40
Chapter 4
Reaction Model for Photo Chlorination of
Toluene in a Bubble Column/Sparged Reactor
Currently, most industrial processes for photochlorination of toluene are carried out in
bubble column/sparged reactors. The chlorine gas bubbles through liquid toluene in the
reactors. The liquid phase is vigorously agitated by the bubbling gas. To achieve higher
selectivity to benzyl chloride, the first chlorinated product from consecutive reactions,
the conversion of toluene is limited to a relatively low level, normally less than 50 %.
For example, in the Solutia plant (Figure 3-15), four sparged reactors are connected in
series. Toluene passes through the reactors from one to another continuously and
chlorine gas is bubbled into the reactor from the bottom of each reactor. The overall
conversion of toluene is about 30 % per pass of toluene. In this continuous process, the
reaction products are pumped from the fourth reactor to the distillation column for
further separation. The unreacted toluene from the distillation column is recycled to the
first chlorination reactor.
41
Figure 4-1 shows the schematic diagram of an industrial photo reactor. The total
volume of the cylindrical reactor is 750 gallons. The dimensions of the reactor were
estimated based on the assumption that the diameter and length are identical. There are
eleven light wells immersed in the reactor. Three mercury vapor lamps in each well are
used to catalyze the free radical reactions.
L=1.535 m
D=1.535 m
FIGURE 4 - 1. Schematic diagram of industrial photo reactor.
The chlorination of toluene involves a series of consecutive reactions, which occur in
the toluene side-chain, and competitive reactions, which occur in the toluene aromatic
ring. The competitive reactions are electrophilic substitution reactions, which are
second order reactions, first order in chlorine and first order in toluene. The consecutive
reactions are free radical substitution reactions, which are also second order reactions,
first order in chlorine free radical and first order in toluene/chlorinated products. The
42
concentration of chlorine free radicals is a function of chlorine concentration, local light
intensity, and flow patterns of the reaction medium. In order to model the chlorination
process properly, we have to take into account the effects of flow patterns of gas and
liquid, and of the light intensity distribution in the reactors.
4.1 Model equations for chlorination of toluene in the bubble
column/sparged reactors
For model development purposes, extreme mixing patterns were assumed so as to
bracket the real flow patterns in the reactors. Since the reactor diameter is almost
identical with reactor length and the bubbling gas vigorously agitates the liquid phase,
the extent of liquid back mixing is significant. The flow pattern of the liquid can
normally be treated as well mixed. As gaseous reactant, chlorine, and product, hydrogen
chloride, go upward, some of the gas in form of small bubbles may be carried out by the
liquid and may be back mixed. Therefore, the flow pattern of the gas phase could lie
between well mixed and plug flow. In order to evaluate the reactor performance and
determine a flow pattern needed in achieving the best performance, three extreme flow
patterns are considered:
1) The liquid phase is well mixed and the gas phase is in plug flow (henceforth referred
to as m1);
2) Both liquid and gas phase are well mixed (henceforth referred to as m2);
3) Both liquid and gas phase are in concurrent plug flow (henceforth referred to as m3).
43
Under the well mixed condition of the liquid, the isothermal conditions in the reactor
