rm .2:
,3. k.
..

3;»!

£33.51 v

V.‘
n. A: .

.3 4
‘ ,‘x
¢ ‘

s...

..~n .
5:12.: > u. I
. 1‘ “ ‘htan
..

. :2
3.3%.»

 

THE“

, I
.‘s
I I

1):,

A

(7

This is to certify that the

thesis entitled

AQUEOUS PHASE ADSORPTION OF BENZALDEHYDE,
BENZOIC ACID AND BENZYL ALCOHOL AND THEIR MIXTURES

presented by

Chirag Ashok Shah

has been accepted towards fulfillment
of the requirements for

M. S . degree in Chemical Engineering

 

 

Mflq

Major professor

Date //Z/0/

0-7639 MS U i: an Affirmative Action/Equal Opportunily Institution

 

 

LIBRARY
Michigan State
University

 

 

 

PLACE IN RETURN Box to remove this checkout from your record.
To AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/01 cJCIRCIDateDuepss-pJS

AQUEOUS PHASE ADSORPTION OF BENZALDEHYDE, BENZOIC ACID AND
BENZYL ALCOHOL AND THEIR MIXTURES.

By

Chirag Ashok Shah

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

MASTER OF SCIENCE

Department of Chemical Engineering

2001

ABSTRACT

AQUEOUS PHASE ADSORPTION OF BENZALDEHYDE, BENZOIC ACID AND
BENZYL ALCOHOL AND THEIR MIXTURES.

By
Chirag Ashok Shah

Aqueous-phase adsorption equilibria of benzaldehyde, benzyl alcohol and benzoic
acid are measured in the concentration range 0.2-2.0 moi/m3 at room temperature on
synthetic SP-850 resin. Several two and three parameter isotherm equations are tested.
Among the models tried, the two-parameter Langmuir equation is found to be the most
satisfactory. Experiments are presented to study the fixed-bed breakthrough behavior for
single component liquid adsorption. The Thomas model based on the Langmuir equation
is found to predict the breakthrough behavior satisfactorily for process design, and is
superior to the Hougen and Marshall model using linear model. Multicomponent liquid-
phase adsorption was studied for systems comprising benzaldehyde, benzyl alcohol and
benzoic acid. Ideal Adsorbed Solution theory is used to model the breakthrough patterns
for multicomponent systems using a linear driving force model where intraparticle
diffusion is the rate-limiting step.

The process to recover benzaldehyde and benzyl alcohol from cherry pits by
adsorption was studied. Various experimental variables were optimized. Knowledge of
equilibrium adsorption data and breakthrough behavior data of benzaldehyde, benzyl
alcohol and benzoic acid was used to predict the adsorption behavior of hydrolyzate

obtained from cherry pits.

I dedicate this finished product to my Dad, who taught me to value education.

iii

ACKNOWLEDGMENTS

I would like to thank my advisor Dr. Carl Lira for his patience and support as I have
struggled to produce this finished product from my education. I would also like to thank
Joel Dulebohn and Xiao Ning for their support and guidance in conducting the cherry pit

experiments.

Regards to all the friends who have come and gone over the years. You will see me

again.

Credit goes to Jamie Morgan, Khanghy, Amber and Shannon for assisting in the

adsorption runs and issues related to the GC.

iv

TABLE OF CONTENTS

List of Tables ..................................................................................... vi
List of Figures .................................................................................... ix
Chapter 1: Introduction .......................................................................... 1

Chapter 2: Adsorption Behavior of Benzaldehyde, Benzyl Alcohol and Benzoic
Acid ................................................................................... 4

Chapter 3: Production of Benzaldehyde and Benzyl alcohol from Cherry Pits. . . .....70

Appendix A: Physical Properties ............................................................ 93
Appendix B: GC Calibration ................................................................. 96
Appendix C: Procedures and Raw Data ................................................... 103
Appendix D: Computer Program ............................................................ 140

LIST OF TABLES

2-1: Results of Best Linear Fit .................................................................... 25
2-2: Results of best Langmuir Fit ................................................................. 25
2-3: Results of Best Freundlich Fit ................................................................ 25
2-4: Results of best Langmuir—Freundlich Fit .................................................... 26
2-5: Summary of Experimental Conditions for Single Component Adsorption ............ 31
2-6: Comparision of KD from Fitting Equilibrium Adsorption Data and Fitting
Breakthrough Behavior Data using Hougen and Marshall Model ....................... 37
2-7: Results of Data-fitting for Benzaldehyde using Hougen and Marshall Model ....... 37

2-8: Results of Data-fitting for Benzyl-Alcohol using Hougen and Marshall Model.. . ..37

2-9: Results of Data-fitting for Benzoic acid using Hougen and Marshall Model ........ 37

2-10: Values of {I and 7M1) Function for Experiment 1 for Single Component

Adsorption ..................................................................................... 38
2-11: Results of Data-fitting for Benzaldehyde using Thomas Model ...................... 45
2-12: Results of Data-fitting for Benzyl-Alcohol using Thomas Model .................... 45
2-13: Results of Data-fitting for Benzoic acid using Thomas Model ....................... 45
2-14: Results of Best Freudlich Fit for the Desired Concentration Range .................. 54
2-15: Adsorbate Concentrations in Feed ....................................................... 56
2-16: Summary of Experimental Conditions for Multicomponent Adsorption... ......... 57
3-1: Hydrolyzate Composition for Different Temperatures ................................. 73
3-2: Hydrolyzate Composition for Different Water- Dry Pit Ratio ......................... 74
3-3: Hydrolyzate Composition for Different Water- Wet Pit Ratio ........................ 75
3-4: Desorption Results ............................................................................ 89
A-l: Physical Properties of Benzaldehyde ...................................................... 93

vi

A-2: Physical Properties of Benzyl Alcohol ................................................... 94

A—3: Physical Properties of Benzoic Acid ...................................................... 95
B-l: Sample Solutions of Benzaldehyde/Water ................................................. 97
B-2: Sample Solutions of Benzyl Alcohol /Water ............................................. 97
B-3: Benzaldehyde/Water GC Calibration Data ................... 98
B4: Benzyl Alcohol/Water GC Calibration Data .............................................. 98
C-1.1: p, and p, Calculations .................................................................. 104
C-2.1.A: Results of Experiments to Detect the Ratio of Dry Resin Weight / Wet Resin
Weight ........................................................................................ 106
C-2.1.B: Calibration of pH-meter ............................................................... 107
C-2.2: Equilibrium Adsorption Data for Benzaldehyde Solution .......................... 108
C-2.3: Equilibrium Adsorption Data for Benzyl Alcohol Solution ......................... 109
C-2.4: Equilibrium Adsorption Data for Benzoic Acid Solution ........................... l 10
C-3. 1: Benzaldehyde Breakthrough Behavior, Experiment-l ................................ 112
C-3.2: Benzaldehyde Breakthrough Behavior, Experiment -2 ............................... 113
C-3.3: Benzaldehyde Breakthrough Behavior, Experiment -3 ............................... 114
C-3.4: Benzaldehyde Breakthrough Behavior, Experiment -4 ............................... 115
C-3.5: Benzyl Alcohol Breakthrough Behavior, Experiment-5 .............................. 116
C-3.6: Benzyl Alcohol Breakthrough Behavior, Experiment—6 .............................. 117
C-3.7: Benzyl Alcohol Breakthrough Behavior, Experiment-7 .............................. 118
C-3.8: Benzyl Alcohol Breakthrough Behavior, Experiment-8 .............................. 119
C-3.9: Benzoic Acid Breakthrough Behavior, Experiment-9 ................................ 120
C-3.10: Benzoic Acid Breakthrough Behavior, Experiment-10 ............................. 121
C-3.1 l: Benzoic Acid Breakthrough Behavior, Experiment-11 ............................. 122

vii

C-3. 12: Benzoic Acid Breakthrough Behavior, Experiment-12 ............................. 123

C-4.l: Constant Parameters in Multicomponent Adsorption Experiments. . . . . . . . ...125
C-4.2: Multicomponent Breakthrough Behavior, Experiment-1 ............................ 126
C-4.3: Multicomponent Breakthrough Behavior, Experiment-2 ............................. 127
C-4.4: Multicomponent Breakthrough Behavior, Experiment-3 ............................ 128
C-4.5: Multicomponent Breakthrough Behavior, Experiment-4 ............................. 129
C46: Multicomponent Breakthrough Behavior, Experiment-5 ............................ 130
C-4.7: Multicomponent Breakthrough Behavior, Experiment-6 ............................ 131
C48: Multicomponent Breakthrough Behavior, Experiment-7 ............................. 132
C49: Multicomponent Breakthrough Behavior, Experiment-8 ............................. 133
C-4.10: Multicomponent Breakthrough Behavior, Experiment-9 ............................ 134
C-5. 1: Data for Volume Determination of Sample Loop .................................... 137

C-5.2: Results for Checking precision of Sampling loop .................................... 138

viii

LIST OF FIGURES

2-1: Brunauer’s Classification of Adsorption Isotherms, Showing Amount Adsorbed

Versus Normalised Concentration in the Outlet ......................................... 5
2-2: Classification of Isotherms from Solution ............................................... 11
2-3: Benzaldehyde Equilibrium Adsorption Isotherm and Linear Fit ..................... 13
2-4: Benzaldehyde Equilibrium Adsorption Isotherm and Langmuir Fit .................. 14
2-5: Benzaldehyde Equilibrium Adsorption Isotherm and Freudlich Fit ................... 15

2-6: Benzaldehyde Equilibrium Adsorption Isotherm and Langmuir-Freundlich Fit. . . . 16

2-7: Benzyl Alcohol Equilibrium Adsorption Isotherm and Linear Fit ..................... 17
2-8: Benzyl Alcohol Equilibrium Adsorption Isotherm and Langmuir Fit ................. 18
2-9: Benzyl Alcohol Equilibrium Adsorption Isotherm and Freudlich Fit ................. 19

2-10: Benzyl Alcohol Equilibrium Adsorption Isotherm and Langmuir-Freundlich Fit..20

2-11: Benzoic Acid Equilibrium Adsorption Isotherm and Linear Fit ........................ 21
2-12: Benzoic Acid Equilibrium Adsorption Isotherm and Langmuir Fit ................... 22
2-13: Benzoic Acid Equilibrium Adsorption Isotherm and Freudlich Fit .................... 23

2-14: Benzoic Acid Equilibrium Adsorption Isotherm and Langmuir-Freundlich Fit. ....24

2-15: Benzaldehyde Breakthrough Behavior using Linear Isotherm ........................ 33
2- 16: Benzyl Alcohol Breakthrough Behavior using Linear Isotherm ...................... 34
2-17: Benzoic Acid Breakthrough Behavior using Linear Isotherm ......................... 35
2-18: Benzaldehyde Breakthrough Behavior using Langmuir Isotherm .................... 42
2-19: Benzyl Alcohol Breakthrough Behavior using Langmuir Isotherm .................. 43
2-20: Benzoic Acid Breakthrough Behavior using Langmuir Isotherm. .. . . . . .. . . .....44

