Metodos Diferencial y Otros

Chemical Reactor Design
Y
Youn-Woo
W L Lee
School of Chemical and Biological Engineering
Seoul National University
y
155-741, 599 Gwanangro, Gwanak-gu, Seoul, Korea  ywlee@snu.ac.kr  http://sfpl.snu.ac.kr
第5章
Collection and Analysis of Rate Data
Chemical Reactor Design
化學反應裝置設計

Objective
★ Determine the reaction order and specific reaction rate from

experimental data obtained from either batch or flow reactors.
★ Describe how to use equal-area differentiation, polynomial

fitting,
g, numerical difference formulas and regression
g to analyze
y
experimental data to determine the rate law.
★ Describe how the methods of half lives, and of initial rate, are
used to analyze rate data.
★ Describe two or more types of laboratory reactors used to

obtain rate law data along with their advantages and disadvantages.
★ Two common type reactors for obtaining rate data:
((1)) Batch reactor : Conc. vs. time

- homogeneous reaction during transient operation
- Concentration (or pressure) are usually measured and
recorded at different times during the course of reaction.
(2) Differential reactor : Conc. @ steady state

- Solid-fluid
Solid fluid heterogeneous reactions
- Product concentration is usually monitored
f different
for diff t sets
t off feed
f d conditions.
diti
★ Six different methods

methods of analyzing the data collected
(1) the differential method

primarily in analyzing
(2) the integral method batch reactor data
(3) the method of half-lives

(4) method of initial rates
(5) linear regression
(6) nonlinear regression
(least squares analysis)
5.1 Algorithm for Data Analysis
Steps in Analyzing Rate Data


A. Differential Analysis

B. Integral Analysis


5. 2 Batch Reactor Data
t CA
0 50 0
50.0 Measuring concentration
50 38.0 as a function of time
100 30.6
.. ..
. .
Differential,
Data
integral
analysis
or nonlinear regression method
i i  and
Determining
D dk
in -rA = kCA

5.2 Batch Reactor Data
 Irreversible reaction
Determine  and k
-Determine
by either nonlinear regression or by numerically
diff
differentiating
i i concentration
i versus time
i data
d
For example
-for decomposition reaction (only one reactant)
Assuming that the

rate la
law is of the
A  products
-rArA= kkAACCAA
form
-rA = kACA
(5-1)
then differential method may be used.

5.2 Batch Reactor Data
Excess Experiments
Consider the irreversible reaction : A + B  p
products
 rA  k AC A C B (5-2)
Excess B experiments: Excess A experiments:

CB remains unchanged CA remains unchanged
during the reaction (CBo) during the reaction (CAo)
 rA  k AC A C B  k AC B0C A  k C A  rA  k AC A C B  k AC A0C B  k C B

(5-3) (5-4)
O
Once  andd  are determined,
d t i d kA can be
b calculated
l l t d from
f the
th measurementt off -rA
at known concentration of A and B
r
k A   A
dm 3
/ mol 
   1
C A CB s (5-5)
5.2.1 Differential Method of Analysis
Consider a reaction carried out isothermally in a constant-volume batch reactor

and the concentration recorded as a function of time. By combining the mole
balance with the rate law given by Equation (5-1), we obtain
dC A
  k AC A (5-6)
dt
After taking the logarithm of both sides of Eq. (5-6)
 dC A 
ln    ln k A   ln C A (5-7)
 dt 

Differential Method to determine α and k
 dC A 
ln    ln k A   ln C A
 dt 
ln ln
P
P kA=
slope= (CAp)
ln ln
CA CAp

How to get –dCA/dt ?
dC A
T obtain
To b i theh derivative
d i i –dCA/dtdCA/d usedd iin this
hi plot,
l
dt
we must differentiate the concentration-time data
either numerically and graphically. These methods are:
★ Graphical differentiation
★ Numerical differentiation formulas
★ Differentiation of a polynomial fit to the data

How to get –dCA/dt ? (Graphical Method)
Time (min) t0 t1 t2 t3 t4 t5
1 Tabulate the (ti, Ci)
1. C
Concentration
t ti ((mol/dm
l/d 3) C0 C1 C2 C3 C4 C5
2. For each interval, calculate

t and C
ti Ci t C C/t dC/dt
3. Calculate C/t as an estimate
g slope
of the average p in an interval.
t0 C0 (dC/d )0
(dC/dt)
4. Plot these values(C/t) t1-t0 C1-C0 C/t)1
t1 C1 (dC/dt)1
as a histogram versus ti.
t2-tt1 C2-C
C1 C/t)2
t2 C2 (dC/dt)2
5. Next draw in the smooth curve
that best approximates the area t3-t2 C3-C2 C/t)3
t3 C3 (dC/dt)3
under
d the
h histogram.
hi
t4-t3 C4-C3 C/t)4
t4 C4 (dC/dt)4
6. Read estimates of the dC/dt
from this curve at the data points t5-tt4 C5-C
C4 C/t)5
t5 C5 (dC/dt)5
t1, t2, … and complete the table.

