UNIT 8: MODELS FOR NON-IDEAL REACTORS

Recommended Reading:
OL: 11.2, 12, 13, 14
HSF: 16.5, 17.1-17.4, 18.1-18.5, 18.8 (Exclude Maximum Mixedness model)
LEARNING OBJECTIVES

◼ Describe quality of mixing and state of aggregation.

◼ Predict performance of non-ideal reactors using models.

MODELS FOR NON-IDEAL REACTORS 2

OUTLINE
◼ Performance of Non-Ideal Reactors
◼ Degree of Mixing
◼ State of Aggregation
◼ Models to predict conversion
❑ Zero Parameter – Segregation Model
❑ Single Parameter – Tank-In-Series & Dispersion Models
❑ Two Parameters – Compartment Model

MODELS FOR NON-IDEAL REACTORS 3

PERFORMANCE OF NON-IDEAL REACTORS

➢ Reaction Kinetics: How the reaction progresses?

REACTION KINETICS
−𝑟𝐴 = 𝑘(𝑇) 𝑓(𝐶𝑖 )

➢ Residence Time Distribution: How long a fluid element stays in a reactor?

RTD
𝐸 𝑡 , 𝐹 𝑡 ⟹ 𝑡𝑚 , 𝜎 2
Referred as MACROMIXING information as it does not describe how the fluid elements interact with
each other.
➢ Quality of mixing: How the fluid elements move through the reactor?
MODELS
IS THERE ANY EXCEPTION?

MODELS FOR NON-IDEAL REACTORS 4

DEGREE OF MIXING
MACROMIXING
❑ Describes the distribution of residence time
❑ Does not specify how molecules of different ages interact with one another in a reactor
MICROMIXING
❑ Describes how molecules of different ages encounter with one another in a reactor
❑ Complete Segregation: Molecules of same age move together and does not interact with
molecules of different age.
❑ Complete Micromixing: Molecules of different age group are mixed completely as soon as they
enter the reactor.
For a given state of macromixing the extremes of micromixing provide the upper and lower limits
on conversion

MODELS FOR NON-IDEAL REACTORS 5

STATE OF AGGREGATION
Flowing stream tends to aggregate as globules (fluid particle comprising of
molecules of same age)
MACROFLUID MICROFLUID
Globules of given age do not mix Molecules are not constrained to
with other globules. remain in a globule and are free to
move anywhere.

Macrofluid Microfluid

Molecules aggregate Molecules are

to form globules uniformly mixed

MODELS FOR NON-IDEAL REACTORS 6

MODELS TO PREDICT CONVERSION

➢ Zero Parameter Models

❑ Segregation model
❑ Maximum Mixedness model
➢ One Adjustable Parameter Model
❑ Tank-In-Series model
❑ Dispersion model
➢ Two Adjustable Parameter Model
❑ Compartment model
Which model is the best to predict conversion in a non-ideal reactor?

MODELS FOR NON-IDEAL REACTORS 7

SEGREGATION MODEL

➢ Fluid flows through the reactor as globules.

➢ Globule comprises of molecules of same age. Tiny batch reactor

➢ Each globule spend a definite amount of time, different from the other
globules.
➢ No molecular interchange between the globules.
➢ Each globule behaves as a single batch reactor.
➢ Reaction time for tiny batch reactor will be the time the globule spends
Well Mixed
in the reactor.
➢ At the exit of the reactor globules of different ages are mixed.

MODELS FOR NON-IDEAL REACTORS 8

 Conversion in the exit stream will be average of conversion achieved in the
globules exiting the reactor

Mean Conversion of
Conversion achieved Fraction of globules that
globules spending
in a globule after spend between 𝑡 and
between time 𝑡 and
spending time 𝑡 t + 𝑑𝑡 in the reactor
t + 𝑑𝑡 in the reactor

𝑑𝑋ത = 𝑋(𝑡) × 𝐸 𝑡 𝑑𝑡
∞
ഥ = න 𝑿 𝒕 𝑬 𝒕 𝒅𝒕
𝑿
𝟎

ഥ : Mean Conversion in the exit stream

𝑿
Solving Batch Reactor
𝑿 𝒕 : Conversion in a batch reactor design equation using
reaction kinetics
RTD of Real
𝑬 𝒕 : Exit Age Distribution
Reactor
MODELS FOR NON-IDEAL REACTORS 9
EXAMPLE 1
A first order liquid decomposition reaction is carried out in a plug flow reactor. Pulse tracer tests on the
reactor showed deviations from ideality.
Find the fraction of reactant unconverted in the real reactor and compare this with the fraction
unconverted in an ideal PFR of the same size.
Additional Data
Reaction Kinetics: −𝑟𝐴 = 𝑘𝐶𝐴 , 𝑘 = 0.307 𝑚𝑖𝑛−1
RTD:
Time 𝑡, 𝑚𝑖𝑛 0 5 10 15 20 25 30 35
𝐸 𝑡 , 𝑚𝑖𝑛−1 0 0.03 0.05 0.05 0.04 0.02 0.01 0

MODELS FOR NON-IDEAL REACTORS 10

SOLUTION
For a first order reaction, RTD and reaction kinetics is sufficient to predict the conversion.
a) Fraction of reactant unconverted in a real reactor
Segregation model,
∞
𝐶𝐴
= 1 − න 𝑋 𝑡 𝐸 𝑡 𝑑𝑡
𝐶𝐴0 0

