Sec. R5.

125

Least-Squares Analysis

R5.1 Least-Squares Analysis

R5.1.1 Linearization of the Rate Law

If a rate law depends on the concentration of more than one species and it is
not possible to use the method of excess, we may choose to use a linearized
least-squares method. This method of data analysis is also useful to determine
the best values of the rate law parameters from a series of measurements when
three or more parameters are involved (e.g., reaction order,
; frequency factor,
A; and activation energy, E).
A mole balance on a constant-volume batch reactor gives
dC


 ---------A-  r A  kC AC B
dt

(R5-1)

If we now use the method of initial rates, then

dC A



 ----------  r A0  kC A0C B0
dt 0
Taking the log of both sides, we have
dC
ln  ---------A-  ln k 
ln C A0   ln C B0
dt 0

Used when C A0
and C B0 are varied
simultaneously

(R5-2)

Let Y  ln ( dC A  dt )0 , X 1  ln C A0 , X 2  ln C B0 , a0  ln k , a1 
, and

a2  . Then
Y  a0  a1 X 1  a2 X 2

(R5-3)

If we now carry out N experimental runs, for the j th run, Equation (R5-3)
takes the form
Y j  a0  a1 X 1 j  a2 X 2 j

(R5-4)

where X 1 j  ln C A0 j , with C A0 j being the initial concentration of A for the j th

run. The best values of the parameters a0 , a1 , and a2 are found by solving
Equations (R5-5) through (R5-7) simultaneously.
For N runs, 1, 2, ..., N,
N

Y j  Na0  a1

j1

Three equations,
three unknowns
( a0 , a1, a2 )

j1

X 1 j Y j  a0

X 1 j  a2

j1
N

j1

X 1 j  a1

X2 j

(R5-5)

j1
N

j1

X 1 j  a2

X1 j X2 j

j1

(R5-6)

126

Chap.

X 2 j Y j  a0

X 2 j  a1

j1

j1

X 1 j X 2 j  a2

X2 j

(R5-7)

j1

j1

We have three linear equations and three unknowns which we can solve for:
a0 , a1 , and a2 . A detailed example delineating the kinetics of the reaction
H2( g )  C8H16( g ) C8H18( g )
using linear least-squares analysis can be found in Example 10-2. If we set
a2  0 and consider only two variables, Y and X, Equations (R5-5) and (R5-6)
reduce to the familiar least-squares equations for two unknowns.
Example R55 Using Least-Squares Analysis to Determine Rate Law Parameters
The etching of semiconductors in the manufacture of computer chips is another
important solidliquid dissolution reaction (see Problem P5-12 and Section 12.10).
The dissolution of the semiconductor MnO2 was studied using a number of different
acids and salts. The rate of dissolution was found to be a function of the reacting
liquid solution redox potential relative to the energy-level conduction band of the
semiconductor. It was found that the reaction rate could be increased by a factor of
105 simply by changing the anion of the acid1!! From the following data, determine
the reaction order and specific reaction rate for the dissolution of MnO2 in HBr.

A 105 fold increase

in reaction rate!!!

C A0 ( mol HBr/dm3 )

0.1

0.5

1.0

2.0

4.0

rA0 ( mol HBr/m2  h ) 102

0.073

0.70

1.84

4.86

12.84

Solution
We assume a rate law of the form

rHB r  kC HBr

(RE5-5.1)

Letting A  HBr, taking the ln of both sides of (RE5-5.1), and using the initial rate
and concentration gives
ln ( rA0 )  ln k 
ln C A0

(RE5-5.2)

Let Y  ln ( rA0 ) , a  lnk, b 
, and X  ln C A0 . Then

Y  a  bX

(RE5-5.3)

The least-squares equations to be solved for the best values of a and b are for N runs
N

Y i  Na  b

i1

Xi

(RE5-5.4)

i1

i1

i1

i1

X iY i  a
X i  b
X i

S. E. Le Blanc and H. S. Fogler, AIChE J., 32, 1702 (1986).

(RE5-5.5)

Sec. R5.1

127

Least-Squares Analysis

where i  run number. Substituting the appropriate values from Table RE5-5.1 into
Equations (RE5-5.4) and (RE5-5.5) gives
21.26  5a  0.92b

(RE5-5.6)

15.10   0.92a  8.15b

(RE5-5.7)

TABLE RE5-5.1
2

Run

C A0

Xi

rA0

Yi

X iY i

Xi

1
2
3
4
5

0.1
0.5
1.0
2.0
4.0

2.302
0.693
0.000
0.693
1.380

0.00073
0.007
0.0184
0.0486
0.128

7.22
4.96
4.00
3.02
2.06

16.61
3.42
0.00
2.09
2.84

5.29
0.48
0.00
0.48
1.90

X i  0.92

Y i  21.26

i1

i1

X i Y i  15.1

i1

2

X i  8.15

i1

Solving for a and b yields

b  1.4

therefore

 1.4

and
a  3.99

k  1.84 102 ( dm3/mol )0.4 /m2  h

1.4

rHBr  0.0184C HBr

(RE5-5.8)