You are on page 1of 96

Dr.

Shaker Haji, University of Bahrain 4/30/2013

CHEMICAL REACTOR DESIGN


CHENG 323 Dr Shaker Haji

Lecture 20

Prepared by: Dr Shaker Haji - CHENG 323

CHENG 323 1
Dr. Shaker Haji, University of Bahrain 4/30/2013

4 Isothermal Reactor Design

In this Chapter

Chapters 1 & 2 discussed mole balances on reactors & the


manipulation of these balances to predict reactor sizes.
Chapter 3 discussed reactions, its rates, & stoichiometry.
In Chapter 4, we combine reactions and reactors to design various
types of reactors.
We divide the chapter into two parts: Part 1 "Mole Balances in
Terms of Conversion," and Part 2 "Mole Balances in Terms of
Concentrations & Molar Flow Rates"
In Part 1, we will concentrate on CSTRs & PFRs where X is the
preferred measure of a reaction's progress for single reactions.
We will also discuss batch reactors.

CHENG 323 2
Dr. Shaker Haji, University of Bahrain 4/30/2013

In this Chapter

In Part 2, we will analyze membrane reactors, the startup of a


CSTR, semibatch reactors, & also liquid-phase rxns in batch reactor
which are most easily analyzed using Cj & Fj as the variables rather
than X.
We will again use mole balances in terms of these variables (Cj &
Fj) for multiple reactors (chap. 6 & end of 8).
Chap. 4 focuses attention on rxns that are operated isothermally.
In this chap., we achieve our learning objective through designing a
multi-reactor chemical plant that produces 200 million lb/y of
ethylene glycol from a feedstock of ethane.

4 Isothermal Reactor Design

CHENG 323 3
Dr. Shaker Haji, University of Bahrain 4/30/2013

4.1 Design Structure for Isothermal Reactors

One of the primary goals of this chap. is to solve chemical


reaction engineering (CRE) problems by using logic rather
than memorizing which equation applies where.
This logical procedure is summarized in the following five
steps:
1. Choosing the applicable mole balance.
2. Writing the correct rate law.
3. Utilizing the stoichiometry.
4. Combining all of the above.
5. Evaluating the parameters & equations.

4.1 Design Structure for Isothermal Reactors

We now will apply this procedure/algorithm to a specific


situation.
In the following Figure, we will see how the
procedure/algorithm is used to formulate the equation to
calculate the PFR volume for a first-order gas phase
reaction.

CHENG 323 4
Dr. Shaker Haji, University of Bahrain 4/30/2013

PFR volume for a first-


order gas phase reaction
The algorithm for the CRE
design problem is:
1. Mole balance, choose
species A reacting in a PFR.
2. Rate law, choose the
irreversible first-order rxn.
3. Stoichiometry, choose the
gas-phase concentration.
4. Combine Steps 1,2, & 3 to
arrive at Equation A.
5. Evaluate. The resultant
equation can be evaluated
by different methods.

4.1 Design Structure for Isothermal Reactors

The algorithm for the CRE


design problem is:
1. Mole balance
2. Rate law
3. Stoichiometry
4. Combine
5. Evaluate. The combine step can be evaluated either:
a. Analytically (Appendix A1)
b. Graphically (Counting the unit areas under the curve)
c. Numerically (Appendix A4), or
d. Using software (Polymath, MATLAB)

CHENG 323 5
Dr. Shaker Haji, University of Bahrain 4/30/2013

4.1 Design Structure for Isothermal Reactors

Notes:
Unless the parameter values are zero or facilitates
manipulation, we typically don't substitute numerical
values for parameters in the combine step until the very
end.
In the majority of situations, analytical solutions to the
ordinary differential equations appearing in the
combine step are not possible.

4.2 Batch Reactors for Labs & Plants

One of the jobs in which chemical engineers are Involved is


the scale-up of laboratory experiments to pilot-plant
operation or to full-scale production.
In this section, we discuss more about the use of batch reactors
in laboratories & plants, and conduct calculations for cases
where X is the preferred variable.
In Part 2 of this Chap. we show how to analyze a laboratory-
scale batch reactor to determine the specific rxn rate, k, which
can be used in the design of a full-scale flow reactor.
Here, calculations are conducted for cases where Cj is the
preferred variable (e.g. liquid-phase rxn).

CHENG 323 6
Dr. Shaker Haji, University of Bahrain 4/30/2013

4.2.1 Batch Operation

Batch reactors are usually analyzed with a constant-volume


assumption (V=V0), why?
Liquid-phase rxn with negligible density change.
Gas-phase in a constant-volume, sealed reactor.
Write a mole balance for A in a constant-volume batch reactor.

Manipulate the obtained equation to obtain the time necessary


to achieve a given conversion X for the irreversible second-
order reaction A B.

4.2.1 Batch Operation

The time calculated from the batch reactor design equation is the
reaction time (i.e., tR) needed to achieve a given X for a given rxn in
a batch reactor with a given V.
It could be in the order of seconds, minutes, or hours.
The total cycle time in any batch operation is considerably longer
than the rxn time, tR. Why?
The total time to process a batch includes the time necessary to:
fill the reactor

heat and/or pressurize the reactor

carry out the reaction

clean the reactor between batches.

CHENG 323 7
Dr. Shaker Haji, University of Bahrain 4/30/2013

4.2.1 Batch Operation

In some cases, the rxn time may be only a small fraction of the total
cycle time.
Typical cycle times for a batch polymerization process vary from 3 to
6 hours excluding the rxn time; the rxn time may vary between 5 and
60 hours.
As mentioned earlier, the rxn time in a batch reactor could be in the
order of seconds, minutes, or hours.
Usually, flow reactors are used for rxns with characteristic rxn times,
tR, of minutes or less.
However, there are cases where flow reactor are used to carry out
rxns with longer tR (e.g. heterogeneous rxns with mass transfer
limitations or of refractory reactants).

In the following examples, we will describe the various reactors needed


to produce 200 million lb/year of ethylene glycol from a feedstock
of ethane. (lost)

Example 4-3: PFR


(lost)

Example 4-6: PBR

(lost)

Example 4-2: CSTR

Example 4-1:
Batch
(kinetics)

CHENG 323 8
Dr. Shaker Haji, University of Bahrain 4/30/2013

Timeout with Damkhler number

Gerhard Damkhler (1908-1944) was a professor at the


University of Gttingen, Germany, between 1936 and 1944.
He was one of the founders of Chemical Reaction Engineering.
Damkhler number, which is a dimensionless number, was named
after him.
Damkhler number (Da), is a quick estimate of the degree of
conversion that can be achieved in continuous flow reactors.

The Damkhler number, is the ratio of the rate of rxn of A to the


rate of convective transport of A at the entrance to the reactor.
Reaction Rate of A
Da
Convection Rate of A reactor enterance

Rate of rxn at enterance -rA0V


Da
Entering flow rate of A FA0

The Da for a first-order irreversible rxn is: Da =

For a second-order irreversible rxn, the Da =

Or of nth-order irreversible rxn, Da = k(CA0)n -1

CHENG 323 9
Dr. Shaker Haji, University of Bahrain 4/30/2013

Time-out with Damkhler number

It is important to know what values of the Damkhler number,


Da, give high and low conversion in continuous-flow reactors.

The rule of thumb is:

Week 6 - Lecture 21

Prepared by: Dr Shaker Haji - CHENG 323

CHENG 323 10
Dr. Shaker Haji, University of Bahrain 4/30/2013

Design Problem 1:
(lost)

PFR
(lost)

(lost)

4.4 Tubular Reactors

What is the tubular reactor typically used for?


gas-phase rxn.
Under which assumptions we can model the tubular reactor as a
plug-flow reactor?
no axial dispersion
no radial gradients in either T, velocity, or concentration (turbulent flow)

Laminar reactors are discussed in Chap. 13 and dispersion effects in


Chap. 14.

CHENG 323 11
Dr. Shaker Haji, University of Bahrain 4/30/2013

4.4 Tubular Reactors

Write the differential form of the PFR design equation in terms of X


dX
FA0 - rA
dV

In the absence of P drop and heat exchanging, we evaluate the


integral form of the PFR design equation to obtain V:
X
F dX
dV A0 dX V FA0
- rA 0
-rA

Consider the second order rxn, A B, with -rA kC A2 to take


place as a:
1st case: liquid-phase rxn
2nd case: gas-phase rxn

FA0
Liquid Phase case: =0 -rA kC A2 dV
-rA
dX

Evaluate the reactor volume for an isothermal operation.