How to Get –dCA/dt ? (Graphical Method)
★ Graphical Method (Equal-Area Graphical Differentiation)

Draw smooth curve
that best approximates
 dC A  the area under histogram
 
 dt
d 0
 C A 
 
 t 1
 dC A 
 
 dt 1
A  C A 
 
 dC  B  t 2
 A
 dt  2  C A 
 
 t 3  C A 
 dC A   
   t 4 C
CA 
 dt
d 4  
 t 5
0 1 2 3 4 5
t
How to Get –dCA/dt (Numerical Method)
Concentration (mol/dm3) CA0 CA1 CA2 CA3 CA4 CA5
★ Numerical Method (Independent variables are equally spaced)

 dC A   3C A0  4C A1  C A 2
Initial po int :   
 dt  t 0 2 t
 dC  C  C A0
Interior po int s :  A   A 2
 dt  t1 2 t
 dC A  C  C A1
   A3
The three-point
differentiation
 dt  t 2
 dC A 
2 t
C  C A2
 dC A 
  
1

C A(i1)  C A(i1)  
formulas    A4  dt ti 2t
 dt
d t 3 2 t
 dC A  C  C A3
   A5
 dt  t 4 2 t
 dC A  C  4C A 4  3C A5
Last po int :    A3
 dt  t 5 2 t
How to Get –dCA/dt ? (Polynomial Fit)
★ Polynomial
y Fit
Concentration (mol/dm3) CA0 CA1 CA2 CA3 CA4 CA5
Polynomial fit with software program to get best value of ai
n-th order polynomial
C A  a0  a1t  a2t 2  ...  ant n (5-11)

Differential equation
dC A
 a1  2a2t  3a3t 2  ...  nant n 1 (5-12)
dt
How to Get –dCA/dt (Polynomial Fit)
3rd-order
3rd order 5th-order
5th order Negative
N i
polynomial polynomial derivative
t5
Care must be taken in choosing the order of the polynomial. If the order is too
low, the polynomial fit will not capture the trends in the data. If too large an
order is chosen,
chosen the fitted curve can have peaks and valleys as it goes through
most all of the data points, thereby producing significant errors when the
derivatives, dCA/dt, are generated at the various points. Seoul National University
Finding the Rate Law Parameter

Finding the Rate Law Parameter

Example 55--1: Determining the Rate Law
The reaction of triphenyl methyl chloride (trityl) (A) and methanol (B) was
carried out in a solution of benzene and pyridine at 25oC. Pyridine reacts with
HCl that
h then
h precipitates
i i as pyridine
idi hydrochloride
h d hl id thereby
h b makingki theh reaction
i
irreversible. The concentration-time data was obtained in a batch reactor. The
initial concentration of methanol was 0.5 mol/dm3.
(C6H5)3CCl (A) + CH3OH (B)  (C6H5)3COCH3 (C) + HCl (D)
Time (min) 0 50 100 150 200 250 300

CA (mol/dm3) x 103 50 38 30 6
30.6 25 6
25.6 22
22.2
2 19
19.5
5 17
17.4
4
CAo=0.05 M
CBo=0.5 M
Part (1) Determine the reaction order with respect to trityl (A)
Part (2) In a separate set of experiments, the reaction order with respect to
methanol was found to be first order. Determine the specific
p reaction
rate constant.
Concentration of triphenyl methyl chloride

as a function
f ti off time
ti
55
50
103 (mol/dm3)
45
40
35
30
CA x 1
25
20
15
0 50 100 150 200 250 300
t (min)

Solution
Part (1) Determine the reaction order with respect to trityl (A)
Step 1 Postulate a rate law
 rA  k AC A C B (E5-1.1)
Step 2 variable, which is this case is CA.

Process your data in terms of the measured variable
Step 3 Look for simplifications. Because concentration of methanol is 10 times

p y methyl
the initial concentration of triphenyl y chloride,, its concentration is
essentially constant
CB  CB 0 (E5-1.2)
Substituting for CB in Equation (E5-1.1)
 rA  k AC B0C A  k C A (E5 1 3)
(E5-1.3)

Step 4 Apply the CRE algorithm
dN A
Mole Balance  rAV (E5-1.4)
dt
Rate law  rA  k C A (E5-1.3)
V  V0
Stoichiometry: Liquid NA
CA 
V0
dC A
Combine   k C A (E5-1 5)
(E5-1.5)
dt
Taking the natural log of both sides of Equation (E5-1.5)

 dC A 
ln    ln k    ln C A (E5-1.6)
 dt 
 dC 
The slope of plot of ln  A  versus ln C A will yield the reaction order 
 dt 
with respective to triphenyl methyl chloride (A).
(A)
dC A
Step 5 Find  as a function of CA from concentration-time
concentration time data.
dt
dC
We will find  A by each of the three methods just discussed, (1) the graphical
dt
(2) finite difference, and (3) polynomial methods.