2-21: Benzaldehyde Breakthrough Behavior using Langmuir Adsorption Isotherm and
Constant K. .................................................................................... 47

ix

2-22: Benzyl Alcohol Breakthrough Behavior using Langmuir Adsorption Isotherm and
Constant Ka ................................................................................... 48

2-23: Benzoic Acid Breakthrough Behavior using Langmuir Adsorption Isotherm and
Constant K, ..................................................................................... 49

2-24: Prediction of Equilibrium Adsorption Data for Benzaldehdye by Freundlich Model

for 0-0.9 mol/m3 Concentration Range ................................................... 54
2-25: Prediction of Equilibrium Adsorption Data for Benzyl Alcohol by Freundlich

Model for 0-0.5 mol/m3 Concentration Range .......................................... 54
2-26: Breakthrough Behavior of Mixtures(15 ml resin at 10 ml flow rate) ................ 58
2-27: Breakthrough Behavior of Mixtures(15 ml resin at 15 ml flow rate) ................ 59
2-28: Breakthrough Behavior of Mixtures(15 ml resin at 20 ml flow rate) ................ 60
2-29: Breakthrough Behavior of Mixtures(20 ml resin at 10 ml flow rate) ................ 61
2-30: Breakthrough Behavior of Mixtures(20 ml resin at 15 ml flow rate) ................ 62
2-31: Breakthrough Behavior of Mixtures(20 ml resin at 20 ml flow rate) ................ 63
2-32: Breakthrough Behavior of Mixtures(25 ml resin at 10 ml flow rate) ................ 64
2-33: Breakthrough Behavior of Mixtures(25 ml resin at 15 ml flow rate) ................ 65
2-34: Breakthrough Behavior of Mixtures(25 ml resin at 20 ml flow rate) ................ 66
3-1: Chemical Structures of Benzaldehyde, Mandelonitrile and Amygdalin ............... 72
3-2: Comparision of Wet Cherry Pits with Dry Cherry Pits .................................. 76

3-3: Breakthrough Behavior(l60 ml SP-850 resin) at 2.6 cm/min Superficial Velocity..80

3-4: Schematic Layout of the Regeneration Setup ............................................. 83
3-5: Schematics of the Sampling Loop .......................................................... 85
3-6: Schematic of Sampling Valve Operation in Sampling Section ........................ 86
3-7: Solvent Regeneration Pattern ............................................................... 88

3-8: Carbon Dioxide Regeneration Results .................................................... 9O

B-l: Benzaldehyde Concentration Versus GC Response .................................... 99
B-2: Benzyl Alcohol Concentration Versus GC Response ................................. 99
B-3: Column Operating Conditions ............................................................ 100
B-4: A Sample Chromatograph ................................................................. 101

xi

Chapter 1:
INTRODUCTION

The use of solids for removing substances from either gaseous or liquid solutions
has been widely used since biblical times. This process, known as adsorption, involves
simply the preferential partitioning of substances from the gaseous or liquid phase onto
the surface of a solid substrate. From the early days of using bone char for decolorization
of sugar solutions and other foods, to the later implementation of activated carbon for
removing nerve gases from the battlefield, to today’s thousands of applications, the
adsorption phenomenon has become a useful tool for purification and separation.

Adsorption phenomena are operative in most natural physical, biological, and
chemical systems, and adsorption operations employing solids such as activated carbon
and synthetic resins are used widely in industrial separations and for purification of
waters and waste-waters.

The process of adsorption involves separation of a substance from a fluid phase
accompanied by its accumulation or concentration at the surface of a solid phase. The
adsorbing phase is the adsorbent, and the material concentrated or adsorbed at the surface
of that phase is the adsorbate(Suzuki et a1). Adsorption is thus different from absorption,
a process in which material transferred from one phase to another (e.g. liquid)
interpenetrates the second phase to form a "solution". The term sorption is a general
expression encompassing both processes. Physical adsorption is caused mainly by van

der Waals forces and electrostatic forces between adsorbate molecules and the atoms

which compose the adsorbent surface. Thus adsorbents are characterized first by surface
properties such as surface area and polarity(Slejko et al).

A large specific surface area is preferable for providing large adsorption
capacity, but the creation of a large internal surface area in a limited volume inevitably
gives rise to large numbers of small sized pores with increased diffusivity resistance. The
size of the nricropores also determines the accessibility of adsorbate molecules to the
internal adsorption surface, so the pore size distribution of micropores is another
important property for characterizing adsorptivity of adsorbents. Especially materials
such as zeolite and carbon molecular sieves can be specifically engineered with precise
pore size distributions and hence tuned for a particular separation.

Surface polarity corresponds to affinity with polar substances such as water or
alcohols. Polar adsorbents are thus called "hydrophillic" and alurninosilicates such as
zeolites, porous alumina, silica gel or silica-alumina are examples of adsorbents of this
type(Ruthven et al). On the other hand, nonpolar adsorbents are generally "hydrophobic".
Carbonaceous adsorbents, polymer adsorbents and silicalite are typical nonpolar
adsorbents. These adsorbents have more affinity with oil or hydrocarbons than water.

This thesis looks at some principles and considerations for separation processes
using adsorption. Chapter 2 of the thesis involves a comprehensive study of the
adsorption behavior of benzaldehyde, benzoic acid and benzyl alcohol from water onto a
synthetic resin. Their equilibrium adsorption behavior and their breakthrough behavior
are studied as single components and also as a mixture.

The third chapter is about the experiments and subsequent changes done to the

process of manufacturing benzaldehyde from cherry pits. The process holds significance

because the natural cherry flavoring is much more valuable than the artificial flavoring.
Various experimental variables such as water to pit ratio, physical condition of pits,
temperature of hydrolysis were studied and modified. Several modifications were made
to the filtration and the regeneration process. The desorption patterns were also studied

and efforts were made to improve the desorption yields.

REFERENCES

Ruthven, D.M.; Principles of Adsorption and Adsorption Processes, Wiley Interscience,
New York, 1984

Slejko, F.L.; Adsorption Technology, Marcel Dekker, New York, 1985.

Suzuki, M.; Adsorption Engineering, Elsevier,Amsterdam, 1990.

Chapter 2:
ADSORPTION BEHAVIOR OF
BEN ZALDEHYDE, BENZYL ALCOHOL AND
BENZOIC ACID

2.1 Introduction to Adsorption Isotherms

Brunauer et al.(l940)classified adsorption equilbria into five types as shown in
Figure 2.1. Type I (“favorable”) and Type III (“ unfavorable”) are concave downward
and upward, respectively, while the remaining three types are of an inflecting type. The
type I isotherm represents monolayer adsorption and also applies to microporous
adsorbents with small pore sizes. Adsorbents with type II or III isotherms are
characterized by a wide range of pore sizes such that adsorption may extend from
monolayer to multilayer and ultimately to capillary condensation. An isotherm of type IV
suggests that adsorption causes the formation of two surface layers while type V isotherm
behavior is found in the adsorption of water vapor on activated carbon. Linear isotherms
are usually identified by their slope, which equals the Henry constant, H. Basically, all
the types of isotherms behave as linear isotherms at sufficiently low concentration.
Linear isotherms are linear for a limited concentration range.

Isotherm equations can be derived using the thermodynamic approach or the
kinetic approach or using the potential theory or capillary condensation theory.
Expressions from the Gibbs Isotherm Equation and the Vacancy Solution Theory are
based on the thermodynamic approach while Langmuir expression, Freundlich expression

and Langmuir-Freundlich equations are all based on kinetic theory approach. Although

the Langmuir equation can be derived thermodynamically or from a statistical approach
(Ruthven 1984), this expression is commonly derived through a kinetic approach. The

BET equation for multilayer adsorption is also based on the kinetic theory approach.

 

 

 

 

 

 

 

 

 

 

C 1.0 C 1.0
Figure 2.1 Bnmaner’s classification of adsorption isotherms, showing amatmt adsorbed

versusnonmlisedconeenu'ationintheoutlet.

2.2.1 The Langmuir Model.

The earliest model of gas adsorption was suggested by Langmuir. The model is
limited to monolayer adsorption. It is assumed that gas molecules striking the bare
surface have a given probability of sticking, i.e. adsorbing. Molecules already adsorbed
similarly have a given probability of leaving the surface, i.e. desorbing. At equilibrium a
steady state exists in which as many molecules desorb as adsorb at any time. The
probabilities are related to the strength of the interaction between the adsorbent surface
and the adsorbate gas.

This model leads to the following isotherm:

 

0 = fl— : hp 2.1
gm 1 + bP
where 0 = fractional coverage, i.e. the fraction of the maximum coverage possible.

q = volume of gas adsorbed at pressure P usually expressed in cm’/ g at STP,

qm = volume of the maximum gas adsorbed usually taken to be a monolayer,

b = a constant characteristic of the system. b is related to the strength of the

interaction between the adsorbing gas and the surface.

As the strength of the interaction between the adsorbent (the surface of the solid)
and the adsorbate (the gas adsorbing on the surface) increases the value of b increases
and the surface coverage increases faster as the pressure is increased.

In practice it has been found that the Langmuir model is rarely a useful model to
calculate the surface area from gas adsorption data. The Langmuir model is useful when
there is a strong specific interaction between the surface and the adsorbate so that a single
adsorbed layer forms and no multi-layer adsorption occurs. Strongly held adsorption from

solution may fit the Langmuir model.

2.2.2 The BET Model.

Brunauer, Emmett and Teller developed several models of gas adsorption on
solids, which have become the effective standard for surface area measurements.

The models were generalisations of Langmuir theory monolayer adsorption to multilayer
adsorption.

The assumptions underlying the simplest BET isotherm are: Gas adsorbs on the
flat, uniform surface of the solid with a uniform heat of adsorption due to van der Waals
forces between the gas and the solid. There is no lateral interaction between the adsorbed
molecules. After the surface has become partially covered by adsorbed gas molecules
additional gas can adsorb either on the remaining free surface or on top of the already
adsorbed layer. The adsorption of the second and subsequent layers occurs with a heat of
adsorption equal to the heat of liquefaction of the gas. There is no limit to the number of
layers which can adsorb.

Other isotherms developed by Brunauer, Emmett and Teller were based on more
complex models which included the assumptions that the thickness of the adsorbed layers
cannot exceed some finite number of layers n and the second adsorbed layer has a heat of
adsorption intermediate between the first layer and the heat of liquefaction.

The standard 2-parameter BET isotherm based on the simplest BET model may
be written in various ways. The isotherm form, which gives the amount of gas adsorbed

as a function of the relative pressure of the adsorbing gas is:

_q__ a(P/P,) 22
q... '(l-P/P.)(1+(a-1)(Ple» ‘

 

Of

i = 0“ 2 3
q", (l-x)(l+(a-l)x) '

 

where, q = Volume of gas adsorbed at pressure P
qm = Volume of gas which could cover the entire adsorbing surface with a
monomolecular layer
P8 = Saturation pressure of the gas, i.e. the pressure of the gas in equilibrium with
bulk liquid at the temperature of the measurement.
x = P/Ps = relative pressure.
or = a constant for the gas/solid combination.