can be easily maintained by using a heating or cooling coil. Since all the chlorination
reactions are exothermic, the cooling coils are used in the industrial reactors to remove
the heat of reactions. Thus, the temperature in the reactors is maintained around 80C.
At this temperature, the vapor pressures of toluene, benzyl chloride, benzal chloride,
benzotrichloride, and o-chlorotoluene are 0.3882105 Pa, 0.3451104 Pa, 0.1391104
Pa, 0.1376104 Pa, 0.7443104 Pa, respectively. The vapor pressure of pure toluene
accounts for more than 35% of the total pressure of 1 atm (1.01325105 Pa). The
volatility of toluene should be taken into account in the model development. The vapor
pressures of the chlorinated products only account for less than 8 % of the total reactor
pressure. The actual partial vapor pressures of the chlorinated products in the liquid
mixture are lower than the vapor pressure of pure liquid due to their lower molar
fractions. Therefore, the volatility of the chlorinated products can be neglected.
The following assumptions are employed for modeling the photo chlorination of toluene
in the bubble column/sparged reactors:
1) The liquid flow rates are constant throughout the reactors.
2) Gas holdup is a constant in the reactor.
3) The evaporation of the chlorinated products is negligible.
4) The toluene vapor is always in equilibrium with the liquid phase.
5) The chlorination of toluene mainly take place in the liquid phase.
44
Under the conditions of well mixedness for the liquid phase and plug flow for the gas
phase (model m1) the model equations can be derived as follows:
Mass balance for toluene:
s[uL (cL, j 1 x1j 1  cL, j x1j )  u g ,0cg ,0 y01  u g , j cg , j y1j ]  VR (1   g )rj1  0
(4-1)
Mass balance for chlorinated products in the liquid phase:
su L (cL, j 1 x ij 1  cL, j x ij )  VR (1   g )rji  0 ,
for i = 2, 5
(4-2)
for i = 6,7
(4-3)
for i = 6,7
(4-4)
Mass balance for gas species dissolved in the liquid phase:
L
uL scL, j 1 ( xij 1  xij ) 
k a c
i
i
L B L, j
( xij*  xij ) sdz  VR (1   g )rji  0 ,
z 0
Mass balance for gas species in the gas phase:
d (u g , j cg , j y ij )
dz
 k Li aB cL , j ( x ij*  x ij )  0 ,
Heat balance:
7
7
7
7
i 6
i 6
i 1
i 1
s[u g ,0  c g ,0 y 0i e gi ,0  u g , j  c g , j y ij e gi , j  u L ( c L , j 1 x ij 1e Li , j 1  c L , j x ij e Li , j )]
(4-5)
 hw, j a w, j (T j  Tw )  0
where, c L , j and c g , j are the liquid and gas molar concentration in reactor j,
respectively. x ij and y ij are molar fractions of species, i, in reactor, j, in the liquid and
gas phase, respectively. The superscripts, i from 1 to 7, refer to toluene, benzyl chloride,
benzal chloride, benzotrichloride, o-chlorotoluene, chlorine, and hydrogen chloride,
respectively. x ij* is the molar fraction of gas species i at the gas and liquid interface. L,
45
s, and VR are the length, cross-sectional area and the total volume of the reactor. VR =
sL. r ji is the reaction rate of species i in reactor j. e Li and e gi are the enthalpies of gas
and liquid species, respectively. The notation used for the other variables is listed in the
nomenclature.
By introducing mean residence time, t L and t g , for the liquid and gas phase,
respectively, and dimensionless length,  (  
z
), equations (4-1) to (4-5) can be
L
rewritten as:
1
1
1
1
[ (cL, j x1j  cL, j 1x1j 1 )  cg , j y1j 
cg ,0 y01 ]  rj1  0
(1   g , j ) t L
tg
t g ,0
1
1
(cL, j x1j  cL, j 1 x1j 1 )  rji  0 ,
(1   g , j ) tL
1
(1   g , j )t L
(c L , j x  c
i
j
x )
i
L , j 1 j 1
1  g, j
i
1 c g , j dy j
 k Li aB cL, j ( x ij*  x ij )  0 ,
t g , j d
s[
1
t g ,0
7
c
i 6
i i
g ,0 0 g ,0
ye

