that corresponding to a
carbon disulfide
rate of (0 + 0.0079)/2 = 0.0040 g mole/hr.
At this point the molal rates of the other components will be
CS, = 0.0040 g mole/hr
S. = 0.238 — 2 x 0.0040 = 0.230
CH, = 0.119 — 0.0040 = 0.1150
H.S = 2 X 0.0040 = 0.0079
Total = 0.357 g mole/hr
The partial pressures will be
0.1150
Pecu, — 0.357 = 0.322 atm
0.230
= — = (0.645 at Pes ~ 0.357 aie,
From the rate equation, using the value of r obtained in (1),
r eee
Powy Ps. (0.322 X 0.645
1.08 <X 107~* g mole/(ml) (atm?) (hr)
3. If the variations in rate through the reactor are taken into
account, the
integral design expression from Eq. (4-8) should be used.
Vr _ [ &
F im 0 ip
If the conversion is based upon methane instead of the total feed,
this equation
may be written
Ver Ee dx 7 = i = (A)
A au
where F', is the methane feed rate.
At a point in the reactor where the conversion of methane is x, the
molal flow
rate of each component will be
C8. — 0.1192
S2 = 0.238(1 — 2)
CHa = 0.119(1 — 2)
HS = 0.2382z
Total = 0.357 g mole/hr
Using this information, the rate equation may be written in terms of
z:
0.119 X 0.238(1 — 2)?
r = kycuys2= k 0.3572 ee
Equation (A) may be integrated from the entrance to the exit of the
reactor
using Eq. (B).
T- ae Sebi ”
Solving for k,
pa ae
J r/F 1 — ota
The conversion of methane at the exit of the reactor is
0.0079
1: = 0.119 = 0.0664
Hence the specific reaction rate is
4.5 0.0664
ie 35.2/0.119 1 — 0.0664 = 1.08 X 10~* g mole/(ml) (atm?) (hr)
It is apparent in this instance that the change in composition in the
reactor is
sufficiently small that a rate corresponding to the average
composition may be
used to evaluate the specific reaction rate. In other words, the
concept of a
differential reactor used in (1) and (2) to evaluate k is satisfactory. If
the
conversion had been considerably larger than x = 0.066, this would
not have
been true.
While the differential-reactor system just described was _
satisfactory,
in many cases difficulties arise in connection with this method of
approach. If the small differences in concentrations between
entering
and exit streams cannot be determined accurately, the method is
not
satisfactory. Also, because of a high heat of reaction, or particularly
high rate, or a combination of both, it may not be possible to
operate
even a small reactor at conditions approaching constant
temperature
and concentration.
When the conditions of approximately constant rate cannot be met,
the measured conversion data will represent the integrated value of
the
rates existing in all parts of the reactor. A reactor operated in this
way
is commonly termed an integral reactor. The problem of obtaining a
rate equation is essentially one of differentiating the measured
over-all
conversion to give point values of the rate. To give this
differentiation
process a reasonable chance of success, it is clear that as many
variables
as possible should remain constant. The ideal situation is one in
which
the only variable is composition, i.e., in which the temperature and
pres134
CHEMICAL ENGINEERING KINETICS
sure are the same in all parts of the reactor. In essence this type of
integral operation is but one step removed from a differential
reactor.
If the temperature as well as composition changes significantly in
the
integral reactor, the process of differentiating the experimental data
to
obtain a rate equation becomes of doubtful accuracy and, indeed, is
seldom
successful. These conclusions concerning the ease of interpretation
of laboratory data in terms of a rate equation are summarized in
Table 6-1.
TaBLE 6-1. ExpERIMENTAL MeTuops OF OBTAINING RaTE DATA FOR
TUBULAR
FLow SYSTEMS
2. SS SE Eee ee
Type of reactor Characteristics Interpretation of data
Differential........| Constant temperature, compo-| Rata data obtained
directly.
sition, and pressure (i.e., con- | Interpretation simple
stant rate)
Integral (A).......| Constant temperature and Interpretation of integral
data
pressure (rate depends upon | usually satisfactory by
graphicomposition
only) cal differentiation or fitting of
integral-conversion curves
Integral (B)....... Constant pressure only Interpretation complicated
by
temperature variation. If
effect of temperature on rate is
known from independent
measurements, interpretation
is theoretically possible
A complete set of integral-reactor data consists of measurements of
the conversion for different flow rates through the reactor, each run
being made at constant reactants ratio, pressure, and temperature
(if
possible). Then an additional set of conversion vs. flow runs is made
at a different reactants ratio but at the same pressure and
temperature.
This procedure is continued until data are obtained over the entire
reactants-ratio, temperature, and pressure ranges that may be used
in
the commercial reactor. A useful means of summarizing the results
is
to prepare graphs of Vr/F (analogous to the reciprocal of the space
velocity) vs. conversion zx’ at constant values of reactants ratio, 7’,
and p.
Such graphs for representing the experimental data are suggested
by the
form of the basic design equation for flow reactors [Eq. (4-8)],
Lae. de!
(ae I (6-1)
The shape of these curves is determined by the nature of the rate
equation
and hence they are useful in a qualitative way for telling something
about the reaction mechanism.
Two procedures for treating such integral-reactor data are the
integration
and differential methods first described in Chap. 3 and illustrated
there for batch systems (Examples 3-1 and 3-2). In the integration
approach a rate equation is assumed, and then the design
expression
[Eq. (6-1)] integrated. This integration gives a relationship between
Vr/F and x’, which may then be compared with the experimental
data.
The final step is the choice of the rate equation which gives the best
agreement with the experimental Vz/F vs. x’ curves under all
conditions
of reactants ratio, temperature, and pressure.
The differential method involves differentiating graphically the Vz/F
vs. x’ curves and thus obtaining the rate of reaction as a function of
composition. Various assumed rate equations can then be tested for
agreement with the rate-composition data.
It may not be immediately evident under what conditions the slope
of a Vr/F vs. x’ plot is equal to the rate r. In applying Eq. (6-1) the
integration is to be carried out at constant feed rate and variable
reactor
volume. Hence if experimental data are obtained by measuring
conversions
at various volumes for constant Ff, Eq. (6-1) may be directly
differentiated, giving
Ot aie FP.
~ dVr/F- d(Vr/F)
re = slope of Vez/F vs. x’ graph (6-2)
However, it is much easier experimentally to determine the data by
varying the feed rate and using one reactor volume. Now, to obtain
the rate, Eq. (6-1) must be differentiated at constant volume but
variable
feed rate. The left-hand side may still be written as d(Vr/F), but,
to obtain dx’/r for the right side, the rate must not depend upon the
feed rate, since this is a variable. The rate will be a function of #
when
diffusion resistances are an important contribution to the over-all
reaction
process. This may be the case in fixed-bed catalytic reactors or in
homogeneous reactions where the rate is very high (e.g., some
combustion
reactions). Under these conditions the slope of a Vz/F vs. x’ plot is
not
equal to the rate. For most homogeneous reactions the reactor
design is
such that the rate is not a function of the feed rate and Eq. (6-2) is
valid.
The role of diffusion in fixed-bed catalytic reactions is considered in
detail
in Chap. 9. .
Perhaps the clearest way of explaining these methods of
interpreting
laboratory data is to carry out specific examples in some detail,
pointing
out the features of the method which are general and applicable to
any
reaction. In Example 6-2 the integration and differential methods
are
applied to a simple system. Example 6-3, which is concerned with
the
homogeneous reaction of sulfur vapor and methane to produce
carbon
disulfide, is a more complex illustration. It is complicated by the fact
136 CHEMICAL ENGINEERING KINETICS
that the sulfur vapor is present as three different species, Ss, Se,
and S».
The temperature is constant in both examples so that they are
illustrations
of an integral reactor, case A, described in Table 6-1.
Example 6-2. A kinetic study is made of the decomposition of

acetaldehyde
at 518°C and 1 atm pressure in a flow apparatus. The reaction is
CH;CHO — CH, + CO
Acetaldehyde is boiled in a flask and passed through a reaction tube
maintained
by a surrounding furnace at 518°C. The reaction tube is 3.3 em ID
and 80 cm
long. The flow rate through the tube is varied by changing the
boiling rate.
Analysis of the products from the end of the tube gives the results in
Table 6-2.
TABLE 6-2
Rate of flow, Fraction of acetaldehyde
g/br decomposed
130 0.05
50 oe
21 0.24
10.8 0.35
What is a satisfactory rate equation for these data?
Solution. A second-order rate equation, r = k.C%, will be tested by
both the
integral and the differential method. To utilize either, it is necessary
to express
the rate in terms of the conversion of acetaldehyde, x (which is
equal to 2’ in this
ease). This may be accomplished by applying material balances and
the perfect-
gas law.
The molal flow rate of acetaldehyde entering the reaction tube is F.
Ata point
where the conversion is x it will be
Na =F —Fx
The molal rates of the other components will be
Neu; = oF
Neco =H
Total flow rate N, = F(14+ 2)
From the perfect-gas law
C = Na a Na a Na Pt
“"V - NRT/p, =N, RT
Oy eele—z e7e;
Vea ee
The second-order rate expression in terms of conversion is
BW ee el Ces eh
acs (in) ¢ =) (A)
_ INTEGRATION METHOD. Equation (A) may be substituted in the

design expression,
Eq. (6-1), to evaluate the second-order assumption.
Le eet de
a ks Sy 1-2)’
ne 0 l+z2
Integrating yields
2 Pt Ver 4
ma(2 ) Ye 4 pmo te-3 (B)
Equation (B) provides the required relationship between Vz/F and x
which
should now be tested with the experimental data. Noting that
2
Ve =n (22) x80 = 084
the experimental data may be expressed in terms of Vz/F as shown
in Table 6-3.
TABLE 6-3
Conver- Feed rate
sion — Vr/F, ko,
oe G/hr ombienvies (liters) (sec) /g mole | liters/(g mole) (sec)
0.05 130 0.000825 828 0.32:
Oris 50 0.000316 2,160 0.33
0.24 21 0.000131 5,210 0.32
0.35 10.8 0.0000680 10,000 Ores
Equation (B) may be used to compute a value of k for each of the
sets of x and
Vr/F values given in the table. For example, at x = 0.13, substituting
in
Eq. (B),
1 : Any ee 4 eet Te 1) PL oak
‘ieF eo a = oe Pon 6s het
k. = 0.33 liter/(g mole) (sec)
Values of ke for the other three sets, shown in the last column of
Table 6-3, are
in good agreement with each other.
DIFFERENTIAL METHOD. The experimental Vr/F data are plotted vs.

conversion
in Fig. (6-1). According to Eq. (6-2) the slope of this curve at any
conversion
gives the rate of reaction at that point. The slopes of the curve at
the experimental
conversions are given in the second column of Table 6-4.
These slopes or experimental rates may be compared with the
assumed rate
1‘,
equation by (1) plotting In r vs. In ee a according to Eq. (A), and
noting
whether or not a straight line of slope = 2 is obtained; or (2) by
computing values
of ke directly from Eq. (A). Following this second procedure, at « =
0.13,
‘ary 1 ) (; = ae)
r= 4.9 x 10 = %2\ 9 og x 7917 \i + 0.13
ky = 0.35 liter/(sec)(g mole)
The results at other conversions are shown in the third column of
Table 6-4.
While there is some variation from point to point, a significant trend
does not
TABLE 6-4
ee ee
Conver | slope from Fig. (6-1), he op
on g moles/(liter)(sec) | liters/(sec)( g mole) secr—
0.05 632 <-105" 0.32 0.0045
0.13 4.9 X 1075 0.35 0.0041
0.24 2.8 X 10-* 0.33 0.0030
0.35 PH Ue <5) UI ce. 0.35 0.0027
exist. Hence the differential method also confirms the validity of a
second-order
rate equation. The variation is due to errors associated with the
measurement
of slopes of the curve in Fig. 6-1.
0.35
Cxo- nversion
0 2000 4000 6000 8 \,/F ~liters (sec) /(g, mole) oe 10,000
Fic. 6-1, Conversion data for decomposition of acetaldehyde.
FIRST-ORDER RATE EQUATION. For comparison the differential
method will be
applied to a first-order mechanism. The rate expression, replacing
Eq. (A), will
be
l—@z
Ue = hi1UCV A Sy‘ ePtTes ide (C‘)
This mechanism may be tested by computing k; values from the
experimental
rates of reaction, using Eq. (C). For example, at « = 0.13,
1 1 — 0.13
0.082 X 7911+ 0.13
r Il
ky
slope = 4.9 X 10-5 = ky
I 0.0041 see!
The k, values for other conversions (fourth column of Table 6-4)
show a distinct
trend toward lower values at higher conversions. This indicates that
the firstorder
rate is not a likely one.
Example 6-3. A flow reactor consisting of a 1-in. stainless-steel pipe

6 in. long
and packed with inert rock salt is used to study the noncatalytic
homogeneous
reaction
4
b Ss + He —-> 2H.S — CS.
The measurements are carried out at atmospheric pressure in the
vapor phase at
600°C. From available data on the rate of dissociation of the sulfur
species it is
reasonable to assume that the reactions
Ss are 45,
Se —? 352
are very fast with respect to the combination of sulfur vapor with
methane.
Accordingly, assume that equilibrium exists between Ss, Ss, and Ss.
The void
volume, measured by benzene displacement, was 35.2 ml. The
conversion of
methane to carbon disulfide (2) was measured for various flow rates
and initial
reactants ratios, and the results are as shown in Table 6-5.
The symbol (S2) refers to the total amount of sulfur vapor present,
expressed
as So; 1.e.,
(S2) = Ns, + 3Ns, + 4Ns,
where the N values refer to the number of moles of each sulfur
species.
TABLE 6-5
| Flow rates, g moles/hr :
Reactants ratio, Conoan
CH (Ss) cs _ moles (82)/moles (CH,) version
422
55 0.02975 0.0595 0.0079 2 0.268
58 0.0595 0.119 0.0086 2 0.144
57 0.119 0.238 0.0078 2 0.066
59 0.119 0.238 0.0072 2 0.060
56 0.238 0.476 0.0059 2 0.025
715 0.0595 0.119 0.0079 2 0.133
76 0.02975 0.0595 0.0080 2 0. 269
fis 0.119 0.238 0.0069 Z 0.058
78 | 0.0893 0.0893 0.0087 I 0.0975
79 0.119 0.0595 0.0096 0.5 0. 0807
2 — ——
1. Assuming first that the only species of sulfur which exists is Se
(this is
equivalent to choosing b = 2 in the chemical reaction), test the
assumptions of
first- and second-order rate equations with the experimental data.
2. Repeat part 1, assuming that equilibrium exists between Se, Ss,
and Ss, but
that only S. reacts with CH.
3. Using a second-order rate expression, determine which
assumption regarding
the reactive sulfur species best fits the experimental data.
Solution. The tabulated experimental data should be converted to a
table of
Ve/F and 2’. For run 55
F = molal feed rate = 0.02975 + 0.0595
= 0.0892 g moles/hr
Vr 39.2
“F 0.0892 — ae
The conversion x is the fraction of the methane that is actually
converted to
CS... The moles of CS. per mole of feed, x’, is
Initial reactants ratio
moles(Se) /CHa
OF20
x 10
oO O35
200 300 4
Ve /F -(ml)(hr)(g.mole)
e
Fig. 6-2. Experimental conversion data for the sulfurization of
methane.
also plotted in Fig. 6-2. Note that
pe moles product _ r moles CH, in feed 1
moles total feed moles total feed _ % l+a
or xz = (a+ 1)z’
where the moles of sulfur in the feed, per mole of methane, is a.
1. Two first-order mechanisms may be assumed, one with respect to