The constant at is related to the difference between the heat of adsorption of the
first layer (q: ) and the heat of liquefaction (q; ) in the form

0‘ = “PKG: ~q2 )/RT} 14

or (q; -q2)= RT ln 0t

q,“ - q; is also known as the net heat of adsorption.

R = gas constant (8.31447 J K" moi")

T = temperature (K)

2.2.3 F reundlich Model.
The Freundlich isotherm, frequently described as the classical equation, is widely

used, particularly in the low to intermediate pressure range. It is expressed as

q=rxpfii 25
where q = volume of gas adsorbed at pressure P usually expressed in cm’lg at STP,

K = a constant characteristic of the system.

n: another constant restricted to values greater than unity.
The Freundlich equation, in many ways is the simplest equation for data representation.
The bi ggest limitation of the Freundlich model is, it fails to observe Henry’s law behavior

in the limiting situations of P——>0.

2.2.4 Langmuir-Freundlich Expression.
The Langmuir-Freundlich expression combines the Langmuir and Freundlich

Equations and is given as

%
9=JL= bP 26
q.

l+bP%
This equation follows the same asymptotic behavior as the Langmuir equation as

P—>oo. And the expression translates to Langmuir equation at n=1.

2.3 Isotherm Expressions for Liquid Adsorption

In contrast to gas phase adsorption, the density of pure adsobate in liquid phase is
essentially invariant. For liquid systems, the term singleocomponent adsorption isotherm
refers to the adsorption of a single adsorbate from liquid solutions in which the activity of
the solvent is constant. Giles et al.(l960, 1962) examined several liquid adsorption
isotherms and classified them into four categories- S, L, H and C types with subdivisions
for each time. Their classification is based on the initial curvature of the isotherm curve at
the origin. The S type is convex and the L type is concave which correspond to types [[1
and 1 respectively, in the BET classification for gas adsorption isotherms. The C type
exhibits linear behavior at least part of the adsorption range, while the H-type isotherms
show a strong preferential adsorption of the adsorbate and are steep at low concentration.

Most of the gas phase isotherm expression can be extended to liquid systems by
replacing the pressure term with concentration and with corresponding changes in the
units of the various parameters. In liquid phase adsorption, it is not easy to assume a
monolayer coverage as the adsorbed molecules are not necessarily tightly packed with
identical orientation. This and other complications such as the presence of solvent
molecules and the formation of micelles from adsorbed molecules make the liquid phase
adsorption much more complex than the gas phase adsorption.

The following isotherm expressions can however be used for liquid phase

 

 

adsorption.
Linear Isotherm q = Kc 2.7
Langmuir Isotherm _q_ = bc or q = 2.8
q 1 + be 1+ be

10

 

 

 

 

 

 

. I/ / /
r r /
if /

4 I

 

 

 

 

 

—> —»
Concentration of Solute in Solution

Figure 2.2 Classification of isotherms for adsorption from solution.

11

Freundlich Isotherm q = K(c)% 2.9

. . q be};
Langmurr-Freundlrch Isotherm - -

_. 2.10
q,,, 1+ beX

 

2.3.1 Benzaldehyde, Benzyl-alcohol and Benzoic Acid Liquid Phase Adsorption

Equilibrium adsorption behavior of benzaldehyde, benzyl alcohol and benzoic
acid at room temperature was studied in a liquid phase adsorption from water. The initial
liquid concentrations were varied as well as the quantities of SP-850 resin. The
procedure followed in finding the equilibrium adsorption behavior and the raw data is
given in Appendix C-2. All the experiments were done at room temperature.

The data were fitted using a C-H- code to find the best fit for linear model,
Langmuir model, Freundlich model and the Langmuir-Freundlich model. The
concentration of the adsorbate in the solution at equilibrium was provided in moi/m3
while the amount of adsorbate adsorbed was given in gm adsorbate/gm resin. Since the
code was written in C++, which is not a very powerful language for mathematical
calculations, the code fails for extremely low concentrations in order of E8. The
concentrations of benzaldehyde, benzyl alcohol and benzoic acid in the feed for single
component and multicomponent fixed bed adsorption experiments were around 0.95
moi/m3 , 0.45 mol/m3 and 0.15 mol/m3 respectively. The linear adsorption model was
made to fit data points around these concentrations. The least square method for linear
model was applied for data points up till 1 mon3, l mon3 and 0.25 mon3 for
benzaldehyde, benzyl alcohol and benzoic acid respectively. The fits are summarized in

Figures 2.3 to 2.14.

12

gm benzaldehyde! gm resin

 

0.16

0.14 -

.0
d
N

P

.0.
1

O

p
8

p
8

 

0 Experimental data
-—- Linear fit

9
2
o
e

 

 

 

0.02 - :

 

 

 

O V I T I I I V
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4

Concentration (mollmca)

Figure 2.3 Benzaldehyde Equilibrium Adsorption Isotherm and Linear Fit

13

gm benzaldehyde! gm resin

 

0.14

 

 

 

 

 

 

 

o
O
0.12 4
o
O
0.1 q .
0.08 -
. O
0.06 ~
9 Experimental data
0-04 ‘ -- Langmuir tit
0.02 4
o I 1 V T I T r I U
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5

Concentration (mollm‘3)

Figure 2.4 Benzaldehyde Equilibrium Adsorption Isotherm and Langmuir Flt

14

5.0

 

0.16

0.14 4

0.12 *

  

.0

..A

1
O

p
8

 

0 Experimental data
——- Freundlich fit

gm benzaldehyde! gm resin
8
Q

 

 

 

 

 

 

0.04 r
0.02 -
0 j r I r v T r f r
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5

Concentration (mollm‘3)

Figure 2.5 Benzaldehyde Equilibrium Adsorption lsotherrn and Freundlich Fit

15

gm benzaldehyde] gm resin

 

0.16

0.14 -

0.12 i

0.1 a

0.08 ~

0.06 a

0.04 '1

0.02 ~

 

 

 

0 Experimental data
--- Langmuir-Freundlich fit

 

 

 

 

0.0

0.5

1.0

1.5

I

2.0

I

2.5

T

3.0

Concentration (mollmca)

3.5 4.0 4.5

5.0

Figure 2.6 Benzaldehyde Equilibrium Adsorption Isotherm and Langmuir-Freundlich Fit

l6

gm benzyl alcohol! gm resin

 

0.08

0.07 -1 '

0.06 1

0.05 .

0.04 .

0.03 4

0.02 -l

0.01 1

 

O

 

 

0 experimental data
--- Linear tit

 

 

 

 

0.0

A
V] j

I V Y

0.2 0.4 0.6 0.8 1.0 12 1.4
Concentration (mollm‘3)

1.6 1.8

Figure 2.7 Benzyl Alcohol Equilibrium Adsorption Isotherm and Linear Fit

17

2.0

 

0.08

0.07 ~

,0
8

p
8

gm benzyl alcohol! gm resin
9 o
8 2

p
8

0.01 .

’0

 

 

0 Experimental data
-—- Langmuir iii

 

 

 

 

 

0.0

I fi I I

0.5 1.0 1.5 2.0 2.5
Concentration (mollm‘3)

3.0

3.5 4.0

Figure 2.8 Benzyl Alcohol Equilibrium Adsorption isotherm and Langmuir Fit

18

4.5

gm benzyl alcohol! gm resin

0.08

0.07 <

0.06

0.05

0.04 ~

0.03

0.02 -

0.01 1

 

0

1

i

 

 

0 Experimental data
-- Freundlich fit

 

 

 

 

0.0

r

0.5 1.0 1.5 2.0 2.5
Concentration (moi/m’ia)

Figure 2.9 Benzyl Alcohol Equilibrium Adsorption Isotherm and Freundlich Fit

19

3.0

3.5 4.0

 

4.5

gm benzyl alcohol] gm resin

 

0.08

0.07 .

.o
8

 

 

 

 

 

 

 

0.05 -
O
0.04 -
0.03 - o
0 Experimental data
0.02 ~ 0 —-Langmuir-Freundlich iii
0
0.01 « . ’
O I’ 1 I T I V I 1
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5

Concentration (mollm‘3)

Figure 2.10 Benzyl Alcohol Equilibrium Adsorption Isotherm and Langmuir-Freundlich Fit

20

gm benzoic-acid! gm resin

 

0.04

0.035 r

0.03 4

0.025 e

0.02 4

0.015 .

0.01 1

0.005 ~

 

. 0 Experimental data
-- Linear fit

 

 

 

 

 

 

I T I

0.05 0.10 0.15 0.20 0.25 0.30 0.35
Concentration (mollm'ta)

Figure 2.11 Benzoic Acid Equilibrium Adsorption Isotherm and Linear Fit

21

0.40

gm benzoic-acid! gm resin

 

0.05

0.045 -

0.04 '1

0.035 -

0.03 ~

0.025 -

0.02 -

0.015 -1

0.01 -1

 

 

0 Experimental data
--— Langmuir tit

 

 

 

 

 

0.0

I I I I I I I I I

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
Concentration (mollm‘3)

Figure 2.12 Benzoic Acid Equilibrium Adsorption Isotherm and Langmuir Flt

22

1.0

gm benzoic-acid] gm resin

 

0.05

0.045 -

0.04 -

0.035 1

0.03 ~

0.025 -

0.02 a

0.015 .

0.01 4

0.005 ~ ’

 

 

0 Experimental data
-- Freundlich tit

 

 

 

 

 

I I I

0.1 0.2 0.3 0.4 0.5 0.6
Concentration (mollmca)

Figure 2.13 Benzoic Acid Equilibrium Adsorption Isotherm and Freundlich Fit

23

0.7

0.8 0.9

1.0

gm benzoic-acid! gm resin

 

0.05

0.045 4

0.04 -

0.035 .

0.03 ~

0.025 -

0.02 a

0.015 ~

0.01 a

0.005 -

 

 

 

0 Experimental data
--Langmuir-Freundlich tit

 

 

 

 

0.0

I I I I I

0.1 0.2 0.3 0.4 0.5 0.6
Concentration (mollm’ia)

0.7 0.8 0.9

Figure 2.14 Benzoic Acid Equilibrium Adsorption Isotherm and

Langmuir-Freundlich Flt

1.0

The concentrations were provided in mol/m3 and the adsorbed amount in mass
unit adsorbate per mass unit adsorbent. The constants corresponding to the best fits are
tabulated below. The mean error for all the fits are calculated. The mean error(e) is given

by the following equation :
2
. = .1. ZFLL]
n X C
where, e: Dimensionless mean error for every data point.

n: Number of data points
XE=Experimental Value of Amount adsorbed for a particular y

XC=Calculated Value of Amount adsorbed for the same y

Table 2-1: Results of Best Linear Fit.