Find the X = f(Da2), where Da2 is the Da for a second-order rxn

CHENG 323 12
Dr. Shaker Haji, University of Bahrain 4/30/2013

FA0
Gas Phase case: 0 -rA kC A2 dV
-rA
dX

Evaluate the reactor volume for an isothermal operation with


negligible pressure drop for unknown values of yA0

Plot X along the length (volume) of the reactor for different values
of . Assume:
yA0 = 1
0/kCA0 = 2 L

Gas Phase case: 0


- rA kCA2
yA0 1
yA0

X increases as decreases, why?

CHENG 323 13
Dr. Shaker Haji, University of Bahrain 4/30/2013

Gas Phase case:


-rA = kCA2, T=T0, P=P0
For a given V, higher X is
achieved for lower , why?
= 0 (1+ X)
= A.u

When < 0 ( < 0),


decreases as rxn
progresses, thus,
residence time
increases and ,
therefore, X
increases

Gas Phase case:


-rA = kCA2, T=T0, P=P0
= 0 (1+ X)
CA = FA /

CHENG 323 14
Dr. Shaker Haji, University of Bahrain 4/30/2013

Example 4-3. Producing 300 MM Pounds per Year of ethylene


in a PFR: Design of a Full-Scale Tubular Reactor

Ethylene ranks 4th in the United States in total pounds of chemicals


produced each year, and it is the number one organic chemical
produced each year. Over 50 billion pounds were produced in 2000,
and it sold for $0.27 per pound (225 fils/kg).

PFR
(lost)

Example 4-3. Producing 300 MM Pounds per Year of ethylene


in a PFR: Design of a Full-Scale Tubular Reactor

Approximately 90% of ethylene is used to produce the following four


products:
Ethylene Glycol
Vinyl Chloride Ethylene Dichloride Ethylene Oxide

Polyethylene Ethylbenzene Styrene

Main industrial uses of ethylene [Ref: wikipedia.org].

CHENG 323 15
Dr. Shaker Haji, University of Bahrain 4/30/2013

Example 4-3. Producing 300 MM Pounds per Year of ethylene


in a PFR: Design of a Full-Scale Tubular Reactor

Determine the PFR volume necessary to produce 300 million


pounds of ethylene a year from cracking a feed stream of pure
ethane. The rxn is irreversible and fallows an elementary rate
law.
We want to achieve 8O% conversion of ethane, operating the
reactor isothermally at 1100 K at a pressure of 6 atm.
Data from literature:
k = 0.072 s-1 at 1000 K.
Activation energy, Ea = 82 kcal/gmol

Solution:
m B 300 106 lb/y, X 0.8, T 1100K, P 6 atm
Data from literature: k = 0.072 s-1 at 1000 K, Ea = 82 kcal/gmol
The algorithm for the CRE design problem is:
1. Mole balance
2. Rate law
3. Stoichiometry
4. Combine
5. Evaluate

Answer: V = 80.7 ft3

CHENG 323 16
Dr. Shaker Haji, University of Bahrain 4/30/2013

Answer: V = 80.7 ft3

If we use 2-in. schedule 80 pipe


(AC=0.0205 ft2):
L = 80.7/0.0205 = 4,000 ft
Is this reasonable configuration?
Suggested length 40 ft.
No. of parallel tubes needed =
4,000/40 = 100 tubes

Fig. Representative Pipe Sizes from


the Nominal Pipe Size (NPS) system.
Ref: James O. Wilkes, Fluid Mechanics for Chemical
Engineers, Prentice Hall PTR, NJ, 1999.

The concentration and conversion profiles down any one of the pipes
are shown above.

CHENG 323 17
Dr. Shaker Haji, University of Bahrain 4/30/2013

Submit Homework
#5
one week from today

Design Problem 2:
(lost)

(lost)

PBR

(lost)

CHENG 323 18
Dr. Shaker Haji, University of Bahrain 4/30/2013

4.5 Pressure Drop in Reactors

For the purpose of sizing reactors:


Do we have to account for the effects of P drop in liquid-phase rxns?

In liquid-phase rxns, the Ci is insignificantly affected by even relatively


large changes in the total pressure.
Consequently, we can totally ignore the effect of P drop on the rate of
reaction when sizing liquid-phase chemical reactors.
Do we have to account for the effects of P drop in gas-phase rxns?

In gas-phase reactions, the Ci are proportional to the total pressure.

Consequently, proper accounting for the effects of P drop on the rxn


system can, in many instances, be a key factor in the success or failure
of the reactor design & operation.
This fact is especially true in micro-reactors packed with solid catalyst.

Here the channels are so small that P drop can limit the throughput and
conversion for gas-phase reactions.

4.5.1 Pressure Drop and the Rate Law

Consider the gas-phase second order rxn that is to be carried out


in an isothermal packed-bed reactor:
2A B + C
Write the differential form of the design equation in terms of
catalyst weight. dX
FA0 - rA'
dW

The rate law is: -r ' kC A2


A

From stoichiometry, for gas-phase flow reactive systems, CA is:

CHENG 323 19
Dr. Shaker Haji, University of Bahrain 4/30/2013

dX
FA0 - rA' -rA' kC A2
dW

What is the effect of the pressure drop on the reaction rate?


The larger the pressure drop from frictional losses, the smaller the
reaction rate!
Combine the above equations:

dX
Which implies that for isothermal operation: F1 (X , P )
dW

In order to find X = f(W), we first need to find P = f(W)


dP
In other words, we need to find: F2 (X , P )
dW

4.5.2 Flow Through a Packed Bed

The majority of gas-phase rxns are catalyzed by passing the


reactants through a packed bed of catalyst particles.

The equation used most to calculate pressure drop in a packed


porous bed is the Ergun equation:

Term 1 is dominant for laminar flow, while Term 2 is dominant for


turbulent flow.

CHENG 323 20
Dr. Shaker Haji, University of Bahrain 4/30/2013

Term 1 is dominant for laminar flow, while Term 2 is dominant for turbulent flow.

Fig. Ergun Equation.


Ref: James O. Wilkes, Fluid Mechanics for
Chemical Engineers, Prentice Hall PTR, NJ, 1999.

= equivalent spherical diameter of the particle defined by

CHENG 323 21
Dr. Shaker Haji, University of Bahrain 4/30/2013

In calculating the P drop, the only parameter that varies


significantly with P on the RHS of the Ergun Eq. is:
the gas density, .

Find = f(P).
From the continuity eq. From ideal gas law
(conservation of mass)
FT RT P
m m0 FT 0 T 0 P
0 FT 0 RT 0 P 0 0
00 FT T P0

0 FT 0 T 0 P
Substitute in Ergun Eq.
0 0 FT T P0

Simplifying yields

where 0 is a constant that depends only on the properties of the


packed bed and the entrance conditions.

But are we interested in dP/dZ ? Or are we interested in dP/dW?


Then, find the relationship between dW and dZ.
W V b ( b is bulk density =m catalyst / V bed )

W V (1 - ) c ( c is density of solid catalyst particles)

CHENG 323 22
Dr. Shaker Haji, University of Bahrain 4/30/2013

W V (1 - ) c ( c is density of solid catalyst particles)

W Ac Z (1 - ) c or dW Ac (1 - ) c dZ

Further simplifying yields

where

Let y=P/P0 , then

When is this Eq. used?


When there is P in multiple reactions or membrane reactor
systems.
However, for single rxns in packed-bed reactors, it is more
convenient to express the Ergun Eq. in terms of X.
How shall we do that?
From the stoichiometric table:
Therefore,

CHENG 323 23
Dr. Shaker Haji, University of Bahrain 4/30/2013

Upon substitution:

Notes:
when is -ve, the P will be less than that for =0.

when is +ve, the P will be greater than when =0.

For isothermal operation, the above Eq. is only f(X,P):


dP
F2 (X , P )
dW
From the combined mole balance, rate law, and
stoichiometry: dX
F1 (X , P )
dW

dX dP
F1 (X , P ) F2 (X , P )
dW dW

So, we have two coupled first-order differential equations


that must be solved simultaneously.
How do we solve them?
A variety of software packages and numerical integration
schemes are available for this purpose.

