Collection and Analysis of Rate Data

Objectives
 Methodology for determining the rate law
Two types of reactor:
 batch reactor: for homogeneous reaction
 differential reactor : solid-fluid heterogeneous reaction
Six methods of data analysis: differential method, integral method, method
of half-lives, method of initial rates, linear and nonlinear regression
Outline
 The Algorithm for Data Analysis
 Batch Reactor Data
 Method of Initial Rates
 Method of Half-Lives
 Differential Reactors
1
 5.1 The Algorithm for Data Analysis
Table 5-1

2
3
 5.2 Batch Reactor Data
General Introduction
 For determination of rate law parameters for homogeneous rxn.
 Achieved by measuring Conc. vs. Time
 Data Analysis for determining reaction order and rate constant
Differential Method; Integral Method; Nonlinear Regression Method

For irreversible rxn. with one reactant

 A → Products -rA = kA CAα
 Batch Reactor Data: Conc. vs. Time + Data Analysis → Determining kA and α.
Method of Excess for rate laws as function of more than one
reactants
 A + B → Products -rA = kA CAα CBβ
 In an excess of B, CB ~ CB0 = constant, -rA = kA’ CAα → Determining α.
 In an excess of A, CA ~ CA0 = constant, -rA = kA” CBβ → Determining β.
 At known CA and CB, -rA = kA CAα CBβ → Determining kA.
4
5.2.1 Differential Method of Analysis
 Consider a constant-volume batch reactor running a reaction
with rate law of -rA = kA Caα,
dN dC dC 
design equation: dt A  V dtA  rAV →  dtA  rA  k AC A
dC A
→ ln( )  ln k A   ln C A
dt

 From the exp. data of CA vs. t, find (-dCA/dt) vs. t by appropriate methods
→ sec. 5.2.1A-C
 Plot ln(-dCA/dt) vs. ln(CA) → the slope of plot is reaction order α.

5
 Methods for obtaining (-dCA/dt) from exp. data of CA vs. t :
Graphical differentiation
Numerical differentiation formulas
Differentiation of a polynomial fit to the data
 5.2.1A Graphical Method
Refer to Appendix A.2. (skip here)
 5.2.1B Numerical Method (Finite difference)
Specific formula applicable if data points are equally-spaced (i.e. with constant time interval)
eg.

with t1 – t0 = t2 – t1 = Δt

The three-point differentiation formulas:

dC  3C A0  4C A1  C A 2
Initial point : ( A ) t0 
dt 2t
dC 1
Interior point : ( A ) ti  [(C A( i 1)  C A(i 1) )]
dt 2t
dC 1
Last point : ( A ) t5  [C A3  4C A 4  3C A5 ]
dt 2t

6
 5.2.1C Polynomial Fit
First Fit the data of conc. vs. time to an nth-order polynomial:
CA = a0 + a1t + a2t2 + … + antn
- This could be accomplished by commercial software.
- By inputting CA at each time point and choose order n to get best-fit of ai.
Then Determine (dCA/dt) by differentiate the polynomial w.r.t. time
(dCA/dt) = a1 + 2a2t + 3a3t2 + … + nantn-1
Caution must be taken in choosing the order of polynomial.
- order too low: not able to capture the trend of data
- order too high: peaks and valley to force the curve pass most data points
but results in significant errors in differentiation
Figure 5-2

7
 5.2.1D Finding the Rate Law Parameters
After data analysis of CA vs. t, the following table can be set up

The plot of ln(-dCA/dt) vs. ln(CA) can be used to obtain reaction order α and
rate constant kA
dC A
ln( )  ln k A   ln C A
dt
Review the steps to determine the reaction rate law from a set of data points

Example 5-1 8
5.2.2 Integral Method
 A trial-and-error procedure to find reaction order
 Guess the reaction order integrate the differential equation of batch reactor
If guess is correct, the appropriate plot of the concentration-time data should be linear.
 If the reaction order is known, the integral method can be used to evaluate
reaction rate constants at different temperature to determine the activation
energy.
 Appropriate linear plots for reactions of zero, first, and second order.
 Zero-order, -rA = kA
for constant-volume batch reactor,
dC A
 rA  k AC A   k A
dt
with initial condition of CA = CA0 at t = 0,
C A  C A0  kt
 First-order, -rA = kA CA
dC A C A0
 rA  k AC A   k AC A ln  kt
dt → CA
 Second-order, -rA = kA CA2
dC A 1 1
 rA  k AC A   k AC A2   kt
dt → C A C A0 9
 Figure 5-3 ~ 5-6
 0  1

 2  2

 Example 5-2 10
5.2.3 Nonlinear Regression
 Search for parameter values to minimize the sum of the squares of the
differences between the measured values and the calculated values for all data
points.
 Many computer softwares are available to find these parameters.
 Initial estimates of these parameters obtained from linear-least-squares
techniques can facilitate efficient search.
 Nonlinear Least-Square Analysis
Calculate the rate of reaction, rc, with estimates of parameters (eg. reaction order and rate
constant)
The measured reaction rates, rm, obtained from experimental data
Data analysis: to minimize the sum of the squared differences, (rm – rc)2, for all data points
If we carry out N experiments to determine K parameters, we would like to minimize the
quantity
s2 N
(rim  ric ) 2
 
2

N  K i 1 N  K
where rim and ric are the measured and calculated reaction rate for run i.
A number of optimization techniques or software packages are available to find the optimal
parameters that give the minimum of the sum of squares, σ2.