Step 5A.1a Graphical method
t C A  103  C A / t 10 4  dC A / dt 10 4
(min) (mol / dm 3 ) (mol / dm 3  min) (mol / dm 3  min)
0 50 3.0

C 2.40
t
 38  50 
50 38 1.86
1.48
3
    10
 50  0 
 2.4  10  4 100 30.6 1.2
1.00
150 25.6 0.8
0.68
200 22.2 0.5
0.54
250 19.5 0.47
0 42
0.42
300 17.4
Step 5A.1a Graphical method
 dC A 
  3.0 3.0
 dt  0
 C A 
2.4  
 t 1
 dC A  2.0
 
 dt 1 1.86
 C A 
 
1.48  t  2
 dC A 
  12
d  2 1.2
 dt C
CA 
1.0 1.00  
 dC A   t 3  C A 
   
 dt  4 0.8 0.68  t  4
0.5  C A 
 
0 47
0.47 0.54  t
t 5
0.42
0.0
0 50 100 150 200 250 300

Step 5A.1b Finite Differential Method
 dC A   3 C A 0  4 C A1  C A 2 [  3 ( 50 )  4 ( 38 )  30 . 6 ]  10  3
t 0       2 . 86  10  4
 dt  t  0 2t 2 ( 50 )
 dC A  C A2  C A0 [ 30 . 6  50 ]  10  3
t  50       1 . 94  10  4
 dt  t  1 2t 2 ( 50 )
 dC A  C A 3  C A 1 [ 25 . 6  38 ]  10  3
t  100       1 . 24  10  4
 dt  t  2 2t 2 ( 50 )
 dC A  C A4  C A2 [ 22 . 2  30 . 6 ]  10  3
t  150       0 . 84  10  4
 dt  t  3 2t 2 ( 50 )
 dC A  C A5  C A3 [19 . 5  25 . 6 ]  10  3
t  200       0 . 61  10  4
 dt  t  4 2t 2 ( 50 )
 dC A  C A6  C A4 [17 . 4  22 . 2 ]  10  3
t  250       0 . 48  10  4
 dt  t  5 2t 2 ( 50 )
 dC A  C A 4  4 C A 5  3C A 6 [ 22 . 2  4 (19 . 5 )  3 (17 . 4 )]  10  3
t  300       0 . 36  10  4
 dt  t  6 2t 2 ( 50 )

Step 5A.1c Polynomial method
t CA dCA/dt
Another method to determine (dCA/dt) is to fit 0 50 -0.299
the concentration of A to a polynomial in time
and then to differentiate the resulting 50 38 -0.189
polynomial.
l i l WeW first
fi choose
h the
h fourth
f h 100 30.6 -0.120
polynomial degree
150 25.6 -0.081
C A  a0  a1t  a2t 2  ...  ant n 200 22.2 -0.061
dC A 250 19.5 -0.049
 a1  2at  3a2t 2  ...  nant n 1
dt 300 17 4
17.4 -0.034
0 034
C A  0.04999  2.978  10 4 t  1.343  10 6 t 2  3.485  10 9 t 3  3.697  10 12 t 4

dC A
 10 3   2 .978  0 .002686 t  1 .045  10  5 t 2  1 .479  10 8 t 3
dt
Step 5A.1c Polynomial method
C A  0.04999  2.978  10 4 t  1.343  10 6 t 2  3.485  10 9 t 3  3.697  10 12 t 4

SUMMARY: There is quite a close agreement between

the graphical technique
technique, finite difference
difference, and polynomial methods
min mol/dm3 Graphical Finite Polynominal

Difference
t CAx103 -dC
dCA/dt -dC
dCA/dt -dC
dCA/dt
0 50.0 3.00 2.86 2.98
50 38.0 1.86 1.94 1.88
100 30.6 1.20 1.24 1.19
150 25.6 0.80 0.84 0.80
200 22.2 0.68 0.61 0.60
250 19.5 0.54 0.48 0.48
300 17.4 0.42 0.36 0.33

From Figure E5-1.3, we found the slope to be 1.99 so that the

reaction is said to be second order w.r.t.
w r t triphenyl methyl chloride.
chloride
To evaluate k’, we can evaluate the derivative and CAp=20x10-3
mol/dm3, which is
 dC A  4
   0.5  10 mol/dm  min
3
 dt  p
then
 dC A 
 
 dt  p 0.5  10  4 mol/dm
l/d 3  min
i d 3
dm
k    0.125
2
C Ap  
20  10 mol/dm
3 3
2
mol  min

Excel plot to determine  and k

Coefficients:
10.00
b[0] -6.6685984257
dC A b[1] 1.9958592737
  0.0013C 1A.99
dt r² 0.9942392165
  1.99
dC A
  10 4
dt 1.00 k   0.13 dm 3 /mol  min
 mol 
 dm 3  min 
  0.5
Graphical
Finite
Regression again
Polynomial
0.10
10 20 100  2
 mol 
C A  103  dm 3 
  k   0.122 dm 3 /mol  min
Part (2) The reaction was said to be first order wrt methanol, =1,
k   kCB0  C B 0 k
Assuming CB0 is constant at 0.5 mol/dm3 and solving for k yields
dm 3
k
k