Conclusion: Where y=P/P0

CHENG 323 24
Dr. Shaker Haji, University of Bahrain 4/30/2013

Analytical Solution

If = 0, or if we can neglect X with respect to 1.0 (i.e. 1X),


we can obtain an analytical solution to the above Eq. for
isothermal operation.
For isothermal operation with = 0, the pressure drop
differential Eq. becomes:

Rearranging:

Taking y inside the derivative:

Integrating from y = 1 (P=P0) at W =0

Be sure not to use this equation if 0, or the rxn is not carried


out isothermally.
This Eq. can be used to substitute for the P in the rate law in which
case the mole balance can be written solely as a function of
conversion and catalyst weight.
dX dX
F1 (X , P ) F1 (X ,W )
dW dW
The resulting equation can readily be solved either analytically or
numerically.
Furthermore, If we wish to express P(Z), we can use the W-Z
relationship, W (1 - ) c Ac Z , and manipulate the above Eq.:

CHENG 323 25
Dr. Shaker Haji, University of Bahrain 4/30/2013

Example 4-4 Calculating the Pressure Drop in a Packed Bed

Plot the pressure drop in a 60 ft length of 1 -in. schedule 40 pipe


packed with catalyst spherical particles. There is a 104.4 lbm/h
of gas passing through the bed.
The temperature is constant along the length of pipe at 260C. The
void fraction is 45% and the properties of the gas are similar to
those of air at this temperature. The entering pressure is 10 atm.
Additional Data:
1 -in. schedule 40 pipe has OD of 1.9 in & wall thickness of 0.145 in
Dp= in = 0.0208 ft
For air at 260oC & 10 atm,

Solution
L = 60 ft
.
m = 104.4 lbm/h T = 260C.
Pipe ID =1.61 in = 0.45 P0 = 10 atm.
Ac = 0.01414 ft2
.
G = m/Ac = 7383.3 lbm/h.ft2
Dp= in = 0.0208 ft gc = 4.17 x 108 lbm.ft/lbf.h2 For air (260oC & 10 atm)

What should be
the units of 0 ?

Turbulent Flow term is dominant

CHENG 323 26
Dr. Shaker Haji, University of Bahrain 4/30/2013

P0
P P0 1 - 0.0155 z 0
0.5

Example 4-4 Calculating the Pressure Drop in a Packed Bed

CHENG 323 27
Dr. Shaker Haji, University of Bahrain 4/30/2013

4.5.4 Analytical Solution for Reaction with Pressure Drop

First, lets examine the effect of the pressure drop on the rxn:

No P

4.5.4 Analytical Solution for Reaction with Pressure Drop

Second, lets discuss how the pressure drop affect our CRE algorithm.
Consider the 2nd order isothermal The algorithm for the CRE
design problem is:
gas phase rxn: A B, occurring in a PBR.
1. Mole balance, Design Eq.
Design Equation: 2. Rate Law
3. Stoichiometry (to find Ci)
Rate Law:
4. Combine (all of the above)
5. Evaluate
Stoichiometry:

where y for isothermal & = 0 rxn is

Combine:

CHENG 323 28
Dr. Shaker Haji, University of Bahrain 4/30/2013

4.5.4 Analytical Solution for Reaction with Pressure Drop

Combine:

Evaluate: upon separation of var. & integration

solving for X gives:

solving for W gives:

Isothermal PBR with P for 2nd order rxn with = 0

Example 4-5 Effect of Pressure Drop on the Conversion Profile

Reconsider the packed bed in Example 4-4 for the case where a second-order rxn,
2A B + C, is taking place in 20 meters of a 1 schedule 40 pipe packed with catalyst.
The flow and packed-bed conditions in the example remain the same except that they are
converted to metric units; that is:
P0 = 10 atm = 1013 kPa, T = T0 = 260oC, k = 12 m6/(kmol.kgcat.h)
CA0 = 0.1 kmol/m3, 0 = 7.15 m3/h, P parameter: 0 = 25.8 kPa/m
Reactor length: L = 20 m, Cross-sectional area of the pipe: Ac = 0.0013 m2
Cat. particle size: Dp = 0.006 m, Solid catalyst density: c = 1923 kg/m3

(a) First, calculate X in the absence of pressure drop.


(b) Next, calculate X accounting for pressure drop.
(c) Finally, determine how your answer to (b) would change if the catalyst particle diameter
were doubled. Assume that the specific rxn, k, is unaffected by particle size, an assumption
we know from Chap. 12 is valid only for small panicles.

CHENG 323 29
Dr. Shaker Haji, University of Bahrain 4/30/2013

2A B + C

P0 = 10 atm = 1013 kPa, T = T0 = 260oC, k = 12 m6/(kmol.kgcat.h)

CA0 = 0.1 kmol/m3, 0 = 7.15 m3/h, P parameter: 0 = 25.8 kPa/m

Reactor length: L = 20 m, Cross-sectional area of the pipe: Ac = 0.0013 m2

Cat. particle size: Dp = 0.006 m, Solid catalyst density: c = 1923 kg/m3

(a) First, calculate X in the absence of pressure drop.

Solution

By substitution:
X = 0.82

2A B + C
P0 = 10 atm = 1013 kPa, T = T0 = 260oC, k = 12 m6/(kmol.kgcat.h)
CA0 = 0.1 kmol/m3, 0 = 7.15 m3/h, P parameter: 0 = 25.8 kPa/m
Reactor length: L = 20 m, Cross-sectional area of the pipe: Ac = 0.0013 m2
Cat. particle size: Dp = 0.006 m, Solid catalyst density: c = 1923 kg/m3

(a) First, calculate X in the absence of pressure drop. X = 0.82


(b) Next, calculate X accounting for pressure drop.

Solution

For isothermal PBR with P for 2nd order rxn with = 0 :

By substitution:
X = 0.693

CHENG 323 30
Dr. Shaker Haji, University of Bahrain 4/30/2013

P parameters: 0 = 25.8 kPa/m,

(b) Next, calculate X accounting for pressure drop. X = 0.693


(c) Finally, determine how your answer to (b) would change if the catalyst particle
diameter were doubled. Assume that the rate constant, k, is unaffected by particle
size, an assumption we know from Chap. 12 is valid only for small particles.

Solution

, Since DP,2 = 2DP,1


By substitution:
0,2 = 0,1
X = 0.774
0,2 = 0,1 = 0.0185 kg-1

By increasing the catalyst particle size (doubling: DP,2 = 2DP,1),


conversion (X) increased from 0.69 to 0.77

Why X increased by increasing the particle size?


Increasing the DP decreased the P and thus -rA & X increased.
However, Chapters 10 & 12 explain that when diffusion effects are
important in the catalyst pellet, this increase in X with increasing
particle size will not always be the case.
For larger particles, it takes a longer time for a given number of
reactant and product molecules to diffuse in and out of the catalyst
particle where they undergo reaction.
Consequently, k decreases with increasing
particle size (k~1/DP), which in turn
decreases X.

CHENG 323 31
Dr. Shaker Haji, University of Bahrain 4/30/2013

Insight into the steps involved in a catalytic reaction:


Transport & Kinetic Processes in Catalytic Reactions

external diffusion of A

bulk gas phase


Reactant A

hydrodynamic
boundary
layer

porous catalyst particle

Courtesy of Prof. David Rockstraw, New Mexico State University.

Transport & Kinetic Processes in Catalytic Reactions

external diffusion of A

internal diffusion of A

bulk gas phase

hydrodynamic
boundary
layer

porous catalyst particle

Courtesy of Prof. David Rockstraw, New Mexico State University.

CHENG 323 32
Dr. Shaker Haji, University of Bahrain 4/30/2013

Transport & Kinetic Processes in Catalytic Reactions

external diffusion of A

internal diffusion of A

adsorption of A
A + S AS bulk gas phase

hydrodynamic
boundary
layer

porous catalyst particle

S catalyst adsorption site


Courtesy of Prof. David Rockstraw, New Mexico State University.

Transport & Kinetic Processes in Catalytic Reactions

external diffusion of A

internal diffusion of A reaction of A to B


AS BS
adsorption of A
A + S AS bulk gas phase

hydrodynamic
boundary
layer

porous catalyst particle

Courtesy of Prof. David Rockstraw, New Mexico State University.

CHENG 323 33
Dr. Shaker Haji, University of Bahrain 4/30/2013

Transport & Kinetic Processes in Catalytic Reactions

external diffusion of A

internal diffusion of A reaction of A to B


AS BS
adsorption of A desorption of B
A + S AS bulk gas phase BS B + S

hydrodynamic
boundary
layer

porous catalyst particle

Courtesy of Prof. David Rockstraw, New Mexico State University.

Transport & Kinetic Processes in Catalytic Reactions

external diffusion of A

internal diffusion of A reaction of A to B internal diffusion of B


AS BS
adsorption of A desorption of B
A + S AS bulk gas phase BS B + S

hydrodynamic
boundary
layer

porous catalyst particle

Courtesy of Prof. David Rockstraw, New Mexico State University.

CHENG 323 34
Dr. Shaker Haji, University of Bahrain 4/30/2013

Transport & Kinetic Processes in Catalytic Reactions

external diffusion of A external diffusion of B

internal diffusion of A reaction of A to B internal diffusion of B


AS BS
adsorption of A desorption of B
A + S AS bulk gas phase BS B + S

hydrodynamic
boundary
layer

porous catalyst particle

Courtesy of Prof. David Rockstraw, New Mexico State University.