methane
and one with respect to sulfur vapor (Sz). Starting with the former
and applying
Kq. (6-1),
Ye foe. dx’ -[* dx’ A
F 0 r 0 kipeus 0 kiyeus (A)
Assuming perfect-gas behavior at 600°C and 1 atm total pressure,
Pocus = YousPt = You,
The mole fraction methane ycn, varies with the conversion. Hence,
to relate
Ycu, to x’ requires a knowledge of the number of moles of each
component present
as a function of conversion. Since it is to be assumed in part (1) that
all the
TABLE 6-6
Conversion Ve —log (1 — (a + 1)21]
Run |- --—7——-— F or ki X 10*| kis X 104
Ly ry (ml) (hr) /g mole — log (1 — 2)
55 (0.268 |0.0890 395 0.135 (8 3.93
58 |0.144 |0.0480 197 0.0675 7.90 3.95
57 |0.066 |0.022 98.6 0.0297 6.93 3.46
59 |0.060 |0.020 98.6 0.0269 6.29 3.15
56 |0.025 |0.00833 49.3 0.0110 5.14 2.57
75 |0.133 |0.0443 197 0.0620 7.20 3.62
76 (0.269 |0.0895 395 0.136 7.95 3.97
77 |0.058 |0.0193 98 .6 0.0269 6.05 3.02
78 |0.0975/0.0487 197 0.045 5.30 5.45
79 |0.0807|/0.0538 197 0.0368 4.31 9.90
sulfur present is Se, the number of moles of S2 and CH, are simple
to ascertain.
Thus entering the reactor, for a mole ratio of S2 to CH, of 2,
1
You, = e—a— = 0.03. 333
If a basis of 1 mole of CH, and 2 moles of 8: is chosen, at a point in
the reactor
where the conversion is x’, the number of moles of each component
will be
CS; = 4 = 32’
HS a ee 6a’
CH, =1—2=1- 37’
S. = 2 — 62’
Total = 3
a ig 2 — 62’ " ;
You, = —= = 0.333 — 2’ Ys, = 3 = 0.6607 — 22
More generally, if the initial mole ratio of sulfur to CH, is a,
1—(+a)r* 1 ;
Ya ge one (B)
a — 2(1 + a)v’ a ;
i Dee eee ee
Using the integral method, Eq. (A) can now be integrated with Eq.
(B),
Vr “ae dir! —1, V/dtay—-2 1
FO ky i ee Te,
1
= ae at! (C)
In [1 — (a + 1)a4]
where x; is the conversion leaving the reactor. The Vr/F vs. 2’ data
shown in
Table 6-6 can now be used to test Eq. (C), and hence the
assumption of a firstorder
reaction. Rather than plot curves obtained from Eq. (C) with
different k
values for comparison with the experimental curve (Fig. 6-2), it is
more effective
to use one of two other approaches:
a. Plot the experimental Vr/F vs. « data given in Table 6-6 as Vr/F vs.
log
(le).
+ 8 Initial reactonts ratio
moles (S2)/CH4
Fia. 6-3. Test of first-order mechanism with respect to CH.
This method of plotting should result in a straight line if the
assumed rate
expression is correct. Figure 6-3 shows that this is not true if the
rate is firstorder
with respect to methane. At a constant reactants ratio of a = 2.0 the
points fall on a straight line, but the data for other ratios deviate
widely from
the line.
b. From the data in Table 6-6 compute values of k; required by Eq.
(C) for
each of the runs. If these values are essentially constant, the
assumed mechanism
fits the data. For this case the resulting k; values are given in
column 6 of
Table 6-6. Again the runs for different reactants ratios show that the
assumed
rate equation does not fit the data.
As will be seen later in this example, the nature of the assumed rate
equation
may be complex enough to prevent analytical integration of the
design equation
(6-1) or Eq. (A). However, in such cases it is still possible to evaluate
the
integral fdx’/r graphically and then apply either method a or method
b to determine
the suitability of the rate assumption. Hence either method a or
method 6
is general and as such is applicable to any reaction.
The design equation for an assumed mechanism which is first-order
with respect
to sulfur vapor, Ss, will be similar to Eq. (C), i.e.,
Vr = i dx’ 7 1 Ihe dx’ oon 1 a a S 2(a + 1)z;
F o kisys, kis Jo a/(+a) — 22’ 2k1,8 a
RY ee _ 2a = Diss In (: - ) (D)
For a = 2 Eq. (D) becomes the same as Eq. (C) if kis = 3k,. Hence
the test of
this mechanism is whether or not the two runs for a = 1.0 and a =
0.5 (1) are
brought in line with the other data on a plot similar to Fig. 6-3 or (2)
give the
same value of ky,3. The values of ki,s are given in the last column of
Table 6-6.
It is clear that the results for the runs with a = 1.0 and 0.5 are far
removed from
the rest of the data. Hence neither first-order mechanism is
satisfactory.
If a second-order mechanism is tried, again assuming all the sulfur
vapor is
present as Se, the design equation is
Vr dx’ 1 dx’ ae dx’
fo - kopou,Ps, ko YcusyyYs8. ko ( 1 ‘) (
This expression can be simplified and integrated to become
Ve _a+i1 |” dat aye te (2 )-ina- E tt) | eam aan mG Bee len
3 T1 = for a = 2
J 2ke1 — x1 ithe t
Now it is possible to use this equation and the experimental data to
determine
the suitability of a second-order rate equation, just as Eqs. (C) and
(D) were
employed, to test first-order assumptions.
+ Equation (E) becomes indeterminate when a = 2. Hence it is
necessary to
substitute a = 2 prior to integration and obtain the particular
solution given.
Figure 6-4 shows a plot of Vr/F vs. = ; | m (: _ ® 21) —In(i— »)|,
30 24
ye Li— oy
tion is an improvement over the previous ones, since the point for a
= 1.0 is
close to the straight line. However, the data point for a = 0.5 is still
not satisfactorily
correlated.
‘ONT
for a = 2, for the experimental data. It is apparent that this
correlarand
s /nitiol moles (S2)/CH4
“0 0.2 °o e2s0
Fia. 6-4. Test of second-order reaction mechanism.
From this preliminary investigation, neglecting the Ss and Ss
species, it is
concluded that a second-order rate is more appropriate than a first-
order one.
The next step is to refine the analysis by considering the distribution
of sulfur
vapor between Ss, and Sz and 8s.
2. To take into account the presence of the Ss and Ss species,

equilibrium will
be assumed between S., Ss, and Ss, utilizing the following
equilibrium constants
at 600°C:
aSs—>S. K, = 0.930 (atm)?
3856S. K, = 0.669 (atm)}
This information along with the stoichiometry of the reaction ean be
used to
develop expressions analogous to Eqs. (B) for the relation between
2 s, and you,
and the conversion x. As in part 1, suppose that entering the reactor
there are a
moles of sulfur vapor per mole of CHy. However, let a refer to the
total moles
of sulfur considered as Ss, not to the sum of the moles of Se, Se,
and Ss. If
ao, Bo, and Yo represent the moles of Ss, Ss, and Ss entering the
reactor, the total
moles of sulfur, considered as Ss, will be “)
a@ = ao + 380 + 470
At a point in the reactor where the conversion is z, the moles of
each component
will be
CS. Sle
HS = 22
CH,z=1-—-2
Se = a
Se = 8
iy
n
Ss
Total = =atBt+yt+2x+1
The conversion x has been defined in terms of the methane reacted.
It can
be written also in terms of the sulfur used, thus providing a
relationship between
x, a, B, and y. If (S:) represents the total moles of sulfur (considered
as S.)
present at any point in the reactor,
x = 3[a — (S:)]
but S:=a+ 368+ 4y
so that
a—a— 36 —4y
a,
1G ih
Two additional relationships between z, a, 8, and y must exist
because of the
dissociation equilibria. These are as follows:
Ps Ys» a
AaE 0.930 = Po
(ps,)* (ys,)* (vy)(at+tB+y+ 2x +1)?
and
3: 0.669 == =. & _ =
(ps)! (ys)? (8)"a+B+y+t 2¢ + 1)!
Equations (1), (2), and (3) permit the evaluation of a, 6, and y at any
conversion
and reactants ratio a. Then the mole fractions are immediately
obtainable
from the expressions
a
Ys, = CEG yo 26 yA)
1—z2z
one at BEY + Qe t1
To illustrate the method of calculation, take a = 2, the
stoichiometric ratio.
Entering the reactor, z = 0.0, and Eqs. 1, 2, and 3 become
2=a-36 ++ 47
a taPy ta = 5o s
a
(B)i(a+6B+7 +1! = 5G,
Solving these by trial gives
a = 0.782 Ys, = male = 0.360
B = 0.340 Ys_ = — = 0.156
y = 0.049 Ys, = on = 0.023
CH, = 1.000 Ycu, = aa = 0.461
Total = 2.171 zy = 1.000
In a similar manner the mole fractions corresponding to any
conversion can be
evaluated. The results of such calculations are shown in Fig. 6-5,
where comfM-
royal cet ion
fo) DS
0.3
0.16
fMroalcet ion-y
oO ~%
0.08
0.04
O 0.08 0.16 0.24 0.32 0.40 048
Conversion of methane - x
ee 6-5. Composition of reaction mixture vs. conversion for the
sulfurization of
methane.
position of the reaction mixture is plotted vs. the conversion x.
Similar information
for other values of a are obtained in an analogous fashion.
The necessary data are now at hand for testing the second-order
rate expression.
However, the calculational procedure used in part 1 is not applicable
in this
case because the design equation
\euaee / a de!
Fo ky Jo you ds, a cannot be integrated analytically. Two
alternatives are possible. The integral
method involves graphical evaluation of the integral in Eq. (F) for
each experimental
point, using the measured value of the conversion as the upper limit
in
each case. This procedure leads to values of fs dx'/ycu.ys, for each
value of
V2/F shown in Table 6-6. The results are given in the fourth column
of Table
6-7 and plotted in Fig. 6-6, where Ver/F is the abscissa and ie
dx'/Yycu.ys, the
ordinate. If the assumed mechanism is a satisfactory interpretation
of the data,
10
Pe etiell (el ||
. ee ae
oe
2 oe
eg ale ee
ae 04 ak hee Initial reactonts ratio
ee ee ee
ieee a2
°% 100 20O0 300 400
Vg/F - mi. (hr)/g .m ole
Fia. 6-6. Test of rate equation r = ke'pcu,Pso-
Fig. 6-6 should be a straight line with a slope equal to kj. This is
apparent if
Eq. (F) is rearranged as follows:
if dx’ - i! Vr
0 YcH.Ys, wi
It is evident from the plot that this condition is essentially satisfied.
The other approach, the differential method, involves graphical
differentiation
of the data for comparison with the differential form of Eq. (F). If the
data are
plotted as 2’ vs. Vr/F, as in Fig. 6-2, the slope of the curve is equal to
the rate of
reaction. Thus from Eq. (F)
F (72) __
F koycuYss
da!
aa d(V r/F)
= slope = kiycu Ys,
If the slope is taken at various values of the conversion from Fig. 6-
2, and the
corresponding values of you, and ys, read from Fig. 6-5, and this is
done for each
reactants ratio, complete information on the effect of composition
on the rate is
available. The equation r = kjycuyys, then can be tested by
observing the
constancy in computed values of ky.
3. The results of part 2 indicated that the assumption of S» as the
active sulfurvapor
species in the second-order mechanism gave a satisfactory
interpretation of
the conversion data. However, it is important to know whether
equally good
TaBLe 6-7. Test or REAcTION MECHANISMS: SULFURIZATION OF
METHANE
nae eS sO cee MT a ee = ee
zy’ dx’
V ie Ycu Ys
Run | Reactants ratio, a re aa
b= 2 b=6 b=8
55 2 395 0.762 | 2.11 17.0
58 2 197 0.347 | 0.890 6.64
57 2 98.6 0.144 0.350 2.48
59 2 98.6 0.130 02315 2.23
56 2 49.3 0.052 O2123 0.885
res) 2 197 0.316 0.800 5.88
76 2 395 0.765 2.0% iv
77 2 98.6 0.126 0.303 2.15
78 1 197 0.315 1.27 12-1
79 0.5 197 | 0.388 2.97 36.8
correlations would be obtained by assuming that Sx or Ss is the
reactive form.
To do this, the design equations, corresponding to (F), for the other
species may
be written as follows:
Ve _ 1 ila ds
pees, Cees! goes G
F ke Jo yous, (G)
Ve 1 | da!
F ks 0 YcH,Yss (H)
Figures 6-7 and 6-8, which show Vr/F plotted vs. the integrals in (G)
and (H),
were prepared by graphical integration, using Fig. 6-5 for the
relations between
x’ and ys, and ys,. The values of the integrals are given in the last
two columns
of Table 6-7. In contrast to Fig. 6-6 based upon &- it is seen that the
data do not
fall on a straight line as required by Eqs. (G) and (H). Also, the
points for
various reactants ratios are not in agreement. It is concluded that a
secondorder
mechanism based upon either Ss or Ss as the reactive sulfur species
does
not agree with the experimental data. Note that the alternate
method of differentiating
curves of Vr/F vs. x’ to obtain the rate of reaction could also have
been used in this part of the problem, as in part 2.
The preferred equation for the rate of reaction is
i,
(DS kopcu,Ps,
149 HOMOGENEOUS FLOW REACTORS
/nitial reactonts ratio
(So) /CHg
OF:
x 710
o0 05
200 300 “400
V_/F - ml. (hr) /g. mole
Fic. 6-7. Test of rate equation r = ke’ pou,psegy’VSeg
Ha‘f
/nitial reactants ratio
pant
(So) /CHg
(ROL SX}
0. x
O53
% 100 200 300 400
Ve/F - ml (hr)/g.mole
Fia. 6-8. Test of rate equation r = ks'/ponypss.
From Fig. 6-6 the slope of the straight line is 2.2 x 107% g mole/(hr)
(ml)(atm)?.
Since this number is equal to #3, the final expression for the rate of
reaction at
600°C! is
y= 22 X 10 “pon?
1 By evaluating the specific-reaction rate kz at other temperatures
an activation
energy can be determined. For this aspect of the problem see Fisher
and Smith,
loc. cit.
It should be emphasized that it is not advisable to make a firm
statement
regarding the mechanism of a reaction from a limited amount of
data such as was available in Example 6-3. It is perhaps best to
consider
that the above rate equation is a satisfactory interpretation of the
kinetic data. More knowledge concerning the actual mechanism of
the
reaction could be obtained if the pressure had been one of the
variables
studied as well as reactants ratio and conversion.
The problem of interpreting laboratory kinetic data in terms of a rate
equation will be considered again in Chap. 9, for catalytic reactions.
6-2. Isothermal Design (Tubular Flow). Once a satisfactory
expression
is known for the rate of reaction, the problem of determining the
size of a flow reactor as a function of operating variables is one of
combining
and solving the rate and design equations. Since a number of
variables
such as temperature, pressure, and mass velocity can be changed,
it is necessary to solve the equations for a number of sets of
operating
conditions and then choose the conditions which will result in the
most
practical and economical reactor. While this optimizing of the design
is of great importance in practice, the repetition of the calculations
is
unnecessary for an understanding of the principles of reactor design
and no attempt will be made to illustrate economic-balance
processes.
Instead, design calculations will be carried out for single sets of
operating
conditions for several different types of flow reactors.
As pointed out earlier, the design procedures are the simplest when
temperature and pressure changes within the reactor can be
neglected.
They become more complex for adiabatic and pseudo-isothermal
reactors
where significant temperature differences exist.
The isothermal design calculations are essentially the reverse of the
procedures illustrated in Example 6-3 for establishing the rate
equation
from integral-reactor data. They are illustrated in Example 6-4.
Example 6-4. Hougen and Watson! in an analysis of Kassell’s data

for the
vapor-phase dehydrogenation of benzene in a flow reactor
(homogeneous)
reported the following rate equations:
if 2C6H6(g) > Ci2Hio(g) + He(g)
Ty = 14.96 X 10%e~15.200/7 (05 - bePe)
r, = lb moles benzene reacted/(hr)(cu ft)
2: CeHe(g) + CizHi0(g) > CisHis(g) + He
rT. = 8.67 & 10%e715,200/T (moro = PIPE
Ke
r2 = lb moles triphenyl produced (or diphenyl reacted)/(hr)(cu ft)
10, A. Hougen and K. M. Watson, “Chemical Process Principles,”’ 5
. . M. Watson, af: es,’ vol. 3, p. 84
John Wiley & Sons, Inc., New York, 1947. : ‘
’
where ps = partial pressure of benzene, atm
Pp = partial pressure of diphenyl, atm
pr = partial pressure of triphenyl, atm
Pu = partial pressure of hydrogen, atm
Wee SK
Ky, AK. = equilibrium constants for the two reactions in terms of
partial
pressures
The data upon which the rate equations are based were obtained at
a total
pressure of 1 atm and at temperatures of 1265 and 1400°F in a 0.5-
in. tube, 3 ft
long.
It is now proposed to design a tubular reactor which will operate at 1
atm pressure
and 1400°F. Determine the total conversion of benzene to di- and
tripheny|
as a function of space velocity. Also, determine the reactor volume
required to
process 10,000 lb/hr of benzene as a function of the total
conversion. In order
to obtain a relatively simple picture of the method of approach, first
carry out
the solution, assuming that only reaction 1 occurs, and then
proceed to the complete
solution, considering both reactions.
Assume that the reactor will be operated isothermally and that no
other reactions
are significant.
Solution. Since the reactor is isothermal, the equilibrium constants
K,; and Ke
will have fixed values. They may be estimated at 1400°F from
equations developed
by Hougen and Watson,! using methods analogous to those
described in
Chap. 2. The results are
K, = 0.312
K» = 0.480
As these values are not large, the reverse reactions may well be
important.
Hence the negative terms in the two rate equations cannot be
neglected. At
1400°F (1033°K) in terms of the rate of disappearance of benzene
the two rates are
Oeide (tee se PPR ee 9 _ Poppa (14.96 X 10°%)e (03 a 6.23 (05
0.312 (A)
- 6) p—14.7 ae tg eee B
rz = (8,07 X 10 Je (vs, 0.480 3.61 \ papo 0.480 (B)
Ti
The design expression for the reactor can be written
Vr _ dx
Fr
where x’ represents the moles of benzene reacted per mole of feed.
Since reactions 1 and 2 occur simultaneously in the same reactor,
an equation
of the preceding type must hold for each reaction individually.
These expres-
Veef [ ae ©C) sions are
1 Loc. cit.
rVrerce s |d ax s (DD)
where x, = Ib moles benzene disappearing by reaction 1 per lb mole
feed
ay = lb moles benzene disappearing by reaction 2 per lb mole feed
An expression could also be written for the total rate of
disappearance of
benzene by reactions 1 and 2, and it would be
a) potee ap (B)
F +s 1+ Ta
However, this equation by itself cannot be used to obtain the
conversion as a
function of space velocity, because it does not take into account
that each reaction
is occurring separately in the reactor and, therefore, that the design
equation
must be satisfied for both reactions. Equations (C) and (D) do
express this
restriction and can be used for the required solution, once the rates
r; and r, are
expressed in terms of the conversions x; and zy. These latter
relationships
depend only upon the stoichiometry of the reactions and may be
developed in the
following way:
Take as a basis 1.0 mole of entering benzene. When the conversion
by the first
reaction is x; and by the second is x3, the moles of each component
will be
Moles hydrogen = $2, + x5
Moles Ci2Hio = $2, — 2
Moles CeHe = 1 — 2} — 25
Moles CisHis = 25
Total moles = 1.0
Since the total moles are equal to 1.0 regardless of x} and x4, the
mole fractions
of each component are also given by the above quantities. If the
components
are assumed to behave as perfect gases, the partial pressures will
be
i eg / Lay, / Pu = 1.02; + ate) = 32 + Ze
PDO oh ies
/ / Pap=1l1—2,— 2,
/
Pr= X
With these relationships the rate equations (A) and (B) can be
expressed in
terms of x} and x4.
fs Ang! = / oy , /
"1 = 6.23 K = a oe x5)? _ (3 1 sai + 2 (F)
ef onl / Te = 3.61 a =f eee aoe ee vei +) | (G)
In theory Eqs. (F) and (G) can be substituted in the design equations
(C) and
(D) and values of exit conversions a, and x computed corresponding
to various
values of Vr/F. Actually a difficulty arises which prevents this
procedure:
both x; and x3 oceur in Eq. (F) or (G), making the integration of Eq.
(C) or (D)
impossible. It is apparent that this difficulty is due to the fact that
two chemical
reactions are occurring simultaneously. If only the first reaction is
considered,
only x; is involved and direct integration of the design equation is
possible. To
obtain a concept of the method of approach for this type of reactor-
design problem,
the solution will first be carried out, making the assumption that just
the
first reaction is significant. This means that v5 = 0.
Simplified Solution (reaction 2 assumed unimportant). Since ry = 0,
Kq. (F)
for r simplifies to
r = 6.23[(1 — x})? — 0.801(x1)3]
and involves only x}. Hence it is possible to substitute it in design
Eq. (C) and
integrate directly.
y ae dx’ el ie dx'
F 6.23 Jo (1 — xj)? — 0.801(21)? 6.23 Jo (1 — 0.10524)(1 — 1.8952’)
V_ 0.559 (; - oe
F 1 — 1.8952’ 6.23 e
Table 6-8 gives the values of Vr/F, corresponding to successive