11
 Concentration-Time Data analyzed by nonlinear regression
for constant-volume batch reactor,
dC A
 rA   kC A → C 1A0  C 1A  (1   )kt
dt

→ C A  [C 1A0  (1   )kt ]1/(1 )

 Use software packages to find k and α that gives minimum of
N N
s   (C Ami  C Aci )   [C Ami  [C 1A0  (1   )kti ]1/(1 ) ]2
2 2

i 1 i 1

If Polymath is used, one should use absolute value for the CAci, i.e.
N N
s   (C Ami  C Aci )   [C Ami  {(abs[C 1A0  (1   )kti ]}1/(1 ) ]2
2 2

i 1 i 1

 Another way is to use time rather than concentrations

C 1A0  C 1A
tc 
k (1   )

in this case the sum of squared differences is

N N
C 1A0  C 1A 2
s   (t mi  tci )   [t mi 
2 2
]
i 1 i 1 k (1   )

 Example 5-3 12
 5.3 Method of Initial Rates
General Introduction
 For reversible reactions (render the differential method ineffective)
 It requires a series of experiments carried out at different initial concentrations,
CA0.
 The initial rate of reaction, -rA0, is determined for each run by differentiating the
data and extrapolating to zero time.
 After data analysis to obtain –rA0 vs. CA0
consider a reaction with the rate law of
 rA0  kC A0 → ln( rA0 )  ln k   ln(C A0 )
the slope of the plot of ln(–rA0) vs. ln(CA0) gives the reaction order .

 Example 5-4

13
 5.4 Method of Half-Lives
General Introduction
 The half-life of a reaction, t1/2, is defined as the time it takes for the
concentration of the reactant to fall to half of its initial value.
 The reaction order and specific reaction rate are determined by the plot of t1/2 vs.
CA0.
 If more than one reactant are involved in the rate law, the method of excess in
conjunction with the method of half-lives
 Consider a constant-volume batch reactor with rate law of rA = -kCAα,
dC A
   rA  kC A
dt
integrating with the initial condition of CA = CA0 at t = 0,
→ 1 1 1 1 C
t (  1   1 )   1 [( A0 ) 1  1]
k (  1) C A C A0 kC A0 (  1) C A

at t = t1/2, CA = (1/2)CA0, thus

2 1  1 1 n 1  1 1
t1/ 2  ( ) → t1/ n  ( ) 14
k (  1) C A01 k (  1) C A01
 2 1  1
→ ln(t1/ 2 )  ln[ ]  (1   ) ln(C A0 )
k (  1)
 The plot of ln(t1/2) vs. ln (CA0) gives (1-α)

 Another way is to use nonlinear regression on the half-life data

15
 5.5 Differential Reactors
General Introduction
 For catalytic (heterogeneous) reaction.
 For determination of the rate of reaction as a function of either concentration or
partial pressure.
 It consists of a tube containing a very small amount of catalyst usually arranged
in the form of a thin wafer or disk.

 Criterion for a differentia reactor:

- the conversion of reactant is extremely small
- so as the change in temperature and concentration
- the concentration is essentially constant (~CA0)
 The reactor is gradientless, and the reaction rate is spatially uniform within the
bed. 16
 Some precautions:
The reactant liquid or gas should not bypass or channel through the packed catalyst.
If the catalyst decays rapidly, the reaction rate parameters would vary significantly from the
start to the end of a run.
Sampling and analysis of product stream may be difficult for small conversions in
multicomponent systems.
 Differential Reactor Design Equation
consider the mole balance of reactant A for the catalyst bed,

solving for –rA',

FA0  FAe
 r'A 
W
in terms of concentration, conversion, or product flow rate (FP)
FA0 X F
 r'A   P ~ Differential Reactor Design Equation
W W

(note that stoichiometric coefficients of A and P should be considered.) 17

 For constant volumetric flow rate, v = v0 = constant,
v0 (C A0  C Ae ) v0C P
 r'A  
W W
 By using very little catalyst and large volumetric flow rate, the concentration
difference, (CA0 – CAe), can be made quite small.
The rate of reaction can be obtained as a function of the reactant
concentration in the catalyst bed, CAb, by varying the inlet concentration
 r ' A   r ' A (C Ab )
 One approximation of CAb would be the arithmetic mean of the inlet and
outlet concentrations
C A0  C Ae
C Ab 
2
 However, due to the very little reaction takes place within the bed, the bed
concentration is essentially equal to the inlet concentration,
C Ab  C A0 →  r ' A  r ' A (C A0 )
 Various numerical and graphical techniques can be used to determine the
appropriate algebraic equation for the rate law.
 Use high flow rates and small catalyst particle sizes to avoid mass transfer
limitation.
 If data show reaction order to be 1 with a low activation energy, suspect the
18
data is being collected in the mass transfer limited regime (chap. 10-12).