Expression for 𝑋 𝑡 : Batch reactor design equation

𝑋
𝑑𝑋𝐴
𝑡 = 𝐶𝐴0 න 𝜀𝐴 = 0
0 −𝑟𝐴
1 𝑋 𝑑𝑋𝐴
𝑡=− න
𝑘 0 (1 − 𝑋𝐴 )
𝑋 𝑡 = 1 − 𝑒 −𝑘𝑡
∞
𝐶𝐴
= න 𝑒 −𝑘𝑡 𝐸 𝑡 𝑑𝑡
𝐶𝐴0 0

MODELS FOR NON-IDEAL REACTORS 11

Using RTD data,
Time 𝑡, 𝑚𝑖𝑛 0 5 10 15 20 25 30 35
𝐸 𝑡 , 𝑚𝑖𝑛−1 0 0.03 0.05 0.05 0.04 0.02 0.01 0
𝑒 −𝑘𝑡 𝐸 𝑡 , 𝑚𝑖𝑛−1 0 0.0064 0.0023 0.0005 0 0 0 0
∞
𝐶𝐴
= න 𝑒 −𝑘𝑡 𝐸 𝑡 𝑑𝑡 = 0.0469
𝐶𝐴0 0

Fraction of reactant unconverted in real reactor: 4.7%

Area under the curve
b) Fraction of reactant unconverted in an ideal PFR
PFR design equation,
𝐶𝐴
𝑑𝐶𝐴
𝜏 = −න 𝜀𝐴 = 0
𝐶𝐴0 −𝑟𝐴
1 𝐶𝐴0 𝐶𝐴
𝜏 = ln ⟹ = 𝑒 −𝑘𝜏
𝑘 𝐶𝐴 𝐶𝐴0

MODELS FOR NON-IDEAL REACTORS 12

Mean residence time: RTD data
∞
Using, 𝑡𝑚 = න 𝑡𝐸 𝑡 𝑑𝑡
0
𝑡𝑚 = 15 𝑚𝑖𝑛
𝐶𝐴
= 𝑒 −(0.307)(15) = 0.01
𝐶𝐴0
Fraction of reactant unconverted in an ideal PFR: 1%
c) Conclusions
➢ Conversion achieved in real reactor (95.3%) is less than that of the ideal PFR (99%).
➢ Most of the unconverted reactant leave the reactor early (Plot, 𝑒 −𝑘𝑡 𝐸 𝑡 𝑣𝑠 𝑡)

MODELS FOR NON-IDEAL REACTORS 13

SINGLE PARAMETER MODEL
Tank-In-Series Model
❑ Real reactor modeled as a series of identically sized CSTRs
❑ Model Parameter: N, no of ideal CSTRs

Dispersion Model
❑ Real reactor modeled as a PFR superimposed with axial dispersion
❑ Model parameter: D, Dispersion coefficient

MODELS FOR NON-IDEAL REACTORS 14

TANK-IN-SERIES MODEL 𝑣, 𝐶𝐴
𝑽
𝑣, 𝐶𝐴0

𝑽𝟏 𝑽𝟐 𝑽𝟑 𝑽𝟒 𝑽𝑵

𝑣, 𝐶𝐴0 𝑣, 𝐶𝐴

𝟏 𝟐 𝟑 𝟒 𝑵
𝑉1 = 𝑉2 = ⋯ 𝑉𝑁 = 𝑉𝑖 (or) 𝑉 = 𝑁𝑉𝑖
𝜏1 = 𝜏2 = ⋯ 𝜏𝑁 = 𝜏𝑖 (or) 𝜏 = 𝑁𝜏𝑖

Conversion achieved in the real reactor can be predicted by solving ideal

CSTR in series performance equation, if 𝑁 is known.
How to find the number of equal sized ideal CSTRs that behave similar to the
real reactor?
MODELS FOR NON-IDEAL REACTORS 15
TANK-IN-SERIES MODEL 𝑣, 𝐶𝐴
𝑽
𝑣, 𝐶𝐴0

𝑽𝟏 𝑽𝟐 𝑽𝟑 𝑽𝟒 𝑽𝑵

𝑣, 𝐶𝐴0 𝑣, 𝐶𝐴

𝟏 𝟐 𝟑 𝟒 𝑵
𝑉1 = 𝑉2 = ⋯ 𝑉𝑁 = 𝑉𝑖 (or) 𝑉 = 𝑁𝑉𝑖
𝜏1 = 𝜏2 = ⋯ 𝜏𝑁 = 𝜏𝑖 (or) 𝜏 = 𝑁𝜏𝑖

Conversion achieved in the real reactor can be predicted by solving ideal

CSTR in series performance equation, if 𝑁 is known.
How to find the number of equal sized ideal CSTRs that behave similar to the
real reactor?
MODELS FOR NON-IDEAL REACTORS 16
RTD of equal sized identical CSTRs in series
Stimulus-Response Technique