values of xj
obtained by substituting in Eq. (H). A particular Ve/F figure
represents the
TABLE 6-8
eee ee ayes pecomiponed/ Mes cu ft/(Ib mole feed) (hr) | Space
velocity hr7!
Ib mole feed F
0.0 0 sa
0.1 0.0179 21,200
0.2 0.0409 9,280
0.3 0.0725 5 , 230
0.4 0.123 3,080
0.5 0.260 1,460
0.52 0.382 990
0.528 2% 0
+ Equilibrium conditions.
ratio of volume of reactor to feed rate necessary to give the
corresponding conversion.
The space velocity (per hour) is related to Ve/F in the following way,
volumetric feed rate — Fy,
piacere ety EY olime of reactor Vg
or By = We (I)
where vy is the volume per mole of the feed. If the basis for
measuring this volume
is taken as 60°F and 1 atm pressure (standard conditions for gas
volumes)
and benzene behaves as a perfect gas at these conditions,
_ RT _ 10.73(400 + 60)
ea 14.7
= 379 cu ft/lb mole
The third column in Table 6-8 shows the space velocities from Eq.
(1) with
ve = 379. Figure 6-9 is a plot of the space-velocity—conversion data
given in
the table. The equilibrium conversion is 52.8 per cent, as that is the
value corresponding
to infinite time. The same result could have been obtained by using
,
oCbfo-e nnvzeernsei on Simplitied solution
(Dipheny/ formation only)
0.2
Vp/F -cu ft (hr) /|lb mole
a a a Ee
10,000 6 4 3000 2000 1000
Space velocity - hr!
Fia. 6-9. Dehydrogenation of benzene (simplified solution).
the equilibrium constant K, = 0.312 to compute the equilibrium yield
by the
method described in Chap. 2. The space velocity is related to the
time the
reaction gases remain in the reactor (i.e., the contact time). In this
isothermal
problem the contact time would be equal to the reciprocal of a
space velocity
based upon vr measured at 1400°F. Thus according to Table 6-8 a
contact time
of 990(530/1,860) hr, or about 1.0 sec, would be required to obtain
a conversion of
52 per cent. The Vr/F data shown in Table 6-8 can be used to
determine the
reactor volume Ver required for any conversion x, at a given feed
rate F. This
calculation will be carried out for a feed of 10,000 lb/hr when the
more accurate
Vr/F vs. x’ results are obtained in the following treatment, where
both reactions
1 and 2 are taken into account. |
General Solution (both reactions 1 and 2 considered). The rate
equations (F)
and (G) and the design equations (C) and (D) provide the means for
determining
t, and «2 as a function of Ve/F. As direct integration is not possible,
it is
necessary to rewrite (C) and (D) in difference form and accomplish
the integration
by a stepwise numerical approach. Thus in a small element of
volume AV
. / / the conversions Ar, and Az, are
AV 7
Az; = a m, =A (a) ae (C’)
Az, = “e toh Ce) To, (D’)
where 7, and ro, are average values of the rate for the conversion
range x} to
t, + Ar} and 2} to ty + Az.
The method of solution of (C’) and (D’) will be illustrated by starting
at the
entrance to the reactor, where x; and x are both zero, and carrying
out several
stepwise calculations.
The initial rates of reaction r,, and rs, are given by Eqs. (F) and (G)
with
t4 =2,=0.
7, = 6.23(1 — 0) = 6.23 lb moles CgHe converted/(hr)(cu ft)
ro=, 3. 61 X0= 0
Choose an interval of A(Vr/F) = 0.005 cu ft/(Ib mole feed) (hr).
If the initial rates are assumed to be constant over the interval of
A(Vr/F) = 0.005
the conversions at the end of the interval are, according to Eqs. (C’)
and (D’),
zt, = 0+ 0.005 X 6.23 = 0.03115
ry = 0+ 0.005 X 0 =0
Now more accurate values of 71, and rs, can be obtained by
substituting these
values of the conversion in the rate equations.
1 0.031152
4. 6312
0.03115
2
r, = 6.23 ja — 0.03115)? | = 6.23 X 0.938 = 5.84
T, = 5.61 a — 0.03115) - 0| = 0.0544
The average values of the rates in the first increment of A(Vr/F) may
be taken
as the arithmetic average of the rates entering and leaving the
increment; Le.,
= T, +71, = 6.23 + 5.84 = 6.04 mae 2 2
ro, = + ie: T° Oot = 0.0272
Using these revised values of the average rates, the conversions at
the end of
: 4 / / the first increment are, again according to Eqs. (C’) and (D‘),
zi, = 0 + 0.005 x 6.04 = 0.0302
’ = (0+ 0.005 < 0.0272 = 0.000136
£9,
Proceeding to the second increment of A(Vr/F), the computations
can be
speeded up by anticipating the average rates in the increment,
rather than by
using the values entering the increment. Estimate that the average
rates will be
r1, = 6.04 — (6.23 — 5.84) = 5.65
ro, = 0.0272 + (0.0544 — 0) = 0.0816
Using these values, the conversion occurring within the second
increment is
given by Eqs. (C’) and (D’).
Ar = 0.005 x 5.65 = 0.0282
and xv, = 24, + Axi = 0.0302 + 0.0282 = 0.0584
Ax, = 0.005 < 0.0816 = 0.000408
19, = 0.000136 + 0.000408 = 0.000544
The estimated values of the average rates can be checked by
evaluating rates
at the beginning and end of the increment.
At the end of the increment
Vigo 6.23 |(1 soe 0.0584 — 0.0005)? re (o92 = 0.0005 (aug
+ 0.0005 = |
: 2 )0 .312
6.23 X 0.883 = 5.50
m4
1
+ 0.0005 )|
= 3.61 X 0.0270 = 0.0975
At the beginning of the increment
r1, = 6.23 cK — 0.0302 — 0.0001)? — Ge — 0.0001 ) Gee oto
0.0001)
aa | Bice 0.312
pe 3.01]( — 0.0302 — 0.0001) Cee = 0.0001) — 0.000136 ue
+ 0.0001) = ae|
= 0.0529 0.480
The average rates are
m1, +m, 5.86 + 5.50
ao . = 5.68 vs. 5.65 estimated
oe 2, + 1% = 0.0529 + 0.0975
fer ae 9 = 0.0752 vs. 0.0816 estimated
1, =
a
The comparison with the estimated average rates is sufficiently
close that a
second calculation is unnecessary. The final values for the
conversion leaving
the second increment (Vr/F = 0.005 + 0.005 = 0.01) will be
ae = 0.0302 + 0.005 x 5.68 = 0.0586
to, = 0.000136 + 0.005 x 0.0752 = 0.000512
This stepwise procedure can be repeated until any desired value of
x{ + 2 is
reached. The results from 0 to equilibrium are shown in Table 6-9
and Fig. 6-10.
The required relationship between total conversion of benzene and
space velocity
is given in columns 2 and 5. Asin Table 6-8 the space velocity is
based upon the
feed volume measured at 60°F and 1 atmosphere pressure.
‘
+Xo
,
0.3 Fi
vy,
ocTb4-fooe t nnavzlee rnsei on a 0.1
‘4 Space
10,000 00 3 2
% 010 0.20 0.30 0.40
Yp/F ~cu tt (hr) /Ib mole
10 20 30 40
Reactor volume (cu ft) for feed rate of 10,000 Ib/hr
Fia. 6-10. Dehydrogenation of benzene.
It is also required to determine the reactor volume required to
process 10,000
lb/hr of benzene. This may be computed from the Vr/F data in the
table.
For example, for a total conversion of 41.7 per cent, Ve/F = 0.12,
and the reactor
volume is
10,000
0.112 7— 8
15.4 cu ft
Ve = 0.12F
If a 4-in.-ID tubular reactor were used, a length of about 175 ft
would be required.
The complete curve of volume vs. total conversion is also shown in
Fig. 6-10.
The equilibrium conversion of 49.6 + 9.1 = 58.7 per cent and the
equilibrium
composition are shown as the values corresponding to Vr/F = ».
These
results could also have been obtained from Eqs. (F) and (G). Thus at
equilib158
rium 7; = f2 = 0, and
/ a ee (Ax — x9)(o01 + 22)
aaa 0.312
| dig oe Capea! § /
L(x, + £5)
and (1 — xi — 2)($21 — £9) 7 ere
./
Simultaneous solution of these two equations leads to values of x; =
0.496
and 2, = 0.091 as shown in Table 6-9.
TABLE 6-9. CONVERSION vs. Vr/F FOR THE DEHYDROGENATION OF
BENZENE
on
Ve Space Total Composition of mixture (molal)
r velocity | 2 Le conversion,
hr7! t, + 22 CoHe | Ci2Hi0 CisHis He
0 0 0 0 0 1.000} O 0 0
0.005 | 75,800 |0.0302)0.00014) 0.0303 0.970} 0.0150 | 0.00014 |
0.0152
0.010 | 37,900 |0.0586}0.00051 0.0591 0.941} 0.0288 | 0.00051 |
0.0298
0.020 | 18,950 |0.1105)0.00184) 0.112 0.888} 0.0534 | 0.00184 |
0.0571
0.040 9,500 |0.197 |0.0062 0.203 0.797} 0.0923 | 0.0062 |0.1047
0.060 6,320 |0.264 |0.0119 0.276 0.724] 0.1201 | 0.0119 {0.1439
0.080 4,740 |0.316 |0.0180 0.334 0.666] 0.140 | 0.0180 {0.176
0.100 3,790 |0.356 |0.0242 0.380 0.620) 0.154 | 0.0242 {0.202
0.120 3,160 |0.387 |0.0302 0.417 0.583] 0.163 | 0.0302 |0.224
0.140 2,710 |0.411 |0.0359 0.447 0.533| 0.170 | 0.0359 {0.241
0.180 2,110 |0.445 |0.0459 0.491 0.509] 0.177 | 0.0459 {0.268
0.220 1,720 |0.468 |0.0549 0.523 0.477| 0.179 | 0.0549 |0.289
0.260 1,460 |0.478 |0.0617 0.540 0.460} 0.177 | 0.0617 |0.301
0.300 1,260 |0.485 [0.0673 0.552 0.448} 0.175 | 0.0673 {0.310
0.400 950 |0.494 |0.0773 0.571 0.429} 0.170 | 0.0773 |0.324
co f 0 |0.496 |0.091 0.587 0.413} 0.157 | 0.091 0.339
{ Equilibrium conditions.
The rates of each reaction as a function of Vr/F are shown in Fig. 6-
11. In
both cases the rates fall off toward zero as Vr/F approaches high
values. The
rate of reaction 2 also approaches zero at very low values of Vr/F
because
dipheny] is not present in the benzene entering the reactor.
The last four columns in Table 6-9 give the composition of the
reaction mixture.
With a fixed feed rate the increasing Vr/F values correspond to
moving through
the reactor, i.e., increasing reactor volume. Hence these columns
show how the
composition varies with position in the reactor. It is observed that
the mole
fraction benzene decreases continuously as the reaction mixture
proceeds through
the reactor, while the diphenyl content reaches a maximum at a
Vr/F value of
wbout 0.22, This may be explained by the fact that diphenyl is
produced by
reaction 1 and disappears by reaction 2. At low conversions the rate
of reaction
2 is small, but at high conversion this rate becomes greater than
that of the first
reaction, so that diphenyl is destroyed more rapidly than it is
formed.
Example 6-4 illustrates the design procedure for a homogeneous
flow
reactor operating at constant temperature and pressure. When the
pressure
is not constant and when the rate of reaction varies significantly
with the pressure, the pressure drop in the reactor must be taken
into
account. The method of accomplishing this involves no new
principles
or procedures. Comparing such a case with Example 6-4, the
differences
are as follows:
1. The effect of pressure on the rate of reaction must be accounted
for,
i.e., rate equations (A) and (B) would be used rather than (F) and
(G),
which are based upon | atm pressure.
/of|rR1mctbfhe-oou a rtl c eet si on
oRrm2-fcfe|tao ub a tl ceet si/onh r
0
O 0.04 O08 O12 0.16 020 O24 028 032
Ve/F - cu ft (hr) / |b mole
Fic. 6-11. Rates of reaction, dehydrogenation of benzene.
2. The same stepwise numerical procedure is applicable, except that
the pressure at any point in the reactor must be computed and the
appropriate value used in Eqs. (A) and (B) to estimate average rates
of
reaction.
The changes in pressure are due to friction, potential-energy
changes,
and kinetic-energy changes in the flow process in the reactor. Hence
the
pressure drop may be computed from Bernoulli’s equation, using
suitable
auxiliary relationships, such as the Fanning equation for the friction.
Since the pressure drop is rather small for flow in empty pipes,
which is
the usual case for homogeneous reactors, the effect of pressure
sometimes
may be neglected. However, this is not invariably the case. Thermal
cracking of hydrocarbons is accomplished by passing the fluid
through
pipes which may reach several hundred feet in total length. In such
instances the pressure drop may be of the order of 100 psi at an
absolute
pressure of 500 psia, corresponding to a 20 per cent change from
entrance
to exit of the reactor.
6-3. Nonisothermal Design (Tubular Flow). It is not always possible
or desirable to operate a commercial-scale reactor at a constant
temperature.
High heats of reaction or poor heat-transfer rates may prevent
isothermal operation. In addition the equilibrium yield, rate, or both
may be more favorable if the heat of reaction is absorbed by the
reaction
mixture rather than transferred completely to the surroundings.
When the temperature varies in the flow reactor, the rate equation
must include the temperature effect before the design can be
accomplished.
In comparison with the isothermal case (Example 6-4) Eqs. (A)
and (B) must show the effect of temperature before they can be
used for
nonisothermal conditions. If the temperature is known in each part
of
the reactor, it can be substituted in the rate equations and then the
numerical procedure of integrating the design equations is the same
as
used in Example 6-4. The temperature as a function of Vz is
determined
by an energy balance around the reactor. This can be written in
general
form as follows:
rdVr(—AH) — dQ! = » Mily, at (6-3)
t=1
The first term represents the rate at which energy is evolved in the
reactor
volume dV as a result of the heat of reaction. The second term is the
rate of transfer of energy from the element, by heat transfer to the
surroundings.
The term on the right represents the effect of this net rate
of addition of energy on the temperature ¢ of the reaction mixture.
The
coefficient Lmic,, is the total heat capacity per unit time of the
material
flowing through the volume increment. The mass rate of flow of any
one component is ™,.
The heat transferred to the surroundings will depend upon the over-
all
heat-transfer coefficient U and the mean temperature difference (£
— be)m
between the surroundings and the reaction mixture. If Ap represents
the heat-transfer area per unit length of reactor (dz),
dQ, = UARk(t — t.)m dz
If A, represents the cross-sectional area, Eq. (6-3) becomes
rA,de(—AH) — UAp(i — inde = > Mil, at (6-4)
i=1
By noting that rdVp = F dx’ the energy balance can be written so
that
all reference to the rate or kinetics of the reaction is eliminated.
Thus
Kq. (6-4) may be expressed in the following way:
Port An) — UAi— t\e ds = > mil, dt (6-5)
t=1
The energy balances expressed in the form of Eqs. (6-4) and (6-5)
are
valid for any tubular flow reactor, provided steady-state conditions
are
obtained and provided radial (perpendicular to direction of flow)
temperature
variations are negligible. This last requirement means that the
reaction mixture at any longitudinal position in the reactor can be
assumed to be at an average temperature t.
If the operation is adiabatic, the heat exchange with the
surroundings
is zero and the energy balance reduces to
F dx'(—AH) = =mep, dt (6-6)
It is observed that this equation for an adiabatic reactor is
completely
independent of the reactor size and the reaction rate, in that it may
be
integrated separately to give a relationship between the
temperature and
conversion. Then the design equation can be integrated without
simultaneous
consideration of the energy balance.
On the other hand, for reactors where the heat-transfer dQ! is
important
the temperature is a function of the reactor length z as well as the
conversion, as indicated by Eq. (6-5). Under these conditions the
design
equation and energy-balance expression must be integrated
simultaneously
by stepwise numerical methods. It may be noted that these same
conclusions were valid for batch reactors and were illustrated in
Example
5-3, with the variable time replacing the reactor length.
The methods of designing adiabatic and nonisothermal tubular
reactors
will be illustrated by considering the substitution reaction between
propylene and chlorine to give allyl chloride. The reaction is
important
commercially because allyl chloride is a useful starting material for
production
of such substances as allyl alcohol, cyclopropane, epichlorohydrin,
trichloropropane, and allyl isothiocyanate.
Because of competing side reactions, particularly the additive
chlorination
to give dichloropropane, it is advantageous to operate at essentially
constant temperature as well as adiabatically. This may be
accomplished
by constructing the reactor in a form of tank to which reactants
are continuously added at a low temperature and products
continuously
withdrawn at a high temperature. In this case the excess thermal
energy
leaving in the products over that supplied with the reactants is
equivalent
to the heat of chemical reaction. However, to indicate the type of
calculations
involved in tubular nonisothermal reactors, a long-tube nonadiabatic
reactor involving considerable temperature change will be
considered first (Example 6-5). Then, after the general subject of
tanktype
reactors has been discussed, the constant-temperature adiabatic
reactor used commercially will be taken up (Example 6-6).
Example 6-5. It is proposed to design a pilot plant for the production