Transport & Kinetic Processes in Catalytic Reactions

external diffusion of A external diffusion of B

internal diffusion of A reaction of A to B internal diffusion of B


AS BS
adsorption of A desorption of B
A + S AS BS B + S

One of these seven transport and kinetic processes occurs the


slowest. We say that step is rate-limiting
It is necessary to determine the rate-limiting step to analyze
the kinetics.

Courtesy of Prof. David Rockstraw, New Mexico State University.

CHENG 323 35
Dr. Shaker Haji, University of Bahrain 4/30/2013

Insight into the steps involved in a catalytic reaction:


1. Reactants External Diffusion: Mass transfer (diffusion) of reactants from the bulk
fluid to the external surface of the catalyst pellet.
2. Reactants Internal Diffusion: Diffusion of reactants from the pore mouth through
the catalyst pores to the immediate vicinity of the internal catalytic surface.
3. Reactants Adsorption: Adsorption of reactant(s) onto the catalyst surface.
4. Reaction: Reaction on the surface of the catalyst (e.g., A B)
5. Products Desorption: Desorption of the products from the surface.
6. Products Internal Diffusion: Diffusion of
the products from the interior of the pellet
to the pore mouth at the external surface
7. Products External Diffusion: Mass
transfer of the products from the external
pellet surface to the bulk fluid.

Active site

CHENG 323 36
Dr. Shaker Haji, University of Bahrain 4/30/2013

At small particle diameters, k is large, and at its maximum value, but


P is also large, resulting in a low rate of rxn & conversion.

At small particle diameters, k is large, and at its maximum value, but


P is also large, resulting in a low rate of rxn & conversion.
At large particle diameters, the P is small, but so is the overall rate
constant, kr , and the rate of rxn, resulting in low X.
Thus, we see a low X at both large and small particle diameters with
an optimum in between.

CHENG 323 37
Dr. Shaker Haji, University of Bahrain 4/30/2013

Would you do anything just to reduce P?


if P is to be minimized, why not pack the catalyst into a larger
diameter tube to decrease the mass flux, G, thereby reducing P?
There are two reasons for not increasing the tube diameter:
(1) there is an increased chance the gas could channel and bypass
most of the catalyst, resulting in little X.

(2) the ratio of the heat transfer surface area to reactor volume
(catalyst weight) will be decreased, thereby making heat
transfer more difficult for highly exothermic and endothermic
rxns.
We now proceed to combine P with rxn in a PBR when we have
volume change with rxn & therefore cannot obtain an analytical
solution.

Design Problem 2:
(lost)

(lost)

PBR

(lost)

CHENG 323 38
Dr. Shaker Haji, University of Bahrain 4/30/2013

Example 4-6 Calculating X in a Reactor with Pressure Drop

Approximately 7 billion pounds of ethylene oxide were produced in


USA in 1997. The 1997 selling price was $0.58 a pound (480
fils/kg)*, amounting to a commercial value of $4.0 billion. Over
60% of the ethylene oxide produced is used to make ethylene
glycol. The major end uses of ethylene oxide are antifreeze (30%),
polyester (30%), surfactants (10%), and solvents (5%).

*Compare to ethylene price: 225 fils/kg.

We want to calculate the catalyst weight necessary to achieve 60% conversion


when ethylene oxide is to be made by the vapor-phase catalytic oxidation of
ethylene with air.

Ethylene and oxygen are fed in stoichiometric proportions to a PBR operated


isothermally at 260oC. Ethylene is fed at a rate of 0.30 Ibmol/s at a total
feed pressure of 10 atm. It is proposed to use 10 banks of 1-in.-diameter
schedule 40 tubes packed with catalyst with 100 tubes per bank.
Consequently, the ethylene molar flow rate to each tube is to be 3 X 10 -4
Ibmol/s. The properties of the reacting fluid are to be considered identical to
those of air at this T & P. The density of the -in. catalyst particles is 120
Ibm/ft3 and the bed void fraction is 0.45.
The rate law is -rA' k PA1/3 PB2/3 lbmol/(lbmcat h)
with k 0.0141 lbmol/(lbmcat h atm) at 260oC

CHENG 323 39
Dr. Shaker Haji, University of Bahrain 4/30/2013

X=0.6, T = 260oC, P0 = 10 atm, FA0 = 3 X 10-4 Ibmol/s = 1.08 lbmol/h

Reactor: 1-in.-diameter schedule 40 tube packed with catalyst, Ac = 0.01414 ft2

Dp = 0.0208 ft, c = 120 lbm/ft3, = 0.45

-rA' k PA1/3 PB2/3 lbmol/(lbmcat h)

k 0.0141 lbmol/(lbmcat h atm) at 260oC

Solution:
Design Equation:

Rate Law: -rA' kRT C A1/3 C B2/3

Stoichiometry:

Combine:

Combine:

Evaluate:

CHENG 323 40
Dr. Shaker Haji, University of Bahrain 4/30/2013

Combine:

Evaluate:

Yeah ho! We already have 0


(from example 4-4)

From Example 4-4, 0 = 0.0775 atm/ft; & Example 4-5, = 0.0166 lbm-1

Evaluate:

Use any ordinary differential equation


(ODE) solver software packages to solve the
two ODEs simultaneously.

CHENG 323 41
Dr. Shaker Haji, University of Bahrain 4/30/2013

60% conversion is achieved with 44.5-Ib catalyst in each tube.


Its easier to increase the X by one point in the beginning of the reactor
compared to its end.
If we neglect P, 60% X can be achieved with only 35.3 Ib of catalyst.
Therefore, if we had used this catalyst weight (35.5 lb) in our reactor we
would have had insufficient X (only 53%).

4.5.5 Spherical Packed-Bed Reactor

Let's consider carrying out this rxn in a spherical reactor


In a spherical reactor, the Ac varies as we move through the reactor & is
greater than in a normal packed-bed reactor.
Therefore, the superficial velocity (mass flux, G) will
be smaller.
The smaller value of G will give a smaller P and
thus a greater X.
If 40,000 lb of catalyst in the PBR in Example 4-6
had been used in spherical reactor, 67% X would
have been achieved instead of 60%.
The equations for calculating X in spherical reactor
is given in the CD-ROM.

CHENG 323 42
Dr. Shaker Haji, University of Bahrain 4/30/2013

4.5.3 Pressure Drop in Pipes

Which one is more significant, P due to flow through empty


pipes or through packed-beds?
Normally, P for gases flowing through pipes without packing can
be neglected.
For flow in pipes, the pressure drop along the length of the pipe
is given by

where,

What is the friction


factor, f, function of?
Its a function of the
Reynolds number and
pipe roughness.

Fig. Fanning friction factor


for flow in pipes.
Ref: James O. Wilkes, Fluid Mechanics
for Chemical Engineers, Prentice Hall
PTR, NJ, 1999.

CHENG 323 43
Dr. Shaker Haji, University of Bahrain 4/30/2013

Is the mass velocity (mass flux), G, constant along the length of the
pipe?
Yes. Why?
How can I reduce the No. of variables in the above equation?
du G d
By replacing u with G/ and du/dL with - 2
dL dL
By substitution:
2
dP G d 2fG 2
-
dL dL D

The pressure drop along the length of the pipe

The simplest case of P in flow in pipes: FT = FT0 & T = T0

2
F T P dP G d 2fG 2
0 T 0 0 -
FT T P0 dL dL D

For the case of constant T & FT, 0 P P0


By substitution:

Multiplying by dL and integrating with limits P = P0 when L = 0 &


assuming f does not vary, we have

CHENG 323 44
Dr. Shaker Haji, University of Bahrain 4/30/2013

The simplest case of P in flow in pipes: FT = FT0 & T = T0

Recall Example 4-4 with 60 ft PBR where FT = FT0 & T = T0, the
calculated P was 73.5% (P0=10P=2.65 atm).
For the same flow conditions given in Example 4-4 for an unpacked
pipe with a length of up to 1000 ft, the P is less than 10%.
Hence, normally P for gases flowing through pipes without
packing can be neglected.
However, for high volumetric flow rates through microreactors, the
P may be significant.

Week 8 - Lecture 29

Prepared by: Dr Shaker Haji - CHENG 323

CHENG 323 45
Dr. Shaker Haji, University of Bahrain 4/30/2013

Design Problem 3:
(lost)

(lost)

(lost)

CSTR

4.3 Design of Continuous Stirred Tank


Reactors (CSTRs)
What is the CSTR typically used for?
Liquid-phase rxn.
FA0
Write the design equation for CSTR. V X
-rA
exit

Derive the space time, , required to achieve a X in a CSTR.

C A0
X

-rA exit
This equation applies to a single CSTR or to the first reactor
of CSTRs connected in series.