Pulse Tracer 𝑣
𝐶 𝑡=0
𝐶𝑖𝑛 = ቊ 0
0 𝑡>0 𝐶1 𝐶2 𝐶3 𝐶4
𝐶𝑁
𝟏 𝟐 𝟑 𝟒 𝑵

𝐶𝑁 (𝑡)
Exit Age Distribution, 𝐸 𝑡 = ∞
‫׬‬0 𝐶𝑁 𝑡 𝑑𝑡
MB on Reactor 1
𝐼𝑛 – 𝑂𝑢𝑡 = 𝐴𝑐𝑐𝑢𝑚𝑢𝑙𝑎𝑡𝑖𝑜𝑛
𝑑𝐶1
0 − 𝑣𝐶1 = 𝑉1 𝜏1 = 𝑉1 Τ𝑣
𝑑𝑡
−𝑡Τ𝜏1 𝑒 −𝑡Τ𝜏1
Integrating, 𝐶1 = 𝐶0 𝑒 ⟹ 𝐸1 (𝑡) =
𝜏1
MODELS FOR NON-IDEAL REACTORS 17
MB on Reactor 2
𝑑𝐶2
𝑣𝐶1 − 𝑣𝐶2 = 𝑉2
𝑑𝑡
𝑑𝐶2
𝑣𝐶0 𝑒 −𝑡Τ𝜏1 − 𝑣𝐶2 = 𝑉2
𝑑𝑡
𝑑𝐶2 𝐶2 𝐶0 −𝑡Τ𝜏
+ = 𝑒 1 𝜏2 = 𝑉2 Τ𝑣
𝑑𝑡 𝜏2 𝜏2
Integrating,
𝑡 −𝑡 Τ 𝜏 𝑡 −𝑡 Τ𝜏 𝜏1 = 𝜏2 = ⋯ = 𝜏𝑖
𝐶2 = 𝐶0 𝑒 𝑖 ⟹ 𝐸2 (𝑡) = 2 𝑒 𝑖
𝜏𝑖 𝜏𝑖
MB on Reactor 3
𝑑𝐶3
𝑣𝐶2 − 𝑣𝐶3 = 𝑉3
𝑑𝑡 Generalizing for 𝑁 reactors,
Using 𝐶2 and solving for 𝐶3
𝑡 2 −𝑡 Τ 𝜏 𝑡 2 −𝑡 Τ 𝜏 𝒕𝑵−𝟏 − 𝒕 Τ 𝝉𝒊
𝐶3 = 𝐶0 𝑒 𝑖
⟹ 𝐸3 (𝑡) = 𝑬(𝒕) = 𝒆
3𝑒
𝑖 𝑵
2𝜏𝑖 2 2𝜏 𝑵 − 𝟏 ! 𝝉𝒊
𝑖

MODELS FOR NON-IDEAL REACTORS 18

RTD of equal sized identical CSTRs in series
𝒕𝑵−𝟏 − 𝒕 Τ 𝝉𝒊
𝑵(𝑵𝜽)𝑵−𝟏 −𝑵𝜽
𝑬 𝒕 = 𝑵 𝒆 𝑜𝑟 𝑬(𝜽) = 𝒆
𝑵 − 𝟏 ! 𝝉𝒊 𝑵−𝟏 !
𝜃 = 𝑡Τ𝜏

𝑁=∞

𝜽 = 𝒕Τ𝝉
MODELS FOR NON-IDEAL REACTORS 19
Determine Number of Tanks using RTD

∞
𝜎2 = න 𝑡 − 𝑡𝑚 2 𝐸 𝑡 𝑑𝑡
0

∞
𝑡 𝑁−1 −𝑡Τ𝜏𝑖
𝜎2 = න 𝑡−𝜏 2
𝑁 𝑒 𝑑𝑡
0 𝑁 − 1 ! 𝜏𝑖
𝟐
𝝉 𝟏
𝝈𝟐 = 𝑜𝑟 𝝈𝜽 𝟐
=
𝑵 𝑵
Real Reactor behaves similar to,
CSTR when ?
PFR when ?

MODELS FOR NON-IDEAL REACTORS 20

How to predict conversion using TIS Model?
a) Evaluate RTD parameters: 𝑡𝑚 and 𝜎 2
b) Evaluate Reaction kinetics: 𝑘, 𝐶𝐴0
c) Evaluate reactor parameters: 𝑣, 𝑉
d) Calculate the number of identical CSTRs required to mimic real reactor.
𝜏 2 𝑡𝑚 2 N should be an integer (except
𝑁= 2= 2 for 1st order reaction)
𝜎 𝜎
e) Predict conversion in real reactor by solving ideal CSTR in series performance equation.
𝐶𝐴0 −𝐶𝐴1 𝜏𝑖 = 𝑉 Τ𝑁𝑣
Reactor 1: 𝜏𝑖 =
−𝑟𝐴1 (or)
𝐶𝐴1 −𝐶𝐴2 𝜏𝑖 = 𝑡𝑚 Τ𝑁
Reactor 2: 𝜏𝑖 =
−𝑟𝐴2
𝑿𝑨
𝐶𝐴(𝑁−1) −𝐶𝐴𝑁 For first order reaction,
Reactor 𝑁: 𝜏𝑖 = 𝟏
−𝑟𝐴𝑁
𝑿=𝟏−
𝟏 + 𝝉𝒊 𝒌 𝑵
MODELS FOR NON-IDEAL REACTORS 21
EXAMPLE 2
A first order liquid decomposition reaction is carried out in a plug flow reactor. Pulse tracer tests on the
reactor showed deviations from ideality. Find the fraction of reactant unconverted in the real reactor
using Tank-In-Series Model.
Additional Data
Reaction Kinetics: −𝑟𝐴 = 𝑘𝐶𝐴 , 𝑘 = 0.307 𝑚𝑖𝑛−1
RTD:

Time 𝑡, 𝑚𝑖𝑛 0 5 10 15 20 25 30 35
𝐸 𝑡 , 𝑚𝑖𝑛−1 0 0.03 0.05 0.05 0.04 0.02 0.01 0

MODELS FOR NON-IDEAL REACTORS 22

a) RTD parameters
(i) Mean Residence Time (ii) Variance
∞
Using, 𝑡𝑚 =
∞
‫׬‬0 𝑡𝐸 𝑡 𝑑𝑡 Using, 𝜎 2 = ‫׬‬0 𝑡 − 𝑡𝑚 2 𝐸 𝑡 𝑑𝑡
𝑡𝑚 = 15 𝑚𝑖𝑛 𝜎 2 = 47.5 𝑚𝑖𝑛2