of allyl
chloride. The reactants will consist of 4 moles of propylene per mole
of chlorine
and will enter the reactor at 200°C. The reactor will be a vertical
tube, 2 in. ID.
If the combined feed rate is 0.85 Ib moles/hr, determine the
conversion to allyl
chloride as a function of tube length. The pressure may be assumed
constant
and equal to 29.4 psia.
The reactants will be preheated separately to 200°C and mixed at
the entrance
to the reactor. At this low temperature explosion difficulties on
mixing are not
serious. The reactor will be jacketed with boiling ethylene glycol so
that the
inside-wall temperature will be constant and equal to 200°C. The
inside-heattransfer
coefficient may be taken as 5.0 Btu/(hr)(sq ft)(F).
ADDITIONAL DATA AND NoTES. The basic development work for the
allyl
chloride process has been reported by Groll and Hearne? and
Fairbairn, Cheney,
and Cherniavsky.? It was found that the three chief reactions were
as follows:
1. Allyl chloride formation,
Cl. + C3Hs ~ CH.—CH—CH.Cl + HCl
2. Substitution to 1,2-dichloropropane,
Cl, + C;Hs — CH.CI—CHCI—CH:;
3. Further chlorination of allyl chloride to 1,3-dichloropropene,
Cl, + CH.—CH—CH.Cl — CHCI—CH—CH.Cl + HCl
In order to simplify the kinetic treatment of the problem, only the
first two
reactions will be considered. The heats of reaction (in calories per
gram mole)
available are:
Dial 355°K
Reaction AH eee — 26,800 — 26,700
Reaction 2, AH........ —44,100 — 44,000
The molal-heat capacities will be assumed constant and equal to the
values
given in Table 6-10.
TABLE 6-10
Cp, Btu/(lb mole) (°R)
Propylene (9) eevee areca PARES
Chlorine (gi SiaAe eee 8.6
Hydrogen chloride (g).......... 7.2
Allyl chloride (ty). ...2.4.+ae . 28.0
1,2-dichloropropane (g)......... 30.7
1 Ind. Eng. Chem., 31:1530 (1939).
? Chem. Eng. Prog., 43 :280 (1947).
No information is published for the rate-of-reaction equations,
although it is
known that reaction 2 takes place at temperatures as low as 100°C,
while reaction
1 has an insignificant rate below 200°C. As the temperature
increases above
200°C, the rate of reaction 1 increases rapidly, until at 500°C it is
several times
as fast as that of reaction 2. From this general information and some
published
data on the effect of residence time and temperature on the
conversion the following
rate equations! are proposed:
Reaction 1: 7, = 206,0270.2000 keT y-
Reaction 2: Py = 11.8760/8R T:D y y
where 11, 72 = lb moles Cl. converted/(hr)(cu ft)
T =°R
Pp = partial pressure C3He¢, atm
Poa = partial pressure Clo, atm
Solution. Since there will be heat transfer from the reaction gases to
the glycol
jacket, the temperature in the reactor will depend upon the length z
and the
design calculations have to be carried out by the stepwise
integration of the rate
and energy equations. As in Example 6-4 the conversion resulting
from each of
the two simultaneous reactions must be obtained separately.
Defining the conversion,
moles of chlorine reacted per pound mole of total feed, as x} for
reaction 1
and x4 for 2, the design equations are
r,dVr = F dr,
redVr = F dx,
Ree (2) (A)
(B)
or, in difference form,
/ Az, |= iy >
re|rae e e re
The energy balance defining the temperature is given by either Eq.
(6-4) or
Eq. (6-5). Choosing the former and writing it also in difference form,
Dmicp; At = —(Tiay AH1 + 72, AH 2) Ac Az + Ul(ts — t)mAr Az (C)
Along with the rate expressions these three equations can be used
to determine
x, t, and ¢ as a function of V r/F, or reactor length. The reactor
length is
related to the volume of reactor by the expression zA, = Vr; so
Vee 2A,
FF
or, in difference form,
Vr\ _ Ae A (2) Fp 4 (D)
1 Insufficient data are available in the literature to determine the
order of the
reactions. The second-order irreversible functions as well as the
numerical values
suggested by the equations are unverified assumptions.
This last relationship may be used to eliminate either Az or A(Vr/F)
from
Eqs. (A) to (C). The choice depends upon whether reactor length z or
Vr/F is
chosen to be the independent variable. If the former is taken,
expressions (A)
and (B) become,
A,
Az} = Tay Az (A’)
A. '
Ax} = Toay FP Az (B )
The solution procedure will be as follows:

1. From a knowledge of the initial conversion (x, = 0, x, = 0) and
temperature
(200°C) the rates 7; and r, entering the reactor can be computed.
2. Then an increment of reactor length is arbitrarily chosen. The
smaller this
increment, the more accurate will be the solution (and the more
time-consuming
the calculations).
3. For the chosen Az first estimates of the conversion occurring
within the
increment can be obtained from Eqs. (A’) and (B’). To do this, it is
convenient
to assume that the average values of the rates r,, and r2,, are equal
to the initial
values evaluated in step 1.
4. The change in temperature, At, within the increment can be
determined
from Eq. (C). In making this calculation an estimate of the mean
temperature
difference (t, — ) in the increment is required in order to evaluate
the heat-loss
term. Once At has been computed, the estimate of (t, — t)m can be
checked.
Hence, a trial-and error calculation is necessary in order to evaluate
Af.
5. From the conversion and temperature at the end of the first
increment, as
determined in steps 3 and 4, the rate of reaction can be computed
at this position
in the reactor. Then steps 3 and 4 can be repeated, using values for
ri,, and
ro,, based upon arithmetic average values of the rates at the
beginning and end
of the increment. This, in turn, will give more precise values of the
conversion
and temperature at the end of the first increment and permit a third
estimate of
the average values of the rates. If this third estimate agrees with
the second, the
next increment of reactor length is chosen and the procedure
repeated.
The steps will be illustrated by carrying out the numerical
calculations for the
first two increments. The rate equations should be converted to a
form involving
the conversions rather than partial pressures. Four moles of
propylene per mole
of chlorine enter the reactor. Then at a point where the conversions
are 2;
and 2, the moles of each component will be as follows:
Na (chlorine)t} = 1 — 52, — 5x5
N, (propylene) = 4 — 5x} _ ay
N 4 (allyl chloride) = 52}
N p (dichloropropane) Bas
Nu (hydrogen chloride) = 52}
Total = 5(1 — 24)
I
} The coefficient 5 appears in this equation because 2’ is the
conversion based upon
« 1A 7 / ° 1 mole of total feed. Note that the maximum value for the
sum of zx, and x, is 0.2,
as this corresponds to complete conversion of the chlorine to either
allyl chloride or
dichloropropane.
Assuming that all the components behave as perfect gases at p, =
29.4 psia,
the partial pressures of chlorine and propylene (in atmospheres) are
given by
_ 29.44 — 5x, — 524 x 08 = 81 = 2
Pia, a ee ar
_ 29.41 — 52, — 524 0.2 — x, — 2
on DER, oa ea Ada let tt
These expressions can be substituted in the rate equations stated in
the problem
to relate the rates to the conversions and the temperature. The
results are
1 — %)(0.2 — x, — 24)
(1 — 24)?
ry = 824,000e—13.700/T Stel (E)
and
A6.8e73:460/T (0.8 — x} — 25)(0.2 — x} — 24) a — 2) (F) Te
Equations (A’) to (F) can be employed to carry out the previously
enumerated
steps in the solution.
1. The rates of reaction at the reactor entrance are given by Eqs. (E)
and (F),
with T = (200 + 273) X 1.8 = 852°R and x; = 2 = 0.
fy = 824,000e—!1*-! < 0.16
= 0.0135 lb mole/(hr)(cu ft)
f2 = 46.8¢e—4-° & 0.16
= 0.129 lb mole/(hr) (cu ft)
2. Choose an increment of reactor length, Az = 4.0 ft.
3. Assuming that the rates computed in step 1 are average values
for the
increment, the first estimates of the conversion in the increment are
given by
Eqs. (A’) and (B’).
i 2h a (=)
= 0.0218 sq ft
tA.e e 0p.0e21 8 =, 00.02
F 0.85 ee
Az, = 0.0135 X 0.0257 x 4.0 = 0.0014
z, = 0+ 0.0014 = 0.0014
In a similar manner for the second reaction
Ar, = 0.129 X 0.0257 X 4.0 = 0.0133
2, = 0 + 0.0133 = 0.0133
4. To use Eq. (C) for estimating At for the increment, it is necessary
to know
AH, AH:, t,, Ar, Ac, U, Zmicp, and (t, — t)m. The heat-of-reaction
data given
are for low temperatures but show little change with temperature.
Hence AH;
and AH, will be assumed constant.
AH, = —26,700 X 1.8
—48,000 Btu/lb mole
AH, = —44,000 X 1.8
= —79,200 Btu/lb mole
T, = (200 + 273) X 1.8
= 852°R
Ar =7D X1=1%5X1Xe
= 0.524 sq ft/ft length
A, = 0.0218 sq ft
h = 5 Btu/(hr)(sq ft) (°F) Tf
YmMicp, is the heat capacity per unit time of the reaction mixture.
Since the conversion
is small during the first increment, this heat capacity is essentially
that
of the chlorine and propylene in the feed to the reactor,
ZMNily = “= 8.6 + : X 0.85 X 25.3
= 0.85 X 21.8 = 18.5 Btu/(hr) (°F)
If the mean temperature difference (t, — t)m for the first increment
is estimated
to be —20°F,
18.5 At = —(0.01385 X —48,000 + 0.129 X —79,200) x 0.0218 x 4
+5 X —20 X 0.52x4 4
ere 946 — 210
18.5
= 40°F
T, = temperature at end of first increment = 852 + 40
7 ROD?
t, — t at entrance = 0
t, — t at end of increment = 852 — 892
= —40°F
(iio Ost (40)
2
= —20° vs. —20° assumed
tf; = 852 + 40
5502
5. At the end of the first increment, the first estimate of the
conversions and
temperature is
a, = 0.0014
ty = 0.0133
t; = 852 + 40
= 892°R
} Since the inside-wall temperature is known, the heat transfer with
the surroundings
can be evaluated from the expression h(ts — t)mAr Az, where h is
the inside-film
Substitution of these conditions in rate equations (E) and (F) yields
71 = 0.0256
T2 = 0.143
A second, and more accurate, estimate of the average values of the
rates for the
first increment can now be made.
_ 0.0135 + 0.0256
lay — 5
_ 0.129 + 0.143
Ty = es = (IRIE
= 0.0195
A second estimate of the conversion occurring in the increment is
obtained from
these rates and Eqs. (A’) and (B’).
Ar, = 0.0195 X 0.0257 x 4.0 = 0.0020
z, = 0 + 0.0020 = 0.0020
Ars = 0.136 X 0.0257 X 4.0 = 0.0140
zo = 0+ 0.0140 = 0.0140
The second estimate of At is determined from Eq. (C), as in step 4.
18.5 At = —(0.0195 X —48,000 + 0.136 X —79,200) x 0.0218 x 4
+ 5 X —20 X 0.52x4 4
At = 41°F
T; = 852 + 41
= 893°R
At the revised values of x}, ry, and ¢ (at the end of the first
increment) the rates
are
rT, = 0.0258
tT. = 0.144
Since these values are essentially unchanged from the previous
estimate, the
average rates will also be the same and no further calculations for
this increment
are necessary.
For the second increment choose again Az = 4 ft, and estimate the
average
rates to be
T1,, = 0.032
Wes Qhalls:
Then from Eqs. (A’) and (B’),
Ar, = 0.032 X 0.0257 xX 4 = 0.0033
zx, = 0.0020 + 0.0033 = 0.0053
Ax} = 0.15 X 0.0257 X 4.0 = 0.0154
x, = 0.0140 + 0.0154 = 0.0294
coefficient and t, is the inside-wall surface temperature. The over-all
coefficient U
is not needed in this case.
From Eq. (C), estimating (t, — t)m = —55°F,
18.5 At = —(0.032 X —48,000 + 0.15 X —79,200) X 0.0218 x 4
+5 xX —55 X 0.524 x 4
At = 32°F
T, = 893 + 32
= 925°R
The estimated (t, — t), for the second increment can be checked by
evaluating
the temperature difference at the beginning and end of the
increment.
t; —¢t = 852 — 893
= —41°F beginning of increment
t, —t = 852 — 925
= —73°F
(t, — t)m = —57°F vs. —55°F assumed
The conversions and temperature, according to the first estimate, at
the end of
the second increment are
t, = 0.0053
ry = 0.0294
T, = 925°R
Using these quantities in the rate equations,
r, = 0.0400
tT. = 0.145
A second estimate of the average rates during the increment is
given by the
expressions
Ti = e0.025e e0.0 400 = 0.0329
0.144 + 0.145
LS aR SR aia 0.144
Since these values are close to the first estimates, additional rate
calculations
for the increment are unnecessary. The revised conversion values
leaving the
increment are
2, = 0.0020 + 0.0329 x 0.0257 x 4.0 = 0.0054
vo = 0.0140 + 0.144 x 0.0257 x 4.0 = 0.0288
The revised temperature is given by the equation
ZmMicp, At = —(0.0329 X —48,000 + 0.144 xX —79,200) x 0.0218
x4
+5 X —57 x 0.524 Xe
Since the total conversion is becoming significant at the end of the
second
increment, it may be necessary to evaluate Ymjc,, using the actual
mixture
existing in the reactor rather than assuming it equal to that for the
feed. The
rate of flow of each component at the end of the second increment
will be
Allyl chloride = Fr, = 0.85 X 0.0054 = 0.0046 moles/hr
Dichloropropane = Fry = 0.85 X 0.0288 = 0.0245
Chlorine = F(0.2 — x, — x4) = 0.85 X 0.1658 = 0.1410
Propylene = F(0.8 — x; — zy) = 0.85 X 0.7658 = 0.6510
HCl = Fx; = 0.85 X 0.0054 = 0.0046
Dmicp, = 0.0046 X 28 + 0.0245 x 30.7 + 0.1410 x 8.6
+ 0.6510 X 25.3 + 0.0046 x 7.2
= 18.6 Btu/(hr)(°R)
This result turns out to be not appreciably different from that based
upon the
feed but will become so when the conversion reaches larger values.
Hence
536
= 29°F
and the revised temperature leaving the increment is
T2 = 893 + 29
= 922°R
The results of these calculations indicate two trends. First, the rate
of reaction
1 is relatively low with respect to reaction 2 at low temperatures but
increases
rapidly with temperature (from 0.0258 at the end of the first
increment to 0.0400
at the end of the second), while that of reaction 2 is not very
sensitive to temperature
changes. This conclusion, considered by itself, would suggest that
there would be some conversion to dichloropropane at the
beginning of the reactor.
However, as the heat of reaction raised the temperature, the
reaction to
allyl chloride would become increasingly important. The second
trend is the
increase in heat-transfer rate to the surroundings with an increase
in reactor
length. This is due to the increase in (¢ — ts)m. Its effect is to
combat the rise
in temperature due to the exothermic nature of the reactions and
results ultimately
in a decrease in temperature with reactor length. This point is
reached
when the heat transferred from the reactor tube to the surroundings
is greater
than the heat evolved from the reactions and would show up in the
calculations
in the form of a negative At determined from Eq. (C). From the
results of additional
incremental calculations (Fig. 6.12 and Table 6-11) it is observed
that
the point at which the temperature starts to decrease is about 18 ft
from the
entrance to the reactor. The temperature vs. length profile has a
maximum
value at this point and has the shape shown in Fig. 6-13.
This maximum temperature occurs at too low a level for the rate of
reaction 1
to become large with respect to that of reaction 2. Therefore the
conversion to
allyl chloride never reaches a high value, most of the reaction being
to the dichloropropane.
A reactor of this type is not well suited for the production of allyl
chloride. An adiabatic reactor would eliminate the maximum in the
temperature
ys. reactor-length curve and would result in higher temperatures.
This in turn
would favor the production of allyl chloride at the expense of
reaction 2. The
——— Adiobotic
Non-odiobatic
Complete conversion of
chlorine in 21 ft long
reactor (x; + x2 =/0)
Conversion to dichloropropane - Xo
Cocoxx-fohr 2 ln ovreirnsei on
Fic. 6-12. Conversion curves, tubular reactor for allyl chloride
production.
ple 6-5.)
a
O
Pee? Conversion to allyl chloride - x,
O
4
16
Reactor length -ft
(Examonly
change in the calculations for such an adiabatic case is that the
term
U(t; — t)mA, Az would be zero. For comparison purposes the
temperature and
conversion results for such an adiabatic reactor are also shown on
Figs. 6-12 and
6-13.
TaBLE 6-11. CoNVERSION vs. Reactor LENGTH FOR CHLORINATION
OF PROPYLENE
Reactor |} Reactor Conversion per mole of feed Conversion per mole
of Cl.
volume, | length, ae
cu ft ft zy Lo Total
0 0 0.0 0.0 0.0 852
0.087 4 0.0020 | 0.0140 | 0.0160 893
0.174 8 0.0054 | 0.0288 | 0.0342 922
0.261 12 0.0098 | 0.0435 | 0.0533 940
0.349 16 0.0147 | 0.0575 | 0.0723 949
0.436 20 0.0193 | 0.0698 | 0.0891 949
Adiabatic Operation
0 0 0.0 0.0 0.0 852
0.087 4 0.0023 | 0.0144 | 0.0167 905
0.174 8 0.0088 | 0.0315 | 0.0403 981
0.261 12 0.0283 | 0.0505 | 0.0788 | 1,093
0.349 16 0.0779 | 0.0667 | 0.1446 | 1,261
0.436 20 0.1204 | 0.0739 | 0.1943 | 1,380
21 X2 Total
0.0 0.0 0.0
0.0100 | 0.070 | 0.080
0.0270 | 0.145 0.172
0.0490 | 0.217 0.266
0.0735 | 0.287 0.361
0.0964 | 0.349 0.445
0.0 0.0 0.0
0.0115 | 0.0720 | 0.0835
0.0440 | 0.157 0.201
0.142 0.252 0.394
0.389 0.334 0.723
0.602 0.369 0.971 ed ee ee a
Even adiabatic operation results in the formation of considerable
amounts of
the undesirable dichloropropane. This occurs in the first part of the
reactor,
where the temperature of the flowing mixture is low. The low-
temperature see HOMOGENEOUS
FLOW REACTORS 171
Non-odiabatic operation
—-—-— Adiabatic operation
R°te-Re a mcpteorra ture
0 4 8 12 16 20 24 28
Reactor length - ft
Fic. 6-13. Temperature profiles in tubular reactor, production of allyl
chloride.
(Example 6-5.)
tion can be eliminated by using a tank reactor operated under
isothermal adiabatic
conditions. This type of equipment is considered in the next section.
TANK FLOW REACTORS!