0. 122 3

mol  min  0.244 dm /mol
2

CB 0 mol min
0.5 3
d
dm
The rate law is

 rA  0.244(dm3 /mol)2 / min C A2C B 
Integral Method
★ The integral method uses a trial-and-error procedure to find rxn order:
We gguess the reaction order and integrate
g the differential equation
q used to
model the batch system. If the order we assume is correct, the appropriate
plot of the concentration-time data should be linear.
★ It is important to know how to generate linear plots of functions of CA

versus t for zero-, first-, and second-order reactions.
★ The integral method is used most often when the reaction order is
known and it is desired to evaluate the specific reaction rate constants
att different
diff t temperatures
t t t determine
to d t i the
th activation
ti ti energy.
★ Finally we should also use the formula to plot reaction rate data in
terms of concentration vs. time for 0, 1st, and 2nd order reactions.
zero order first order second order
dC A dC A 2
dC A
 k   kC A   kC A
dt dt dt
@ t  0, C A  C A0 @ t  0, C A  C A0 @ t  0, C A  C A0
C A  C A0  kt ln
CA
 kt 1 1
  kt
C A0 C A C A0
C A0
slope = -k
k
CA C 1
ln A slope = k slope = k
C A0 CA
t t t
The rxns are zero, first, and second order respectively since the plots are linear.
second order
dC A 2
  kC A
dt
1
@ t  0, C A  C A0
CA
non linear
1 1
  kt
C A C A0
t
If the plots of concentration data versus time had

turned out not to be linear, we would say that the
proposed reaction order did not second order.
Example 55--2
Use the integral method to confirm that the reaction is second order with respect
to trityl(A) as described in example 5-1 and to calculate the specific reaction rate k'.
Trityl(A) + Methanol (B)  Products
dC A 
  k C A   k C A
2
dt
Integrating with @ t  0, C A  C A0
1 1
  k t
C A C A0
Time (min) 0 50 100 150 200 250 300
( l/d 3)
CA (mol/dm 0.05
0 05 0.038
0 038 0.0306
0 0306 0.0256
0 0256 0.0222
0 0222 0.0195
0 0195 0.0174
0 0174
1/CA (dm3/mol) 20 26.3 32.7 39.1 45 51.3 57.5
Example 55--2
1
second order  0.12t  20.12
CA
60
Experiments
50 Regression
ol)
1/CA (dm /mo
3
40
1 1
30
  k t
C A C A0
20
0 50 100 150 200 250 300 350
t (min)
Example 55--2
 
Note that differential method gives  rA  0.244(dm 3 /mol)2 / min C A2 C B

Example 55--2
Linear Regression Report

by SigmaPlot
Function values: Function values:
x f(x) x f(x)
Coefficients: 0
6
20.1142857143
20.8634285714
156
162
39.592
40.3411428571
12 21.6125714286 168 41.0902857143
a0 20.1142857143 18
24
30
22.3617142857
23.1108571429
23 86
23.86
174
180
186
41.8394285714
42.5885714286
43 33 1428
43.3377142857
36 24.6091428571 192 44.0868571429
k’ 0.1248571429 42
48
25.3582857143
26.1074285714
198
204
44.836
45.5851428571
54 26.8565714286 210 46.3342857143
r² 0 999903135
0.999903135 60
66
27.6057142857
28.3548571429
216
222
47.0834285714
47.8325714286
72 29.104 228 48.5817142857
78 29.8531428571 234 49.3308571429
84 30.6022857143 240 50.08
90 31.3514285714 246 50.8291428571
96 32 1005714286
32.1005714286 252 51 5782857143
51.5782857143
102 32.8497142857 258 52.3274285714
108 33.5988571429 264 53.0765714286
114 34.348 270 53.8257142857
120 35.0971428571 276 54.5748571429
k=0 25 dm3/mol/min
k=0.25 126
132
35.8462857143
36.5954285714
282
288
55.324
56.0731428571
138 37.3445714286 294 56.8222857143
144 38.0937142857 300 57.5714285714
150 38.8428571429
zero order first order

CA
C A  C A0  kt ln  kt
C A0
A0
55 1.2
50
1.0
CA x 103 (mol/dm3)
4
45
0.8
40
CA/CA0)
35 0.6
ln (C
30
0.4
25
0.2
20
15 0.0
0 50 100 150 200 250 300 0 50 100 150 200 250 300
t ((min)) t ((min))

Comparison
The differential method tends to accentuate the

uncertainties in the data,
data while the integral method
tends to smooth the data, thereby disguising the
uncertainties in it.
In most analyses,
y , it is imperative
p that the engineer
g
know the limits and uncertainties in the data.
This prior knowledge is necessary to provide for a

safety factor when scaling up a process from
l b t
laboratory experiments
i t to
t design
d i either
ith a pilot
il t plant
l t
or full-scale industrial plant
Accentuate : 강조하다 . 두드러지게하다.
Disguise: 위장하다. 변장하다.
5.2.3 Nonlinear Regression
★ In nonlinear regression analysis,

analysis we search for those parameter
values that minimize the sum of squares of the differences between the
measured values and the calculated values.
values Not only can nonlinear
regression find the best estimates of parameter values, it can also be used
to discriminate between different rate law models, such as Langmuir
Langmuir-
Hinshelwood models.
★ Many software programs are available to find these parameter values

so that all one has to do is enter the data.
data In order to carry out the search
efficiently, in some cases one has to enter initial estimates of the
parameter values close to the actual values. These estimates can be
obtained using the linear-least-squares technique.
5.2.3 Nonlinear Regression
★ We will now apply nonlinear least-squares analysis to reaction rate data to