b) Reaction kinetics c) Reactor parameters

−𝑟𝐴 = (0.307 𝑚𝑖𝑛−1 )𝐶𝐴 Mean residence time is sufficient
d) Number of identical CSTRs
𝑡𝑚 2 152
𝑁= 2 = ⟹ 𝑁 = 4.737
𝜎 47.5
e) Conversion in ideal CSTRs in series
1
𝑋 =1−
1 + 𝜏𝑖 𝑘 𝑁
𝜏𝑖 = 𝑡𝑚 Τ𝑁 ⟹ 𝜏𝑖 = 3.167 𝑚𝑖𝑛
1
𝑋 =1− 4.737
⟹ 𝑿 = 𝟗𝟔%
1 + (3.167)(0.307)
MODELS FOR NON-IDEAL REACTORS 23
AXIAL DISPERSION MODEL
𝑫
𝑽
𝑣, 𝐶𝐴
𝑽
𝑣, 𝐶𝐴0 𝑣, 𝐶𝐴
𝑣, 𝐶𝐴0

Dispersed Plug flow

➢ Plug flow superimposed with some degree of backmixing.
➢ Axial dispersion is characterized by Dispersion coefficient 𝐷.
➢ Dispersion coefficient 𝐷 can be estimated from RTD or mass transfer correlations.

Molecules move
ahead or lag behind
due to molecular and
turbulent diffusion

MODELS FOR NON-IDEAL REACTORS 24

RTD of Dispersed Plug Flow

𝑡1 𝑡2 𝑡3

𝑡4 𝑡5

Molecular diffusion causes the pulse to broaden and spread. The spreading is given by
dispersion coefficient 𝐷.
Large 𝐷: Rapid spreading
Small 𝐷: Slow spreading
What should be the value of 𝑫?
CSTR: ?
PFR: ?

MODELS FOR NON-IDEAL REACTORS 25

As the tracer flows through the reactor at superficial velocity 𝑈, its concentration varies along the
reactor. Consider a differential element of length ∆𝑧 within the reactor with cross-sectional area 𝐴𝑐 .
Mole Balance: 𝐹𝑙𝑜𝑤 𝑖𝑛 − 𝐹𝑙𝑜𝑤 𝑜𝑢𝑡 + 𝑔𝑒𝑛𝑒𝑟𝑎𝑡𝑖𝑜𝑛 = 𝐴𝑐𝑐𝑢𝑚𝑢𝑙𝑎𝑡𝑖𝑜𝑛
𝜕𝐶𝑇
𝐹𝑇 ቚ − 𝐹𝑇 ቚ = 𝐴𝑐 ∆𝑧
𝑧 𝑧+∆𝑧 𝜕𝑡 𝐶𝑇𝐿
𝑈
𝜕𝐹𝑇 𝜕𝐶𝑇
− = 𝐴𝑐
𝜕𝑧 𝜕𝑡
𝐹𝑇 : Molar Flow Rate of Tracer 𝐴𝑐 𝑧 𝑧 + ∆𝑧

𝐹𝑇 = 𝐴𝑐 𝑊𝑇
𝑊𝑇 : Molar Flux of the Tracer
𝑊𝑇 = 𝑀𝑜𝑙𝑒𝑐𝑢𝑙𝑎𝑟 𝐷𝑖𝑓𝑓𝑢𝑠𝑖𝑜𝑛 + 𝐵𝑢𝑙𝑘 𝐹𝑙𝑜𝑤 𝝏𝟐 𝑪𝑻 𝝏 𝑼𝑪𝑻 𝝏𝑪𝑻
𝑫 − =
𝑊𝑇 = 𝐽𝑇 + 𝐵𝑇 𝝏𝒛𝟐 𝝏𝒛 𝝏𝒕
𝜕𝐶𝑇
𝐽𝑇 = −𝐷 𝐵𝑇 = 𝐶𝑇 𝑈
𝜕𝑧
𝜕𝐶𝑇
𝐹𝑇 = 𝐴𝑐 −𝐷 + 𝐶𝑇 𝑈
𝜕𝑧

MODELS FOR NON-IDEAL REACTORS 26

 𝜕 2 𝐶𝑇 𝜕 𝑈𝐶𝑇 𝜕𝐶𝑇
𝐷 − =
𝜕𝑧 2 𝜕𝑧 𝜕𝑡
𝐶𝑇 𝑧 𝑡𝑈
𝜓= 𝜆= 𝜃=
𝐶𝑇0 𝐿 𝐿
1 𝜕 2 𝜓 𝜕𝜓 𝜕𝜓
− =
𝑃𝑒𝑟 𝜕𝜆2 𝜕𝜆 𝜕𝜃
𝑈𝐿 Reactor Peclet Number
𝑃𝑒𝑟 =
𝐷
𝑅𝑎𝑡𝑒 𝑜𝑓 𝑡𝑟𝑎𝑛𝑠𝑝𝑜𝑟𝑡 𝑏𝑦 𝑐𝑜𝑛𝑣𝑒𝑐𝑡𝑖𝑜𝑛
𝑃𝑒𝑟 = Bodenstein Number, 𝐵𝑜
𝑅𝑎𝑡𝑒 𝑜𝑓 𝑡𝑟𝑎𝑛𝑠𝑝𝑜𝑟𝑡 𝑏𝑦 𝑑𝑖𝑓𝑓𝑢𝑠𝑖𝑜𝑛/𝑑𝑖𝑠𝑝𝑒𝑟𝑠𝑖𝑜𝑛
Analytical solution for partial differential equation (Closed-Closed System)
𝐶𝑇 = 𝐼𝑛𝑓𝑖𝑛𝑖𝑡𝑒 𝑠𝑒𝑟𝑖𝑒𝑠
Evaluating 𝑡𝑚 and 𝜎 2
𝒕𝒎 = 𝝉
𝟐
𝝈𝟐 𝟐 𝟐 −𝑷𝒆𝒓
𝝈𝜽 = = − 𝟏 − 𝒆
𝒕𝒎 𝟐 𝑷𝒆𝒓 𝑷𝒆𝒓 𝟐