In the flow reactors considered up to this point it has been supposed
that no mixing occurs in the direction of flow. This assumption is
justified
in any vessel where the forced velocity in the direction of flow is
large with respect to secondary mixing velocities, due to entrance-
velocity
gradients, natural convection, or turbulent diffusion. In reactors
consisting
of long tubes, such as those studied in Examples 6-2 to 6-5, the
velocity in the direction of flow is usually high enough to prevent
appreciable
mixing along the length of the tube.” In large-diameter (several
feet) reactors longitudinal mixing is still not significant as long as
the
rate of movement in the direction of flow is high.
On the other hand, when the reactor consists of a vessel whose
diameter
is of the same order of magnitude as its length and when the flow
rate is
low, considerable mixing may occur. For example, suppose a reactor
10 ft long and 5 ft in diameter has a 1-in.-ID inlet line. If the velocity
in the inlet line is 100 ft/sec, that in the reactor proper will be only
1 This is the third type discussed in the classification in Chap. 4 and
illustrated in
Fig. 4-1B.
2 Note, however, that in all such homogeneous reactors complete
radial mixing
(mixing across the tube, perpendicular to the direction of flow) has
been assumed.
This assumption would not be valid in streamline flow but is
reasonable under turbulent-
flow conditions in long, small-diameter tubes.
100 (@s)2 or 0.03 ft/sec (neglecting temperature, pressure, and
volume
changes of the reaction mixture). As a result the inlet line will act as
a
jet causing turbulence in the reactor. The design of a reactor under
these circumstances of partial mixing is difficult because the extent
of
mixing cannot be predicted at present.
The most common example of reactors where mixing is important is
the tank-flow, or continuous-stirred-tank, type commonly used for
liquid
systems.! With properly designed mechanical stirring equipment
essentially
complete mixing can be obtained. Under these conditions the
reactor-design problem is greatly simplified, because the mixing
results
in uniform composition, temperature, and pressure throughout the
reactor.
This means that the rate of reaction is constant and that the design
equations do not require integration. This is true regardless of the
mechanism of the reaction and may be expressed in the simple
manner
illustrated by Eq. (4-12).
In summary, satisfactory design methods are available for
conditions
of no longitudinal mixing (as illustrated in the previous examples of
tubular flow reactors) and also for conditions of complete mixing as
will
be shown in Examples 6-6 to 6-8. For partial mixing, design
procedures
have yet to be developed.
6-4. Comparison of Tank Flow and Tubular Flow Reactors. The tank
flow reactor has certain advantages because of the uniform
temperature,
pressure, and composition attained as a result of mixing. First of all
it
is possible to operate such a reactor under isothermal conditions
even
when the heat of reaction is high—an impossibility in a long, tubular
type. When operation within a small temperature range is desired,
for
example, in order to minimize side reactions or avoid unfavorable
rates,
the opportunity for isothermal operation at the optimum
temperature is
a real advantage. Thus in the manufacture of allyl chloride
considered
in Example 6-5 the tube-type reactor results in the production of
considerable
amounts of undesirable dichloropropane because the temperature
is low in part of the reactor. In the tank type equipment considered
in Example 6-6 this disadvantage is overcome by isothermal
operation at an optimum temperature level.
Stirred-tank reactors, by virtue of their large volumes (and therefore
large Vz/F values) represent an economical method of providing for
a
long residence time. This combined with the isothermal nature of
the
reactor permits operation at the optimum temperature for a long
reaction
time.
Against these advantages are a number of adverse features.
Efficient
stirred-tank reactors are difficult to construct for gaseous systems
because
'The terms tank flow and continuous stirred tank are both used to
designate the
steady-state tank-type reactor.
of the mixing problem. Fixed baffles and mechanical stirrers can be
used,
but these do not ensure complete mixing in gaseous-reaction
mixtures.
Hence stirred-tank equipment is generally restricted to liquid
systems,
although the allyl chloride reactor is an exception.
For high-pressure reactions it is usually necessary, because of cost
considerations,
to use small-diameter tubular reactors rather than tank types.
To operate the latter at high pressures requires a large wall
thickness
and also necessitates complex sealing arrangements for the mixer
shaft.
These factors increase both initial and maintenance costs.
The rate of heat transfer per unit mass of reaction mixture is
generally
lower in the conventional tank-type than in a small-diameter tubular
reactor. Two factors are primarily responsible: (1) a lower ratio of
surface
area (available for heat transfer) to volume in the tank reactors;
(2) lower heat-transfer coefficients. In instances where the heat of
reaction
is high it may be desirable for this reason to use a tubular reactor.
For example, in the thermal cracking of hydrocarbons it is necessary
to
supply significant amounts of thermal energy at an elevated
temperature
level. This would be difficult to accomplish with a large-diameter
reactor
because of the limited external heat-transfer surface (per unit mass
of
reaction mixture) and the low coefficient of heat transfer from the
oil in
the tank to the tank wall. In the tubular reactors (pipe stills) used in
industry the coefficient of heat transfer can be increased by forcing
the
oil through the tubes at a high velocity. It is also apparent that
severe
difficulties would arise in attempting to provide for efficient stirring
under
cracking conditions (600 to 900°F, 300 to 800 psia). It is possible to
approach the complete-mixing characteristic of a tank-type reactor
in
other equipment than a tank-shaped vessel. For example, a tubular
loop
operated at high circulation rates can give complete mixing in a
smalldiameter
tube. In shape this reactor is similar to a tubular flow reactor,
and high rates of heat transfer can be attained. By introducing small
solid particles, free to move, a type is obtained in which there is
considerable
mixing. Such fluidized-bed reactors also give greatly improved
heat-transfer coefficients between the fluid and the wall.
Finally it should be noted that in stirred-tank equipment the reaction
occurs at a rate determined by the composition close to that of the
exit
stream from the reactor. If mixing is complete, the composition in
the
reactor is the same as that in the exit stream. Since the rate
generally
decreases with the extent of conversion, the tank reactor operates
at the
lowest point in the range between the high rate corresponding to
the
composition in the reactor feed and the low rate corresponding to
the exit
composition. In the tubular type maximum advantage is taken of
the
high rates corresponding to low conversions in the first part of the
reactor.
Practically, this means that the tank reactor would have to have a
larger
volume for a given feed rate (i.e., larger Vz/F value). Of course, this
reasoning does not take into account the effects of side reactions or
temperature
variations. These may offset this disadvantage of the tank
reactor, as illustrated in the production of allyl chloride. Also, the
total
volume required in a tank flow reactor can be reduced by using
several
small units in series.
The points that have been discussed indicate why stirred-tank flow
reactors have been used mainly for liquid-phase reaction systems at
low
or medium pressures. They can be used when the heat of reaction is
high, but only if the temperature level obtained in their isothermal
operation
is satisfactory from other points of view. In general if the reactions
involved are endothermic and a high temperature is required,
tubular
reactors are indicated. On the other hand, a tank type may
frequently
be employed for a highly exothermic reaction. For example, the
production
of hexamethylenetetramine by reacting ammonia and formaldehyde
(in aqueous solution) is highly exothermic, but the rate of reaction
is rapid and 100 per cent conversion is possible over a range of
temperature
of at least 80 to 100°C. Hence, by adjusting the rate of feed and
reactor volume, it is possible to add the feed at 20°C and remove
enough
heat to maintain the reaction mixture at below 100°C.
Denbigh and coworkers in a series of papers! have discussed the
advantages
and disadvantages of continuous-stirred-tank reactors, especially in
comparison with batch-operated tank reactors. In another paper
Stead,
Page, and Denbigh? describe experimental techniques for
evaluating rate
equations from stirred-tank data.
6-5. Design Procedures for Single Tank Flow Reactors. The design
concepts for tank flow reactors were developed in Sec. 4-4 and
resulted in
Kq. (4-12). As written there and repeated as Eq. (6-7), the feed and
conversion can be expressed in either mass or molal units.
Fay — ty) =1Ve (6-7)
where F = total feed rate, mass/unit time (also, F equals exit mass
rate)
x’ = conversion of reactant, mass of reactant/mass of total feed
(the subscript F refers to feed stream and E to exit stream)
r= constant rate of reaction at uniform temperature, pressure,
and composition of reaction mixture [the rate is based upon
same reactant as the conversion and has the units (mass of
reactant converted)/(unit volume) (unit time)]
I
‘K. G. Denbigh, Trans. Faraday Soc., 40:352 (1944), 43:648 (1947).
K. G,
Denbigh, M. Hicks, and F. M. Page, T’rans. Faraday Soc., 44:479
(1948).
aa Stead, F. M. Page, and K. G. Denbigh, Discussions Faraday Soc.,
(2):263
i
It has been mentioned that tank flow or continuous-stirred-tank
reactors
are most commonly employed for liquid systems. In such cases the
density changes are often negligible so that the feed and exit flow
rates
are equal in volume as well asin mass units. Then Eq. (6-7) can be
written
in a simple form in terms of the concentrations (moles per volume)
of
reactant in feed and exit streams. The volumetric flow rate of feed
and
exit stream is Pvp, where F is the molal flow rate and vp the volume
of the
feed per mole. The moles of reactant entering per unit time is (Fur)C
F;
and leaving, FurCz. Then, according to Eq. (4-1),
(Fur)Cr — (Fur)Cz = rVp
rV or Cr — Cz = Fup (6-8)
Eldridge and Piret' have developed design equations based upon
constant
density for single and multiple tank flow reactors.
Equation (6-8) gives the exit concentration Cg as a function of the
reactor volume. The ratio Vg/Fv, is the holdup time in the reactor
and
has the physical meaning of the average time a slug of liquid is in
the
vessel. The reciprocal, Fur/Vz is the familiar space velocity as
defined
by Eq. (4-13).
When the temperature in the reactor is the same as that of the
entering
reactants, Eq. (6-7) [or Eq. (6-8) for constant-density liquid systems]
can be used directly to determine the reactor volume required to
obtain
a given conversion x;. This will be true when (1) the heat of reaction
is
egligible with respect to the heat capacity of the reactor contents or
(2)
here the heat of reaction is significant but the heat-transfer rate
with
he surroundings is sufficient to balance the heat of reaction.
When neither of these conditions is fulfilled, an energy balance must
e made over the system in order to determine the reaction
temperature
nd, therefore, the value of the rate that is to be used in Eq. (6-7). If
, represents the temperature of the entering reactants and é: that of
the
eaction mixture, the energy balance may be written
Lepli(ti a to) aie rVp(—AH) = UMC p;(te = to) =i UA (te a A) Be (6-9)
he left-hand term represents the thermal energy added in the feed
bove an arbitrary base temperature ¢); the second term? gives the
nergy evolved as a result of the heat of reaction. The right-hand
rms represent the energy removed from the system. The first is the
1J. W. Eldridge and FE. L. Piret, Chem. Eng. Prog., 46 :290 (1950).
2The minus sign preceding AHp is necessary, because AH»
represents the energy
bsorbed by the reaction, not the energy evolved. Note that AHo is
the heat of
eaction at lo.
energy removed with the products leaving the system.' The second
represents the heat transfer from the reactor to the surroundings.
The summation terms in the equation correspond to the heat
capacity
per hour of the feed and the products; F’; is the mass feed rate of
component
7 in the feed and m; the mass rate of product 7 leaving the reactor.
The surroundings temperature is ¢,, and U and A are the heat-
transfer
coefficient and area.
Equation (6-9) may be simplified by choosing the base temperature
fy
equal to the feed temperature ¢; or reactor temperature fs, since
each
choice eliminates one term of the equation. If fo is chosen equal to
ty,
the necessity of evaluating Dm,cp, is avoided, and the energy
balance
becomes
LepFi(t: — t2) + rVe(—AH) = UA(he — A= (6-10)
Equations (6-7) and (6-10) are sufficient to complete the design of
the
reactor, i.e., to determine the exit temperature and the reactor
volume
required for a given conversion. As Eq. (6-10) is written, a trial-
anderror
solution of the two equations is necessary, since the unknown
volume
and temperature are both involved in each equation. However, if
Eq. (6-7) is used to eliminate the term rVz, a simultaneous solution
is
avoided.2. Thus Eq. (6-10) may be written
Lepii(t: — te) + F(ae — 2p)(—AH) = UA(te — te)m (6-11)
This expression can be solved directly for the exit temperature ¢,
from a
knowledge of the conversion required in the reactor. Then the rate r
can be evaluated at this temperature f2 and used in Eq. (6-7) to
compute
the reactor volume.
On the other hand, if the reactor volume is known and the exit
conversion
x is the desired quantity, a trial-and-error simultaneous solution
of the design equation and energy balance is required regardless of
the form of the energy balance. However, it is still more convenient
to use Eq. (6-11) rather than (6-10). A method of approach is to
assume
a series of exit conversion values, x;, and compute the
corresponding exit
temperatures fz from Eq. (6-11). Then for each x} and fz evaluate r,
and
use it in Eq. (6-7) to determine a reactor volume. When the
computed
volume agrees with the known volume, the temperature and
conversion
are correct. This variation of the design problem would be
encountered
when the performance of an existing reactor is to be predicted.
Numerical illustrations of stirred-tank flow reactors are illustrated in
Kxamples 6-6 and 6-7,
' The reaction-mixture temperature tz is equal to the temperature of
the exit stream
because of the assumption of complete mixing.
pert ce . ; ‘
* When only one reaction occurs. With reversible, simultaneous, or
consecutive
reactions a simultaneous solution may be required.
Example 6-6. Consider the design of a tank-type flow reactor for the
production
of allyl chloride from propylene, using the reaction-rate data given
in Example
6-5. In order to compare the different types of reactors for this case,
the same
feed condition will be employed: total rate = 0.85 mole/hr,
temperature = 200°C,
pressure = 29.4 psia; molal ratio of propylene to chlorine = 4.0.
The equipment should be insulated so that the operation is
essentially adiabatic.
Also, suitable baffles and entrance nozzles will be used so that the
contents
of the reactor, though gaseous, will be of uniform temperature,
pressure, and
composition.
Using the heats of reaction and heat capacities given in Example 6-
5, determine
the conversion of chlorine to allyl chloride expected for a range of
sizes of
reactor (i.e., reactor volumes).
Solution. The rate of each reaction (allyl chloride and
dichloropropane formation)
will be a constant and should be evaluated at the temperature and
composition
of the stream leaving the reactor. To determine this temperature,
the
energy balance Eq. (6-11) will be employed. Since the operation is
adiabatic
and the entering conversion zero, this expression may be written
ZF icp,(ti — te) + Flx}(—AHi) + 23(—AH.)] = 0 (A)
where x; and «ry refer to exit conversions for reactions 1 and 2.
If only one reaction were involved (for example, if 2, = 0), this
energy balance
could be used to calculate an exit temperature f. for a series of
arbitrarily
shosen conversions x}. Then each value of az, would fix a rate 71.
These rates,
substituted in the design equation (6-7), would give the required
reactor volumes
sorresponding to the conversions. In the present case the procedure
is complisated
by the existence of the two simultaneous reactions. Two equations
of the
form of (6-7) must be satisfied, i.e.,
V
ti —-O=7 oa (B)
ee ete (C) Vs = 712 P
The rate expressions are obtained by substituting 7’. = t. + 460 for
the temerature
in Eqs. (E) and (F) of Example 6-5,
(0.8 — x; — ©2)(0.2 — 2, — 29)
Suath —13,700/T2 : D r; = 824,000e Gd — 21) (D)
(0.8 — x; — 23)(0.2 — x; — x4) = —3,460/T2 = E rz = 46.8¢ a — 2)
(E)
The essence of the problem is the solution of the five equations (A),
(B), (C),
D), and (FE) for the five unknowns 21, 9, ts, 71, and ry at different
values of the
actor volume Vz. One procedure which is not tedious is to choose a
value of
Then from the ratio of Eqs. (B) and (C), using (D) and (E) for r; and
rq,
btain the ratio of the conversions 2;/z5. Employing this ratio in Eq.
(A) will
ive separate values for each conversion. Finally the corresponding
reactor
olume can be obtained from either Eq. (B) or Eq. (C). This approach
will
be illustrated by including the numerical calculations for an exit
temperature of
1302°R (450°C),
Lo I's 46.8e7—3) 460/1, 302
4 1 824, 000e= 18) 700/2,308
= 6.77
Using this ratio in Eq. (A), and noting that the hourly heat capacity
of the
feed SF cp, was determined as 18.5 Btu/(hr)(°F) in Example 6-5,
18.5(t, — te) + 0.85(6.772, X 48,000 + 2 X 79,200) = 0
18.5(852 — 1,302) + 343,0002, = 0
ty = 0.0243
z, = 6.77 X 0.0243 = 0.164
Since the feed rate is 0.85 lb moles/hr, the reactor volume required
for these
conversions is, from Eq. (B),
Vz ' (0.8 — 0.164 — 0.024)(0.2 — 0.164 — 0.024)
0.164 = 9 = 824,000¢ 13,700/1,302 @ — 0.0243)
y, — 0-164 X 0.85
teeta tGy
= 0.83 cu ft
The corresponding values of the conversions and volume for other
temperatures
are summarized in Table 6-12. In comparing the results with those
of Example
TABLE 6-12. CONVERSION vs. REACTOR VOLUME FOR ADIABATIC
TANK REACTOR:
ALLYL CHLORIDE PRODUCTION
Reactor (and Conversion /mole Conversion of
exit) temp of feed chlorine in feed esner
vol.,
Cutt
ny al ry Lo ry Xe
960 | 260 | 0.0098 | 0.0237 | 0.049 | 0.119 | 0.12
1032 300 0.0282 0.0324 0.14 0.162 O15
1122 350 0.0660 0.0341 Ons OR At 0.18
PAP 400 0.114 0.0298 OLo7 0.149 0.24
1257 425 0.138 0.0273 0.69 0.136 0.34
1302 450 0.164 0.02438 0.82 O72 0.83
——$LL—ii——————————————————————
6-5 it is seen that the adiabatic tank reactor is better adapted to the
productior
of allyl chloride than the tubular type. In the latter equipment
considerable
dichloropropane is formed in the initial sections of the reactor,
where the tem
perature is relatively low. This is avoided in the adiabatic tank
reactor b>
operating at a constant temperature high enough to favor allyl
chloride rathe:
than dichloropropane formation. For example, if the adiabatic tank
reactor i
operated at 450°C, 0.164 X 5 X 100, or 82 per cent, of the chlorine
is convertee
to allyl chloride and 0.0243 X 5 X 100, or 12.1 per cent, converted to
dichloro
propane. The total conversion is (0.0243 + 0.164) x 5 x 100, or 94.1
per
cent. In the adiabatic tubular reactor of Example 6-5 the products
contained
much greater amounts of dichloropropane for all reactor volumes.
These conclusions
are summarized in Fig. 6-14 where the fraction of the chlorine
converted
to each product is shown plotted vs. reactor volume for both the
tubular and the
tank reactor.
Tank reactor, adiabatic (Example 6)
——-—— Jubulor reactor, adiabatic (Example 5)
0.1 0.2 0.3 04 0.5 0.6 0.7 0.8 09
Reactor volume -cu ft
Fra. 6-14. Comparison of tubular and tank reactors for allyl chloride
production.
Example 6-7. Eldridge and Piret! have investigated the hydrolysis of