d t
determine
i the
th rate
t law
l parameters
t (k , ..).
(k, ) We
W then
th search
h for
f those
th values
l
that will minimize the sum of the squared differences of the measured reaction
rates,
t rm, and
d the
th calculated
l l t d reaction
ti rates,
t rc. That
Th t is,
i we wantt the
th sum off (r
( m-
rc)2 for all data points to be minimum. If we carried out N experiments, we
would
ld wantt find
fi d the
th parameter
t values
l th t minimize
that i i i the
th quantity
tit
s 2 N
( r  r ) 2
2    im ic (5-34)
N  K i 1 N  K
N = number
n mber of runs
r ns
K = number of parameters to be determined
rim = measured reaction rate for run i (i.e.,
(i e -rrAim)
ric = calculated reaction rate for run i (i.e., -rAic)
Nonlinear Regression
To illustrate this technique, let’s consider the first-order reaction
A Product
for which we want to learn the reaction order, , and the specific reaction rate, k,
r  kC A
Th reaction
The ti rate
t will
ill be
b measured
d att a number
b off different
diff t concentrations.
t ti W
We
now choose values of k and  and calculate the rate of reaction (ric) at each
concentration at which an experimental point was taken.
taken We then subtract the
calculated value (ric) from the measured value (rim), square the result, and sum
the squares for all the runs for the values of k and  we have chosen.

This procedure is continued by further varying  and k until we find their best
values,, that is,, those values that minimize the sum of the squares.
q Many
y well-
known searching techniques are available to obtain the minimum value  min
2
.
g
Figure 5-7 shows a hypothetical
yp plot of sum of the squares
p q as a function of the
parameters  and k:
 2  2  57
  2.0  2  13.25
k  5.0 dm / ls
d 3 /mol  2  7. 3
 2  1.85
5

  f (k ,  )
2
 min
2
 0.045
2
k Seoul National University

In searching to find the parameter values that give minimum of the

sum of squares 2, one can use a number of optimization
techniques or software packages.
A number of software packages are available to carry out the

procedure to determine the best estimates of the parameter values
and corresponding confidence limits.
All on has to do is to type the experimental values in the computer,

specify the model, enter the initial guesses of the parameter values
alongg with 95% confidence limit appear.
pp

Concentration-time data. We will now use nonlinear regression to
determine the rate law parameters from concentration-time data obtained in
batch experiments. We recall that the combined rate law-stoichiometry-mole
balance for a constant-volume batch reactor is
dC A
 kC A (5-6)
dt
We now integrate Eq.
Eq (5-6)
(5 6) to give
C 1A0  C 1A  (1   )kt


C A  C 1A0  (1   )kt 
1 /(1 )
(5-18)
W wantt the
We l off  and
th value d k that
th t will
ill make
k s2 a minimum.
i i
 
N N
s   (C Ami C Aci )    C Ami  C 
1 1 /(1 )
 (1   )kti
2 2 2
A0 (5-19)
i 1 i 1
Nonlinear Regression (Nonlinear Least
Least--Squares Analysis)

Example 5-
5-3 Use of Regression to find the Rate Law Parameter
We shall use the reaction and data in Ex. 5-1 to illustrate how to use regression
to find  and k’
(C6H5)3CCl (A) + CH3OH (B)  (C6H5)3COCH3 (C) + HCl (D)
dC A 
 
 k CA
dt
Integrating with the initial condition when t=0 and CA=CA0 for  ≠1.0
1 C A(10 )  C A(1 )
t
k (1   )
Substituting for the initial concentration CA0=0.05 mol/dm3
(1 ) (1 )
1 ( 0 . 05)  C
t
A
k (1   )
Let’s do a few calculations by hand to illustrate regression.

Example 5-
We now first assume a value of  and k and then calculate t for the
concentrations of A given in Table E5
E5-1.1
1.1 (pp 261). We will then calculate the sum
of the squares of the difference between the measured time, tm and the calculated
times (i.e., s2). For N measurements,
1 1 2
N

N
C C 
s   (t mi  tci )    t mi 
2 2 A0 A

t 1 t 1  k (1   ) 
Our first guess is going to be  = 3 and k’ = 5, with CA0 = 0.05.

Equation (E5-3.2) becomes (1 )
 C A(1 )
1 (0.05)
t
k (1   )
1  1 1  1  1 
tc1    2    2  400 
2k   C A C A0  10  C A
2

We now make the calculations for each measurement of concentration and fill in
column 3 and 4 of Table E5-3.1. For example, when CA=0.038 mol/dm3 then
Example 5-
1  1 1  1  1 
tc1    2    400   29.2 min
2k   C A C A0  10  (0.038)
2 2

Which is shown in Table E5-3.1 on line 2 for guess 1. We next calculate the squares
of difference (tm1-tc1)2=(50-29.2)2=433. We continue in this manner for points 2, 3, and
4 to calculate the sum s2=2916.
After calculating s2 for  = 3 and k = 5, we make a second guess for  and k’. For
our second guess we choose  = 2 and k = 5; Equation (E5-3.2) becomes
1 1 1  1 1 
tc        20 ((E5-3.2))
k   C A C A0  5  C A 
We now proceed
W d with
i h our secondd guess to find
fi d the ( m1-tc1)2 to be
h sum off (t b s2=49,895,
49 895
which is far worse than our first guess. So we continue to make more guesses of 
and k and find s2. Let’s stop and take a look at tc for guesses 3 and 4.
Example 5-
Table E5-3.1. Regression of data