1
tg
7
c
i 6
1
 hw, j aw, j (T j  Tw )  0
L
g, j
y ij egi , j 
for i = 2, 5
1
1
(4-6)
k

i
L
(4-7)
a B cL , j ( x ij*  x ij )d  r ji  0 ,
0
for i = 6,7
(4-8)
for i = 6,7
(4-9)
7
1 7
( cL , j 1 xij 1eLi , j 1  cL , j xij eLi , j )]
t L i 1
i 1
(4-10)
46
In addition to the above mass and heat balance equations, the following relationships
hold in the gas and liquid phase:
Phase equilibrium relationship:
y1j  K 1j x1j
(4-11)
y ij  K ij x ij*
(4-12)
Summation constraint for molar fractions in the liquid and gas phase:
x
i
j
 1.0
(4-13)
 1.0
(4-14)
i
y
i
j
i
The overall molar concentration of the gas species can be determined by the following
equation if the gas phase is assumed to be an ideal gas.
cg , j 
Pj
RT j
(4-15)
Under the condition of well mixedness for both the gas and liquid phase, the mass
balance for the hydrocarbon species (i=1,5) and heat balance are the same as in flow
pattern m1. The model equations for chlorine and hydrogen chloride can be derived as
follows:
47
Mass balance for gas species dissolved in the liquid phase:
uL scL, j 1 ( xij 1  xij )  kLi aBcL, j ( xij*  xij )VR  VR (1   g )rji  0
for i = 6,7
(4-16)
for i = 6,7
(4-17)
Mass balance for gas species in the gas phase:
ug ,0cg ,0 y0i  ug , j cg , j yij  kLi aBVRcL, j ( xij*  xLi , j )  0
The model equations (4-16) and (4-17) can be expressed as follows:
1
(1   g , j )t L
1
tg, j
(cL, j xij  cL, j 1 xij 1 ) 
c g , j y ij 
1
t g ,0
1
1   g, j
kLi aBcL, j ( xij*  xij )  rji  0 ,
c g ,0 y ij ,0  k Li aB cL, j ( x ij*  x ij )  0 ,
for i = 6,7
(4-18)
for i = 6,7
(4-19)
Under the conditions of concurrent plug flow for both gas and liquid phase (model m3),
the model equations can be written as follows:
Mass balance for toluene:
1
1
u L d (cL, j x j )
1 d (u g , j cg , j y j ) 1

 rj  0
1 g
dz
1 g
dz
(4-20)
Mass balance for chlorinated products in the liquid phase:
i
u L d (c L , j x j ) i
 rj  0
1  g
dz
for i = 2,5
(4-21)
48
Mass balance for gas species dissolved in liquid phase:
i
u L d (c L , j x j )
1

k Li aB cL, j ( x ij*  x ij )  r ji  0 ,
1  g
dz
1  g
for i = 6,7
(4-22)
for i = 6,7
(4-23)
Mass balance for gas species in the gas phase:
i
1 d (u g c g , j y j )
 k Li a B cL, j ( x ij*  x ij )  0
tg
dz
Heat balance:
 7

i i
i i
 d  (uL cL , j x j eL , j  u g , j cg , j y j eg , j )  1
  hw, j aw (T j  Tw )  0
s  i 1
dz

 L


By introducing dimensionless gas velocity, Ug ( U g 
ug
u g ,0
(4-24)
), and other variables used
above, equations (4-20) to (4-24) can be rearranged as:
1
d (cL, j x1j )
(1   g )t L
d
1
d (cL, j x ij )
(1   g )t L
d
1
d (cL, j x ij )
(1   g )t L
d

1
d (U g , j c g , j y1j )
(1   g )t g ,0
d
 r j1  0
 r j1  0 ,

1
k Li a B cL, j ( x ij*  x ij )  r ji  0 ,
1  g
(4-25)
for i = 2,5
(4-26)
for i = 6,7
(4-27)
49
i
j
1 d (U g , j c g , j y )
 k Li a B cL, j ( x ij*  x ij )  0 ,
tg
d
for i = 6,7




d  c L , j  x ij eLi , j 
d U g , j c g , j  y ij egi , j 
s 
i
i
 s 
  1 h a (T  T )  0
w, j w
j
w
tL
d
t g ,0
d
L
(4-28)
(4-29)
B.C. :   0 , xij  xij 1
y ij  y0i
(4-30)
T j  T j 1
The reaction rates, r ji , in the above equations can be evaluated by using the following
appropriate kinetic expression:
r  k e
1
j
r  (k e
2
j

1
0
r  (k e
3
j

3
0
r k e
5
j

2
0
r k e
4
j
Ea1
RT j

1
0

4
0
Ea1
RT j
Ea2
RT j
Ea3
RT j
Ea4
RT j
r  ( k e
6
j
rj7  rj6
1
0

c x [Cl] j  k e
1
L, j j
4
0
c x k e
1
L, j j

2
0
cL, j x  k e
2
j

3
0
Ea2
RT j
Ea3
RT j

Ea4
RT j
cL, j x1j c6j
(4-31)
cL, j x 2j )[Cl] j
(4-32)
cL, j x3j )[Cl] j
(4-33)
cL, j x3j [Cl] j
(4-34)
cL, j x1j cL, j x6j
(4-35)
Ea1
RT j
c x k e
1
L, j j
2
0