 

 

 

 

 

Compound K(m3/mol) 8(unitless)
Benzaldehyde 0.095 0.140
Benzyl-alcohol 0.031 0.084
Benzoic-acid 0.095 0.070

 

 

 

 

Table 2-2: Results of Best Langmuir Fit.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Compound a(m3/mol) b(m3lmol) 8(unitless)
Benzaldehyde 0.165 1.036 0.014
Benzyl-alcohol 0.036 0.270 0.032
Benzoic-acid 0.094 1.040 0.005
Table 2-3: Results of Best Freundlich Fit.

Compound K((m3/mol)“) n(unitless) 8(unitless)
Benzaldehyde 0.077 2.393 0.048
Benzyl-alcohol 0.029 2.234 0.041
Benzoic-acid 0.048 1.558 0.006

 

 

 

 

25

 

 

Table 2-4: Results of best Langmuir-Freundlich Fit.

 

 

 

 

 

Compound a((m3/mol)“) b((n?/mol)“) n(unitless) 8(unitless)
Benzaldehyde 0.175 1.102 0.986 0.013
Benzyl-alcohol 0.037 0.270 1.008 0.031
Benzoic-acid 0.101 1.1 12 0.983 0.004

 

 

 

 

 

 

Comparing the mean error between the experimental values and the calculated
values, the Langmuir fit was found to be the best fit for equation up to two constants for
all the three compounds. Langmuir-Freundlich isotherm provided almost the same model
as Langmuir Isotherm. The value of n for Langmuir-Freundlich model for all the three
components was almost unity. And with n as l, the Langmuir-Freundlich model reduces
to a Langmuir model. It can be realized from equation 2.8 that

qm=alb 2.11

Comparing the Langmuir constants obtained by fitting, the values of qm were
found to be 0.159, 0.133 and 0.090 g adsorbate/g adsorbent for benzaldehyde, benzoic
acid and benzyl alcohol respectively. Since qm is a measure of the amount of adsorbate
needed for saturation of the adsorbent, it can be understood that the resin adsorbs a lot
more of benzaldehyde than benzyl alcohol and benzoic acid for a saturated coverage,

when adsorbing from aqueous solutions.

26

2.3.2 Dynamic Behavior of Fixed Bed Adsorption of Single Component.

Generally, practical applications of adsorption for separation and purification are
carried out in the fixed bed mode. The type of adsorption processes carried out in fixed
bed operations include saturation, desorption and their combinations. The important
feature of the dynamic behavior of fixed bed adsorption is the history of effluent
concentration. The effluent history is generally depicted as concentration-time curve.
This concentration-time curve is known as the breakthrough curve. The time at which the
effluent concentration reaches a particular threshold value, which makes it impractical to
continue further is known as the breakthrough time. Accurate predictions of the
breakthrough curves are essential for the rational design of the adsorption system.

Breakthrough behaviors of aqueous benzaldehyde, benzyl alcohol and benzoic
acid were studied on SP-850 resin. The resin was washed with 20 % ethanol solution in
water and then washed with water. The feed solution containing a single adsorbate was
passed through the resin bed at a superficial velocity of 2.7 cm/min. A detailed procedure
is in Appendix 03. Breakthrough patterns for benzaldehyde, benzyl alcohol and benzoic
acid were followed for different amounts of resin, different heights of bed and different
flow rates. However, the superficial velocity Us of the feed flowing through the bed was
kept constant. The superficial velocity is a velocity averaged over the entire cross section
of the preform. In a one-dimensional flow, superficial velocity is defined as:

Us = Volumetric Flow Rate/ Cross Section Area of Bed
The superficial velocity is related to the interstitial velocity, Uz by U s = iSUz , where is is

the void fraction ratio of the adsorbent.

27

2.3.1.1 Prediction of Breakthrough Curves using Linear Adsorption Iisotherms.

Hougen and Marshall (1947) deve10ped a model to predict the breakthrough
curves for cases in which the mass transfer of the solid adsorbent was controlled by a
fluid phase mass transfer coefficient. Their model was based on the assumptions that the
equilibrium between the fluid and solid phases could be expressed by a linear isotherm
equation.

The general mass balance for solute A through a differential column section
neglecting the transport by axial diffusion can be given as:

(rate of A in) — (rate of A out) = (rate of A accumulation)

 

maul)“ - 55(CAU2)2+A2.1 = SAZKEa ac ’42] +-[(1 5%] ] 2.12
Where,

2: Bed height, m.

8: Void fraction ratio for the SP-850 resin.

Ur: Interstitial velocity, m/min.

S=' Cross sectional area of the column, m2

t: Time, minutes.

CA: Concentration of adsorbate in the fluid phase, moi/m3

C As: Average concentration of adsorbate in the solid phase, moi/m3

Applying the limiting process for AZ after dividing equation 2.12 by £SAZ gives

3.0;?) {Lg ] [lg-effi—g—t ] 2.13

28

 

 

Introducing p, = pl, /(1— 6) and 4.1 = C A, l ,0, into equation 2.13 and assuming that the

fluid content of the bed is small compared to the total volume of fluid throughput, the

mass balance for the bed can be expressed by a partial differential equation:

—£UZ{BCA = i]; 2.14
p, iaZ , a: z

 

where,

pb= Bulk adsorbent density, kg/m3.
p5: Skeletal adsorbent density, kg/m3.

q ,1 =Equi1ibrium uptake, g adsorbed/g resin.

The change in the adsorbate content of the adsorbent and the fluid can be given by the

rate expression.

an Kfa ‘
— =— C —C 2.15
[ a: )2 p, ( A A)

While the equilibrium fluid-phase concentration for a linear isotherm can be related to

the adsorbate concentration in the solid by the following equations.
0.. = K DC; 2.16
q: = K D C 10 2.17
Where,
Kfa= Rate constant, (min’l)(m3 adsorbate/m3 bed).

Kn: Distribution coefficient, m3/mol.

CA0: Concentration of adsorbate in the inlet feed stream, moi/m3

29

q: =Saturation capacity of the bed corresponding to concentration CA0 of the

adsorbate in the influent
C; =Concentration of A in the fluid phase that is in equilibrium with uptake q A in

solid.

The entire model can be expressed in terms of dimensionless variables:

 

 

 

K

r: ’0 r-1 2.18
Kpr U2

Y=CA 2w
C10
ZK

g: I“ 2.20
5U

Where,
C: Dimensionless bed length parameter.

1:: Dimensionless time parameter.

X = Dimensionless concentration parameter.

A solution can be obtained using the Laplace transforms. The solution is

1'

3? = 1- ie"‘*"J,(i,/4r;)d; = 7(r,;) 2.21

0
Where,

Jo = Zero Order Bessel solution of the first kind.

30

An approximation of the 7(7, {) function for large values of 1.’ and C was given by

Thomas in 1944. The approximation in terms of the error function is given as under.

2.22

 

_ 1 e-(Jf—F)’
J(§9T)='2— l-e’f(JZ-—J;)+J77(J;+W)]

The above equation is accurate to within 1% when {I .>_ 36 (Vermeulen et al.). When
{2 2 3600 , the last term may be neglected.

The procedure followed in performing experiments for studying the breakthrough
behavior is given in Appendix C-3. 0,, oh and void ratio 8 were calculated by performing

experiments shown in Appendix C-l. Table 2.5 summarizes the experimental conditions.

Table 2-5. Summary of Experimental Conditions for Single Component Adsorption.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

N 0 Compound Z(m) CA0 3 Diameter of Amount of Figure
("ml/m ) bed(m) resin(ml) no
1 0.0635 0.83 0.0222 13.0
2 Benzaldeh de 0.0853 1.13 0.0222 18.0
3 y 0.1219 1.05 0.0603 172.0 2.15
4 0.1117 1.09 0.0603 152.0
5 0.0508 0.63 0.0222 1 1.0
6 0.0660 0.48 0.0222 13.5
7 Benzyl 31mm] 0.0812 0.48 0.0603 112.0 2'16
8 0.0838 0.48 0.0603 125.0
9 0.0508 0.26 0.0222 1 1.0
10 Benzoic acid 0.0808 0.26 0.0222 16.5 2.17
11 0.1219 0.19 0.0603 170.0
12 0.0750 0.19 0.0603 100.0

 

The superficial velocity, Uz was kept constant at 2.7 cm/min for all the runs. The
experimental data obtained by running the synthetic feeds containing benzaldehyde,

benzyl alcohol or benzoic acid were analyzed using the Hougen and Marshall model. The

31

 

experimental data was entered in a Microsoft® Excel sheet along with the equations and

was fitted by iterating the parameter Kfa. The results are shown in Figures 2.15, 2.16 and

2.17.

32

 

 

 

 

 

 

 

 

 

12
1 .
0.8 ~
0 1-Experimentaldata
8 06- 16alculateddata
D ' A 2-Experimental data
------ 2-Calcuiateddata
- 3-Experimentaldata
04 ‘ ----3—Calculateddata
' I 4-Erperimerrtal data
--- 4-Caiculated data
0.2 ~
0 iii-I ‘ ' r
0 200 1200

 

Figure 2.15 Benzaldehyde Breakthrough Behavior using Unear Adsorption Isotherm.

33

1.2

 

 

o 5-Experimentai data
5-Calculated data

a 6-Expenmental data
------ 6-Calculated data

- 7-Expenmentai data
- - - - 7-Calculaied data

I B-Experimantal data
- - - B—Caicuiatod (hta

 

 

 

 

 

 

 

 

400

Murine)
Figure 2.16 Benzyl Alcohol Breakthrough Behavior using Linear Adsorption Isotherm.

34

 

1.2

0.8 4 [I ‘o.
I

 

 

 

I/
° II .5
0.4< I] ;
I
l 4
1"!
,‘x
0.2« ’7 I
.0 I I
f" 1
4" a .I'
0 40-43—05—941:- 4—-"'
0 200 400 6(1)
tirna(mlna)

1200

Figure 2.17 Benzoic Acid Breakthrough Behavior using Linear Adsorption lsotherrn.

35

 

0 9—Experimentaldata
-——9—Calcuiatad(hta

a 10-Experimentaldata
------ 10-Calculateddata
- 11-Experinentaldata
----11-Calculateddata

I 12-Experirmntal data
1" - - 12~Calculaied data

 

 

 

It can be realized from equation 2.22 that the value of 7(§,r) becomes 0.5
when; = z . So, the necessary condition to reach a 50 % breakthrough is that both these

values are equal. Comparing equation 2.18 and equation 2.20 for these dimensionless

variables, the time required for a 50 % breakthrough (to; ) can be given as under.

10.5 = “(ll—[1 ‘1' __Kpr] 2.23
E
2

Since 105, is not a function of Kfa , the time required to reach 50 % breakthrough actually
doesn’t change by changing Kfa.

It was observed that Kfa decides the slope of the breakthrough sigmoid curve
while the value of KD decides the time at which the breakthrough occurs. Increasing the
value of KD increases the time interval at which the breakthrough occurs. While, lowering
its value shortens the time period when the breakthrough occurs. The other parameter Kfa
doesn’t have any hold on the time frame but just determines the slope. As the value
decreases, the slope of the sigmoid curve gets closer to 0 while increasing the value
makes it closer to vertical line. And, hence decides the time required for running from a
20 percent breakthrough to a 80 percent breakthrough. However, the sigmoid curve just
keeps pivoting around a 50 % breakthrough time.