MODELS FOR NON-IDEAL REACTORS 27

 Numerical solution for partial differential equation at different 𝑃𝑒𝑟

MODELS FOR NON-IDEAL REACTORS 28

Conversion in Dispersed Plug Flow
Mole Balance on a small differential element,
𝐹𝑙𝑜𝑤 𝑖𝑛 − 𝐹𝑙𝑜𝑤 𝑜𝑢𝑡 + 𝑔𝑒𝑛𝑒𝑟𝑎𝑡𝑖𝑜𝑛 = 𝐴𝑐𝑐𝑢𝑚𝑢𝑙𝑎𝑡𝑖𝑜𝑛
𝐹𝐴 ቚ − 𝐹𝐴 ቚ + 𝑟𝐴 𝐴𝑐 ∆𝑧 = 0
𝑧 𝑧+∆𝑧 𝑈
1 𝑑𝐹𝐴
− + 𝑟𝐴 = 0
𝐴𝑐 𝑑𝑧
𝐴𝑐 𝑧 𝑧 + ∆𝑧
𝐹𝐴 : Molar Flow Rate of Species A
𝐹𝐴 = 𝐴𝑐 𝑊𝐴
𝑊𝐴 : Molar Flux of Species A 𝑫 𝒅𝟐 𝑪𝑨 𝒅𝑪𝑨 𝒓𝑨
𝑊𝐴 = 𝑀𝑜𝑙𝑒𝑐𝑢𝑙𝑎𝑟 𝐷𝑖𝑓𝑓𝑢𝑠𝑖𝑜𝑛 + 𝐵𝑢𝑙𝑘 𝐹𝑙𝑜𝑤 − + =𝟎
𝑼 𝒅𝒛𝟐 𝒅𝒛 𝑼
𝑊𝐴 = 𝐽𝐴 + 𝐵𝐴
𝑑𝐶𝐴
𝐽𝐴 = −𝐷 𝐵𝐴 = 𝐶𝐴 𝑈
𝑑𝑧
𝑑𝐶𝐴
𝐹𝐴 = 𝐴𝑐 −𝐷 + 𝐶𝐴 𝑈
𝑑𝑧
MODELS FOR NON-IDEAL REACTORS 30
𝐷 𝑑 2 𝐶𝐴 𝑑𝐶𝐴 𝑘𝐶𝐴
2
− + =0 𝒓𝑨 = 𝒌𝑪𝑨
𝑈 𝑑𝑧 𝑑𝑧 𝑈 2
𝐷 𝑑 𝜓 𝑑𝜓
𝐶𝐴 𝑧 2
− − 𝑘𝜏𝜓 = 0
𝜓= 𝜆= 𝑈𝐿 𝑑𝜆 𝑑𝜆
𝐶𝐴0 𝐿
1 𝑑 2 𝜓 𝑑𝜓
− − 𝐷𝑎𝜓 = 0
𝑃𝑒𝑟 𝑑𝜆2 𝑑𝜆
Damköhler number
𝐷𝑎 = 𝑘𝜏
𝑅𝑎𝑡𝑒 𝑜𝑓 𝑡𝑟𝑎𝑛𝑠𝑝𝑜𝑟𝑡 𝑜𝑓 𝐴 𝑏𝑦 𝑟𝑒𝑎𝑐𝑡𝑖𝑜𝑛
𝐷𝑎 =
𝑅𝑎𝑡𝑒 𝑜𝑓 𝑡𝑟𝑎𝑛𝑠𝑝𝑜𝑟𝑡 𝑜𝑓 𝐴 𝑏𝑦 𝑐𝑜𝑛𝑣𝑒𝑐𝑡𝑖𝑜𝑛
Solution for differential equation (Closed-Closed System)
𝑷𝒆𝒓
𝑪𝑨𝑳 𝟒𝒒𝑬𝒙𝒑
𝝍𝑳 = = 𝟏 − 𝑿𝑨 = 𝟐
𝑪𝑨𝟎 𝒒𝑷𝒆𝒓 𝒒𝑷𝒆𝒓
𝟏 + 𝒒 𝟐 𝑬𝒙𝒑 − 𝟏 − 𝒒 𝟐 𝑬𝒙𝒑 −
𝟐 𝟐
𝒒 = 𝟏 + 𝟒𝑫𝒂/𝑷𝒆𝒓
For reactions other than first order kinetics, iterative solution is required.

MODELS FOR NON-IDEAL REACTORS 31

 Conversion in a dispersed plug flow reactor
𝐴 → 𝐵 𝑟𝐴 = 𝑘𝐶𝐴

MODELS FOR NON-IDEAL REACTORS 32

 Conversion in a dispersed plug flow reactor
2𝐴 → 𝐵 𝑟𝐴 = 𝑘𝐶𝐴 2 𝑜𝑟 𝐴 + 𝐵 → 𝐶 𝑟𝐴 = 𝑘𝐶𝐴 𝐶𝐵 , 𝐶𝐴0 = 𝐶𝐵0

