Chapter 5

Collection & Analysis of Rate Data

There are some basic steps in Analysis of Rate Data

1. Postulate a rate law:
a) Power law model for homogeneous rxns:
α β
− rA = k ⋅ C A ⋅ C B
b) Langmuir – Hinshelwood for heterogeneous rxns:

k ⋅ PA ⋅ PB
− rA ' =
(1 + K A ⋅ PA + PB ) 2
2. Select reactor type; i.e.; Batch, CSTR, PFR, PBR......
3. Process your data in term of measured variables (NA, CA or PA)
4. Do simplifications (assumptions), if –rA = k CACB and CA>>CB
→ -rA ≈ k CB

5. For batch reactors, find reaction order

i. Find (- dCA / dt) from CA vs t data.

Take ln.

Find the order of rxn by ln (- dCA / dt) = ln k + α ln CA

Find the rate constant, k.

ii. Integral Form:
dC α
− A = k ⋅ CA
dt
Integrate

[
1 1−α
t = C A0 ........
k
]
6. For PBR, find –rA’ as a function of CA or PA.

1
SUMMARY
Differential Method:
dC A α ⎛ dC ⎞
− = k A ⋅ CA ln⎜ − A ⎟ = ln k A + α ln C A
dt ⎝ dt ⎠
slope = α
(dC A / dt )P
kA =
⎛ dC ⎞
ln⎜ − A ⎟ ⎛ dC ⎞ (C AP )α
ln⎜ − A ⎟
⎝ dt ⎠ ⎝ dt ⎠ P

ln C A ln C AP
We should differentiate the concentration-time data either
graphically or numerically.
1. Graphical differentiation
2. Numerical differentiation formula
3. Differentiation of a polynomial fit to the data.

1. Graphical
Tabulate (CAi, ti) observations and for each interval calculate
∆Cn = Cn – Cn-1 ∆tn = tn – tn-1

ti Ci ∆t ∆CA ∆CA/ ∆t (dCA / dt)

t0 C0
t1 – t0 C1 – C0 (∆CA/ ∆t)1
t1 C1 (dCA / dt)1
t2 – t1 C2 – C1 (∆CA/ ∆t)2
t2 C2 (dCA / dt)2
t3 – t2 C3 – C2 (∆CA/ ∆t)3
t3 C3

2
Plot these values as a histogram
Draw a smooth curve
Read estimates of (dCA/dt) at t1, t2, ......

∆x = xn – xn-1
∆y = yn – yn-1
Plot these
n
∆y
y n − y1 = ∑ ⋅∆ xi
i=2 ∆ xi
xn
dy
A
y n − y1 = ∫ dx ⋅ dx
x1
B

C Try to find
D
Area(A) ≈ Area(B)
A+C≈B+D

Numerical Methods
If the data points are equally spaced, i.e., t1 – t0 = t2 – t1 = ∆t
time (min) t0 t1 t2 ...........
Conc’n (mol/dm3) CA0 CA1 CA2 ...........

⎛ dC A ⎞ − 3C A0 + 4C A1 − C A2
For initial point ⎜ ⎟ =
⎝ dt ⎠ t0 2 ∆t

interior point
⎛ dC A ⎞
⎜ ⎟ =
1
∆t
[
C A( i +1) − C A( i −1) ]
⎝ dt ⎠ ti 2

End point
⎛ dC A ⎞
⎜ ⎟ =
1
[
C A( n − 2 ) − 4C A( n −1) + C A( n ) ]
⎝ dt ⎠ tn 2∆t
Polynomial fit:
Fit the conc’n – time data to an nth order polynomial as:
C A = a o + a1t + a 2 t 2 + a 3 t 3 + a 4 t 4 ......a n t n
Find best values for {a o , a 1 , a 2 ,........, a n }
dC A
= a 1 + 2 a 2 t + 3a 3 t 2 + ......n a n t n -1
dt

3
Integral Method

We first guess the rxn order and integrate the differential equation used to
model the batch system. If the order is correct, the plot of conc’n time data
should be linear.

This method is used when rxn order is known but EA and kA are unknown.