Since the time required for a 50 % breakthrough was obtained by experiments, the
value of KD was calculated using equation 2.23 and the value found for each fit is given

in Table 2.6. The values should have been identical to those listed in Table 2.1.

36

Table 2-6: Comparision of KD from Fitting Equilibrium Adsorption Data and Fitting
Breakthrough Behavior Data usirLg Hougen and Marshall Model

 

 

 

 

 

Compound KD (m3/mol) from KD (m3/mol) from
equilibrium data(T able 2.1) breakthrough data
Benzaldehyde 0.095 0.18
Benzyl-alcohol 0.031 0.08
Benzoic—acid 0.095 0.16

 

 

 

 

It was found that the values aren’t identical. However, the breakthrough data required KD

almost twice the value predicted from equilibrium data for all the three compounds. This

might have originated from the inefficiency of the linear model to accurately predict the

breakthrough behavior. The results of the breakthrough data fitted by iterating Kfa and

using the value of KD obtained from the breakthrough data are given below.

Table 27 Results of Data-fitting for Benzaldehyde using Hougen and Marshall Model

 

 

 

 

 

 

 

 

 

 

No Height of bed(m) Diameter of bed(m) Kfa (dimensionless) KD(m3/mol)
1 0.0635 0.0222 25.77 0.18
2 0.0853 0.0222 50.07 0.18
3 0.1219 0.0603 62.78 0.18
4 0.1117 0.0603 63.80 0.18

 

Table 2-8 Results of Data-fitting for Benzyl-Alcohol using Hougen and Marshall Model

 

 

 

 

 

 

No Height of bed(m) Diameter of bed(m) Kfa (dimensionless) Kp(m3/mol)
5 0.0508 0.0222 17.62 0.08
6 0.0660 0.0222 17.46 0.08
7 0.0812 0.0603 13.88 0.08
8 0.0838 0.0603 20.00 0.08

 

 

 

 

 

Table 2-9 Results of Data-fitting for Benzoic acid using Hougen and Marshall Model

 

 

 

 

 

 

No Height of bed(m) Diameter of bed(m) Kra (dimensionless) KD(m3/mol)
9 0.0508 0.0222 7.54 0.16

10 0.0808 0.0222 8.92 0.16

11 0.1219 0.0603 12.76 0.16

12 0.0750 0.0603 1 1.60 0.16

 

 

 

 

 

37

 

 

 

It was observed that while fitting the data with the Hougen and Marshall model,

there was a lot of discrepancy between the experimental data and the model fit. It was

observed that the value of Kfa varied a lot for benzaldehyde. The model fitted well for

benzyl alcohol. This could probably be attributed to the lower value of K0. With the

value of Kfa found by iteration, the product {I was greater than 36 and reached as high

as 36000. But since the value was greater than 36, the 7({,1’) function given by equation

2.22 could be used. The values of {1 and the 7({,r) function for experiment I

mentioned in Table 2.6 are given in Table 2.10.

Table 2-10 Values of {I and 7(4' ,1') Function for Experiment 1 for Single Component

 

 

 

 

 

 

 

 

 

 

 

 

 

Adsorption
Time(min) {I 7“,?)
6 37.44 0.00
30 331.30 0.00
60 504.16 0.00
90 763.45 0.00
120 1022.74 0.00
150 1282.08 0.00
270 2319.17 0.00
390 3356.32 0.17
450 3874.90 0.43
510 4393.47 0.70
630 5344.19 0.95

 

 

 

 

It can be realized from the table that the product {1 becomes greater than 36 in the first 6

rninutes.The term follows the same pattern for all the other experiments. However, all the

fits predicted a higher concentration in the outlet stream towards the end of the run.

38

2.3.1.2 Prediction of Breakthrough Curves using Langmuir Adsorption Isotherms.
The solution given by Thomas (1944) stands out to be most widely cited and
used among all the analytical solutions available for fixed bed adsorption. The solution
was originally developed to describe ion exchange in fixed beds in which the exchange
process is described by reversible second-order kinetics. He developed the solution
assuming surface kinetics as the rate-limiting step. However, I-Iiester and Verrneulen
(1952) showed that the solution could be used for cases in which rate controlling steps
other than surface kinetics applied. For this model, the Langmuir isotherm describes
equilibrium between the fluid and solid phase. The kinetic derivation for adsorption

described by this model can be derived by making a mass balance of the adsorbate.

an a 1
—=k C —- -— 2.24
0! 0|: A(b qA) bqn]

where, q A =Equilibrium uptake of adsorbate on the adsorbent, g adsorbed/g resin.
t= time, minutes.
CA: Concentration of adsorbate in the fluid phase, moi/m3
k.= Adsorption rate constant, m3/mol
a,b=Langmuir constants, m3/mol

Saturation capacity of the bed (q: ) corresponding to concentration CA0 of the adsorbate

in the influent can be given by

q; = ——“CA° 2.25
1+ bC,,0

39

where,

C A0: Concentration of adsorbate in the influent fluid phase mol/ m3

Dividing equation 2.24 by equation 2.25 ,

001.. m: ) 1+ bC a 1

Introducing dimensionless variables 0 ,3? , r * and A, into equation 2.26
93% =Aa[x(1-é)—r*é[1-x)]

where,

Q=quqz
Aa=ka.l+bc"°
b
i=3-
A0
III: 1
l+bCAo

2.26

2.27

2.28

2.29

2.30

2.31

Introducing dimensionless time parameter and dimensionless bed length parameter into

equation 2.27,

40

.32: [)}(1_é)_r*é[1—x]] 2.32
at

 

4' = Z“ — WM“ 2.33
UzCAoe
r = A“ r - i 2.34
2
Where,
Z: Bed Height in m.

e: Void Fraction ratio for the SP-850 resin.

U2: Interstitial Velocity in m/min.

Hiester and Venneulen expressed Thomas’ results in terms of 7 Function. The J

function is defined as shown in equation 2. 20. The dimensionless concentration

parameter is expressed in terms of J function as below

7(T,r*§)

_ - _ - 2.35
J(r.r*§)+ll—Jtr.r*;)exp[(r*-1><r-§)1I

3E:

 

It can be observed that for r* = l , the Thomas model and Hougen and Marshall
soltuions are the same. r * can be 1 only when b is equal to zero. And with b=0, the
Langmuir model is same as the Linear model. The same experimental data tabulated in
Table 2.5 was analyzed using the Thomas model. The experimental data was entered in a

Microsoft® Excel sheet along with the equations and was fitted by iterating the parameter

k, . Results are shown in Figures 2.18-2.20.

41

 

1.2

14

0.8 1

0.4 1

0.2 <

 

 

 

 

 

 

 

/"
/. N— . e 1~Experimentai data
0/ . i-Caicuiated data
/ -' ,’ A 2-Experimentai data
'5 ,' ------ 2-cttlcurareddm
a} I - 3—Experlmental data
I’ - - - - 3-Caicuiated data
.1 r - 4-Experimentai data
: . r/ - - - 4-Caicuiated data
.' r
: I
: r
: r
: r ,
'.‘ ’r I
O I
a ' ,
I
/ ,.-* . .l
, [I .’
O ' I g
: ‘..!v.‘ . I ’r I’l’
Ar” " '. Ari-o‘ r __J
400 600 800 1200

 

42

Figure 2.18 Benzaldehyde Breakthrough Behavior using Langmuir Adsorption isotherm.

1.2

 

 

 

 

o 5-Expertnentaidata
—5-Calcuiateddata
a s-Experimentaldata
------ 6-Calculateddata
- 7-Erperimentaldata
-o--7-Calcuiateddata
0 8-Experimentaldata
---8—Caiculateddata

 

 

 

e , .
o" ’3’.’/’.
a . . ,
A .' ' ,’I
0.8 1 ',' 1’!
' I I
' I
A ' I.II
8 081 [.1/
B . I
e -///
.’ /
.4 I
/
0.4 ~ I’ I
II. I
/ I
' I
.I /
O I,
0.2 '-’
0 r 1
0 250 300 350

Figure 2.19 Benzyl Alcohol Breakthrough Behavior using Langmuir Adsorption lsothenrr.

 

43

 

1.2

 

14

 

e 9-Experimentaidata
—9-Calculateddata
A 10-Experirnentaldata
------ 10-Calcuiateddata
I ii-Experlmentaithta
----11-Caiculateddata
I 12-Expenmentaldata
---12-Cal0ulateddeta

 

 

 

 

 

 

1200

 

Figure 2.20 Benzoic Acid Breakthrough Behavior using Langmuir Adsorption Isotherm.

Results of fitting the experimental data with the behavior predicted by Thomas model are

given in Tables 2.11, 2.12 and 2.13

Table 2-11 Results of Data-fittin for Benzaldehyde using Thomas Model

 

 

 

 

 

 

No. Height of bed(m) Diameter of bed(m) ka(m3/mol),
1 0.0635 0.0222 0.014
2 0.0853 0.0222 0.014
3 0.1219 0.0603 0.012
4 0.1117 0.0603 0.013

 

 

 

 

Table 2-12 Results of Data-fitting for Benzyl-Alcohol using Thomas Model

 

 

 

 

 

 

No. Height of bed(m) Diameter of bed(m) ka(m3/mol)
5 0.0508 0.0222 0.027
6 0.0660 0.0222 0.024
7 0.0838 0.0603 0.020
8 0.0812 0.0603 0.022

 

 

 

 

Table 2-13 Results of Data-fitting for Benzoic acid using Thomas Model

 

 

 

 

 

 

No. Height of bed(m) Diameter of bed(m) ka(m3/mol)
9 0.0508 0.0222 0.022
10 0.0808 0.0222 0.022
11 0.1219 0.0603 0.025
12 0.0750 0.0603 0.017

 

 

 

 

The value of the adsorption rate constant (ka ) found by iteration was almost constant in

all the four experiments for benzaldehyde and benzyl alcohol. There was a little

discrepancy in the iterated results of benzoic acid. This could have resulted from the

inaccurate means of quantification used for benzoic acid. The adsorption rate constant is

a function of temperature and it is inversely proportional to the molecular weight of the

4S

 

 

 

compound. Hence the adsorption rate constant for benzaldehyde, benzyl alcohol and
benzoic acid were expected in a descending order. However, the value of kit found for
benzaldehyde was a little lower than benzyl alcohol. Comparing Figures 2.18, 2.19 and
2.20 to 2.15, 2.16 and 2.17 respectively, it can be realised that the Thomas model could
account for the sigmoidal curve better than the Hougen and Marshall model. Effort was
made to find a constant value of k, for each compound, which would be successful in
predicting the breakthrough behavior for all the breakthrough experiments done with that
compound. By observing the model predictions with different values of k3 within the
range given in Table 2.12, 21.3 and 2.14, it was found that the best overall value of k. for
benzaldehyde, benzyl alcohol and benzoic acid were 0.013 m3/mol, 0.022 m3/mol and
0.022 m3/mol respectively. Figures 2.21, 2.22 and 2.23 illustrate the behavior predicted

by Thomas model with the best overall value of ka for the three compounds.