Oi i lD
Original Data
t G
Guess 1 G
Guess 2 G
Guess 3 G
Guess 4
=3 =2 =2 =2
k’=5 k’=5 k’=0.2 k’=0.1
t CA x 103
(min) (mol/L) tc (tm-tc)2 tc (tm-tc)2 tc (tm-tc)2 tc (tm-tc)2
0 50 0 0 0 0 0 0 0 0
1
50 38 29.2 433 1.26 2,375 31.6 339 63.2 174
2
100 30.6 66.7 1,109 2.5 9,499 63.4 1,340 126.8 718
3
200 22.2 163 1,375 5.0 38,622 125.2 5,591 250 2,540
4
s2 = 2,916 s2 = 49,895 s2 = 7,270 s2 = 3,432
We see that (k’=0.2 dm3/mol·min) underpredicts the time (e.g., 31.6

min versus 50 minutes), while (k’=0.1 dm3/molmin) overpredicts
the time (e.g., 63 min versus 50 minute)
Example 5-
  2.04
k   0.147 dm 3 /mol  min
Regression again
  2.0
k   0.125 dm 3 /mol  min
 
 rA  0.25((dm3 /mol))2 / min C A2 C B

5.3 Method of Initial Rates
The use of the differential method of data analysis to determine reaction

orders and specific reaction rates is clearly one of the easiest,
easiest since it
requires only one experiment. However, other effects, such as the
presence of a significant reverse reaction,
reaction could render the differential
method ineffective.
In these cases, the method of initial rates could be used to determine the
reaction order and the specific rate constant. Here, a series of
experiments is carried out at different initial concentrations, CA0, and the
initial rate of reaction, -rA0, is determined for each run.
The initial rate of reaction –rA0 can be found by differentiating the data
and extrapolating to zero time.
5.3 Method of Initial Rates
By various plotting or numerical analysis techniques relating –rA0 to

CA0, we can obtain
bt i the
th appropriate
i t rate
t law.
l
 rA0  kC A0
the slope of the plot of ln(-rA0) versus lnCA0 is the reaction order 

Example 55--4

Example 55--4

Example 55--4
Solution

Example 55--4

Example 55--4

Example 55--4

Method of Half-
Half-Lives
A  products (irreversible)
 The half-life
half life of a reaction
= the time it takes for the
dC A
concentration of the reactant    rA  kC A
t fall
to f ll to
t half
h lf off its
it initial
i iti l value
l dt
 By determining the half-life of a 1  1 1  1  C  1 

t   1   1    1  A0   1
reaction as a function of the initial k (  1)  C A C A0  kC A0 (  1)  C A  

concentration, the reaction order and
p
specific reaction rate can be determined. 1
t  t1/ 2 when C A  C A0
2
ln t1/ 2 Slope=1-
2 1  1  1 
t1/ 2    1 
k (  1)  C A0 
2 1  1
ln t1/ 2  ln  (1   ) ln C A0
ln C A0 k (  1)
Differential Reactors
• Most commonly used catalytic reactor to obtain experimental data

- use to determine the rate of reaction as a function of either
concentration or partial pressure
- the conversion of the reactants in the bed is very small
- reactant concentration is constant: gradientless (~CSTR)
p y uniform ((~CSTR))
- reaction rate is spatially
A  P
FA0 FAe
L Inert filling
catalyst

Steady state mole balance on reactant A (CSTR)
L
 flow  flow
rate    rate off  rate off 
 t   rate t    
in  out     
generation accumuatio n
FA0 FAe
CA0  rate of reaction  

FP [ FA0 ]  [ FAe ]   (mass of cat.)  0
 mass of cat. 
CP

W FA0  FAe  (rA' )(W )  0
FA0  FAe
r  '
A  P A
W
 0 C A0  C Ae
r '
(5-27)
(5 27)
A
W
Reactor Design Equation
FA0 X FP (
(5-28)
)
r  '

W
A
W
For constant volumetric flow known
Measuring
Can be determined the product
 0C A0 C Ae  0 (C A0  C Ae )  0C P concentration
r '
 
A
W W W
known
- using very little catalyst and large volumetric flow rates (C A0  C Ae ) ~ 0

 rA'  rA' (C Ab )
where CAb the concentration of A within the catalyst bed
C A0  C Ae
- the
th arithmetic
ith ti mean off the
th inlet
i l t andd outlet
tl t concentration:
t ti C Ab 
2
- very little reaction takes place within bed: C Ab ~ C A0
 rA'  rA' (C A0 )
Example 55--5 Differential Reactors
The synthesis of CH4 from CO and H2 using a nickel catalyst was carried out at 500oF
in a differential reactor where the effluent concentration of CH4 was measured.
3H 2  CO  CH 4  2 H 2O
a. Relate the rate of reaction to the exit methane concentration.
b. The reaction rate law is assumed to be the p

product of a function of the ppartial
pressure of CO and a function of the partial pressure of H2:
 4  f (CO )  g ( H 2 )
rCH (E5 5 1)
(E5-5.1)
Determine the reaction order w.r.t. CO, using the data in Table E5-5.1.
 4 on PCO is of the form
Assume that the functional dependence of rCH