MODELS FOR NON-IDEAL REACTORS 33

Physical Interpretation of Dispersion Coefficient
REACTOR PECLET NUMBER
𝑈𝐿
𝑃𝑒𝑟 =
𝐷
𝑫
= 𝑰𝒏𝒕𝒆𝒏𝒔𝒊𝒕𝒚 𝒐𝒇 𝑫𝒊𝒔𝒑𝒆𝒓𝒔𝒊𝒐𝒏 (𝑮𝒆𝒐𝒎𝒆𝒕𝒓𝒊𝒄 𝑭𝒂𝒄𝒕𝒐𝒓)
𝑼𝑳
𝐷 𝐷 𝑑
=
𝑈𝐿 𝑈𝑑 𝐿
𝑑: Characteristic Length (𝑑𝑡 or 𝑑𝑝 )
𝐷
= 𝑓 𝑓𝑙𝑢𝑖𝑑 𝑝𝑟𝑜𝑝𝑒𝑟𝑡𝑖𝑒𝑠 𝑓𝑙𝑜𝑤 𝑑𝑦𝑛𝑎𝑚𝑖𝑐𝑠
𝑈𝑑
𝐷
= 𝑓 𝑆𝑐ℎ𝑚𝑖𝑑𝑡 𝑁𝑢𝑚𝑏𝑒𝑟 𝑅𝑒𝑦𝑛𝑜𝑙𝑑𝑠 𝑁𝑢𝑚𝑏𝑒𝑟
𝑈𝑑

𝜐 𝑑𝑡 𝑈 𝜐: Kinematic viscosity
𝑆𝑐 = 𝑅𝑒 =
𝐷𝐴𝐵 𝜐 𝐷𝐴𝐵 : Effective Diffusivity

MODELS FOR NON-IDEAL REACTORS 34

 Dispersion of Fluid Flowing in Pipes Dispersion of Fluid Flowing in Packed Beds

MODELS FOR NON-IDEAL REACTORS 35

How to predict conversion using Dispersion Model?
a) Evaluate RTD parameters: 𝑡𝑚 and 𝜎 2
b) Evaluate Reaction kinetics: 𝑘, 𝐶𝐴0
c) Evaluate reactor parameters: 𝑣, 𝑉
d) Calculate Dispersion Coefficient.
𝜎2 2 2 Correlations for 𝐷 can also be
𝜎𝜃 2 = 2= − 1 − 𝑒 −𝑃𝑒𝑟
𝑡𝑚 𝑃𝑒𝑟 𝑃𝑒𝑟 2 used.

e) Predict conversion in real reactor (first order reaction).

𝑃𝑒𝑟
𝐶𝐴𝐿 4𝑞𝐸𝑥𝑝
𝜓𝐿 = = 1 − 𝑋𝐴 = 2
𝐶𝐴0 𝑞𝑃𝑒𝑟 𝑞𝑃𝑒𝑟
1 + 𝑞 2 𝐸𝑥𝑝 − 1 − 𝑞 2 𝐸𝑥𝑝 −
2 2
𝑞 = 1 + 4𝐷𝑎/𝑃𝑒𝑟
𝐷𝑎 = 𝑘𝜏 For reaction order 𝑛 > 1 or
𝑛 < 1 numerical solution is
𝑃𝑒𝑟 = 𝑈𝐿Τ𝐷 required.
MODELS FOR NON-IDEAL REACTORS 36
EXAMPLE 3
A first order liquid decomposition reaction is carried out in a plug flow reactor. Pulse tracer tests on the
reactor showed deviations from ideality. Find the fraction of reactant unconverted in the real reactor
using Dispersion Model.
Additional Data
Reaction Kinetics: −𝑟𝐴 = 𝑘𝐶𝐴 , 𝑘 = 0.307 𝑚𝑖𝑛−1
RTD:

Time 𝑡, 𝑚𝑖𝑛 0 5 10 15 20 25 30 35
𝐸 𝑡 , 𝑚𝑖𝑛−1 0 0.03 0.05 0.05 0.04 0.02 0.01 0

MODELS FOR NON-IDEAL REACTORS 37

a) RTD parameters
(i) Mean Residence Time (ii) Variance
∞
Using, 𝑡𝑚 =
∞
‫׬‬0 𝑡𝐸 𝑡 𝑑𝑡 Using, 𝜎 2 = ‫׬‬0 𝑡 − 𝑡𝑚 2 𝐸 𝑡 𝑑𝑡
𝑡𝑚 = 15 𝑚𝑖𝑛 𝜎 2 = 47.5 𝑚𝑖𝑛2

b) Reaction kinetics c) Reactor parameters

−𝑟𝐴 = (0.307 𝑚𝑖𝑛−1 )𝐶𝐴 Mean residence time is sufficient
d) Dispersion Coefficient
𝜎2 2 2
2 = − 2 1 − 𝑒 −𝑃𝑒𝑟 ⟹ 𝑃𝑒𝑟 = 8.333
𝑡𝑚 𝑃𝑒𝑟 𝑃𝑒𝑟
e) Conversion in dispersed plug flow
𝐷𝑎 = 𝑘𝜏 ⟹ 𝐷𝑎 = 4.605
𝑞= 1 + 4𝐷𝑎/𝑃𝑒𝑟 ⟹ 𝑞 = 1.792
𝑃𝑒𝑟
4𝑞𝐸𝑥𝑝
𝑋𝐴 = 1 − 2 ⟹ 𝑿𝑨 = 𝟗𝟔. 𝟔%
𝑞𝑃𝑒𝑟 𝑞𝑃𝑒𝑟
1 + 𝑞 2 𝐸𝑥𝑝 − 1 − 𝑞 2 𝐸𝑥𝑝 −
2 2
MODELS FOR NON-IDEAL REACTORS 38
 Conversion in a dispersed plug flow reactor

MODELS FOR NON-IDEAL REACTORS 40

Tank-In-Series Model versus Dispersion Model

Both models are used to predict conversions in tubular reactor using variance from RTD.
Tank-In-Series Model: Mathematically easy to use for reaction orders other than one.
Dispersion Model: Easy to apply when correlations for Dispersion coefficient is available. But for higher
order reactions it is difficult to use.
EQUIVALENCY OF THE MODELS
𝑩𝒐 = 𝟐 𝑵 − 𝟏
𝐵𝑜: Bodenstein Number, 𝑈𝐿Τ𝐷
𝑁: No of Ideal CSTRs
What is the usefulness of this relation?