CHENG 323 46
Dr. Shaker Haji, University of Bahrain 4/30/2013

4.3.1 A single CSTR

Consider the elementary liquid-phase rxn: A B; -rA = kCA.


Express X(), for the rxn if it is to be carried out in a single
CSTR:
FA0 C A0
V X X
-rA
exit -rA exit

Then, express X(Da):

Finally, express CA():

This exercise is to show that starting from the design eq. we can
write: V = f(X), X = f(, k, CA0), V = f(Da), & CA = f( , k, CA0)

4.3.2 CSTRs in Series

A liquid-phase, first-order rxn is to be carried out in two CSTRs


placed in series:

Find CA1(Da) &


CA2(Da)

Do this at home; SKIP

CHENG 323 47
Dr. Shaker Haji, University of Bahrain 4/30/2013

CA1():

CA2():

If both reactors are of equal size (1= 2) and operate at the


same T (k1= k2), then

If instead of two CSTRs in series we had n equal-sized CSTRs


connected in series operating at the same T, the CA leaving the
last reactor would be

For n equal-sized CSTRs connected in series operating at the


same T, the CA leaving the last reactor is

Find -rAn for such case.

Find X(Da) in such case.

CHENG 323 48
Dr. Shaker Haji, University of Bahrain 4/30/2013

4.3.2 CSTRs in Series

You know how to


find the X at the
end of each
reactor, right?

Plot X at the end


of each reactor
for different , k,
& CA0 (i.e diff. Da)

Fig. Conversion as a function of the number of CSTRs in series for


different Damkhler numbers for a first-order rxn with constant .

4.3.3 CSTRs in Parallel

X
FA0

VS.
- rAi

- rA

One Large CSTR CSTRs in Parallel

CHENG 323 49
Dr. Shaker Haji, University of Bahrain 4/30/2013

4.3.3 CSTRs in Parallel

We now consider the case in which equal-


sized reactors are placed in parallel:

The feed is distributed equally among each


of the reactors:
Will the X be the same for each reactor?
Will the rA be the same in each reactor?
Yes: X1 =X2 = =Xi =Xn
Yes: rA1 = rA2 = = rAi = rAn
if Vi, FA0i, 0i, CA0i, & Ti are the same

Compare Xi to X & rAi to rA.


Approach 1:
Since the reaction rate is an
intensive property, rAi = rA.
From the design eq.:

or . Compare to

Since rAi = rA, therefore Xi = X

CHENG 323 50
Dr. Shaker Haji, University of Bahrain 4/30/2013

Compare Xi to X & rAi to rA.


Approach 2:
From the design eq.:

or

Just to clarify the point, assume 1st order


liquid phase rxn:
X Xi
V FA0 i FA0
-rA i ki CA0i (1 - X i )
FA0 Xi
V
kCA0 1- Xi

FA0 Xi
V
kCA0 1- Xi
Compare to the single large reactor D. Eq.:
FA0 X
V
kCA0 1 - X

This proves that Xi =X.


For given T, rA(X) only, therefore, rAi = rA
What does Xi =X mean?
The X achieved in any one of the reactors in
parallel is identical to what would be achieved
if the reactant were fed in one stream to one
large reactor of volume V.

CHENG 323 51
Dr. Shaker Haji, University of Bahrain 4/30/2013

4.3.4 A Second-Order Reaction in a CSTR

For a second-order liquid-phase rxn being carried out in a


CSTR, the combination of the rate law and the design equation,
with manipulation, yields:

Recall that for irreversible second-order rxn, Da = k CA0

Fig. X = f(Da) for a second-order rxn with constant .

Note that at high X (say 67%) a 10-fold increase in Da (e.g. by


increasing V or k) will increase the X only to 88%. Why?
Because CSTR operates under the condition of the lowest value of
the CA , and consequently the smallest value of the rA.

CHENG 323 52
Dr. Shaker Haji, University of Bahrain 4/30/2013

Design Problem 3:
(lost)

(lost)

(lost)

CSTR

Example 4-2. Producing 200 MM Pounds per Year EG in a CSTR

Close to 12.2 billion metric tons of ethylene glycol (EG) were


produced in 2000, which ranked it the 26th most produced
chemical in USA that year on a total pound basis.
About one-half of the ethylene glycol is used for antifreeze while the
other half is used in the manufacture of polyesters [e.g. PET:
poly(ethylene terephthalate)]. In the polyester category, 88% was
used for fibers and 12% for the manufacture of bottles and films.
The 2004 selling price for ethylene glycol was $0.28 per pound
(233 fils/kg).
Structure formula of PET:
Wikipedia.org

CHENG 323 53
Dr. Shaker Haji, University of Bahrain 4/30/2013

It is desired to produce 200 million pounds per year of EG. The reactor
is to be operated isothermally. A 1 lbmol/ft3 solution of ethylene
oxide (EO) in water is fed to the reactor together with an equal
volumetric solution of water containing 0.9 wt% of the catalyst
H2SO2. The specific rxn rate constant is 0.311 min-1 (will be
determined later in Example 4-1). The rate law is rA = kCA.

(a) If 81.5% X is to be achieved, determine the necessary CSTR volume.


(b) If two 800-gal reactors were arranged in parallel, what is the
corresponding X?
(c) If two 800-gal reactors were arranged in series, what is the
corresponding X?

Extract useful Data:

It is desired to produce 200 million pounds per year of EG. The reactor is to be
operated isothermally. A 1 lbmol/ft3 solution of ethylene oxide (EO) in water is fed
to the reactor together with an equal volumetric solution of water containing 0.9 wt%
of the catalyst H2SO2. The specific rxn rate constant is 0.311 min-1 while the rate law
is rA = kCA.

C A01 1 lbmol/ft 3 C A02 0 lbmol/ft 3


01 02 01

CA0
0.5 lbmol/ft 3

CHENG 323 54
Dr. Shaker Haji, University of Bahrain 4/30/2013

Extract useful Data:

It is desired to produce 200 million pounds per year of EG. The reactor is to be
operated isothermally. A 1 lbmol/ft3 solution of ethylene oxide (EO) in water is fed
to the reactor together with an equal volumetric solution of water containing 0.9 wt%
of the catalyst H2SO2. The specific rxn rate constant is 0.311 min-1 while the rate law
is rA = kCA.

- rA kCA
mC 200 106 lb/year
MM C 62 lb/lbmol
The algorithm for the CRE design problem is:
T Constant
1. Mole balance, Design Equation
C A01 1 lbmol/ft 3
2. Write Rate law
C A0 0.5 lbmol/ft 3 3. Stoichiometry (to find Cj)
4. Combine (all of the above)
FC 6.137 lbmol/min
5. Evaluate

(a) If 81.5 % X is to be achieved, determine the necessary CSTR volume.


C A 0 0.5 lbmol/ft 3

FC 6.137 lbmol/min

FA 0 7.530 lbmol/min

V 213.3 ft 3 1596 gal 1600 gal

CHENG 323 55
Dr. Shaker Haji, University of Bahrain 4/30/2013

(b) If two 800-gal (106.7-ft3) reactors were arranged in parallel, what


is the corresponding X ?

Consider the same feed characteristic: FA0 = 7.53 lbmol/min, CA0 = 0.5
lbmol/ft3

FA0 = 7.53 lbmol/min


CA0 = 0.5 lbmol/ft3

X 0.815

(c) If two 800-gal (106.7-ft3) reactors were arranged in series, what are
the corresponding Xs?

Consider the same feed characteristic: FA0 = 7.53 lbmol/min, CA0 = 0.5
lbmol/ft3

FA0 = 7.53 lbmol/min


CA0 = 0.5 lbmol/ft3

X 1 0.688
X 2 0.903

CHENG 323 56
Dr. Shaker Haji, University of Bahrain 4/30/2013

FA0 = 7.53 lbmol/min FA0 = 7.53 lbmol/min


CA0 = 0.5 lbmol/ft3 CA0 = 0.5 lbmol/ft3

X 1 0.688
X 2 0.903
X 0.815

1600-gal Singe-CSTR

FA0 = 7.53 lbmol/min FA0 = 7.53 lbmol/min


CA0 = 0.5 lbmol/ft3 CA0 = 0.5 lbmol/ft3

X 1 0.688
X 2 0.903
X 0.815

800-gal 2nd-CSTR

1600-gal Singe-CSTR

800-gal 1st-CSTR

CHENG 323 57
Dr. Shaker Haji, University of Bahrain 4/30/2013

4.6 Synthesizing the Design of a Chemical Plant

The three reactors designed in examples 4-2, 4-3, & 4-6 can be
arranged to form a chemical plant to produce 200 MM lbm/year of
ethylene glycol from a feedstock of 402 MM lbm/year of ethane.