 4 ~ PCO
rCH (E5-5 2)
(E5-5.2)

Table E5-5.1. Raw Data
Run PCO (atm) PH2 (atm) CCH4 (mol/dm3)

1 1.0 1.0 1.73x10-4
2 1.8 1.0 4.40x10-4
3 4.08 1.0 10.0x10-4
4 10
1.0 01
0.1 1 65 10-44
1.65x10
5 1.0 0.5 2.47x10-4
6 10
1.0 40
4.0 1 75x10-4
1.75x10
The exit volumetric flow rate a differential packed bed containing 10 g of

catalyst was maintained at 300 dm3/min for each run. The partial pressure of H2
and CO were determined at the entrance to the reactor, and the methane
concentration was measured at the reactor exit.
(a) In this example the product composition, rather than the reactant concentration, is
being monitored. rĆH4 can be written in terms of flow rate of methane from the
reaction,
i
FCH 4 0CCH (300 dm 3 / min )(1.73  10 4 mol / dm 3 ) mol CH 4

  rCH
 rCO  4   4
  5.2  10 5
W W 10 g  cat g  cat  min
Table E5-5.1. Raw and Calculated Data

Run
R PCO (atm) PH2 (atm) CCH4 (mol/dm3) r´ĆH4 (mol CH4/cat·min)
1 1.0 1.0 1.73x10-4 5.2x10-3
2 18
1.8 10
1.0 4 40x10-4
4.40x10 13 2x10-3
13.2x10
3 4.08 1.0 10.0x10-4 30.0x10-3
4 1.0 0.1 1.65x10-4 4.95x10-3
5 1.0 0.5 2.47x10-4 7.42x10-3
6 1.0 4.0 1.75x10-4 5.25x10 -3
Determining the rate law dependence in CO
For constant hydrogen concentration, the rate law can be written as
 
 4  kPCO  g ( PH 2 )  k PCO
rCH (E5-5.4)
Taking the log
l (rCH
ln l (k )   ln
 4 )  ln l (PCO )

We now plot ln  ) versus

l (rCH4
l PCO for runs 1,
ln 1 2,
2 and 3.
3
Table E5-5.1. Raw and Calculated Data

Run PCO (atm) PH2 (atm) CCH4 (mol/dm3) rĆH4 (mol CH4/cat·min)
1 1.0 1.0 1.73x10-4 5.2x10-3
2 18
1.8 10
1.0 4 40x10-4
4.40x10 13 2x10-3
13.2x10
3 4.08 1.0 10.0x10-4 30.0x10-3
4 1.0 0.1 1.65x10-4 4.95x10-3
5 1.0 0.5 2.47x10-4 7.42x10-3
6 1.0 4.0 1.75x10-4 5.25x10-3

cat min) 0.1

rCH 4  0.0056 PCO
1.22
n Rate ((mol/gc
0.01
Rxn
0.001
0.1 1 10
Pressure of CO (atm)

Determining the rate law dependence in CO
Had we include more points, we would have found the reaction rate is
essentially first order with =1, that is,
=1.22 → 1
 4  0.0056 PCO
rCH 1.22
 4  k PCO
rCH (E5-5 5)
(E5-5.5)
 4  k PCO f(PH2 )
rCH

Determining the rate law dependence in H2
From Table E5-5.2, it appears that the dependence of r’’CH4 on PH2 cannot be
represented by a power law. The reaction rate first increases with increasing
partial pressure of hydrogen, and subsequently decreases with increasing partial
pressure of hydrogen. That is, there appears to be a concentration of hydrogen at
which the rate is maximum.
Run PCO (atm) PH2 (atm) CCH4 (mol/dm3) rĆH4 (mol CH4/cat·min)
1 1.0 1.0 1.73x10-4 5.2x10-3
2 1.8 1.0 4.40x10-4 13.2x10-3
3 4.08 1.0 10.0x10-4 30.0x10-3
4 1.0 0.1 1.65x10-44 4.95x10-33
5 1.0 0.5 2.47x10-4 7.42x10-3
6 10
1.0 40
4.0 1 75x10-4
1.75x10 5 25x10-3
5.25x10

@ high PH2
@ low PH2
 β1
rCH4 ~ PH2
β1
P
 4 ~
rCH
H2
 2  1
1  bP β2
H2

aPCO PHβ21
 4 
rCH
1  bPHβ22
Polymath
Regression
((E5-5.11))
0.025PCO PH0.61
 4 
rCH 2
1  2.49PH2
Hydrogen undergo
((E5-5.12)) dissociate adsorption
p
on the catalyst surface
→ Dependence of H2 to the ½ power
Polymath
Regression
0.018PCO PH0.5
again  4 
rCH 2
1  1.49PH2
1  12 ,  2  1
0.018PCO PH0.5 PCO PH0.5 1 b

 4 
rCH 2
rearranging 2
  PH2 (E5 5 13)
(E5-5.13)
1  1.49PH2  4
rCH a a
This plot should be a straight line with an intercept of 1/a and a slope b/a.
400
0.018PCO PH0.5
 4 
rCH 2
300 1  1.49PH2
0.5
PCO P H2
 4
rCH 200
Rate law is
100 indeed
consistent with
the rate law
0 data
0 1 2 3 4
PH2 (atm) Seoul National University
5.7 Evaluation of Laboratory Reactor
The successful design

g of industrial reactors lies pprimarily
y with the
reliability of the experimentally determined parameters used in the
p Consequently,
scale-up. q y, it is imperative
p to design
g equipment
q p and
experiments that will generate accurate and meaningful data.
Unfortunately, there is usually no single comprehensive laboratory

reactor that could be used for all types
yp of reactions and catalysts.
y In
this section, we discuss the various types of reactors that can be chosen
to obtain the kinetics parameters for a specific reaction system.
We closely follow the excellent strategy presented in the article by V.W.