MODELS FOR NON-IDEAL REACTORS 41

TWO-PARAMETER MODEL: COMPARTMENT MODEL

𝑣, 𝐶𝐴
𝑽
𝑣, 𝐶𝐴0

➢ Real reactor modeled as combination of ideal reactors.

➢ Two parameters are defined to represent the deviations in the model.
➢ Parameters are evaluated from RTD experiments.
➢ Conversions are predicted by solving ideal reactor performance equations and mole balances.
How to suggest the combination of ideal reactors for a real reactor?

MODELS FOR NON-IDEAL REACTORS 42

How to Suggest Ideal Reactor Combinations?
a) Analysis of Real Reactor Volume and Volumetric Flow Rate.
Plug Flow Region 𝑉𝑝
Active Volume 𝑉𝑎 = 𝑉𝑝 + 𝑉𝑚
Total Volume 𝑉 Mixed Flow Region 𝑉𝑚
Stagnant Region

Active Flow 𝑣𝑎
Volumetric Flow Rate 𝑣 Bypass Flow 𝑣𝑝
Recycle Flow 𝑣𝑟
b) Compare E-curve of real vessel with theoretical curves for combination of ideal reactors and choose
the best fit.
c) Evaluate model parameters (Fractional Flow rate, Fractional Volume)
d) Use ideal reactor performance equation and mole balance to predict the exit conversion.

MODELS FOR NON-IDEAL REACTORS 43

Compartment Models

Ideal PFR Ideal PFR with stagnant region Ideal PFR with bypass

𝑡𝑚 < 𝜏 𝛿(𝑡 = 0)
𝑉𝑎 = 𝑡𝑚 𝑣 𝑣𝑏 = 𝐴𝑟𝑒𝑎 × 𝑣

MODELS FOR NON-IDEAL REACTORS 45

Compartment Models

Ideal CSTR Ideal CSTR with stagnant region Ideal CSTR with bypass

𝑡𝑚 < 𝜏 𝛿(𝑡 = 0)
𝑉𝑎 = 𝑡𝑚 𝑣 𝑣𝑏 = 𝐴𝑟𝑒𝑎 × 𝑣
MODELS FOR NON-IDEAL REACTORS 46
Compartment Models

MODELS FOR NON-IDEAL REACTORS 47

EXAMPLE 4
A gas phase reaction is to be carried out in a plug flow reactor (𝑉 = 1 𝑚3 and 𝑣0 = 1 𝑚3 Τℎ𝑟) using
pure feed A (𝐶𝐴0 = 2 𝑚𝑜𝑙/𝐿).
Reaction Kinetics: 𝐴 → 𝑅 𝑘 = 0.5 𝑚𝑜𝑙 Τ𝐿 − ℎ𝑟
12.5 0 ≤ 𝑡 < 1 ℎ𝑟
Step Tracer Data: 𝐶𝑜𝑢𝑡 (𝑚𝑔Τ𝐿) =
50 𝑡 ≥ 1 ℎ𝑟
Using the kinetics and macromixing information predict the exit conversion using two parameter model.
SOLUTION
a) Plot RTD
Exit Tracer:
12.5 0 ≤ 𝑡 < 1 ℎ𝑟
𝐶𝑜𝑢𝑡 (𝑚𝑔Τ𝐿) =
50 𝑡 ≥ 1 ℎ𝑟

MODELS FOR NON-IDEAL REACTORS 48

Cumulative Distribution
𝐶𝑜𝑢𝑡
𝐹 𝑡 =
𝐶0 𝑠𝑡𝑒𝑝

𝐶𝑇 = 𝐶0 when 𝑡 ⟶ ∞
𝐶0 = 50 𝑚𝑔/𝐿
0.25 0 ≤ 𝑡 < 1 ℎ𝑟
𝐹(𝑡) =
1 𝑡 ≥ 1 ℎ𝑟
Exit Age Distribution 𝑨𝒓𝒆𝒂 = 𝟑Τ𝟒
𝐸 𝑡 = 𝑑𝐹 Τ𝑑𝑡
1 3
𝐸 𝑡 = 𝛿 𝑡 − 0 + 𝛿 𝑡 − 1 , ℎ𝑟 −1
4 4
𝑨𝒓𝒆𝒂 = 𝟏Τ𝟒
b) RTD parameter
∞
Mean residence time, 𝑡𝑚 = ‫׬‬0 𝑡𝐸 𝑡 𝑑𝑡
1 ∞ 3 ∞
𝑡𝑚 = ‫׬‬0 𝑡𝛿 𝑡 − 0 𝑑𝑡 + ‫׬‬0 𝑡𝛿 𝑡 − 1 𝑑𝑡 = 0.75 ℎ𝑟
4 4

MODELS FOR NON-IDEAL REACTORS 49

c) Analysis of RTD 𝑨𝒓𝒆𝒂 = 𝟑Τ𝟒

Pulse Output at 𝑡 = 0 ℎ𝑟: Bypass

Pulse Output at 𝑡 = 1 ℎ𝑟: Ideal PFR
𝑉 𝑨𝒓𝒆𝒂 = 𝟏Τ𝟒
Real reactor space time: 𝜏 = = 1 ℎ𝑟
𝑣𝑜