Read at Home

(lost)

Example 4-3: PFR


(lost)

Example 4-6: PBR

(lost)

Example 4-2: CSTR

CHENG 323 58
Dr. Shaker Haji, University of Bahrain 4/30/2013

(lost)

(lost)

(lost)

4.6 Synthesizing the Design of a Chemical Plant

How is the profit calculated?


The profit from a chemical plant will be the difference between
income from sales and the cost to produce the chemicals.
An approximate formula might be:
Profit = Sales Revenue Cost of Reactants Operating Cost

The operating costs include such costs as energy, labor, overhead, &
depreciation of equipment.
You will learn more about these costs & the process economics in
Chemical Plant Design course (CHENG 422).

CHENG 323 59
Dr. Shaker Haji, University of Bahrain 4/30/2013

4.6 Synthesizing the Design of a Chemical Plant

Profit = Sales Revenue Cost of Reactants Operating Cost

Let's conservatively estimate the operating and other


expenses to be $8 MM/year.
The prices of ethane, sulfuric acid, & ethylene glycol are
$0.04, $0.042, & $0.38 per pound, respectively.
See www.chemweek.com for current prices.

The profit is around $52 million.

Submit Homework
#6
one week from today

CHENG 323 60
Dr. Shaker Haji, University of Bahrain 4/30/2013

Play the Quiz#4


tic-tac-toe 7-10 days
game from today
Concepts: Interaction of the
Ergun, Arrhenius, and
Van't Hoff equations and
other considerations in
isothermal reactor design.

4 Isothermal Reactor Design

CHENG 323 61
Dr. Shaker Haji, University of Bahrain 4/30/2013

4.7 Mole Balances on CSTRs, PFRs, PBRs, and


Batch Reactors
There are many instances when it is much more convenient to work in
terms of number of moles (Nj), concentrations (Ci), or molar flow rates
(Fj) rather than X.
Can you give examples?

Membrane Reactors Multiple Reactions

Ci or Fi

Unsteady State Liquid Phase Reactions


Operation

4.7 Mole Balances on CSTRs, PFRs, PBRs, and


Batch Reactors

Unsteady state operation in batch or semibatch reactors, membrane


reactors, and multiple reactions are common examples of such cases.
We now modify our algorithm by using Cj for liquids & Fj for gases as
our dependent variables.
The main difference between the conversion algorithm and the molar
flow rate/concentration algorithm is that:
in the X algorithm, we needed to write a mole balance on only one specie,
whereas in the Fj & Cj algorithm we must write a mole balance on each and
every species.

CHENG 323 62
Dr. Shaker Haji, University of Bahrain 4/30/2013

A + 2B C

Isothermal reaction design


algorithm:
1. Mole balance
2. Rate law (3)
a)

3. Stoichiometry
b)
4. Combine
5. Evaluate
c)

d)

(4)
&
(5)

4.7.1 Liquid Phase

For liquid-phase rxns in which there is no volume change,


concentration is the preferred variable.
The mole balances for the
generic rxn

where -rA k AC AC B
are shown in this table:
Where are these equations
coming from?

CHENG 323 63
Dr. Shaker Haji, University of Bahrain 4/30/2013

They are coming from the mole balances:

(lost)
Design Problem 4:

(lost)

(lost)

Scale-Up of Liquid-Phase Batch Reactor Data to the Design


of a CSTR

CHENG 323 64
Dr. Shaker Haji, University of Bahrain 4/30/2013

Determining k from batch data

Example 4-1
It is desired to design a CSTR to produce 200 million pounds of
ethylene glycol per year by hydrolyzing ethylene oxide. However,
before the design can be carried out, it is necessary to perform and
analyze a batch reactor experiment to determine the specific rxn
rate constant, k.
Because the reaction will be carried out isothermally, k will need to be
determined only at the rxn T of the CSTR. At high temperatures
there is a significant by-product formation, while at temperatures
below 40C the rxn does not proceed at a significant rate;
consequently, a temperature of 55C has been chosen.

Example 4-1, continued


The rxn is as follows:

Because water is usually present in excess, its concentration may


be considered constant during the course of the reaction. The
rxn is first-order in ethylene oxide.
In the laboratory experiment, 500 mL of a 2 M solution of
ethylene oxide in water was mixed with 500 mL of water
containing 0.9 wt% sulfuric acid, which is a catalyst.
The T was maintained at 55oC. The concentration of ethylene
glycol was recorded as a function of time.

CHENG 323 65
Dr. Shaker Haji, University of Bahrain 4/30/2013

Determining k from batch data

Example 4-1, continued


t , min C C , M
Using the data in the given
0 0
table, determine the
specific reaction rate at
0.5 0.165
55C. 1 0.250
1.5 0.395
2 0.464
3 0.590
4 0.700
6 0.873
10 0.952

Example 4-1, Solution


C
ln A0 k t
CA
FYI only
t , min C C , M C A , M C A0 /C A ln (C A0 /C A) X
0 0 1 1 0 0.000
0.5 0.165 0.835 1.198 0.180 0.165
1 0.250 0.750 1.333 0.288 0.250
1.5 0.395 0.605 1.653 0.503 0.395
2 0.464 0.536 1.866 0.624 0.464
3 0.590 0.410 2.439 0.892 0.590
4 0.700 0.300 3.333 1.204 0.700
6 0.873 0.127 7.874 2.064 0.873
10 0.952 0.048 20.833 3.037 0.952
C A0 = 1 M

CHENG 323 66
Dr. Shaker Haji, University of Bahrain 4/30/2013

Example 4-1, Solution


C
k = 0.311 min-1 ln A0 k t
CA

Example 4-1, Solution


C C
k = 0.311 min-1 ln A0 k t A0 e k t
CA CA

Practical guidelines for reactor scale-up were recently given by Mukesh


and by Warsteel (see textbook for references)

CHENG 323 67
Dr. Shaker Haji, University of Bahrain 4/30/2013

Week 10 - Lecture 33

Prepared by: Dr Shaker Haji - CHENG 323

4.7.2 Gas Phase

For gas-phase rxns, it is preferred to express the mole balance


equations in terms of number of moles (Nj) or molar flow rates (Fj).
As an example, lets design a PBR for the rxn
1. Mole balance F T
Ci CT 0 i 0 y
dFi FT T
ri ' For each species (X4)
dW FT Fi

2. Rate law -rA' k AC AC B

3. Stoichiometry
-rA' -rB' rC' rD' 4. Combine
3 Eqns
a b c d 5. Evaluate

CHENG 323 68
Dr. Shaker Haji, University of Bahrain 4/30/2013

CHENG 323 69
Dr. Shaker Haji, University of Bahrain 4/30/2013

4.8 Microreactors (Micro Process Engineering)

Microreactors are emerging as a new technology in CRE known as


Micro Process Engineering.
Microreactors are characterized by their high surface area-to-volume
ratios in their microstructured regions that contain tubes or channels.

Fig: Glass Microreactor.


Channels Dipth X Width: 150m X150m.
Ref: www.wikipedia.com,

CHENG 323 70
Dr. Shaker Haji, University of Bahrain 4/30/2013

4.8 Microreactors (Micro Process Engineering)

A typical channel width might be 100 m with a length of 20,000 m


(2 cm).
The resulting high surface area-to-volume ratio (~10,000 m2/m3)
reduces or even eliminates heat and mass transfer resistances often
found in larger reactors.
Consequently, surface catalyzed rxns can be greatly facilitated, hot
spots in highly exothermic reactions can be eliminated, and in many
cases highly exothermic reactions can be carried out isothermally.
Other advantages include shorter residence times and narrower
residence time distributions.

Fig: Microreactors: All-in-One Extractor


Ref: web.mit.edu

CHENG 323 71
Dr. Shaker Haji, University of Bahrain 4/30/2013

Fig: Microreactors.
Ref: www.carbogen-amcis.com

Fig: Microplant with reactors, valves, and


mixers.

Fig: Microreactor with Heat Exchanger.


Courtesy of Ehrfeld, Hessel, & Lowe. Microreactors: New Technology for Modern Chemistry (Wiley-VCH. 2000).

4.8 Microreactors (Micro Process Engineering)

Production in microreactor systems can be increased simply by adding


more units in parallel.
For example, the catalyzed reaction

required only 32 microreaction systems in parallel to produce 2000


tons/yr of acetate!
Microreactors are also used for the production of specialty chemicals,
combinatorial chemical screening, lab-on-a-chip, & chemical sensors.
In modeling microreactors, we will assume they are either in plug flow
(dFj/dV = rj) or in laminar flow (segregation model Chap. 13)

CHENG 323 72
Dr. Shaker Haji, University of Bahrain 4/30/2013

Example 4-7 Gas-Phase Reaction in a Microreactor -


Molar Flow Rates

The gas-phase reaction which follows an elementary rate law


2NOCl 2NO + Cl2
is carried out at 425C & 1641 kPa (16.2 atm) in a microreactor
system using a bank of ten microreactors in parallel. Each
microreactor has 100 channels. Each channel has a cross-sectional
area of 0.2 mm2 & a length of 50 mm (volume of 10-5 L).
In this system, 85% conversion is achieved and 20 tons of NO per year
is produced by processing 0.0226 mol/s of pure NOCl, or 2.26X10-5
mol/s per channel. The rate constant is: k = 0.29 L/mol.s at 500 K
with E = 24 kcal/mol.
Plot the Fj = f(V) down the length of the reactor.