46

 

1.2

0.8 1

0.5 1

CICo

0.4 1

0.21

 

 

 

01

 

1200

 

 

o 1-Experimentel data
1¢Calcuiated data

A 2-Experimentai data
------ 2-Calcutated data

- 3-Experimentai data
- ° - - {it-Calculated data

- 4-Experimentel data
- - - 4-Calcuiatod data

 

 

Figure 2.21 Benzaldehyde Breakthrough Behavior using Langmuir Adsorption isotherm and

Constant its.

47

 

12

 

0 /
‘ I 9,11
08 d I, I”
I II
' I
/I
A I' I
8 06 a '. a” I/
6 ' 5 ’/
g ' .’ I
N
4 I
" /o’/
0.4 " ‘r . I. I,
i 'l
. //
/ I
' I
I

 

 

 

e semmmmmma
———5Qmmmumh

A eanmmmamm
...... ficabm m

- 7EmMdeMn
-"-4Hmewdm

e Himmmmfim
---acannnem

 

 

 

 

2% up

 

Figure 2.22 Benzyl Alcohol Breakthrough Behavior using Langmuir Adsorption isotherm and

Constantin.

48

 

1.2

 

 

 

1 1
O ‘ I
O .4,
I i .r'
0.8] / f
01/
' r
8 1‘
0.6 ~ -'
8 IL,"
0 I"
O I".
'I.‘
0.4 ‘ I, e.
I 3 .
‘.O. l
‘1’," 'I
A / .° ' I
0.2 . . / ,' /
./,°' e I,
A / .’ z
I I I
l .' ' l
-/<»' ,.»l
0 a"! “‘?“""Z ‘ . 1"" .
O 200 400 600 800 1000
time(rnins)

1200

 

 

e 9-Experimentaldata
9-Caicuiateddata

A to-Experimenraldata
------ to-Caicutateddata
- 11eExpenmentaldata
—---11-Caicuiateddata
I 12-Experimentaldate
---12-Calcuiateddata

 

 

Figure 2.23 Benzoic Acid Breakthrough Behavior using Langmuir Adsorption lsotherrn using

Constant Its.

49

 

2.4 Multicomponent Adsorption

Adsorption exists because of a naturally occuning attractive force between the
molecules on the surface of the adsorbent and those in the adsorbate phase.
Multicomponent adsorption is when each adsorbate molecule competes for and occupies
several sites, whilst interacting with each other. For optimal process design, accurate
adsorption equilibria data for both single and multicomponent systems are required.
Although adsorption isotherm data for pure components are relatively easy to measure,
multicomponent equilibrium data is generally obtained by numerous, elaborate
experiments for the adsorbate, adsorbent and temperature range of concern. Eiirninating
the reliance on multicomponent experimental data seems to be an excellent method for
economic feasibility for most process industries. This has fuelled a maturing interest in
possibility of efficient analytical models, able to predict multicomponent adsorption
equilibria. Of the various theories which have been proposed to describe the physical
adsorption properties of gases and vapours on solids, the Ideal Adsorbed Solution
Theory, developed by Myers and Prausnitz in 1965, appears to be the most successful.

Myers and Prausnitz(l965) proposed the ideal adsorbed solution (IAS) theory for
the adsorption of gas mixtures. The theory was subsequently extended to multicomponent
liquid phase adsorption by Radke and Prausnitz(l972). The IAS theory for gas mixtures
is based on the assumption that the various fundamental thermodynamic equations of
liquid are applicable to the adsorbed phase and that the solution in the adsorbed phase is
ideal, such that Raoult’s law is valid. Hence, applying the criterion of equilibrium

between the gas and adsorbed phase gives the following relation

50

Pyl. = p?(7t,.,T)x,.. i=l,2, ...... ,N 2.36
where
P=Total pressure of the gas phase.

xi. =Mole fraction of the ith adsorbate in the adsorbed phase
yi. =Mole fraction of the ith adsorbate in the gas phase.
p? =Hypothetical pressure of the ith adsorbate in its pure component adsorption
that yields a spreading pressure It
71.. =Spreading pressure equal to that of the multicomponent adsorption
case(72,. :71]. = 72)

The spreading pressure can be given by Gibbs adsorption isotherm as under
P.” 0
fl". = R_T 11%,? 2.37
A 0 P1

where 0: Pure component adsorption state
q? =Equilibrium adsorbed phase concentration in the pure component adsorption

of the ith adsorbate.

The total adsorbed phase concentration can be calculated by

-1
q, = [i x’ ] 2.38

j=l 93
The amount of adsorption of the ith adsorbate is given as

‘11 = qrxt' 2.39

51

Equations 2.36 , 2.37, 2.38 and 2.39 constitute the system of equations describing the
multicomponent adsorption equilibrium between the gas phase and the adsorbed phase.

The theory extended by Radke and Prausnitz for multicomponent liquid phase adsorption
revolves around the same system of equations. The system constitutes equation 2.38, 2.39

and equations 2.36 and 2.37 with the pressure term replaced by concentration giving

c, = c?(7:i,T)x,.. 2.40
c? 0
7:, = 51[ jq—gdcf] 2.41
A O c,
Intrapellet Mass Transfer.

The linear driving force model for intrapellet mass transfer is based on the
postulate that the uptake rate of adsorbate by a pellet is linearly proportional to a driving
force, defined as the difference between the surface concentration and the average

adsorbed-phase concentration.

 

ch} 150 -

— = ' - 2.42
dt a; q. q

where,

q =Average adsorbed phase concentration at the exterior surface of the pellet, mol
adsorbate/kg resin.

qx =Adsorbed phase concentration of the pellet, mol adsorbate/kg resin.
D. =Effective intrapellet diffusivity, mzlsec.

a, =Particle radius, m.(4.25E-4, from product literature, Mitsubishi Chemicals)

52

Interphase Mass Transfer
The transport of the adsorbate species from the bulk of the fluid phase to the

external surface of adsorbent pellets is given by the following equation.

 

' 3k
5’- = f c, — c, 2.43
dt app},
where,

cb =Adsorbate concentration in the bulk of fluid, mol/m3.

c, = Adsorbate concentration at the fluid-pellet interface, moi/m3.
pp =Pellet density, kg/ m3.

k f =Interphase mass transfer coefficient, m/sec.

Experiments were done to study multicomponent adsorption. A solution of
benzaldehyde, benzyl alcohol and benzoic acid was passed through a fixed bed of SP-850
resin. The procedure observed in doing these experiments and the raw data is given in the
appendix. A FORTRAN code by Moon (1987) was used to predict the multicomponent
breakthrough behavior, which could accept the equilibrium behavior represented in terms
of the piecewise Freundlich equation. However, the one-piece Freundlich equation as
given by Equation 2.9 was used for all the analysis, while the values for Freundlich
constants were taken from Table 2.3. These values predicted a very high adsorbate
capacity. Hence, the equilibrium adsorption data for the concentration range of interest
was fitted to get the Freudlich constants. The experimental and the predicted results are
shown in Figures 2.24 and 2.25 respectively, while the Freudlich constants and the

concentration range of interest are tabulated in Table 2.14.

53

0.08
0.07 a Q 2 O

 

.o
8

0.05 ~

g:
E

 
   

0.03 ~ 0
0.02 -
0.01 ~

 

 

gm benzaldehyde/gm

O

 

I

0 0.2 0.4 0.6 0.8 1
Concentration, moilm"3

Figure 2.24 Prediction of Equilibrium Adsorption Data of Benzaldehyde by Freundlich
Model for 0-0.9 mol/m3 Concentration Range

0.03
0.025 - ’
0.02 -
0.015 ~
0.01 ~
0.005 .

 

 

 

 

gm benzyl alcohol/gm resin

 

I T

0 0.2 0.4 0.6 0.8 1
Concentration, movmca

Figure 2.25 Prediction of Equilibrium Adsorption Data of Benzyl Alcohol by Freundlich
Model for 0-0.5 mol/m3 Concentration Range

54

Table 2.14 Summary of Freudlich Parameters used for Multicomponent Adsorption

 

 

 

Compound Concentration Range, mol/m3 K((m3/mol)“) n(unitless)
Benzaldehyde O-O.9O 0.054 2.43
Benzyl Alcohol 0-0.50 0.021 2.20

 

 

 

 

 

 

In this program, the linear driving force approximation is used and the dispersion
effect is neglected. The multicomponent adsorption equilibrium is assumed to obey the
IAS theory. Adsorption equilibrium is assumed at the fluid-pellet interface as a result of
the fact that surface diffusion is the rate controlling mechanism in intrapellet mass
transfer. The governing equations for this program are obtained by extending equation

2.14 for N adsorbates and by equating equation 2.42 and 2.43

 

 

E; LC» = 9—3:- 2.44
pb 82 , a: z

' 3k -
51—2: I [Cb—CS): 152D, qs-q 245
d! appp 0,,

Finite difference analysis is used to calculate the adsorbate concentration in the bulk of
fluid and the average adsorbed phase concentration at the exterior surface of the pellet,
corresponding to a particular grid point. These values are used to find the surface

concentration of the adsorbate and the adsorbed phase concentration at the surface.

55

Table 2.15 and 2.16 summarises the experimental conditions. The concentrations of the

adsorbate in the feed are given in Table 2.15 while the other experimental variables are

listed in Table 2.16

Table 2.15 Adsorbate Concentrations in Feed

 

 

 

 

 

 

 

 

 

 

 

 

 

 

No Benzaldehyde cone in Benzyl alcohol cone in Benzoic acid cone in
feed, mol/m3 feed, mol/m3 feed, mol/m3
l 0.90 0.48 0.20
2 0.88 0.46 0.20
3 0.88 0.46 0.20
4 0.90 0.50 0.20
5 0.90 0.50 0.20
6 0.90 0.50 0.20
7 0.90 0.46 0.20
8 0.90 0.46 0.20
9 0.90 0.46 0.20

 

 

The concentration of benzoic acid wasn’t actually measured in all the experiments.
However the amount of benzoic acid added to the feed was constant in all the
experiments. The concentration of benzoic acid was found by making a feed stream with
the same concentration and measuring the pH. However, the concentrations of
benzaldehyde and benzyl alcohol were almost constant for all the experiments. The

variation in concentration was less than 3 % for benzaldehyde and 8% for benzyl alcohol.

56

All the experiments were done in a glass bed of 2.22 cm diameter. The experiments were
carried out at room temperature.

Table 2.16 Summary of Experimental Conditions for Multicomponent Adsorption.

 

 

 

 

 

 

 

 

 

 

No Bed height, m Superficial Amount of Flow-rate, ml/min
velocity, m/sec resin, ml
1 0.075 0.00042 15 10
2 0.075 0.00063 20 15
3 0.075 0.00084 25 20
4 0.106 0.00042 15 10
5 0.106 0.00063 20 15
6 0.106 0.00084 25 20
7 0.121 0.00042 15 10
8 0.121 0.00063 20 15
9 0.121 0.00084 25 20

 

 

 

 

 

 

 

The experimentally observed behavior and the predicted behavior by the program are

given in figures 2.26 to 2.34.