Ea2
RT j
cL, j x  k e
2
j
3
0

Ea3
RT j
cL, j x )[Cl] j  k e
3
j
4
0

Ea4
RT j
cL, j x1j cL, j x6j
(4-36)
(4-37)
50
where [Cl] j is the concentration of chlorine free radicals, whose value is a function of
photo radiation flux rate, wavelength, concentration of chlorine, and mixing state of the
free radicals.
4.2 Evaluation of Concentration of Chlorine Free Radicals
The concentration of chlorine free radicals can only be determined with the
knowledgement of the photon radiation field in the reactor. The photon radiation
distribution in the photo reactor can be evaluated by using a proper light emission
source model. Several types of source emission models have been proposed in the open
literature. Those models include the line source models, which have parallel planes
perpendicular to the lamp axis (Harris and Dranoff, 1965), spherical and isotopic
emission (Jacob and Dranoff, 1966), and three-dimensional source models, which have
volumetric (Irazoqui, Cerdá, and Cassano 1973) and superficial emission (Stramigioli,
Santarelli, and Foraboschi 1975). In the industrial photochlorination of toluene, medium
pressure mercury arc lamps are utilized as light sources to catalyze the photo reaction.
The volumetric emission model is the best representation for this kind of lamp.
The volumetric emission (ESVE) model is based on the following assumptions:
1) The emitters of the radiation source are uniformly distributed inside its volume.
51
2) Each elementary volume inside the source has an isotropic emission, that is, the
specific intensity associated with each bundle of radiation coming from the same
element of the column is independent of direction.
3) Any elementary volume inside the source emits, at any frequency, an amount of
energy proportional to its extension1.
4) The energy emitted from any elementary volume of the lamp is due to spontaneous
emission, and each of these differential volumes is transparent to the emission of its
surroundings.
The photon radiation is attenuated according to the Lambert's law. By making use of the
radiation balance (Figure 4-2), Cassano et al. (1986) derived the local volumetric rate of
energy absorption (LVREA) as follows:
v 
ea 

v 0

v
N e Pv hv
e
dv    ( dV
)exp(    v d* )
2
4
  
0
(4-38)
where Ne represents the number of emitters per unit volume of the source, and Pd
gives, per unit time, the probability of emission at a given frequency interval d,  is
the attenuation coefficient at the point of absorption, which is a function of position.
In a homogeneous system, the attenuation coefficient is the product of the absorption
coefficient and the local concentration of absorbing species. In a gas-liquid system,
1
Extension means the property of an object by which it occupies space.
52
z
y


n
(z,r)


r1
rL
r
x
FIGURE 4 - 2. Variables involved in the theoretical analysis on LVREA
Otake et al. (1983) obtained a semi-empirical correlation between the effective
coefficient, e, and absorption coefficient as follows:
 e  c(1   g )  k
6 g
dB
(4-39)
where  is the absorption coefficient of a species in the liquid phase, (l mol-1 cm-1); c is
the concentration of that species in the liquid phase, (mol l-1);  g is the mean gas holdup, dB is the mean bubble diameter, (cm-1); k is an empirical coefficient which is a
function of the relative refractive indices of the gas and liquid, and of the reflectivity
and absorption coefficient of the dispersed phase (k approximately has the value of
0.125). This correlation is used for calculating the attenuation coefficient in this study.
53
Since the absorption coefficient is a function of wavelength (the absorption coefficient
of chlorine varies with wavelength as shown in Figure 3-6) and the actual lamps have
different spectral distribution (e.g. the spectral distribution of medium pressure arc lamp
is shown in Figure 4-3), the reaction behavior should be different if the reaction
medium is exposed to the light of different wavelengths. In order to understand the
effect of wavelength on the photo reaction, we should study the photochlorination of
toluene in the presence of monochromatic light. In reality, the monochromatic light can
be achieved with the aid of an optical filter to cut off the undesired wavelength.
However, the reactor performance in the presence of polychromatic light will also be
examined in this study.
30
Output power, w
25
20
15
10
5
546.1
404.5
334.1
302.5
289.4
275.3
265.2
253.7
240.0
236.0
222.4
0
Wavelength, nm
FIGURE 4 - 3. Spectral distribution of Hanovia 679A0100, medium-pressure,
mercury arc lamp with 450 watts input power (Scaiano, 1989).
54
For a monochromatic light source, the LVREA can be expressed as:

N hv
e
de   e e ,    ( dV
)exp(    e d* )
2
4   
0
a
v
(4-40)
The effective absorption coefficient, e, is a function of position, and its value depends
on the local concentration of absorption species and the flow pattern.
If we assume that the absorbing species is chlorine dissolved in toluene only, the local
volumetric rate of energy absorbed by the dissolved chlorine in toluene can be described
as (see details in Appendix D):

N e , hv
e
( dV
)exp(    e d* )



2
4   
0
deva,Cl2  cCl2 ( 1   G )
(4-41)
Under the well mixed condition for the liquid phase, the local volumetric rate of energy
absorbed by chlorine dissolved in the liquid phase is (see details in Appendix E) :
  sin 1 (
rL
r
)
N h
2
2 1
de ,Cl 2 (r , z )  e c(1   G )
1 rL [rL  (r sin  ) ] 2
2
  sin ( )
  tg 1 [
]

a
r
r cos   r L2  ( r sin  ) 2
z z
0
  tg 1 [
r cos   r L2  ( r sin  ) 2
z
L
z
exp{
 e [r cos   rL2  (r sin  ) 2 ]
}dd
sin 
]
(4-42)
The coordinate employed in evaluating the energy absorption is shown in Figure 4-4.
55
z
y
zL
rL
L
z0
x
(z,r)
r1
r2
FIGURE 4 - 4. Photo reactor geometry near the lamp.
As described by equations (3-4) to (3-20), the reaction mechanism of photo chlorination
of toluene is the substitution of chlorine for hydrogen in the side chain of toluene by the
attachable chlorine free radical. At the well mixed condition for stable species, two
mixing states in the liquid phase may exist for free radicals: state 1 (henceforth referred
to as s1), well mixed for stable species and for free radicals; state 2 (henceforth referred
to as s2), well mixed for stable species and no mixing for free radicals. These two
mixing states are illustrated in Figure 4-5 according to Felder et al. (1970) and Alfano et
al. (1988). In state 1, the characteristic mixing time is shorter than the mean lifetime of
the stable species and free radicals. In state 2, the characteristic mixing time is shorter
than the mean lifetime of the stable species but longer than that of the free radicals. The
56
free radicals during their lifetime are immobile on the time scale of the mixing process
and it can be said that they "are born, live, and die" in the same position.
The termination of chlorine free radicals is a second order reaction. The half-life,
t1 / 2 
1
, of the chlorine free radical is proportional to the reciprocal of the
2k t [Cl]0
product of the termination rate constant, kt, and concentration of local chlorine free
radicals, [Cl]0. At higher light intensity, a higher concentration of chlorine free radicals
would be produced, and the shorter lifetime of the radicals would be experienced. Due
to the non-uniformity of light intensity distribution in the photoreactor, the half-life of
the chlorine free radicals can span the range of about 10-5 to 10-1 seconds. Hence, in a
real system, the mixing state may fall between states s1 and s2. Model s2 is expected to
Stable
Species
Free
Radicals
S2
S1
characteristic mixing time
mean lifetime
represent reality better.
FIGURE 4 - 5. Illustration of mixing states
57
In state 1, the concentration of chlorine free radicals can be calculated based on the total
energy absorbed by chlorine in toluene. The total energy absorbed by chlorine is:
1
z  L r  r2   sin (
E1a,  N e hc( 1   G ) 
r

rL
r
  tg 1 [
)
[ r12  ( r sin  )2 ]
z  0 r  r1   sin 1 ( rL )
r
1
r cos   rL2 ( r sin  )2
z z
0
]
r cos   rL2 ( r sin  )2
z L z
]