Non – Linear Regression

Non – Linear Least-
Least-Square Analysis
We want to find the parameter values (alpha, k, E) for which the sum of
the squares of the differences, the measured rate (rm), and the calculated
rate (rc) is a minimum.

# of parameters to be determined

# of runs
That is we want to be a minimum.
For concentration-time data, we can integrate the mole balance equation
for -rA = k CAalpha to obtain

4
We find the values of alpha and k which minimize S2

dC A α
= k ⋅ CA
dt
−α
k ⋅ dt = dC A ⋅ C A
1 1−α
k ⋅t = C A + Const
1−α
1−α
C A0
Const = −
1−α
1−α
C A = 1−α C A0 − (1 − α )k ⋅ t

Vary α and k, obtain S2 (or use a search technique)

S(α’, k’) is a minimum (optimization methods)

Concentration vs Time Equations for the proposed rate equations can be

done by differential or integration method.

Ex.: Liquid phase rxn btw trimethylamine(A) and n-propyl bromide (B)
was studied by Winkler & Hinshelwood. The results at 139.4oC are
shown below. Initial sol’ns of A & B in benzene, 0.2 molal, were mixed
and placed in constant temperature bath. After certain times, they were
cooled to stop the rxn. Determine the first order and second order
specificrates, k1 and k2, asssuming the rxn is irreversible. Use integration
& differential methods.

Run t, min x (%)

1 13 11.2
2 34 25.7
3 59 36.7
4 120 55.2

5
Sol’n:
dC A
A + B → C+ + D- 1st order - rA = = k1 ⋅ C A
Volume is constant dt
dC A
2 nd order - rA = = k2 ⋅ C A ⋅ CB
dt
Integration Method
1st order
CA
- ln = k1 ⋅ t [1]
C A0
2 nd order
Stoichiome tric coefficien ts are equal and C A0 =C B0 = 0.1 molal
dC A
= k2 ⋅ C A
2
-
dt
1 1
− = k2 ⋅ t [2]
C A C A0

C A0 − C A
x= C A = C A 0 (1 − x )
C A0
For the first run C A =0.1 ⋅ 0.888
Substitute in [1]
1 C 1 0 .1
k 1 = ⋅ln A0 = ln = 1 .54 × 10 − 4 sn −1
t C A 13 ⋅ 60 0 .0888
Substitute in [2]
1 ⎛ 1 ⎞ 0 .112 L
k2 = ⋅⎜ − 1⎟ = = 1 .63 × 10 − 3
t ⋅ C A0 ⎝ 1 - x ⎠ (13 ⋅ 60 ) ⋅ 0 .1 ⋅ (1 − 0 .112 ) gmol ⋅ s

If you repeat for four of the runs:

Run t, sec k1 x 10-4 (s-1) k2 x 10-3 (L/mol s) CD
1 780 1.54 1.63 0.0112
2 2040 1.46 1.70 0.0257
3 3540 1.30 1.64 0.0367
4 7200 1.12 1.71 0.0552

6
 1

-ln(CA /CA0)
0.5

0
0 800 1600 2400 3200 4000 4800 5600 6400 7200
time, sec

15
1/CA -1/CA0

10

0
0 800 1600 2400 3200 4000 4800 5600 6400 7200
time, sec

So, it is SECOND ORDER

Differential Method
CD = CA0 – CA (moles of D produced = moles of A reacted)
CD = x CA0
A plot of CD vs t
0.06

0.05

0.04
[D] moles/L

0.03

0.02

0.01

0
0 1000 2000 3000 4000 5000 6000 7000 8000
time, sec

dC A dC D
r=− = (the slope of the curve)
dt dt

7
Slopes determined from the curve are given as follows

CD CA r = dCD/dt (105 gmol/L)

0.0 0.10 1.58
0.01 0.09 1.38
0.02 0.08 1.14
0.03 0.07 0.79
0.04 0.06 0.64
0.05 0.05 0.45

log rA = log k1 + log CA (from –rA = k1 CA)

log rA = log k2 + log CA2 = log k2 + 2 log CA (from –rA = k2 CA2)
1st: log r vs log CA should yield a straight line with a slope of 1.0.
2nd: log r vs log CA should yield a straight line with a slope of 2.0.