acetic
anhydride in stirred-tank flow reactors of about 1,800 cc volume at
temperatures
from 15 to 40°C. Their experimental results for several volumetric
feed rates
Fvy are shown in Table 6-13.
TABLE 6-13
ar pee a n ee ee
Entering anhydride | Volumetric feed | % hydrolysis | Temp,
conc., Cr, moles/ce | rate, Fur, cc¢/min | of anhydride °C
i ie a 378 25.8 15
aeoce Oma 582 Son L 25
1:37:10“ 395 40.8 25
1,08 10s 555 15.3 10
O.52ex% 105% 490 16.4 10
OCOD eel ae 575 50.0 40
07025)" 10" 540 55.7 40
ihe cwd &,¢ AWE 500 58.3 40
2:02 X10-* 88.5 88.2 40
eae ees Se
1 Loc. cit.
Independent determination of the rate of reaction by these
investigators, from
other experiments in batch-operated reactors resulted in the
following first-order
equations:
At 10°C
r = 0.0567C, g moles/(cc) (min)
At 15°C
r = 0.0806C, g moles/(ec) (min)
At 25°C
r = 0.1580C, g moles/(cc)(min)
At 40°C
r = 0.380C, g moles/(cc) (min)
where C = acetic anhydride concentration in gram moles per cubic
centimeter.
For each run compute the per cent hydrolysis, and compare with the
observed
value given in the table of data. In all cases the feed temperature
was the same
as the temperature of the reaction mixture. Since the
concentrations are low,
the density of the solution may be assumed constant with little
error.
Solution. The hydrolysis reaction,
(CH;CO).,0 + H.O — 2CH;COOH
would be expected to be second-order. In this case, in which dilute
aqueous solutions
were employed, the water concentration is essentially constant so
that a
first-order equation is satisfactory.
Since the process is isothermal, an energy balance need not be
employed.
Also, Eq. (6-8) is applicable because the reaction mixture is a
constant-density
liquid phase.
Substituting the rate equation r = kC into Eq. (6-8)! and solving the
resulting
expression for the exit concentration Cz,
ioe Cr Se eCar
* 1+k(V3/For) 1+ 0
where @” is the holdup time.
The fraction of the anhydride hydrolyzed is equal to (Cr — Cz)/C'r; so
Cre j 1
Frarcatciotni on hhyydrdorloyzleydz ec = 1 — C—, = 1 — —a—y——
(A)
Equation (A) can be employed to determine the per cent hydrolyzed
(or conversion)
from 6’ and the reaction-rate constant. Thus, for the first set of
datav
in the table, at 15°C,
k = 0.0806 min“!
9’ 1,800.
my Fr a
' 1,800 : ko’ = 0.0806 378 = 0.384 (dimensionless)
: Note that this concentration of anhydride in the rate equation is
the same as the
exit concentration C in Eq. (6-8), because of mixing.
Substitution in Eq. (A) gives
oegnho e
1 + 0.384
= 0.277, or 27.7%
Fraction hydrolyzed = 1
This agrees reasonably well with the experimental value of 25.8 per
cent.
Table 6-14 shows the results of similar calculations for the other
sets of data.
TABLE 6-14
hyd is
Vol. feed rate, % hydrolysis
ce/min Experimental | Calculated from Eq. (A)
378 25.8 27.7 st east 32.8
395 40.8 41.8
555 15.3 15.5
490 16.4 17.2
575 55.0 54.4
540 Sy | 55.9
500 58.3 57.8
88.5 88.2 ihe
The computed values are based upon the concept of complete
mixing and,
hence, upon the assumption of uniform composition and
temperature throughout
the mass. The agreement between the calculated results and the
experimental
ones is a measure of the validity of the complete mixing
assumption. Tests with
successively less stirring can be used to study the level of agitation
required for
complete mixing. MacDonald and Piret! have made such studies for
a firstorder
reaction system.
6-6. Tank Flow Reactors in Series. In some cases it may be desirable

to use more than one stirred-tank reactor operating in series, the
exit stream from the first constituting the feed to the second, etc.
The
design problem to determine the final exit concentration or
conversion
ean be solved by consecutive application of Eq. (6-8) to each
reactor,
starting with the first. Eldridge and Piret? have derived equations
which may be solved algebraically for the final concentration for a
number
of types of rate expressions and for systems of reversible,
consecutive,
and simultaneous reactions. Graphical procedures were also
developed.
These are advantageous for a series of reactors when the rate
equation is
simple. The types of calculations involved are illustrated for the
simple
ease of a first-order reaction in Example 6-8 by extending the
results of
Example 6-7.
1R. W. MacDonald and E. L. Piret, Chem. Eng. Prog., 47 :363 (1951).
2 Loc. cit.
182 CHEMICAL ENGINEERING KINETICS ~
Example 6-8. Acetic anhydride is to be hydrolyzed in three stirred-
tank reactors
operated in series. Suppose that each has a volume of 1,800 cc,
that the
temperature is constant and equal to 25°C, and that the feed rate to
the first
reactor is 582 ec/min. Compute the per cent hydrolysis
accomplished in the
three reactors.
Solution. From the results of Example 6-7 the fraction hydrolyzed in
the
stream leaving the first reactor is 0.328. If the anhydride
concentration leaving
the first reactor is designated as C; and that leaving the second is
C2, Eq. (6-8)
applied to the second reactor becomes
On = Ce = 7205
In other words, the same equation as used for the first reactor can
be used for the
second with Cr replaced by Ci, Cz replaced by C2, and r; by re.
Hence Eq. (A)
of Example 6-7 is applicable for the second reactor if it is written as
follows:
1 Crt a
eta 1+ k65
C * = fraction hydrolyzed in reactor 2 = 1 —
1
Since 6; = 6), the fraction hydrolyzed in the second reactor will be
0.328.
Therefore the following equations may be written for the three
reactors:
oGr ri = 0.328 C, = (1 — 0.331)Cr = 0.672Cr
F
ees 0.328 Ce = 0.672C,
Cy
C20
=e = 0.328 C; = 0.672C.
The total fraction hydrolyzed in all reactors is given by the
expression
Cr—Cs _ , _ 0.6720, _ | 0.672 X 0.672 X 0.672Cr
Cr Cro Cr
1 — 0.672* = 1 — 0.303
= 0.697, or 69.7%
Eldridge and Piret developed the following general equation for a
series of ne
reactors in which a first-order irreversible reaction occurs:
1
[1 + (6) JL + &(6)2] = > + (1 + &(O)a
Fraction converted = 1
In the present example this equation reduces to
1
Fraction hydrolyzed = 1 — ———___
[1 + &(6’)]8
i
1— 1.4893 ~ 1 — 0.672* = 0.697
as before.
NOMENCLATURE
A. Cross-sectional area of tubular reactor
Ar Heat-transfer area /unit length of tubular reactor
F Feed rate
F; Feed rate of component 7
h Inside-heat-transfer coefficient for tubular reactor
AH Heat of reaction (AH» applies at ty)
n Number of components in a reaction mixture
m; Mass rate of flow of component i at any point in a tubular reactor
or leaving a
tank reactor
p Partial pressure
Heat-transfer rate from the reactor to the surroundings
vr Volume/mole of feed
Ve Volume of tank reactor
Ver Volume of tubular reactor
t Temperature; t,, ¢2 entering and exit values, respectively
T Absolute temperature
y Mole fraction in gas phase
6’ Holdup, or average residence, time in tank flow reactors
F Subscripts to denote exit and entering values, respectively, of the
conversion
in a tank-flow reactor
PROBLEMS
1. The production of carbon disulfide from methane and sulfur vapor
can be carried
out homogeneously, or with a solid catalyst. Also, some solid
materials act as a
poison retarding the reaction.
The following data were obtained on a flow basis at a constant
temperature of 625°C
and with an initial reactants ratio of 1 mole of CH; to 2.0 moles of
sulfur vapor (considered
as S2). The first set of data were obtained with the reactor empty
(effective
volume 67.0 ml), while the second set were obtained after packing
the reactor with a
granular material (7-mesh) which reduced the void volume to 35.2
ml. Was the
granular material acting as a catalyst or poison in this case?
eee Product rate, | Conversion
Set | Run CS: g moles/hr} of methane
CH, S.
1 1 0.417 0.834 0.0531 0.127
0.238 0.476 0.0391 0.164
3 0.119 0.238 0.0312 0.262
2 1 0.119 0.238 0.0204 0.171
2 0.178 0.357 0.0220 0.123
Mann ee | I ee ee EN AEE eS eee
2. Butadiene and steam (0.5 mole of steam per mole of butadiene)
are fed to a
tubular flow reactor which operates at 1180°F and a constant
pressure of | atm. The
reactor is noncatalytic.
Considering only the reversible polymerization reaction to the
dimer, determine:
(a) the length of 4-in.-ID reactor required to obtain a conversion of
40 per cent of the
butadiene with a total feed rate of 20 Ib moles/hr; (b) the space
velocity, measured
as liters per hour of feed gas (at 1180°F and 1 atm) divided by the
reactor volume in
liters, required to obtain a conversion of 40 per cent.
The polymerization reaction 1s second-order and has a specific
reaction-rate constant
given by the following equation:
log k = — 779 + 8.003
k = g moles Cie polymerized / (liter) (hr) (atm)?
inns
The reverse (depolymerization) reaction is first-order. The
equilibrium constant for
the polymerization reaction is given by the expression
—Rin K = — 2485 4 5.968 + 11.075 log T + 10-7 — 1.288 X 10-*7"
At 1180°F (911°K) the value of K from the preceding equation is
1.27.
3. (a) A mixture of butenes and steam is to be thermally
(noncatalytically) cracked
in a tubular flow reactor at the constant temperature of 1200°F and
constant pressure
of 1.0atm. While the feed consists of a number of different butenes!
and the products
vary from coke to butadiene, the rate of reaction may be adequately
represented by a
first-order mechanism as follows:
6
fee ee — kipa
PB
60,000
log ki = — L575? + 15.27
The rate was determined experimentally in a reactor packed with
inert quartz chips,
and the reactor to be designed in this problem will also be so
packed.
7, = g moles butenes cracked/(g quartz chips) (hr)
5 = void fraction = 0.40
pp = bulk density of bed packed with quartz chips = 1,100 g/liter
Ps = partial pressure of butenes, atm
T =. "K
The ratio of steam to butenes entering the reactor will be 10:1.0 on
a molal basis.
Under these conditions the change in number of moles during the
course of the reaction
can be neglected.
Determine the conversion as a function of size of reactor. Also,
prepare a plot
having two abscissas [(1) pounds of quartz chips per pound moles of
butene feed per
hour; (2) space velocity defined as cubic feet per hour of feed at
1200°F divided by the
void volume in cubic feet] and one ordinate (the conversion of
butenes). Cover a
range of values of the first abscissa of 0 to 3,000.
What total volume of reactor would be required to obtain a 20 per
cent conversion
with a butenes feed rate of 5 lb moles/hr?
(b) In this case the feed will consist of 10 moles of steam per mole
of total hydrocarbons.
The hydrocarbon fraction is 60 mole per cent butenes and 40 mole
per cent
butadiene. Consider that the butadiene may undergo two reactions,
cracking and
polymerization to the dimer. The rate data for these are as follows:
1See Chem. Eng. Prog., 44:229 (1948).
Et a 6 ,
Cracking: ets kop,
_ 30,000
log kz = T5757 + 7.241
Polymerization to the dimer:!. r, = © kis(py )?
B
25,000
log k3 = TTT + 8.063
where r2=£ moles butadiene cracked/(g of quartz chips) (hr)
Pp, = partial pressure of butadiene, atm
rs = g moles butadiene polymerized/(g of quartz chips) (hr)
Determine the conversion of butenes and that of butadiene as a
function of W/F from
0 to 3,000 lb chips/(lb mole feed) (hr).
Assume that the total number of moles is constant, and neglect all
reactions except
those mentioned.
4. Benzene is to be chlorinated in the liquid phase in a kettle-type
reactor operated
on a steady-state basis. Liquid benzene is added continuously; the
liquid product and
gaseous hydrogen chloride are continuously removed. The chlorine
gas is bubbled
continuously into the liquid-reaction mixture in the kettle.
The rate of reaction may be assumed large enough so that there is
no unreacted
chlorine in the reaction products. Also, the concentration of chlorine
in the reaction
mixture will be small. The hydrogen chloride concentration in the
reaction is likewise
small. The density of the liquid mixture may be assumed constant.
At the constant operating temperature of 55°C the significant
reactions are the
three substitution ones leading to mono-, di-, and trichlorobenzene.
Each reaction is
second-order and irreversible. The rate constants for each case are
designated as
follows:
ki: C.He a. Cl, > C.H;Cl, + HG
ko: C.H;Cl a Cl. — C.H;Cl. ah HCl
k3: C.H;Cl. + Cl. — C.H3Cls + HCl
It was noted in Chap. 3 that, at 55°C, the ratios of the rate constants
are
Ky
ee 8.0
ky _
i 30
Under the proposed operating conditions, the composition of the
liquid product will
be constant for any one run. Different products will be obtained for
different ratios
of benzene and chlorine fed to the reactor. Compute the
composition of the liquid
product for the case where 1.4 moles of chlorine per mole of
benzene are fed to the
reactor.
5. Reconsider Prob. 4 for the case where, instead of a single reactor,
a two-reactor
system is used. The liquid stream enters the first reactor (as pure
benzene), flows
from the first to the second, and finally the product is withdrawn
from the second
reactor. Gaseous hydrogen chloride is withdrawn from each reactor.
Plot the composition of the product vs. moles of total chlorine added
per mole of
benzene. Cover a range of the latter variable from 0 to 2.5.
One-half the total chlorine is added to each reactor.
1 This reaction is reversible, but neglect the reverse reaction in this
instance.
CHAPTER 7
SEMIBATCH REACTORS
The semibatch reactor has been defined in Chap. 4 (Fig. 4-1C) as a
tank type, operated on a non-steady-state, or transient, basis.
Semibatch
behavior occurs whenever a tank flow reactor is started up, when
its operating conditions are changed from one steady state to
another,
or when it is shut down, for then the properties of the system
change
with respect to time. Purging processes in which an inert material is
added to the tank reactor can also be classified in the semibatch
division.
In addition to these applications arising from short-period deviations
from steady-state operation the semibatch reactor is used for its
own
particular characteristics. For example, it is sometimes
advantageous
to add all of one reactant initially and then the other continuously.
In
certain instances this modified procedure has advantages over
ordinary
batch operation. In systems with a high heat of reaction the energy
evolution (or absorption) can be controlled by regulating the rate of
addition
of one of the reactants. In this way the disadvantage of batch
reactors due to their poor heat-transfer characteristics can be
partially
eliminated. As mentioned in Chap. 4, this form of semibatch
operation
also allows for a degree of control of concentration of the reaction
mixture,
and hence rate of reaction, that is not possible in batch or flow
reactors. Another illustration occurs when the reactants are all
initially
in the vessel, but one of the products is continuously removed. An
example is the removal of water, by boiling, in esterification
reactions.
In this instance the advantage gained is an increased rate due to (1)
the
removal of one of the products of a reversible reaction and (2)
increased
concentrations of reactants.
The design equations for semibatch operation are complicated
because
each of the four terms in the general material balance [Eq. (4-1)]
may be
significant. The feed and withdrawal streams from the reactor cause
changes in the composition and volume of the reaction mixture in
addition
to the changes due to the reaction itself. These variations so affect
the rate that the design equations cannot be integrated analytically,
except for first- or zero-order reactions and isothermal conditions.
As
in the case of other examples of tank reactors complete mixing
must be
assumed in order to handle the design problem in a quantitative
fashion.
7-1. Isothermal Design. The design equation may be formulated in
terms of the amount of the specific reactant which is used to
measure
the rate of reaction. Thus the rate of addition of reactant A in the
feed
stream will be Fyz;, where F is the mass feed rate and z the weight
fraction.
Writing the other three terms in Eq. (4-1) in a similar way leads
to the result
M, d(V C2)
Fy21 — Foz — rVe = dé (fall)
Allowance must be made for the rate of feed Fiz; and withdrawal
F222
to be different. The subscripts 1 and 2 are used to indicate this
possibility.
The term on the right represents the rate of accumulation of
reactant A in the reactor contents. MM, is the molecular weight of A,
and C2 is the molal concentration of A in the mixture. The
assumption
of complete mixing means that this concentration corresponds to
the
weight fraction z in the exit stream.
Equation (7-1) cannot be integrated analytically, in the general case
where Fj, 21, F2, 22, r, Vz, and p can all vary with time. However, an
important case will be considered in which assumptions are made
concerning
the constancy of the density, temperature, rates of flow, and the
form of the rate equation.
The general equation can always be integrated numerically. Such
procedures will be illustrated for an example in which there is a feed
stream but no product withdrawn from the reactor.
Constant-density First-order Reactions. Piret and Mason*? have
reviewed the subject of transient conditions in tank flow reactors,
under
the conditions of constant flow rates, constant temperature,
constant
density, and first-order rate equations. From a practical point of
view
these conditions are reasonable in many of the semibatch
operations that
correspond to starting up and shutting down of steady-state
reactors.
The assumption of constant density permits the use of the term
concentration
to account fully for changes in amount of reactant. Also,
constant density, along with constant flow rates, means that the
reactor
volume Vz will remain constant. Under these restrictions Eq. (7-1)
may
be written
dC,
ae (7-2) (Fur)C, — (Fur)C2 — 1Vs = Ve
1 Equally satisfactory would be to define F as the molal rate and 24
as the mole
fraction A.
21). R. Mason and E. L. Piret, Ind. Eng. Chem., 423817 (1950).
27). R. Mason and E. L. Piret, Ind. Eng. Chem., 4321210 (1951).
(Fv) is the constant volumetric flow rate through the reactor.
Equation
(7-2) may be rearranged as follows:
dC, , Fur, , _ For
de ee ee
The quantity Vz/Fu,r is the holdup time in the reactor and has been
designated previously as 6’. If the rate may be expressed in a first-
order
form and the temperature is constant, Eq. (7-3) is a linear
differential
equation which can be integrated analytically. In terms of 6’ it may
be
written
ar (t+) = = (7-4)
where k is the first-order rate constant. Mason and Piret have
developed
solutions of Eq. (7-4) for both single reactors and when the equation
is
applied to each reactor vessel in a multistage system. The
Laplacetransform
method is used to solve the differential equations applying
to a variety of transient conditions for the starting up and shutting
down of tank flow reactors.
The application of Eq. (7-4) is illustrated for a single reactor with a
constant feed composition in the example that follows:
Cy (7-3)
Example 7-1. Acetic anhydride is hydrolyzed at 40°C in a semibatch