2
  tg 1 [
exp{ 
 e [ r cos   r12  ( r sin  )2 ]
}dddrdz
sin 
(4-43)
The volumetric average concentration of quanta absorbed by chlorine is:
 Ia 
Ea
 (r2 2  r12 )Lh
(4-44)
The volumetric average concentration of the chlorine free radical is :
 [ Cl ]  s1  [
  Ia 
kt
]
1
(4-45)
2
where  is the quantum yield of chlorine free radicals and kt is the termination rate
constant for the chlorine free radicals.
In mixing state 2, the local concentration of chlorine free radicals is :
dea 12
[ Cl ]s 2  (
)
kt h
(4-46)
The volumetric average concentration of chlorine free radicals is :
2
 [ Cl ]  s 2  2
( r2  r12 ) L
z  L r  r2
  r [ Cl ]drdz
z  0 r  r1
(4-47)
58
Similarly, we can derive the local volumetric rate of quanta absorbed by chlorine in the
liquid phase for the plug flow model is derived as:
  tg 1 [
  sin 1 ( rrL )
N
1
de ,Cl 2 ( r , z )  e c( 1   G ) 
[ rL2  ( r sin  )2 ] 2
2
  sin 1 ( rL )
]
r cos   rL2 ( r sin  )2
z z
L
]

a
r
r cos   rL2 ( r sin  )2
z z
0
1
  tg [

exp{    e d* }dd
0
(4-48)
In the first mixing state (s1), the cross-sectional average concentration of chlorine free
radicals is:
2
2
1
 [ Cl ]  m3,s1  [
dea,Cl 2 rdr ] 2

k t ( r2  r1 ) r1
r
(4-49)
In the second mixing state (s2), the cross-sectional average concentration of chlorine
free radicals is:
2
de ,Cl 2 12
2
[
] rdr
2
2 
( r2  r1 ) r1
kt
r
 [ Cl ]  m3,s 2 
a
(4-50)
Once the concentration of chlorine free radicals is known, one can solve the reaction
model properly.
4.3 Performance of Photo Bubble Column Reactors
With appropriate parameters, the developed reaction model equations in the previous
section can be solved numerically (Details in Appendix C). The industrial photo reactor,
as shown in Figure 4-1, can be treated as eleven annular reactors for simplicity. The
dimension of each identical reactor is determined and listed in Table 4.1. The heat
59
transfer area listed in this Table is evaluated by taking into account the area of the inner
well, where the heat of reaction is removed by the cooling water.
TABLE 4 - 1. Dimension of annular model reactor
Name
Length of reactor, L, m
Outer diameter of inner well of reactor, r2, m
Inner diameter of outer well of reactor, r1, m
Diameter of lamp, rL, m
Cross-section area, s, m2
Heat transfer area, aw, m2
Values
1.5
2.310-1
4.010-2
1.2710-2
4.0310-2
1.8810-1
The lamp's parameters are listed in Table 4-2 according to the spectral distribution (as
shown in Figure 4-3) and commercial catalog of the UV lamp. The arc length accounts
for three times of the arc length of one lamp since there are three UV lamps in one light
well.
TABLE 4 - 2. Dimension of the UV lamp and its emission rates at 313 nm
Name
Length of arc of the lamp,(zL - z0), m
Quanta emitted from the lamp, k Einstein /(m3 s)
Values
0. 328610-1
0.622410-3
Since detailed operating conditions of the industrial photo reactor are not available, we
have to use some published correlations to determine the other model parameters. They
are summarized in Table 4.3. Under the specified reaction conditions, the mole ratio of
toluene to chlorine is 12.63. As we stated before, the toluene is passed through four
60
reactors in series and the chlorine gas is bubbled through the reactors in cross flow. The
overall conversion of toluene is 31.57% if the chlorine is completely consumed.
TABLE 4 - 3. Operating Conditions and Parameters Used in Reaction Model for
Chlorination of Toluene in Bubble Column Reactor
Parameters
Value
Used in flow patterns
m1
m2
m3



















Column pressure, Pa
Column temperature, C
Cooling water temperature, C
Superficial liquid velocity, uL, m/sec
Superficial liquid velocity, ug, m/sec
Gas holdup, g
Mass transfer coefficient for Cl2, k L6 ,m/sec
1.01325105
80
30
0.005
0.1
0.211
1.18210-4
Mass transfer coefficient for HCl, k L7 ,m/sec
Diameter of a bubble, m
1.33110-4



