Non-ideal Reactors

Ideal flow reactor:

1. Backmix reactor - Complete mixing among all the fluid elements in the
reactor.
2. Plug flow reactor - No mixing in the axial direction.
Real reactors may deviate from these assumptions.
 All the fluid elements may not spend the same time in the reactor
 Channeling of fluid, creation of dead zones

Residence time distribution (RTD):

RTD is a characteristic of mixing in the reactor. This may effectively be utilized
in analyzing the non-ideality of flow in reactors.
Residence Time: This is the time a fluid element has spent in the reactor.
 Ideal PFR & Ideal batch Reactor: All the fluid elements have the same
residence time.
 Ideal CSTR: All fluid elements do not have identical residence time. The
feed gets mixed completely with the reaction mixture. Some may leave
the reactor immediately, while some may reside for a long time.
Measurement of RTD:
RTD is determined experimentally by injecting an inert, called tracer into the
reactor.
Feed, c(t)
Effluent, C(t)

Inert tracer Detection

Reactor
The tracer concentration in the effluent stream, C(t), is measured as a function
of time. This data is used to determine RTD.
Tracer:
 Cheap
 Non-reactive
 Easily detectable
 Should have physical properties similar to that of the reaction mixture
There are two common methods of injection of tracer:
1. Pulse input
2. Step input
(1) Pulse input: A pulse of tracer is injected into feed at t = 0. Its concentration
in the exit stream is monitored with time.

delt

N0 = Total Number of moles of tracer injected into the feed

∆N = Amount of tracer leaving the reactor over ∆t, between t and t + ∆t
( )
( )
( )

This represents the fraction of fluid which has a residence time in the reactor
between t and ∆t.
RTD, E(t):
Residence time distribution or Exit age distribution: It describes in a
quantitative manner how much time different fluid elements have spent in the
reactor. E(t)∆t represents the fraction of the material that has a residence time
between t and ∆t.
( )
( )

If N0 is not known
( )

∫ ( )

( )
( ) The shaded area represents the fraction of fluid elements having RT
∫ ( ) between 3 min and 6 min.

∫ ( )

F(T), Cumulative RTD:

( ) ∫ ( ) Fraction of effluent with residence time < t

∫ ( ) ( ) Fraction of effluent with residence time > t

F-curve
e.g. 80 % of the material spends < 8 min in the reactor. 20% spends > 8 min.
This information is directly obtained from F-curve.
Mean Residence Time, tm:
This is the first moment of RTD function.

∫ ( )
∫ ( )
∫ ( )

A closed vessel is the one in which fluid enters and leave entirely by plug flow.
For a closed vessel tm = 
Space time = Mean Residence time

Variance of RTD, σ2:

Variance is the Square of the standard deviation and is an indication of spread
of the distribution. This is obtained by taking the second moment of RTD
function about the mean.

∫ ( ) ( )

Two parametersmost commonly used to characterize the RTD are tm or  and σ.

 Problem:
A tracer is injected as a pulse to a reactor and the effluent tracer concentration
is measured as a function of time. The data is given below.
i. Construct the C-curve and the E-curve
ii. Determine the fraction of material leaving the reactor that has spent
between 3 and 6 min in the reactor.
iii. Do the same between 7.75 & 8.25 min.
iv. Calculate the mean Residence time.
v. Calculate the variance of the distribution.

3 2
t(min) C(t), g/m E(t) tE(t) t-tm (t-tm) E(t)
0 0 0 0 -5.15 0
1 1 0.02 0.02 -4.15 0.34
2 5 0.1 0.2 -3.15 0.992
3 8 0.16 0.48 -2.15 0.74
4 10 0.2 0.8 -1.15 0.265
5 8 0.16 0.8 -0.15 0.004
6 6 0.12 0.72 0.85 0.087
7 4 0.08 0.56 1.85 0.274
8 3 0.06 0.48 2.85 0.487
9 2.2 0.044 0.4 3.85 0.652
10 1.5 0.03 0.3 4.85 0.706
12 0.6 0.012 0.14 6.85 0.563
14 0 0 0 8.85 0

( )
i. ( )
∫ (( )

∫ ( ) ∫ ( ) ∫ ( )
For N + 1 points where N is Even (Equal intervals)

∫ ( ) [ ]

Where

∫ ( ) [( ) ( ) ( ) ( ) ( ) ( )
( ) ( ) ( ) ( ) ( )]
g. min/m3

∫ ( ) [( ) ( ) ( )]

= 2.6 min/m3

∫ ( ) = 50 g.min/m3
( )
( )

10 0.2

C(t) E(t)

0 3 6 t
C curve / E curve

ii. ∫ ( )
51% of the material leaving the reactor spends between 3 to 6 min in the
reactor.

iii. ∫ ( )
 ∫ ( ) ∫ ( ) ∫ ( )

iv. tm= [( ) ( ) ( ) ( ) ( )
( ) ( ) ( ) ( ) ( ) (
)] + [ ( ) ]
= 4.58 + 0.57
= 5.15 min

v. ∫ ( ) ( )

] [ ]
= 4.136 + 1.972
= 6.1 m2

The RTD is Characteristic by

tm = 5.15 min
Convolution Theorem:

By convolution theorem, it may be proved that the output concentration is related to

the input concentration as:
C(t) = ∫ ( ) ( )
Inlet concentration is usually an imperfect or a perfect pulse (Delta-Dirac) or a step
function.
For perfect pulse
δ(t-t0) = ∞ at t = t0
= 0 at t t0
c(t) Area = 1
∫ ( 0)dt = 1

Property of Delta function

∫ ( 0)f(t)dt = f(t0), if a t0 b
= 0, if (a, b) doesn’t contain t0
t0 t
Delta-Dirac function
2) Step Input:

c (t) = 0 at t < 0

= c0 at t ≥ 0

c(t) = c0 C(t) E(t) F(t)

0 t t t t

From convolution theorem

C(t) = ∫0t c0 E(t’) dt’

C(t)/c0 = ∫0t E(t’)dt’ = F(t)

E(t) = d/dt[ C(t)/c0] =d/dt [F(t)]

Comparison between Pulse and Step Test:

Pulse Test:

 Very little amount of tracer used. Suited for expensive tracer.