Example 4-7 Gas-Phase Reaction in a Microreactor - Molar Flow Rates

Solution:
2NOCl 2NO + Cl2
2A 2B + C

T = 425C P = 1641 kPa yNOCl,0 or yA0 = 1


-rA = kCA2 k = 0.29 L/mol.s at 500 K with E = 25 kcal/mol
Each channel:
Ac = 0.2 mm2 Z = 50 mm V = 10-5 L

CHENG 323 73
Dr. Shaker Haji, University of Bahrain 4/30/2013

Solution: 2NOCl 2NO + Cl2


2A 2B + C
1. Mole balance

2. Rate law
-rA = kCA2
k = 0.29*exp[-E/R(1/T-1/500)] where E = 24x103 cal/mol

3. Stoichiometry
Gas phase with T = T0 and P = P0
Relative Rates

Concentrations

Total Flow Rates

4. Combine
Using Polymath
5. Evaluate

CHENG 323 74
Dr. Shaker Haji, University of Bahrain 4/30/2013

2NOCl 2NO + Cl2


2A 2B + C

4.9 Membrane Reactors

The term membrane reactor describes a number of different


types of reactor configurations that contain a membrane.

CHENG 323 75
Dr. Shaker Haji, University of Bahrain 4/30/2013

4.9 Membrane Reactors

List 3 reasons for using membrane reactors.


To increase X when the rxn is thermodynamically limited
CH4 + H2 O 3H2 + CO

To increase the selectivity when multiple rxns are occurring


CH4 + 0.5 O2 2H2 + CO

CH4 + 2O2 2H2 O + CO2

To obtain pure products.

4.9 Membrane Reactors

Thermodynamically limited rxns are rxns where the equilibrium lies


far to the left (i.e.. reactant side) and there is little X.
Before illustrating the usefulness of the membrane reactors for
thermodynamically limited rxns, lets discuss the following:
What will happen if T is increased during an exothermic rev. rxn?
Increasing T will only drive the rxn further to the left (lower the X)
So should I carry out the rxn at low T? What could happen?
Decreasing T will result in a slow reaction rate and, therefore, low X.

CHENG 323 76
Dr. Shaker Haji, University of Bahrain 4/30/2013

4.9 Membrane Reactors

Is it favorable to run an endothermic rev. rxn under high T?


Yes, increasing T will move the rxn to the right to favor a higher X.
Are there any drawback for increasing the T excessively?
Yes, for many rxns, these higher T induce undesired side rxns and cause the
catalyst to become deactivated.
The membrane reactor is a technique for driving reversible rxns to the
right toward completion in order to achieve very high X.
This can be achieved by having one of the rxn products diffuse out of
a semipermeable membrane surrounding the rxn mixture.
As a result, the reverse rxn will not be able to take place, and the rxn
will continue to proceed to the right toward completion.

Mention the main types of the membrane reactors.


Inert Membrane Reactor (IMR): the reactor is packed with catalyst while the
membrane serves as a barrier to the reactants & some of the products.
Catalytic Membrane Reactor (CMR): the catalyst is deposited directly on the
membrane, and only specific rxn products are able to permeate through the
membrane.

Inert Membrane Reactor

Catalytic Membrane Reactor (CMR)

CHENG 323 77
Dr. Shaker Haji, University of Bahrain 4/30/2013

CRE Algorithm for MRs

We use similar algorithm, but


with slight modification.
We use the reactor volume
instead of the catalyst weight.
Write hydrogen mole balance
for the shown segment of the
membrane reactor.

CHENG 323 78
Dr. Shaker Haji, University of Bahrain 4/30/2013

How do we quantify RB?


Through molar flux of B:

Where kC is the overall mass transfer coefficient in m/s and CBS is


the concentration of B in the sweep gas channel (mol/L).
The kC accounts for all resistances to transport:
the tube side resistance of the membrane

the membrane itself

the shell (sweep gas) side resistance.

Membrane Reactors

Why we dont use X when analyzing membrane reactors?


Are the following definitions valid for membrane reactors?

FA0 - FA B
X A C
FA0

FB FA0 B B X

A C
B

CHENG 323 79
Dr. Shaker Haji, University of Bahrain 4/30/2013

Example 4-8 Membrane Reactor

According to the US Department of Energy (DOE), an energy saving of 10 trillion BTU


per year could result from the use of catalytic membrane reactors as replacements
for conventional reactors for dehydrogenation reactions such as the
dehydrogenation of ethylbenzene to styrene:

and of butane to butene:

The dehydrogenation of propane is another reaction that has proven successful with a
membrane reactor:

All the preceding dehydrogenation reactions above can be represented symbolically


as:

and could take place on the catalyst side of an IMR or inside a CMR.

Consider the following dehydrogenation reaction :


which is to be carried out in a catalytic membrane reactor (CMR). The equilibrium
constant for this reaction is quite small at 227C (Kc = 0.05 mol/l). The
membrane is permeable to B which is H2 but not to A and C. Pure gaseous
A enters the reactor at 8.2 atm and 227C at a rate of 10 mol/min.
Consider the rate of diffusion of B out of the reactor per unit volume of reactor, RB ,
to be proportional to the concentration of B (i.e.. RB = kc.CB , where kc represents
the overall mass transfer coefficient). The reactor volume is 400 liters. The
specific reaction rate, k, and the transport coefficient, kc, are k = 0.7 min-1 and
kc = 0.2 min-1, respectively.
(a) Perform differential mole balances on A, B, and C to arrive at a set of coupled
differential equations to solve.
(b) Plot the molar flow rates of each species as a function of the reactor volume.
(c) Calculate the conversion at the exit of the reactor.
(d) What would be the achieved conversion if the reaction was carried out in a
conventional PFR.
(e) Compare the conversions in both cases and explain the obtained results.

CHENG 323 80
Dr. Shaker Haji, University of Bahrain 4/30/2013

CMR
X = 0.567

FA FC
FA
FC
FB
FB

CHENG 323 81
Dr. Shaker Haji, University of Bahrain 4/30/2013

PFR
X = Xeq = 0.447

FA

FB & F C

Week 11 - Lecture 37

Prepared by: Dr Shaker Haji - CHENG 323

CHENG 323 82
Dr. Shaker Haji, University of Bahrain 4/30/2013

Unsteady-State Operations

4.10 Unsteady-State Operations of Stirred


Reactors
Situations, studied in this chapter, where we have unsteady state
operation include:
Batch Reactor: CA vs. t
Startup of a CSTR: to determine t necessary to reach steady-state operation.
Semibatch Reactor: CA or X vs. t

analytical solutions to the differential equations, arising from the mole


balances in the above situations, can be obtained only for zero- and
first-order rxns.
ODE solvers must be used for other rxn orders.

CHENG 323 83
Dr. Shaker Haji, University of Bahrain 4/30/2013

4.10.1 Startup of a CSTR

If a CSTR is to be started up by introducing the feed & without having


an overflow stream (no output), then this is a semibatch operation.
The startup of a fixed volume CSTR, initially filled with pure solvent,
under isothermal conditions is rare, but does occur occasionally.
We can, however, carry out an analysis to estimate the time
necessary to reach steady-state operation.
AB
Feed AB Feed Product
CA0 CA0 CA & CB

Startup of a CSTR in semibatch mode Startup of a constant volume CSTR

4.10.1 Startup of a CSTR

Consider the liquid phase reactive system shown below where:


The reactor is well-mixed (no spatial variation in rA)
= 0 & V = Vi (where Vi is the initial volume)
Assume the reactor is initially filled with the solvent only.