system
operated by initially charging the stirred-tank reactor with 10 liters
of an aqueous
solution containing 0.50 X 10~4 g mole of anhydride per cubic
centimeter. The
vessel is heated to 40°C, and at that time a feed solution containing
3.0 x 1074
g mole of anhydride per cubic centimeter is added at the rate of 2
liters/min.
Product is withdrawn at the same rate. The solution density may be
assumed
constant, and the reaction is kinetically first-order with the following
rate
equation:
r = kC, g moles/(cc) (min) t
k = 0.380 min“!
Determine the concentration of the solution leaving the reactor as a
function
of time.
Solution. Equation (7-4) is applicable, and in this case (,/6’ is a
constant.
9’ = Ve _ 10,000
Four 2,000
= 5 min Cy 3.00 x 10
@Q’ 5
= 6 X 10~° g mole/(ce) (min)
1
ry +k = 4+ 0.380
0.580 min=!
+ Data taken from Example 6-7.
SEMIBATCH REACTORS 189
The integrated solution of Eq. (7-4) is
Cy 1
Co =F De Tk
4+ Ae-(1/0'+k)0
where A is a constant of integration. It may be obtained by noting
that, at
6 = 0, C2 = 0.50 X 10-4,
A = 0.50 x 10-* — 21 *
2 01/0’ +k
perth, see Ts6 Xa 10o-5 = x 10 5 aR 5.34 X 10-5
Then the final expression for the product concentration is
_ 6X 10-5
0.580
10.3 & 10-5 — 5.34 & 10-5e—0-588
C2 — 5.34 X 10-5e—9-588
I
Table 7-1 shows values of C» for increasing time measured from the
instant of
TaBLE 7-1
Time, 6, min Exit anhydride cone., C2,
g moles/ce
ogOOs wner HOeOr N
addition of the feed stream. These results indicate that after 5 min
the reactor
is operating very close to steady-state conditions; i.e., the exit
concentration is
close to that predicted at infinite time. This latter value could also
be obtained
directly from Eq. (6-8) for tank flow
reactors.
Cr 7 Cr = Gi a C2 — ‘Vs — 0.380026’
For
C, — C2 = 0.380 X 5 XK C2
Ci ee ae C2 me 1-00 2.90
aif F Fia. 7-1. Semibatch reactor with a feed
10.3 X 107° g mole/ce put no withdrawal stream.
Non-first-order Reactions. Figure 7-1 shows a special type of
semibatch
system in which there is a continuous feed, no withdrawal of
product,
and a mass mo, of component A initially in the reactor. The
application
of Eq. (7-1) will be considered for a reactor of this type when the
rate
equation is not first-order and when the density of the reaction
mixture
is not necessarily constant. Isothermal operation 1s taken up here
and
the nonisothermal case in the next section. .
Since there is no exit stream, Eq. (7-1) takes the following form:
d(V eC2,) (7-5)
Fiz, ae rVeB — Ma dé
The rate r could be expressed in terms of the concentrations of the
reactants
and this equation integrated numerically for any rate function.
However, in this case the concept of a conversion has meaning, and
concentration
can be expressed in terms of this variable.'! Define x at any 4
as the ratio of the amount of reactant A converted to the total
amount
of A added up to that time. Then the concentration of A in the
reactor
will be related to x as follows:
_ (mo, + Fz1,9) — 2) :
Co, M:iVe (7 6)
and d(C2,Vs) = 7 [—(mo, + Fizi,0) dx + (1 — 2)Fi21, dO]
A
Substituting this expression into Eq. (7-5) and simplifying lead to the
result
Fy21,2 d0 + (mo, + Fiz1,0) dv = rVzp dd (7-7)
This is the desired design expression from which the conversion
may be
computed at any time for this form of semibatch reactor.
Note that, if reactant A were not present in both the initial charge
and the feed stream, one of the two terms on the left side of the
equation
would disappear. Thus if there were no reactant A in the feed
stream, the design equation would simplify to
mo, dx = Ver dé (7-8)
This result is identical with Eq. (5-2) for batch reactors. If no A were
present in the initial charge to the reactor, the expression would
become
Fy2,,(a d0 + 6 dx) = Ver dé (7-9)
The reactor volume varies with time and may be written
Ve = Vot / (a) dé (7-10)
The initial volume Vo is equal to mo/po, where po is the density of
the
initial mass mo.
In general, the volume will change owing to changes in both the
mass
1 rf “A ar 1A c “7 ; ; When there are both feed and exit streams,
conversion is not a useful concept,
because the amount of material to base the definition of conversion
upon is not clear.
and the density of the reaction mixture. However, for most systems
density changes will be small. With this restriction
dV, F
dd ~—p
and Eq. (7-10) gives the simple result
F
Ve — Vo = p 6 (7-11)
The rate of reaction is the other quantity that must be expressed in
terms
of x and 6 before the design equation can be integrated. The rate
may
be expressed formally (for isothermal operation)
in a f(Ca,Cz, oe 3) (7-12)
where A, B, . . . are reactants and products. The conversion is
related
to the concentration (moles per volume) of A and B by Eq. (7-6) and
by
the expression
Cie (1/Ms)(mog + F1z156) — xb(mo, + Piz.1/ )Ma )
Ve
(7-13)
In Eq. (7-13) the first term in the numerator represents the total
moles
of component B added up to time @ (2;, = weight fraction B in the
incoming
stream), and the second represents the moles of B reacted at the
conversion
x (b = moles of B used per mole of A; for example, if the reaction
is A + 2B—-C+ D,b = 2). Similar equations can be written for each
product of the reaction. Combination of these expressions for C4,
Ca,
. with Eq. (7-12) gives the rate r as a function of x and @.
In theory the design equation (7-7) can now be integrated by
substitution
in it of the expressions for Vg andr. However, it is apparent that
the functions are sufficiently complicated to prevent analytical
integration
in most cases. It is worthwhile at this point to note again the reason
for the complicated nature of Eqs. (7-6) and (7-13). It is due entirely
to
the fact that there is a stream entering the reactor. For simple batch
operation F, = 0, and Vz is constant (for the restriction of constant
density).
The general method of approach for designing the reactor can be
explained as follows:
1. The design equation (7-7) is first written in difference form,
(mo, + F121,9) At = [(Var)a — Fi21,%a] 46 (7-14)
and the value of Ax computed for a chosen time increment Aé. The
average values of Vgr and z for the first A@ can be taken equal to
the
known values at @ = 0, as a first estimate.
2. The reactor volume at time A@ is computed from Eq. (7-11).
3. Using the conversion computed in step 1 and the reactor volume
Vz
from step 2, the concentrations at the end of the first time
increment are
determined from Eqs. (7-6) and (7-13).
4. The rate at the end of the first increment is then obtained from
Eq. (7-12).
5. A second estimate of the average value of rV is determined from
the known values at the beginning and end of the increment. |
6. The design equation (7-7) is employed to calculate a second
estimate
of Az.
7. This procedure is continued until the average values of rVz do not
change. Generally the second estimate is sufficient.
8. Second, third, ete., time increments are chosen and the process
in
steps 1 to 7 repeated.
It will be observed that these steps follow the same pattern as those
employed in stepwise solutions of design equations for simple batch
and
flow reactors as illustrated in Chaps. 5 and 6. This is true because,
from
a mathematical viewpoint, the problem is the same regardless of
the type
of reactor. In each case all the quantities in the differential equation
are functions of but two variables (conversion and reactor volume
for
the flow case and conversion and time for the batch).
Example 7-2. The esterification of acetic acid and ethyl alcohol is to

be carried
out in a semibatch tank reactor at a constant temperature of 100°C.
The
alcohol is added to the reactor initially as 400 lb of pure C.H;OH. An
aqueous
solution of acetic acid is then added at a rate of 3.92 lb/min for 120
min. The
solution contains 42.6 per cent by weight acid. The density of the
solution will
increase during the course of the reaction, but to simplify the
computations
assume that it is constant and equal to that of water.
The reaction is reversible, and the specific rates may be taken as
the same as
those used in Example 5-2,
r = kCuCon k = 4.76 X 10~ liter/(g mole) (min)
r’ = k'CsCw k’ = 1.63 X 10~ liter/(g mole) (min)
CH;COOH + C:H;OH = CH;COOC.H; + H.O
Compute the conversion of acetic acid to ester as a function of time
from 0
min until the last amount of acid is added (120 min),
Solution. As the conversion is to be based upon the acetic acid and
there is
no acid in the initial charge to the reactor, the proper design
equation is (7-9).
Since the rate constants are in molal units, the concentrations will
also be
expressed in those units.
If Cu, Cou, Cw, and Cx represent acid, alcohol, water, and ester,
concentrations
(in pound moles per cubic foot), equations similar to (7-6) and (7-13)
are
fe F421,0(1 — 7)
Cu MuVs analogous to Eq. (7-6) with m4 = 0
11
=— Moon — 2(F 12149) ——
Mou My
Cox = V, analogous to Eq. (7-13) with b = 1
i pe (1/Myw)F 1219 ae LF \21,,9(1/ My)
Ve
PP Po nSleai n x(Fa ll1l2 14a8t) B(e1 /BMeu )o l
Ve
Substituting these expressions into the rate equations gives a
relationship for
the net rate of conversion of the acid in terms of x, 6, and Vz.
‘eae K[F 21,91 — £)](Moon/ Mon a LF \2146/Ma)
MaV2
= k! LF 1214 9(F 121y9/ My +- LF \21,,0/ Mu)
MuV?,
f= —T?T
(A)
where F,; = 3.92 lb/min
Ziy = 0.426
Moon = 400 lb
Ziv = 1 — 0.426 = 0.574
Mu = mol. wt acetic acid = 60
Mou = mol. wt ethyl alcohol = 46
My = mol. wt water = 18
The rate constants expressed as cubic feet per pound mole per
minute become
1
28.32
k = 4.76 X 10~* < 454 = 7.63 X 10-%
k! = 2.62 X 107°
Substituting these results in Eq. (A) to obtain a working expression
for the
rate (pound moles per cubic foot per min),
_ 1.28 X 10-20(1 — x)(8.68 — 0.027826) s 4.37 X 107*6?x(0.125 +
0.02782)
Zz 60V} 60V3
r= a [21.3 x 10-50(1 — x)(8.68 — 0.027828)
: — 7.13 X 10-562x(0.125 + 0.0278x)] (A’)
Since the density is assumed constant, Eq. (7-11) is applicable and
the reactor
volume in cubic feet is
Pee Rea mo + FiO _ or = 6.69 + 0.06550 (B)
p p ),
Equation (7-9), in terms of numerical values, may be written
F 1219 dx = ViarMuy dé = F Zit dé
3.92 « 0.426 K Odzx Var X 60d — 3.92 X 0.426 XK x dé
or 6 dx = 35.9V er dé — x dé
I
In difference form this expression becomes
A(z0) = 35.9(Var)a AD (C)
Equations (C), (B), and (A’) can now be employed to carry out a
stepwise
solution. Since the concentration equations have been combined
with the rate
equation in expression (A’), step 3 in the previously listed sequence
can be
omitted. The method of solution will be illustrated by following
through the
procedure step by step.
stEP 1. Choose an initial time increment of 5 min.
The initial volume Vz = 6.69 cu ft (based upon density of water at
100°C).
The initial rate as computed from Eq. (A’) is zero, since 6 and z are
both zero
(i1.e., at zero time there is no acid in the reactor). Hence, to obtain
a reasonable
estimate of (rVs). with which to start the computations, calculate r
for an average
6 = ¥ = 2.5 min.
12
fa e| taee [Zio L—5 O = 0)(8.68 ~—~ 0) = 0]
1.04 & 1074 Ib mole/(cu ft) (min)
Hence the first estimate of (rV 2), is
(rVe)a = (1.04 X 10-4) X 6.69 = 6.94 & 10-4
From Eq. (C),
A(x) = 35.9(6.94 X 10-4) X 5 = 0.1250
or 2, = 0+ 0.0250 = 0.0250
step 2, The reactor volume at the end of the first increment is given
by
Iq. (B),
Va, = 6.69 + 0.0655 X 5
= 7.02 cu ft
step 4. The rate at the end of the increment is obtained from Kq.
(A’), with
x = 0.0250, 6 = 5, and Vz = 7.02.
12
a a) [21.3 X 10-* X 5 X 0.975(8.68 — 0.003)
— 7.13 X 1075 X 0.625(0.125 + 0.0007)]
0.898 & 10-2 — 0.56 x 1075 i Fis =182 3 10-4
STEP 5. At end of increment, rVz = (1.82 x 10-4) X 7.02 = 12.8 x 10-
4.
At beginning of increment (8 = 0), rVs = 0 X 6.69 = 0. Hence the
second
estimate of (rV xg), is
12.8 x 10-4
Va). = ee = 6.40 x 10-4
STEP 6. The second estimate of Az is given by Eq. (C),
A(@x) = 35.9(6.40 x 10+") xo = (0-115
_STEP 7. Recalculation of (rV xg), will not change the value of 6.40 «
10-4 significantly.
Hence at the end of the first increment
21 = 0.0230
Va, = 7.02 cu ft
mr, = 1.82 X 10 lb mole/(min()cu ft)
(FV) = 13.5 C1074
For the second increment of 5 min estimate (rVze)a = 18.0 X 10-4.
Then the
first estimate of Az is
A(6x) = 35.9(18.0 xk 10-4) x 5 = 0.323
Aor =— 6,2, = Boa — 0 115 = 0.323
x2 = 0.0115 + 0.0323 = 0.0438
Ve = 6.69 + 0.0655 < 10
at 7-4 Ci It
1
r= 7343 (1.71 X 10-2 — 0.004 x 10-2) = 3.18 x 10-4
(rVs)2 = 23.3 X 10-4
PMI 1 25.3) x 104
a 8 lO
Since this result is close to the first estimate of 18.0 * 10-4, further
calculations
of the rate are unnecessary. At the end of the second increment (@
= 10 min),
22 = 0.0115 + 0.0323 = 0.044
Vz, = 7.34 cu ft
re = 3.18 X 10-4
(rVz)2 = 23.3 10-*
The results for the entire time interval 0 to 120 min, determined by
continuing
these stepwise calculations, are summarized in Table 7-2.1
I
Where reverse reactions are important, it has been mentioned that
continuous removal of one or more of the products of the reaction
will
increase the conversion obtainable in a given time. Thus one
reactant
could be charged to the reactor, a second reactant added
continuously,
and one of the products withdrawn continuously.* This form of
semibatch
reactor can be treated by modifying the design equation (7-7), the
1Jt should be recognized that this solution of the problem is
approximate because
of the assumption of constant density of the reaction mixture.
However, if density
vs. composition data were available, the same stepwise method of
calculation could be
employed to obtain a more precise solution. The only difference
would be that Eq.
(B) would have to be modified to take into account density
differences.
2 This withdrawal might be accomplished by distillation, for
example.
volume equation (7-10), and the expressions for the concentration,
to
take into account the withdrawal of material.
7-2. Nonisothermal Design. If the heat of reaction is not negligible,