 Difficult to get a perfect pulse as the injection time is very small.
 C-curve has a long tail. Leads to large inaccuracy.

Step Test:

 Easier to carry out.

 Total amount of tracer injection need not be known.
 Difficult to maintain constant tracer concentration over time.
 To get RTD, differentiation required. Errors.
RTD in ideal reactors:

Plug Flow & Batch Reactors (Pulse tracer input):

All the fluid elements spend precisely the same period of time within the reactor, and
will have the same residence time.

If the input pulse is a unit delta-Dirac function, the output will also be a unit delta Dirac
function. RTD function will be a Delta Dirac Function at t= .

c(t) C(t)

In out

0 t

( ) ( )

( ) ( )
( ) ( )
∫ ( ) ∫ ( )

( ) ( )

Mean RT, tm =
Ideal CSTR:
Concentration at any point in the reactor is equal to the concentration of the effluent
stream.
A tracer pulse is injected into a CSTR.
Feed
After a time t
In – Out = Accumulation Tracer Pulse
Detector

CSTR

At [Initial Condition]

Using initial condition,

=- ⁄ + at 

=- ⁄ +

Ideal PFR
=
( )
( )
∫ ( ) ∫
Ideal STR
[ ] 𝑬(𝒕)
[ ]

∫ ( ) ∫

t t+ t t+

Spread of RTD for an ideal CSTR is very high

Reactor Modeling With RTD:
i
In a non-deal rector, 2 aspects play very important roles - should be considered for
modeling.
(1) Residence Time Distribution: The time spent by different fluid elements in the
reactor.
(2) Mixing among the fluid elements with different residence times.
For first-order reactions, the rate of change in conversion is independent of the
reactant concentration.

= k(1-X)

Hence the mixing of the fluid element with the surrounding elements is not important
for the first order reactions. Only RTD information is sufficient to predict conversion.

Reactions of other orders require the knowledge of both RTD and the degree of mixing
among the fluid elements with different residence time.

Two extremes of such mixing are provided by two Zero-parameter models

 Segregation Model
 Maximum Mixedness Model
Segregation Model (Zero Parameter Model):
This Model assumes complete segregation and the fluid elements of different ages do
not mix at all.

Different fluid elements spend different time in the reactor. Each element acts as a
batch reactor with the reaction time equal to the residence time of the element in the
reactor. RTD of these elements will be obtained from the RTD study of the reactor.

To determine the mean conversion in the effluent stream, the conversions of various
elements in the exit stream should be averaged.

[Mean conversion of the elements spending time between t & t+dt in the reactor]
= [Conversion achieved after spending a time t in reactor] x
[Fraction of elements that spends between t & t + dt in the reactor]
̅ ( ) ( )

̅ ∫ ( ) ( )

If we have the batch reactor equation for ( ) and measure RTD, E(t) experimentally,
we can find the mean conversion in the exit stream for a non-ideal reactor.
First order reactions:

( )

∫ ∫

( )

( )

̅ ∫( ) () ∫ ( ) ∫ ( )

̅ ∫ ()
Mean Conversion in an ideal PFR:

( ) ( )

̅ ∫ () () ∫ ()

∫ ( )

∫ ( ) () ( ) , if a < t0 < b [Convolution Theorem]

̅ ... (A)

For an ideal PFR, the design equation derived from mole balance, rate law and
stoichiometric relations:

∫ ( )

( ) ( )

… (B)

(A) & (B) are identical.

Mean conversion in an ideal CSTR
()

̅ ∫ ()

( )
∫

( )
[ ]

[ ]

̅ ( )

Design equation for an ideal CSTR is derived from mole balance, rate law and
stoichiometry:

( )

( )

(c) and (d) are identical.

 We find that conversions predicted by segregation model for ideal PFR and CSTR
are the same as the conversion predicted by the design equations.
Problem: Calculate the mean conversion in the reactor mentioned in the reactor
mentioned in the previous problem for a first order liquid-phase irreversible reaction.

A ➝k Products
k = 0.1 min⁻1

(i) Compare the conversion with that in ideal PFR.

(ii) Compare the conversion with that in ideal CSTR.

Solution: = 1 - e-kt = 1 - e-0.1 t

̅ ∫ ( ) ( )

t(min) E(t) X(t) = 1 - e-kt X(t)E(t)

0 0 0 0
1 0.02 0.095 0.0019
2 0.10 0.181 0.0180
3 0.16 0.259 0.0414
4 0.20 0.330 0.0660
5 0.16 0.393 0.0629
6 0.12 0.451 0.0541
7 0.08 0.503 0.0402
8 0.06 0.551 0.0331
9 0.044 0.593 0.0261
10 0.030 0.632 0.01896
12 0.012 0.699 0.0084
14 0 0.75 0
∫ ( ) ( ) ∫ ( )E(t)dt + ∫ ( ) ( )

= [0 + 4 0.0019 + 2 0.018 + 4 0.0414 + 2 0.066 + 4

] + [0.01896 + 4

084 + 0]
= 0.350+0.035 = 0.385
̅ = 0.385

Ideal PFR: X = 1-
= 1- = 0.402

Ideal CSTR: X = = = 0.34