log r = -2.76 + 2.0 log CA

The data suggest a slope of 2!
log k2 = -2.76
k2 = 1.73 x 10-3 L / mol s

slope = 2.0
log r

log CA

8
 Another Way (Batch Reactor Data)

B + A → products

-rA = kA CAαCBβ if α and β are both unknown, the rxn could be run in an
excess of B so that CB remains essentially unchanged

α β β
- rA = k'⋅C A where k' =k A C B ≈ k A C B0
In determining α , the rxn is carried (excess of A)
β α α
- rA = k' '⋅C B where k' ' =k A C A ≈ k A C A0

Solving with search

( [ ] )
4 2
1 / 1−α
S 2 = ∑ C Am,i − C A0
1−α
− (1 − α )k ⋅ ti
i =1

t CD CA
780 0.0112 0.0888
2040 0.0257 0.0743
3540 0.0367 0.0633
7200 0.0552 0.0448

( [ ] )
n 2
1 / 1−α
S = ∑ C Am,i − abs C
1−α
2
A0 − (1 − α )k ⋅ ti fminsearch
i =1

9
Method of Initial Rates
If there is a reverse rxn, it could render the differential method ineffective.
In these cases, initial rates could be used for k & α.
Carry out a series of experiments with different CA0.
Determine –rA0 for each run.
-rA0 = k CA0 α → ln(-rA0) vs ln(CA0)
Instead of doing at different time steps, repeat the rxn.

Method of Half Lives:

The half life of a rxn, t½, is defined as the ime it takes for the concentration
of the reactant to fall to half of its initial value.
If you know t½, find k & α.
If there are more than 1 reactant, use method of excess.

α
− rA = k ⋅ C A ( A → products )
dC A α
− = − rA = k ⋅ C A
dt
1
@t = 0 C A = C A0 @ t = t1 / 2 CA = C A0
2
1 ⎛ 1 1 ⎞
t= ⎜ α −1 − ⎟
k (α − 1) ⎜⎝ C A C A 0 ⎟⎠
α −1

α −1
⎛⎛ C ⎞ ⎞
t=
1 ⎜ ⎜ A0 ⎟⎟ − 1⎟
k (α − 1) ⎜ ⎜⎝ C A ⎠ ⎟
⎝ ⎠
α −1
2 α −1 − 1 ⎛ 1 ⎞
t1 / 2 = ⎜ ⎟⎟
k (α − 1) ⎜⎝ C A 0 ⎠
α −1
n α −1 − 1 ⎛ 1 ⎞
Similarly t1 / n = ⎜ ⎟⎟
k (α − 1) ⎜⎝ C A 0 ⎠

10
 ln t1/ 2
α −1
2 −1 slope = 1 − α
ln t1/ 2 = ln + (1 − α ) ln C A0
(α − 1)k

ln C A 0
Differential Reactors
A differential reactor is used to determine the rate of a rxn as a function of
either concentration or partial pressure
∆L
FA0 FAe
FA0 FAe

∆L
∆W
Conversion is very small in bed
CA is constant ≈ CA0 (gradientless)
Rxn rate is uniform
(almost isothermal)

Design Equation is similar to a CSTR:

Rate of rxn
FA0 − FAe + (mass of cat ) = 0
mass of cat
FA0 − FAe + rA '⋅∆w = 0
FA0 − FAe
rA ' =
∆w
v C − v ⋅ C Ae
rA ' = 0 A0 (in terms of conc' n)
∆w
F ⋅x F
rA ' = A0 = P (in terms of flow rate)
∆w ∆w
v (C − C Ae ) v0 (C P )
rA ' = 0 A0 =
∆w ∆w

11
Deviations from Ideal Reactors

When the mixing criteria of ideal reactors are not satisfied,

mathematical expressions for the conservation equations become more
difficult.

Deviations from ideal stirred – tank reactors

a) Stagnant regions
b) by – passing

Deviations from tubular-flow

a) Longtidunal mixing due to vortices and turbulence
b) Laminar flow (poor radial mixing)
c) By-passing in fixed-bed catalytic reactor.

12