57

 

 

 

 

 

 

 

 

1.8
1.6<
' .....
....... s...
1.41 ' ,'.
12« g; x’ ‘
.' /
’l
/
1 a f. 1”

3 I o Expemrontalemldahydo
a ,’ ‘ - Expedmentalamzyl-alcohol
osl :,’ —-Ca|eulated Benzaldehyde

5,’ ------ Calculated Benzyl-alcohol
‘ ,’ -—- CalculatodBmoic-aeid
0'6 4 ‘ If: A Experimental Benzoic-Acid
/:'
04 ‘ ’ -'
. 1 I ,
l-'
I .'
I, 5
0.2 " I.
/ .'
I I
1’ I
0 . r'iz"";’ ‘ ° Y ‘. r . . .
050100150200250300350400450500

Tlmilnmlnum

Figure 2.26 Breakthrough Behavlor of Mixtures ( 15 ml resln at 10 ml flow rate)

58

 

1.8

 

 

 

 

 

 

 

 

 

‘. I '~
1.4~ 1.
e'... ...‘I
l t.
1.2- : ’,..-—--\
.. /’ “~";~., o Experlnentel' W
:. ’/ :N~~‘~~‘ . Ema . I18 11 '“lnhal
1 . / CalculatedBarzaldehyde
I _ ------ CeleuletedBenzyi-aleohol
o ' --- CalwlatedBenzoie-ecid
g 0.8-
0.6-
0.4 «
02-
0 ;‘ Y a f t . r . a
0 50 100 150 200 250 300 350 400 450 500

Time In minutes
Figure 2.27 Breakthrough Behavior of Mixtures ( 15 ml resin at 15 ml flow rate)

59

 

1.6

1.4 4
1.21
11

g 0.84
0.6 r
0.4 4

0.2‘

 

O
.6

 

 

 

oExperlmentalBenzaldahyde

 

CalculatadBenzaldohyde
------ CaicdatedBenzyl-alcohoi
---CalculatedBenzoic-acld

 

 

 

 

0 so 100 150 200 250 300 350
Time in minutes
Figure 2.28 Breakthrough Behavior of Mixtures ( 15 ml resin at 20 ml flow rate)

CICo

 

1.8

15*

l.2~

0.81

0.6 -

0.4 ~

0.2 -

 

100

 

 

 

200 300 400 500 600
Tlrnelnmlnutee

r

800900

1000

Figure 2.29 Breakthrough Behavior of Mixtures (20 mi resin at 10 mi flow rate)

61

 

 

e ExperinentalBermldehyde

I ExperhieniaiBenzyi-aleohoi
—CaicuiatedBenzaidehyde
------ CaicdatedBenzyl-alcohoi
---CalculaiedBenzoic-ecid

 

 

1.8

 

 

 

 

 

 

 

 

1.6 r
.'
.0. 0‘
1.4 ‘ t.‘
/—-\‘\' .1 e Experimental Balzaidehyde
124 ,z' ‘\“‘..‘ . I ExperinentalBenzyi-aioohol
/’ \‘l..\ ‘——Geicuiated Benzaldehyde
.- / ‘*- ------ Calculated Benzyl-alcohol
:/I
0.8 « f;
.'I
.‘I
I
0.6 . I
I.‘

I;

’e
0.4 '1 I":-

.' o

/ :
I ,’
0.2 < I .'
I, a". .
/
I .’
0 Aaf..’ A A , A . . .
o 100 200 300 400 500 600
Time in minutes

Figure 2.30 Breakthrough Behavior of Mixtures (20 ml resin at 15 ml flow rate)

62

 

1.6

1.4 4
‘l b..-
e'. .. .
a .\
’ I
12 1 .' ,”"--“~ “~.
I, \‘ \.
I ’ ‘\..-
\‘o
// mg~l
I
1 1 I
I

o
g 0.8+

 

 

 

 

0.6 r
0.4 ~
0.2 1
0 v w v . - a. . v . . ,
0 50 100 150 200 250 300 350 400

Time in minutes
Figure 2.31 Breakthrough Behavior oi Mixtures (20 mi resin at 20 ml flow rate)

63

450

 

 

eExperimentelBemldehyde

 

CalculatedBenzeidehyde
------ CalculatedBenzyl-aleohol
---CalcuiatedBenzoie-ecid

 

 

 

 

 

 

 

 

 

 

1.8
. -'| ---.
1.6' ,o" g ' ..~
e . ‘.
o. l 5'
1.4- , //’“__‘\ a,“
‘ ’/ \\‘. e e ExperimentalBenzeidehyoe
1.21 /’ ‘§-.\. . I ExperhientaiBenzyl-aieohol
.‘ /’ '\ CaicuiatedBenzaidehyde
5”, ‘ ------ CalcdatedBenzyl-eioohoi
o 1~ 5’, -——Caicuiated8enzolc-ecid
3 :7
0.31 I:
I
II:
0.61 .‘
I:
I:
l:
0.4- ll:
,2
IE
0.2 l I
I : n
I e
eer’ rJ’-".' ’. fl ‘ . 4 .
200 300 400 500 600 700 800 900 1000
Timeinminutes

 

 

Figure 2.32 Breakthrough Behavior oi Mixtures ( 25 ml resin at 10 mi flow rate)

 

1.8

 

 

 

1.6 < .----.._.
O I .‘o
1.4 < .4 "t,
a; ’--"\ ‘ .. I
I. ’/ ~\\ ..s
1.2 1 e. I, \\..°~
.' I \-
O. / *h
0' I,
14 p /
. I
r I
,- I
as 3/
[.1
t
. l .
06 /
I-
’1':
0.4 < f .
/
02< ,’
I ..
,’ e
o AJITI'. ‘ V— . t Y Y
0 100 200 300 400 500
‘iimeinrnhtme

soo

Figure2.33 Breakthrough Behavioral Mixtures ( 25 ml resin at15 ml flow rate)

65

 

 

 

------ CalalletedBenzyl-alcohoi
---CaieulatedBenmic-aeid

 

 

0100

1.8

 

 

 

0 Experimental Benzaldehyde
I Experimental Benzyl-alcohol
Calculated Benzaldehyde
------ Calculated Benzyioabohoi

 

 

- - - Calculated Benzoic-acid

 

 

 

Figure 2.34 Breakthrough Behavior of Mixtures ( 25 ml resin at 20 mi flow rate)

66

1.6 «
..' --------
'. " 1..
1.4 « , ' .
’.O .i
1.24 :0 ’/’a-"' ~~“‘...
e I \ ‘I.
/ ‘-
.° 1’
1 " ’
0‘ l
.I I
, I
0. [I
0.8 ‘ '.'/
ol
,7
’1
0.6 < I; .
I.
J.'
I I
0.4 ‘ I r. .
, 0
1’
/ :" °
02 “ I .-
l .'
I, '.
I.‘
0 e V ‘é’gf.:x ‘f‘ O I r Y T j
0 50 100 150 200 250 300 350 400 450
Timelnminutea

 

The intrapellet mass transfer coefficients and the effective surface diffusivity of the three
compounds were the only unknowns. Since the molecular size of benzaldehyde, benzyl
alcohol and benzoic acid is quite close, the coefficients are assumed to be same for all the
three compounds. The mass transfer coefficient and the surface diffusivity were not
varied for different superficial velocity. Hence, the experimental data was fitted to the
IAS theory predicted results by iterating the value of these two coefficients. The
intrapellet mass transfer coefficient was found to be 0.2E-4 m/sec and the surface
diffusivity was found to be 0.26E-ll mzlsec. Wang and Tien(1982) studied
multicomponent breakthrough behavior for a system of p-nitrophenol, p-chlorophenol
and propionic acid. They found that the mass transfer coefficient and surface diffusivity
were 0.17E-4 m/sec and 0.2E—11 mzlsec for all the three compounds. Since their obtained
values were of the same order of magnitude as the values obtained for our experiments,
the values seemed to be realistic. However, with the same values it wasn’t possible to
predict the single component breakthrough behavior. The program predicted a very late
breakthrough for all the experiments.

The code was successful in predicting the breakthrough behavior for
benzaldehyde and benzyl alcohol. Since the gas chromatograph couldn’t accurately sense
the concentration of benzoic acid, it was not possible to check the correctness of the
predicted behavior of benzoic acid. However, for Experiment-l, HPIE was used to
check the correctness of the benzoic acid breakthrough pattern predicted by the computer
program. A Waters 600 solvent delivery system equipped with a model 410 refractive
index detector was used for liquid phase analysis of the experimental samples. Data from

the analysis was interpreted with a Waters WIRC software program version 1.0 on a IBM

67

XT. The data was transferred through a Waters System Interface Module (SIM). The
column used in the analysis was manufactured by Alltech Associates. The Alltima
Narrow-Bore Column contained a C-18 silicon bonded phase which was 5 microns in
size. The column was 250 mm long by 4.6 mm diameter and its catalog number was
88371. The column was protected with cartridge columns of the same composition. The
solvent was 15 weight percent methanol, 25 weight percent ethyl acetate in water.

The HPLC results didn’t coincide with the results predicted by the program. The HPLC
analysis showed that benzoic acid started coming out from the start of the run and the
concentration of benzoic acid in the outlet didn’t surpass the concentration of benzoic
acid in the feed even after 7 hours. This can be attributed to the failure of the Freundlich
model to predict the equilibrium adsorption behavior of benzoic acid. Since the HPLC
was not used for all the other experiments, it is impractical to draw a conclusion from a
single experiment.

It was observed that benzoic acid starts coming out first immediately followed by
benzyl alcohol. The outlet concentration of both benzoic acid and benzyl alcohol exceeds
the concentration of the feed, indicating that some other compound is displacing them.
And after lapse of some more time, the concentration in the outlet for benzoic acid and
benzyl alcohol get close to their concentration in the feed. Benzaldehyde is the last
compound to break through. And as the concentration of benzaldehyde in the outlet

equals its concentration in the inlet, there is nothing being adsorbed on the resin.

68

REFERENCES

Brunauer, S., Emmett, H. and Teller, E.,Amer. Chem. Soc., 60, 309-319 (1938).

Brunauer, S.P.,The Adsorption of Gases and Vapors, vol. 1, Physical Adsorption, p.97,
133, Princeton University Press, Princeton, NJ (1945).

Giles, C.H., MacEwan, T.H., Nakhwa, S.N., and Smith, D., J Chem Soc. 3973 (1960).

Giles, CH, and Nakhwa, S.N., J. Appl. Chem. 12, 266(1960).

Hiester, N .K., and Vermeulen, T., Chem. Eng. Prog., 48, 505 (1952).

Hougen, O. A., and. Marshall, W. R., Chem. Eng. Prog., 43, 197 (1947).