or if it is required to supply or remove thermal energy from the
reactor,
an energy balance must be included in order to solve semibatch-
reactor
problems. Owing to the assumption of complete mixing the
temperature
of the reactor contents will be uniform at any instant. However,
this temperature may vary with time. The method of determining
the
conversion and temperature at any time will be treated for the
specific
type of semibatch reactor that has no product stream, i.e., the type
illustrated
in Fig. 7-1.
TABLE 7-2. SEMIBATCH-REACTOR DESIGN FOR ESTERIFICATION OF
ACETIC
Acip with Eruyt ALCOHOL
Time, Vs, Conversion of acid, | Rate, 7, lb moles acid
min cu ft = reacted /(min) (cu ft)
0 6.69 0 0
5 7.02 0.023 1.82 X 10-4
10 7.34 0.044 Ba LO
20 8.00 0.081 BO per A |s h
40 Oral 0.138 TOD ee
60 10.6 0.178 Vie Be ee 0a a
80 11.9 0.205 Or 0a alts
100 132 0.224 6.3 X 10-4
120 14.5 0: 237 Dit eels
Suppose that the temperature of the continuously added stream is
fp.
The energy balance is a mathematical expression of the fact that
the
energy supplied by the heat of reaction and that in the incoming
stream
must either be absorbed by the reaction mixture or be transferred
to the
surroundings (i.e., cooling or heating surfaces). If the temperature of
the reactor contents at any time @ is ¢, this expression may be
written
—AH
Ws rVz dé + F, dé Cr(tpr — t) = 77:C dt -t- UA(t oe ts) dé (7-15)
The first term represents the amount of energy supplied by the
reaction
in an element of time d@ (AH is the heat of reaction per mole and r
the
rate in mass units). The second represents the amount of energy
that is
introduced in the continuously added stream, above the reaction
temperature
¢ (cp is the specific heat of the feed). The first term on the right is
the part of the supplied energy that remains in the reaction mixture
(c is
the specific heat of the mixture, and m; is its total mass, mo + F,@).
The
last term represents the energy transferred to the surroundings in
the
time dé (t, = surroundings temperature; U = over-all heat-transfer
coefficient
between reaction mixture and surroundings; A = heat-transfer
area).
If the reactor operates adiabatically, the last term on the right-hand
side of the equation is zero. If the temperature of the stream F; is
the
same as that of the contents of the vessel, the second term on the
left side
disappears. Under the latter restriction the energy balance reduces
to
the same form that applies to ordinary batch-operated reactors,
1.€.,
Eq. (5-5).
sp rVe dé = me dt + UA(t — t,) dé (5-5)
Normally the feed rate F,, the feed temperature ¢, and the
surroundings
temperature ¢, are constants in Eq. (7-15). Under these
cireumstances
the energy balance expresses how the temperature of the reaction
mixture varies with time. In conjunction with the rate and design
equations
it can be solved for the conversion and temperature as a function
of time by a numerical, stepwise process.
There are a number of means of controlling the Penperatire in
semibatch
reactors, suggested by the terms in the energy balance. For
example,
it may be desirable to maintain this temperature constant, as in the
case of an exothermic reaction where a temperature rise beyond a
certain
level causes unfavorable side reactions. The restriction of a constant
temperature eliminates the term mc dt, and the simplified result
expresses
the requirement placed on the heat-removal term dQ, = UA(t — ts)
dé.
Since the rate changes with time, Q, likewise must be varied during
the
reaction in order to maintain the reaction temperature constant. The
energy added in the stream F,; may also be used to control the
temperature.
This can be carried out in two ways: (1) by varying the feed rate;
(2) by changing the feed temperature.
In instances where the reaction rate is very fast and the permissible
temperatures are limited, the factors that control the design of the
reactor
are the rate of energy exchange with the surroundings, the feed
temperature,
and the feed rate. Under these circumstances the solution of
the design problem requires only the energy balance. The amount
of
material reacted in time d@ is the total amount of reactant in the
feed
stream, or F',z,,d6. Hence this quantity multiplied by —AH replaces
the first term in Eq. (7-15), and the energy balance may be written
a Fiz, dé = F, dé Cr(tp = t) = MiC dt fe UA(t = ts) dé
A
dt 2{—Al/d M4)F 121, + Ficr(tr —t) —U. Sih) 7-16)
or —_ =
dé mc
Solution of this expression shows how the temperature varies with
time for
various combinations of feed rate, feed temperature, and heat
exchange
rate. The conversion obtained under such conditions is always the
equilibrium
value corresponding to the conditions at the end of the process.
Its application to a practical problem is illustrated in Example 7-3.
Example 7-3. Hexamethylenetetramine (HMT) is to be produced in a

semibatch
reactor by adding an aqueous ammonia solution (25 per cent by
weight
NH;) at the rate of 2 gpm to an initial charge of 238 gal (at 25°C) of
formalin solution
containing 42 per cent by weight formaldehyde. The original
temperature
of the formalin solution is raised to 50°C in order to start the
reaction. The
temperature of the NH,OH solution is 25°C.
The heat of reaction in the liquid phase may be assumed
independent of temperature
and concentration and taken as —960 Btu/Ib of HMT. If the reactor
can be operated at a temperature of 100°C, the rate of reaction is
very fast in
comparison with the rate of heat transfer with the surroundings.
Higher temperatures
than 100°C are not desirable because of vaporization and increase
in
pressure.
It is proposed to cool the reactor by internal coils through which
water is passed.
The over-all heat-transfer coefficient between the stirred reaction
mixture and
the cooling water will be 85 Btu/(hr)(sq ft)(°F). The water rate
through the
coils is such that its temperature varies but little, and an average
value of 25°C
may be used.
Calculate the length of 1-in.-OD tubing required for the cooling coils.
ADDITIONAL DATA AND NOTES
Density of ammonia solution = 0.91 g/ce
Density of formalin (42%) at 25°C = 1.10 g/cc
Specific heat of reaction mixture (assume constant), c = 1.0 Btu/(lb)
(°F)
Specific heat of 25 wt % NHs solution, cr = 1.0 Btu/(lb) (°F)
The rate of the reverse reaction is negligible.
Solution. Since the rate is very rapid and the reaction is irreversible,
the
ammonia in the inlet stream will be completely converted to HMT
just as soon
as it is added to the reactor. According to the reaction
4NH; + 6HCHO — N,(CH2). + 6H:0
4 moles of ammonia are required for 6 moles of formaldehyde.
Hence the total
amount of ammonia required to react with all the charge of formalin
solution
will be
238(8.33 X 1.10) X 0.42 4
30 6 Ammonia required = mile
= 346 lb
From the ammonia feed rate of 2 gpm the total time of reaction will
be
2 346
2(8.33 X 0.91) x 0.25
= 91.3 min
D4
The heat-transfer surface is to be sufficient to prevent the
temperature from
exceeding 100°C. Hence, when t = 100°C, the change in
temperature with time,
or dt/d@ in Eq. (7-16), will be zero. However, at temperatures below
100°C
the driving force t — t, will be insufficient to transfer enough energy
to the cooling
coils to maintain a constant temperature. Thus at the start of the
addition of
ammonia the first term in the numerator of Eq. (7-16), which is
positive for an
exothermic reaction, will be greater than the sum of the second and
third terms,
and the temperature of the reaction mixture will increase. From a
practical
point of view this heating-up period would be reduced to a minimum
by shutting
off the flow of cooling water until f = 100°C.
To determine the required heat-transfer area, Eq. (7-16) may be
used with
di/d@ = 0 and t = 100°C.
AH
UA(t — t,) = Ps ao Fycr(tr — t)
AH
85A(100 — 25) X 1.8 = 6) i Fiz, + Fi X 1.0(25 — 100) X 1.8 (A)
Ma
The heat of reaction = —960 Btu/Ib HMT. On the basis of NH;
AH _ _ 960 X 140
Ms, 4x17
—1975 Btu/lb of NH;
The feed rate is
Fr= 2 xX 60 X 8.33 Xx 0.91
= 910 lb/hr
olga 0.25
Substituting these values in Eq. (A) and solving for the heat-transfer
area,
910 (1,975 X 0.25 — 1 X 75 X 1.8)
ie, Sats pe a)
= 28.3 sq ft
A=
If the heat-transfer coefficient of 85 is based upon the outside area
of the tubes,
the length of 1-in.-OD coil will be
28.3) 028.3°% 12
— rD T
= 108 ft
L
An approximate size of the reactor can be obtained by noting that
the total
mass of mixture at the end of the process will be
910 ae 4 238(8.33 X 1.10) = 3,560 Ib
Assuming the density of the HMT solution to be 72 lb/cu ft, the
minimum reactor
A eylindrical vessel 4 ft in diameter and 6 ft in height would volume
is 50 cu ft.
If the 1-in. tubing were wound into a provide 33 per cent excess
capacity.
3-ft-diameter coil, approximately 12 loops would be needed.
The length of time necessary to raise the reaction temperature from
its initial
value of 50°C to 100°C can be obtained by integrating Eq. (7-16).
With the
water rate shut off UA(t — t,) will be negligible, and the expression
becomes
[ dt - i do
= ci mo + Fi6)c
to on Fi21, +- Ficr(tr — t) 0 ( 5 7 )
LA
where m; has been replaced by mo + F106. If AH and ¢ do not vary
with temperature,
this equation may be integrated to yield
1 (—AH/Ma)F i241, + Ficr(tr —t) _ 1 mo + F180 = ine ee B
Ficr ie (—AH/Ma)F ia, + Ficr(tr = to) Fic : Mo ( )
Equation (B) expresses the temperature as a function of time during
the heating
period. Taking fo = 50°C, try = 25°C, and, expressing @ in hours,
the time
required for ¢ to reach 100°C is given by
1, 449,000 + 910 x 1(25 — 100) X18 ___1_,, 2,180 + 9108
~ 1X 910” 449,999 + 910 x 1225 — 50) X18 910X1 2,180
In (1 + 0.4186) = — In 3F26o,0y00
1 + 0.4180 = 1.25
6 = 0.60 hr, or 36 min
In summary, the reaction temperature would rise from 50°C to
100°C in 36
min after the ammonia feed is started, provided water is not run
through the cooling
coil. After 36 min the water flow would be started in order to
maintain the
reactor temperature at 100°C. After a total time of 93 min sufficient
ammonia
would have been added to convert all the formaldehyde to HMT.
NOMENCLATURE
A Heat-transfer area, semibatch reactor
C Concentration
ce Specific heat of reaction mixture
cr Specific heat of feed
Ff, Feed rate, mass per unit time
F; Withdrawal rate, mass per unit time
AH Heat of reaction
ma Mass of component A in a mixture
m Total mass
mo Initial total mass
M Molecular weight
r rate of reaction in mass or molal units
t Reaction-mixture temperature
tr Feed temperature
t, Surroundings temperature
U_ Over-all heat-transfer coefficient between reaction mixture and
surroundings
Ve Volume of semibatch reactor
Vo Initial volume
ve Volume of feed per unit mass
x Conversion
z Weight fraction of a component
z, Weight fraction in feed stream
z2 Weight fraction in withdrawal stream
p Density of reaction mixture
po Initial density
@ Reaction time in semibatch reactor
6’ Holdup, or average residence, time in semibatch reactor with
equal flow rates in
and out (6 = Vz/Fvr)
PROBLEMS
1. Repeat Example 7-1 in the text with the modification that the
effluent from the
first reactor is fed to the second. The second reactor originally
contains 10 liters of an
anhydride solution of concentration 0.50 X 10-4 g mole/ce. Product
will be withdrawn
from reactor 2 at a constant rate of 2 liters/min. Temperatures in
both will be
40°C, and all other conditions will be the same as in Example 7-1.
a. Determine the concentration of anhydride in the solution leaving
the second
reactor from 0 time until steady-state conditions are reached.
b. Suppose that reactor 2 was originally empty and that its capacity
is 10 liters.
After it is filled, product is withdrawn at the rate of 2 liters/min.
What would be the
concentration of the first anhydride solution leaving the reactor?
2. Ethyl acetate is to be produced in a 100-gal reactor operated in
the following
manner:
a. The reactor originally holds 100 gal (at 100°C) of solution
containing 20 per cent
by weight ethanol and 35 per cent by weight acetic acid. Its density
is 8.7 lb/gal.
Assume that this value remains constant for all compositions.
b. The reactor is heated to 100°C and pure ethanol added at a rate
of 2 gpm (17.4
lb/gal).
c. Solution is withdrawn at the same volumetric rate.
d. The temperature is maintained at 100°C in the reactor. At this
level the rate
data are as follows:
r = kCyCou k = 4.76 X 10~ liter/(g mole) (min)
r’ = k'CrCw k’ = 1.63 X 10~ liter/(g mole) (min)
The subscripts H, OH, W, EF refer to acid, alcohol, water, and ester.
The concentrations
are expressed in gram moles per liter. Determine the concentration
of ester
in the product stream for time values from 0 to Lhr. What is the per
cent conversion
of the total amount of ethanol added?
Assume that no water is vaporized in the reactor.
3. Ethyl acetate is to be saponified by adding a 0.05-normal solution
of sodium
hydroxide continuously to a kettle containing the ethyl acetate.
The reactor is initially charged with 100 gal of an aqueous solution
containing 10 g
of ethyl acetate per liter. The sodium hydroxide solution is added at
a rate of 1.0 gpm
until stoichiometric amounts are present.
The reaction is relatively fast and irreversible, the specifie-reaction
rate being 92
liters/(g mole)(min) at 20°C. Assuming the contents of the kettle are
well mixed,
determine the concentration of unreacted ethyl acetate as a
function of time. At what
time will this concentration be a maximum?

Chemical r-6+7

Uploaded by

Copyright:

Available Formats

Chemical r-6+7

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Chemical r-6+7

Uploaded by

Copyright:

Available Formats

CHAPTER 6

HOMOGENEOUS FLOW REACTORS

6-1. Determination of Rate Equations from Laboratory Data. The

Example 6-2. A kinetic study is made of the decomposition of

_ INTEGRATION METHOD. Equation (A) may be substituted in the

DIFFERENTIAL METHOD. The experimental Vr/F data are plotted vs.

Example 6-3. A flow reactor consisting of a 1-in. stainless-steel pipe

1. Two first-order mechanisms may be assumed, one with respect to

2. To take into account the presence of the Ss and Ss species,

Example 6-4. Hougen and Watson! in an analysis of Kassell’s data

Table 6-8 gives the values of Vr/F, corresponding to successive

Example 6-5. It is proposed to design a pilot plant for the production

The solution procedure will be as follows:

TANK FLOW REACTORS!

Example 6-7. Eldridge and Piret! have investigated the hydrolysis of

6-6. Tank Flow Reactors in Series. In some cases it may be desirable

Example 7-1. Acetic anhydride is hydrolyzed at 40°C in a semibatch

Example 7-2. The esterification of acetic acid and ethyl alcohol is to

7-2. Nonisothermal Design. If the heat of reaction is not negligible,

Example 7-3. Hexamethylenetetramine (HMT) is to be produced in a

You might also like