Write a general mole balance


equation for the given system. AB
Feed Product
0, CA0, FA0 , FA, FB,
CA, CB

Startup of a constant volume CSTR

CHENG 323 84
Dr. Shaker Haji, University of Bahrain 4/30/2013

4.10.1 Startup of a CSTR

The general mole balance equation for the given system is:
dN A
FA0 - FA rAV
dt
Conversion does not have any meaning in startup because one cannot
distinguish between the NA present due to unreacted A molecules &
these due to accumulation within the CSTR due to introduction of A.
Rewrite the balance equation in terms of concentration.
dC A AB
C A0 - C A rA Feed Product
dt 0, CA0 , CA & CB
For most 1st-order systems (not necessarily
1st-order rxn), steady state is achieved in 4
to 5 times the systems time constant.
Startup of a constant volume CSTR


Example: Startup of a Constant Volume CSTR

Consider the first order liquid phase reaction that is carried out in a well-mixed reactor initially
filled with the solvent only. Assume the density of the reaction mixture remains constant during the
course of the reaction, i.e. = 0 & V = Vi.
Additional information: k = 0.311 min-1, V = 213.3 L, CA0 = 0.500 mol/L, 0 = 15.34 L/min.
1. For the startup period, perform mole balances on A and B in terms of CA, CB, & other constants.
2. Use the derived ordinary differential equations to plot the concentrations of A & B as a function
of time.
3. Calculate the steady state values of CA, XA, and CB. Do both CA and CB achieve their steady state
values at the same time?
4. Determine the system order for both species by examining the obtained differential equations
and concentration profiles.
5. For the first order system(s), if any, determine:
I. the time constant, the steady-state gain, and the input (forcing function). Hint: rewrite the
differential mole balance in the standard form of a first-order system.
II. the time necessary for the concentration to achieve 99% of its steady state value. Compare
that time with the time constant; is it reasonable for a first order system?

CHENG 323 85
Dr. Shaker Haji, University of Bahrain 4/30/2013

4.10.2 Semibatch Reactors

Besides the startup of a CSTR in semibatch mode prior to achieving


constant V, there are two basic types of semibatch operations:
Removal of a product (reactive distillation)

Addition of a reactant

A+BC+D

Feed
CA0 & CB0

Startup of a CSTR in semibatch mode

A+BC+D
4.10.2 Semibatch Reactors A+BC+D

The first type of semibatch reactor is reactive distillation.


Here reactants A and B are initially charged together and one of the
products is vaporized and withdrawn continuously.
What are the benefits of removing one of the products (e.g. C) in this
manner?
1. For an equilibrium rxn, this will shift the equilibrium toward the right,
increasing the final X.
2. This will further concentrates the reactants, and
thereby producing an increased rate of rxn & &
decreased processing time.

CHENG 323 86
Dr. Shaker Haji, University of Bahrain 4/30/2013

4.10.2 Semibatch Reactors

Examples of rxns carried out in this type of rxn include acetylation


rxns & esterification rxns in which water is removed.

E.g. (1): Salicylic acid is acetylated to form aspirin (acetyl functional


group )

E.g. (2): Esterification of acetic acid in excess


ethanol to form ethyl acetate (ester)

Rxns from: www.wikipedia.com

4.10.2 Semibatch Reactors A+BC+D

In the second type of the semibatch operation, one of the reactants


(e.g. B) is slowly fed to a reactor containing the other reactant (e.g. A),
which has already been charged to the reactor.
When this type of reactor is generally used?
when unwanted side reactions occur at high CB

when the reaction is highly exothermic/indothermic

In some rxns, the reactant B is a gas and is bubbled continuously


through liquid reactant A.
Examples of rxns used in this type of
semibatch reactor operation include:
Ammonolysis hydrolysis

chlorination

CHENG 323 87
Dr. Shaker Haji, University of Bahrain 4/30/2013

4.10.2 Semibatch Reactors A+BD+U

One of the best reasons to use semibatch reactors is to enhance


selectivity in liquid-phase reactions.
For example, consider the following two simultaneous rxns:
Desired rxn:

Undesired rxn:

The instantaneous selectivity SD/U is defined as the ratio of the


relative rates:

How can we produce the most of our desired product and the least of
our undesired product?
By keeping the CA high and the CB low.
How can this achieved?
This can be achieved through the use of the semibatch reactor, which
is charged with pure A and to which B is fed slowly to A in the vat.

CHENG 323 88
Dr. Shaker Haji, University of Bahrain 4/30/2013

Commercial Break
Now you are ready
to read this article:

Published in Chemical
Engineering Progress, An
AIChE Publication, CEP:
Mar 2001.

Available on Black Board.

4.10.2 Semibatch Reactors

consider the elementary liquid-phase reaction


A+BC
that is to be carried out in the shown semibatch reactor with a
constant molar feed of B.
Write a mole balance on species A.

Three variables can be used to formulate and


solve semibatch reactor problems:
the concentrations, Cj

the number of moles, Nj

the conversion, X.

CHENG 323 89
Dr. Shaker Haji, University of Bahrain 4/30/2013

4.10.3 Writing the Semibatch Reactor Equations in Terms


of Concentrations

For the considered reaction, rewrite the mole balance equations for
species A & B in terms of concentration.
Also find V(t)
A+BC

Remember, @ t = 0, CA=CAi & CB = 0


If the rxn order is other than zero- or first-order, or
if the rxn is non-isothermal, we must use numerical
techniques to solve the above Eqs.

Problem 4-23 production of ethylene glycol from ethylene chlorohydrin

The production of ethylene glycol from ethylene chlorohydrin and sodium


bicarbonate
is carried out in a semibatch reactor. A 1.5 molar solution of ethylene
chlorohydrin is fed at a rate of 30 mole/min to 1500 liters of a 0.75 molar
solution of sodium bicarbonate. The reaction is elementary and carried out
isothermally at 30oC where the specific reaction rate is 5.1 L/mol.h. The
reactor can hold a maximum of 2500 liters of liquid. Assume constant
density. A conversion of 96% of sodium bicarbonate is desired.
1. Perform mole balances on the reactants.
2. Calculate the time required to achieve the desired conversion.
3. Is it possible to achieve the desired conversion with the given reactor and
conditions? Why? If not, then what is the maximum possible conversion.
4. Plot the concentration of reactants, the reaction rate, and the conversion as
a function of time. (Fig. 1: CA, CB, & X vs t; Fig. 2: -rB vs. t)

CHENG 323 90
Dr. Shaker Haji, University of Bahrain 4/30/2013

CHENG 323 91
Dr. Shaker Haji, University of Bahrain 4/30/2013

Example Hydrolysis of acetic anhydride in a semibatch reactor

The kinetics of hydrolysis of acetic anhydride was studied by Haji &


Erkey* using in-situ FTIR spectroscopy.
O O O

C C + H2O 2 C
H3C O CH3 H3C OH

In part of the study, acetic anhydride


was hydrolyzed in a semibatch
operation as shown in the figure.

*Haji, Shaker; Erkey, Can. AN EXPERIMENT FOR UNDERGRADUATE LABORATORY KINETICS OF HYDROLYSIS OF ACETIC ANHYDRIDE BY IN-
SITU FTIR SPECTROSCOPY. Chemical Engineering Education, winter 2005.

In the semibatch experiment, 150 ml of distilled water (W) was


initially charged to the reactor while pure acetic anhydride (A) was
continuously added to the reactor at a rate of 3.55 ml/min over a
period of 8.17 minutes. The reaction was carried out isothermally at
35oC. Water was present in excess, therefore, it concentration was
assumed to be constant during the course of the rxn. Hence the rxn
was considered to be first-order in acetic anhydride.
Additional information:
k = 0.2752 min-1, MMA = 102.09 g/mol, A = 1.082 g/ml
Assume a constant-density system

Solve for the concentration of acetic anhydride and the rate of


reaction as a function of time.

CHENG 323 92
Dr. Shaker Haji, University of Bahrain 4/30/2013

4.10.4 Writing the Semibatch Reactor Equations in Terms


of Conversion

Consider the reversible elementary reaction: A + B C + D in which


B is fed, at constant rate, to a vat containing only A initially.
Write an integral mole balance on A, B, C, & D.

NA - NA0 = - NA0X NB-NBi = FB0dt - NA0X or NB = FB0.t - NA0X


NC = + NA0X ND = + NA0X
A+BC+D
Then, write the rate law:
Then, use the stoichiometry

Finally, combine & evaluate

Finding the Equilibrium Conversion: A + B C + D

For reversible rxns carried out in a semibatch reactor, the maximum


attainable X (i.e., the equilibrium X) will change as the rxn proceeds.
Why?
Because more reactant is continuously added to the reactor.
This addition shifts the equilibrium continually to the right toward more
product.
The Xe could be calculated as follows at equilibrium:

where

CHENG 323 93
Dr. Shaker Haji, University of Bahrain 4/30/2013

Therefore,

Solving for t:

Solving for Xe:

Closure

Heat Effects

Isothermal Design

Combine & Evaluate

Stoichiometry

Rate Laws

Mole Balance

CHENG 323 94
Dr. Shaker Haji, University of Bahrain 4/30/2013

Submit Homework
#7
one week from today

Homework Solutions

CHENG 323 95
Dr. Shaker Haji, University of Bahrain 4/30/2013

CHENG 323 96