510-3



The absorption coefficients of chlorine at various wavelengths are read from Figure 3-7
and listed in Table 4-4.
TABLE 4 - 4. Absorption Coefficients of Chlorine at Various Wavelengths.
Wavelength, nm
Absorption coefficient, , m2/mol
289.1
296.0
302.8
313.1
334.4
366.0
404.4
436.0
495.0
1.36
2.02
3.12
5.01
6.35
2.36
0.432
0.143
8.2910-3
61
Figure 4-6 shows the selectivity to benzyl chloride versus conversion of toluene based
on our model predictions. Each data point represents the value at the reactor outlet. It
can be seen from Figure 4-6 that the selectivity decreases as the conversion of toluene
increases. Similar performance is obtained for flow patterns m1 and m2 since the liquid
phase is well mixed in these flow patterns and reactions take place in the liquid phase.
The backmixing of the gas phase has no effect on the selectivity to benzyl chlorine
since gas phase reactions are neglected (details in Appendix E) and the mean mass
transfer resistance is not important. The effect of mixing states (s1 and s2) on the
reactor
Selectivity to benzyl chloride, %
100.0
99.5
m1, s1
m1, s2
m2, s1
m2, s2
m3, s1
99.0
98.5
98.0
97.5
97.0
96.5
96.0
95.5
95.0
0
8
16
24
Conversion of toluene, mol.%
32
62
FIGURE 4 - 6. Selectivity to benzyl chloride versus overall conversion of toluene
(reaction condition: superficial chlorine gas velocity, 10 cm/sec; superficial toluene
velocity, 0.5 cm/sec; relative light intensity, 1).
performance is not significant under the current reaction conditions. Since the photon
radiation in the photo reactor is very high, the concentration of chlorine and free
radicals in the liquid phase is similar in these two states. It can be seen from Figure 4-6
that the flow patterns m1 and m2 is superior to the flow pattern m3 in order to achieve
high selectivity to the benzyl chloride.
When the gas and liquid are in the cocurrent plug flow through the reactor a significant
amount of chlorine gas is dissolved in the liquid phase near the reactor inlet as shown in
Figure 4-7. In that region, the competitive reactions become significant due to the high
concentration of chlorine in the liquid. The overall chlorination rates are also higher
compared to the reaction rates when the liquid phase is well mixed. Large amount of
heat from the exothermic reactions is released, which causes a temperature rise in that
region as shown in Figure 4-8. It is known from the kinetic studies that the activation
energy increases as the consecutive reactions move from the first, to the second, and to
the third one. The reaction rates for the last two reactions increase faster than for the
first reaction as temperature increases. Therefore, using the cocurrent plug flow results
in the lowest selectivity to benzyl chloride among these three flow patterns. Another
disadvantage of using the cocurrent plug flow for this reaction system is that reaction
run-away could possibly occur. Our simulation results indicate that the superficial gas
velocity could increase to about one and half times of the inlet velocity due to the
63
increased temperature of reaction medium and evaporation of toluene, which is shown
in Figure 4-9.
Molar fraction of Cl2 in liquid phase
9.0E-03
m1,s1
m1,s2
m2,s1
m2,s2
m3,s1
8.0E-03
7.0E-03
6.0E-03
5.0E-03
4.0E-03
3.0E-03
2.0E-03
1.0E-03
0.0E+00
0.0
0.2
0.4
0.6
0.8
1.0
Dimensionless length, z/L
FIGURE 4 - 7. Molar fraction profile of chlorine in the liquid phase along the
reactor length in the bubble column where gas and liquid are cocurrent plug flow.
64
120
Temperature, 'C
100
80
60
m1,s1
m1,s2
m2,s1
m2,s2
m3,s1
40
20
0
0.0
0.2
0.4
0.6
0.8
1.0
Dimensionless length, z/L
FIGURE 4 - 8. Temperature profile along the reactor length in the bubble column
where gas and liquid are cocurrent plug flow.
Dimensionless gas velocity, ug/ug0
1.8
1.6
m3,s1
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0.0
0.0
0.2
0.4
0.6
0.8
1.0
Dimensionless length, z/L
FIGURE 4 - 9. Dimensionless superficial gas velocity profile along the column
where the gas and liquid are in cocurrent plug flow.
65
From the modeling of photon chlorination of toluene in bubble columns, we learned
why the current industrial process is conducted in bubble column/sparged reactors with
identical diameter and length. The liquid phase can be easily mixed and the low mean
concentration of chlorine in the liquid phase is readily obtained. The constant column
temperature can be easily maintained. However, in order to achieve high selectivity to
benzyl chloride, the conversion of toluene has to be limited to a low level. An external
recycle loop always has to be implemented in this process.