𝑡𝑚 < 𝜏: Dead Volume

Real Reactor

Compartment Model

MODELS FOR NON-IDEAL REACTORS 50

d) Model Parameters
Defining model parameters,
Fractional Volume: 𝛼 = 𝑉𝑝 Τ𝑉 and Fractional Flow: 𝛽 = 𝑣𝑝 Τ𝑣0
From RTD,
Area of pulse output at 𝑡 = 1 ℎ𝑟 ⟹ Fraction of fluid entering ideal PFR
⟹ 𝛽 = 3/4 (or) 𝑣𝑝 = 0.75 𝑚3 /ℎ𝑟
Bypass flow rate: 𝑣𝑏 = 𝑣 − 𝑣𝑝 ⟹ 𝑣𝑏 = 0.25 𝑚3 /ℎ𝑟
From RTD,
Time at which delayed pulse occurs ⟹ Space time of Ideal PFR ⟹ 𝜏𝑝 = 1 ℎ𝑟
Active Volume: 𝑉𝑝 = 𝜏𝑝 𝑣𝑝 ⟹ 𝑉𝑝 = 0.75 𝑚3 (or) 𝛼 = 3/4
Dead Volume: 𝑉𝐷 = 𝑉 − 𝑉𝑝 ⟹ 𝑉𝐷 = 0.25 𝑚3
MODELS FOR NON-IDEAL REACTORS 51
e) Exit conversion 𝐶𝐴𝑜

Ideal PFR 𝐶𝐴𝑜 𝐶𝐴𝑝

𝐶𝐴𝑜 𝐶𝐴 =?
𝑉𝑝 𝑋 𝑑𝑋𝐴
Design Equation: = ‫׬‬0 𝐴𝑝
𝐹𝐴0 −𝑟𝐴

Reaction kinetics: −𝑟𝐴 = 𝑘

Solving, 𝐶𝐴𝑝 = 𝐶𝐴𝑜 1 − 𝑘𝜏𝑝 Τ𝐶𝐴0 = 1.5 𝑚𝑜𝑙/𝐿

Exit Concentration

𝐶𝐴0 𝑣𝑏 + 𝐶𝐴𝑝 𝑣𝑝
𝐶𝐴 =
𝑣𝑜
𝐶𝐴 = 1.625 𝑚𝑜𝑙/𝐿
𝑿𝑨 = 𝟏𝟖. 𝟕𝟓 %

MODELS FOR NON-IDEAL REACTORS 52

SUMMARY
◼ Macromixing (RTD)
❑ Describes how long a fluid element stays in a reactor
❑ It is unique for a reactor but does not define specific flow pattern
◼ Micromixing
❑ Describes exchange of matter between fluid elements of different age.
❑ Complete Segregation: Fluid elements of same age move together and do not interact with fluid
elements of different age (Macrofluid).
❑ Complete Micromixing: Complete mixing of fluid elements and are free to move anywhere
(Microfluid).
◼ Segregation Model: Predicts conversion when reacting mixture behave as a macrofluid.
∞
𝑋ത = න 𝑋 𝑡 𝐸 𝑡 𝑑𝑡
0
Reaction Kinetics and RTD is sufficient
MODELS FOR NON-IDEAL REACTORS 53
SUMMARY
◼ Tank-In-Series Model
❑ Models real reactor as a series of identically sized ideal CSTRs.
❑ Number of CSTRs is given by RTD.
𝜏 2 𝑡𝑚 2
𝑁= 2= 2
𝜎 𝜎
❑ Conversion predicted from Ideal CSTR in series performance equation.
𝐶𝐴0 −𝐶𝐴1
Reactor 1: 𝜏𝑖 =
−𝑟𝐴1
𝐶𝐴1 −𝐶𝐴2
Reactor 2: 𝜏𝑖 =
−𝑟𝐴2
𝜏𝑖 = 𝑉 Τ𝑁𝑣
(or)
𝐶𝐴(𝑁−1) −𝐶𝐴𝑁 𝜏𝑖 = 𝑡𝑚 Τ𝑁
Reactor 𝑁: 𝜏𝑖 =
−𝑟𝐴𝑁

1
First order reaction, 𝑋 = 1 −
1+𝜏𝑖 𝑘 𝑁
MODELS FOR NON-IDEAL REACTORS 54
SUMMARY
◼ Dispersion Model
❑ Models real reactor as a dispersed plug flow reactor characterized by dispersion coefficient.
❑ Dispersion coefficient is estimated from RTD or correlations.
𝜎2 2 2
𝜎𝜃 2 = 2= − 1 − 𝑒 −𝑃𝑒𝑟
𝑡𝑚 𝑃𝑒𝑟 𝑃𝑒𝑟 2
❑ Conversion is predicted by solving mole balance for a dispersed plug flow reactor.
First order reaction,
𝑃𝑒𝑟
4𝑞𝐸𝑥𝑝
𝑋𝐴 = 1 − 2
𝑞𝑃𝑒𝑟 𝑞𝑃𝑒𝑟
1 + 𝑞 2 𝐸𝑥𝑝 − 1 − 𝑞 2 𝐸𝑥𝑝 −
2 2
𝑞= 1 + 4𝐷𝑎/𝑃𝑒𝑟 𝐷𝑎 = 𝑘𝜏 𝑃𝑒𝑟 = 𝑈𝐿Τ𝐷

MODELS FOR NON-IDEAL REACTORS 55

SUMMARY
◼ Compartment Model
❑ Models real reactor as a combination of ideal reactors
❑ Model parameters are estimated from RTD by comparing E-curves or analyzing tracer output
data.
❑ Conversion is predicted by solving ideal reactor performance equation and mole balance.

MODELS FOR NON-IDEAL REACTORS 56