Weekman of ExxonMobil.
5.7 Evaluation of Laboratory Reactor
Criteria used to evaluate laboratoryy reactor
1. Ease of sampling and product analysis

2. Degree
g of isothermalityy
3. Effectiveness of contact between catalyst and reactant
4 Handling of catalyst decay
4.
5. Reactor cost and ease of construction

Evaluation of Laboratory Reactor (Types of Reactors)
Differential
Diff ti l reactor
t Integral reactor
(Fixed bed) (Fixed bed)
Stirred-Batch Reactor
catalyst slurry
Continuous-Stirred Stirred Contained

Tank Reactor Solids Reactor Straight-through Recirculating
(CSTR) (SCSR) transport reactor transport reactor
Figure 5-12 Type of Reactors Seoul National University

Evaluation of Laboratory Reactor (Reactor Ratings)
Reactor type Sampling Isothermality F-S contact Decaying Catalyst Ease of construction
Differential P-F F-G F P G

Fixed bed G PF
P-F F P G
Stirred batch F G G P G
Stirred-contained
Stirred contained solids G G F-G
F G P F-G
F G
Continuous-stirred tank F G F-G F-G P-F
Straight-through transport F-G P-F F-G G F-G
Recirculating transport F-G G G F-G P-F
Pulse G F-G P F-G G
G=Good; F=Fair; P=Poor
CSTR and recirculating transport reactor appear to be the best choice, because they are
satisfactory in every category except for construction. However, if the catalyst under
study does not decay, the stirred batch and contained solid reactors appear to be best
choices. If the system is not limited by internal diffusion in the catalyst pellet, larger
pellets could be used, and the stirred-contained solids is the best choice. If the catalyst
is nondecaying and heat effects are negligible,
negligible the fixed-bed (integral) reactor would be
the top choice, owing to its ease of construction and operation. However, in practice,
usually more than one reactor type is used in determining the reaction rate law
parameters. Seoul National University

Metodos Diferencial y Otros

Uploaded by

Document Information

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

Metodos Diferencial y Otros

Uploaded by

Copyright:

Available Formats

Chemical Reactor Design

Chemical Reactor Design

Seoul National University

★ Determine the reaction order and specific reaction rate from

★ Describe how to use equal-area differentiation, polynomial

★ Describe two or more types of laboratory reactors used to

★ Two common type reactors for obtaining rate data:

((1)) Batch reactor : Conc. vs. time

(2) Differential reactor : Conc. @ steady state

★ Six different methods

(1) the differential method

(3) the method of half-lives

Steps in Analyzing Rate Data

Seoul National University

Seoul National University

Seoul National University

Seoul National University

Seoul National University

Seoul National University

Assuming that the

then differential method may be used.

Excess B experiments: Excess A experiments:

 rA  k AC A C B  k AC B0C A  k C A  rA  k AC A C B  k AC A0C B  k C B

Consider a reaction carried out isothermally in a constant-volume batch reactor

After taking the logarithm of both sides of Eq. (5-6)

Seoul National University

Seoul National University

Seoul National University

2. For each interval, calculate

Seoul National University

★ Graphical Method (Equal-Area Graphical Differentiation)

★ Numerical Method (Independent variables are equally spaced)

Polynomial fit with software program to get best value of ai

n-th order polynomial

C A  a0  a1t  a2t 2  ...  ant n (5-11)

Seoul National University

Seoul National University

(C6H5)3CCl (A) + CH3OH (B)  (C6H5)3COCH3 (C) + HCl (D)

Time (min) 0 50 100 150 200 250 300

Concentration of triphenyl methyl chloride

Seoul National University

Step 1 Postulate a rate law

Step 2 variable, which is this case is CA.

Step 3 Look for simplifications. Because concentration of methanol is 10 times

Substituting for CB in Equation (E5-1.1)

Seoul National University

Step 4 Apply the CRE algorithm

Taking the natural log of both sides of Equation (E5-1.5)

Seoul National University

Seoul National University

Seoul National University

C A  0.04999  2.978  10 4 t  1.343  10 6 t 2  3.485  10 9 t 3  3.697  10 12 t 4

C A  0.04999  2.978  10 4 t  1.343  10 6 t 2  3.485  10 9 t 3  3.697  10 12 t 4

Seoul National University

SUMMARY: There is quite a close agreement between

min mol/dm3 Graphical Finite Polynominal

Seoul National University

From Figure E5-1.3, we found the slope to be 1.99 so that the

Seoul National University

Excel plot to determine  and k

The rate law is

★ It is important to know how to generate linear plots of functions of CA

If the plots of concentration data versus time had

Seoul National University