Langmuir, J. Am. Chem Soc., 38, 2219,(1916); 40, 1368, (1918).

Moon, H., and Tien, C., AIChE Symposium Series 84(264), 94 (1988a).

Moon, H., and Tien, C., AIChE Symposium Series 84(264), 23 (1988b).

Radke, C.J., and Prausnitz, J .M. 1972. Thermodynamics of multi-solute adsorption from
dilute liquid solutions. AIChE Journal, 184): 761-768(1972).

Polanyi, M., Z. Phys., 2, 111 (1920).

Prausnitz, J .M., and Myers, A.L., Molecular Thermodynamics of Physical Adsorption,
Chem. Eng. Prog., 54, (1964). Rosen, J. B., Ind. Eng. Chem., 46, 1590 (1954

Ruthven, D.M., Principles of Adsorption and Adsorption Processes, Wiley Interscience,
New York (1984).

Thomas, H. C., J. Amer. Chem. Soc., 66, 1664 (1944).

Vermeulen, T., Klein, G., and Hiester, N.K., J.H. Perry, Ed., Chemical Engineers’

Handbook, McGraw Hill, New York, 1973.

69

Chapter 3:
PRODUCTION OF BENZALDEHYDE AND
BENZYL ALCOHOL FROM CHERRY PITS.

The flavor industry uses both natural and chemically synthesized flavors. In the
case of cherry flavor, the use of synthetic benzaldehyde has grown over the years as the
flavor component in candies, desserts and carbonated beverages such as Cherry Coke and
Dr Pepper. While from a technical point of view there may be a marginal difference, the
public perception and desire is changing in favor of all natural ingredients in food and
drinks. The high demand for natural cherry flavor has driven the price of bitter almond oil
and naturally derived benzaldehyde. Because of the limited supply and high price of the
natural product, synthetic benzaldehyde is often used as a substitute. However, the
application of the less expensive artificial flavor has to be indicated, and it is suspected
that this is not always done. Although artificial flavor components and natural flavor
components are identical in taste quality, they can be distinguished by their l“C content
(Krueger, 1987). The natural form of the cherry flavor is the benzaldehyde in bitter
almond oil, available in the pits of cherries, apricots, peaches, cherry laurel and plums.

Recovery of Natural Benzaldehyde is on-going research that was initially started
in 1990. A Phase I USDA grant was provided for developing a process to recover
benzaldehyde and benzyl alcohol from cherry pits. A Phase II grant was provided in
1999 to optimize experimental variables and improve the yield of benzaldehyde and
benzyl alcohol. This work was performed in conjunction with Natura, Inc. The recovery
of benzaldehyde from cherry pits involve several experimental steps. It was essential to

optimize the hydrolysis conditions to get the maximum yield of benzaldehyde . Changes

70

to the filtration step were impending , since the earlier filtration process wasn’t efficient
and took a lot of time. It was essential to understand the kinetics of the adsorption process
to better the yield results. Changes were also made to the process used for the

regeneration of the adsorbent.

3.1 Hydrolysis and Reaction Kinetics

The earlier experiments were carried out using the hydrolysis conditions
suggested in Aaron Soule’s thesis. The pits were dried and were ground up before adding
to water preheated to 50°C. A 5:1 water-to-pit mass ratio was used. The water and pits
were mixed together for 1 hour and then filtered. However, due to discrepancy in the pit
yields in previous work, various tests were performed analyzing optimal hydrolysis
conditions as well as the chemical processes occurring during hydrolysis. Macroscopic
variables such as water to pit ratio, temperature of the water and the condition of the
cherry pits were studied to find the optimum condition for hydrolysis.

Research at Michigan State University showed that benzaldehyde can be
recovered from cherry kernel by hydrolysis (Grethlein et al, 1990) or mechanical
crushing. Mechanical crushing releases oil containing a lot of triglyceride and some
benzaldehyde. The pits are expected to contain quanitities of amygdalin and
mandelonitrile (Li et al., 1992; Zheng and Poulton, 1995), which break down in water
producing cyanide and benzaldehyde. The chemical structures of benzaldehyde,
mandelonitrile, and amygdalin are shown in Figure 3.1. Soule’s work shows that
mandelonitrile and amygdalin breaks down in water to produce benzaldehyde, but

doesn’t affect the concentration of benzyl alochol.

71

 

Amygdalin
fi’ TH
@c-H @f-CN
Benzaldehyde H

Mandelonitrile

Figure 3-1. Chemical Structures of Benzaldehyde, Mandelonitrile and Amygdalin.

Experiments were done to study the optimum hydrolysis temperature. 250 ml of
water was heated to a particular temperature in a three necks, angled standard wall 500
ml glass flask and 50 gms of air-dry ground cherry pits were added and allowed to
hydrolyze for an hour. The concentrations of benzaldehyde and benzyl-alcohol in the
hydrolyzate were analyzed by running the sample on the gas shromatograph. The peak at
3.09 corresponds to benzaldehyde while the one at 4.62 corresponds to benzyl-alcohol.
The retention time depends on the type of column used in the gas chromatograph.

Currently, we are using the EC-Wax column. Specifications of the column are given in

72

Appendix B. With the earlier column, benzaldehyde and benzyl alcohol used to appear at
2.67 and 4.02 respectively. The response factor also changed with the change of column.
The current response factors are 73.97E-6 til/area count and 53.91E-6 til/area count for
benzaldehyde and benzyl-alcohol respectively. The response factors were calculated by
running sample of known concentrations and observing their peak area. The details of the
experiments are given in Appendix B. The results of hydrolysis carried out at different
temperatures are tabulated below.

Table 3-1: Hydrolyzate composition for different temperatures

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Retention Time (min) Peak Area Parts Per Million
Hydrolysis at 45 C
3.09 (benzaldehyde) 1.32 97.64
4.62 (benzyl alcohol) 0.39 21.02
Hydrolysis at 50 C
3.09 (benzaldehyde) 1.43 105.77
4.62 (benzyl alcohol) 0.41 22.10
Hydrolysis at 55 C
3.09 (benzaldehyde) 1.60 1 18.35
4.62 (benzyl alcohol) 0.42 23.04
Hydrolysis at 60 C
3.09 (benzaldehyde) 1.59 117.61
4.62 (benzyl alcohol) 0.43 23.181
Hydrolysis at 65 C
3.09 (benzaldehyde) 1.54 113.71
4.62 (benzyl alcohol) 0.43 23.181

 

Since all the pits were taken from the same bucket, the change in composition can
be attributed to the temperature of hydrolysis. It was observed that the concentration of
benzaldehyde in the hydrolyzate was higher for hydrolysis carried out at 55 C than one
carried out at 45 C or 50 C temperature. Concentration of benzaldehyde and benzyl
alcohol was same for 55 C and 60 C. However, the concentration of benzaldehyde was
marginally less for hydrolysis carried out at 65 C. The higher temperature might be

inciting some side reaction, but, since the difference is so small, it is impractical to draw

73

such conclusions. It was observed that the benzaldehyde concentration of hydrolyzate
from different pit buckets varied a little. But for all the subsequent experiments, the
hydrolysis was done at 55 C.

Experiments were also conducted to optimize the water to pit ratio for hydrolysis.
A particular amount of water was heated to 55 C and then 25 grns of cherry pits were
added for hydrolysis. The following are the results of hydrolysis done with different
water to pit ratio. Yield is given as the total amount of benzaldehyde or benzyl alcohol
recovered from the cherry pits.

Table 3-2: Hydrolyzate composition for different water- dry pit ratio

 

 

 

 

Retention Time Peak Parts Per Amount of hydrolyzate Yield,
(min.) Area Million (ml) ml
WaterzPit 3:1
3.09 (benzaldehyde) 1.89 139.80 58ml(75ml water used) 8.10E-3
4.62 (benzyl alcohol) 0.49 26.41 58ml(75m1 water used) 1.53 E-3

 

WaterzPit 5:1

 

3.09 (benzaldehyde) 1.60 118.35 103ml(125ml water used) 12.19 E-3

 

4.62 (benzyl alcohol) 0.42 22.64 103ml(125ml water used) 2.33E-3

 

WaterzPit 7:1

 

3.09 (benzaldehyde) 1.42 105.03 152ml(175ml water used) 15.96E-3

 

4.62 (benzyl alcohol) 0.33 17.79 152ml(175ml water used) 2.70E-3

 

WaterzPit 8:1

 

3.09 (benzaldehyde) 1.36 100.59 176ml(200ml water used) 17.7OE-3

 

4.62 (benzyl alcohol) 0.28 15.09 176ml(200ml water used) 26.55E-3

 

WaterzPit 10:1

 

3.09 (benzaldehyde) 0.954 70.56 225ml(250ml water used) 15.87E-3

 

 

 

 

 

 

4.62 (benzyl alcohol) 0.20 10.78 225ml(250ml water used) 2.42E-3

 

The total hydrolyzate obtained after hydrolysis is always found to be close to the
total amount of water added minus the amount of cherry pits added to it. This loss of
hydrolyzate is due to the water lost with the wet cherry pits after the hydrolysis. This
trend is also observed in bigger runs. Only 20 L of hydrolyzate was generally collected in

a bigger run with 25 L of water and 5 kgs of cherry pits. Although the ppm concentration

74

 

of benzaldehyde and benzyl alcohol is higher for 7:1 ratio than 8:1 ratio, the yield
obtained from 8:1 ratio is higher than a 7:1 ratio. Earlier, a ratio of 5:1 was used, because
it reduced the amount of hydrolyzate to be processed.

Experiments were performed with wet cherry pits instead of air dry cherry pits.
The same procedure was followed to analyze the effect on the hydrolyzate composition.
The observations are as under

Table 3-3: Hydrolyzate composition for different water- wet pit ratio

 

Retention Time (min.) Peak Parts Per Amount of hydrolyzate, Yield,
Area Million ml m1

 

WaterzPit 3:1

 

3.09 (benzaldehyde) 1.68 124.92 74ml(75ml water used) 9.24E-3

 

4.62 (benzyl alcohol) 0.34 18.32 74ml(75ml water used) 1.35E-3

 

WaterzPit 5:1

 

3.09 (benzaldehyde) 1.43 105.77 123ml(125ml water used) 13.00E-3

 

 

 

 

 

 

 

4.62 (benzyl alcohol) 0.29 15.64 123ml(125ml water used) 1.93E-3
WaterzPit 7:1

3.09 (benzaldehyde) 1.32 97.64 l72ml(175ml water used) 16.79E-3

4.62 (benzyl alcohol) 0.27 14.55 172ml(l75ml water used) 2.50E-3
WaterzPit 8:1

3.09 (benzaldehyde) 1.27 93.94 197ml(200ml water used) 18.50E-3

4.62 (benzyl alcohol) 0.26 14.01 197ml(200ml water used) 2.75E-3

 

WaterzPit 10:1

 

3.09 (benzaldehyde) 0.89 65.83 247ml(250ml water used) 16.26E-3

 

 

4.62 (benzyl alcohol) 0.19 10.24 247ml(250ml water used) 2.33E-3

 

 

 

 

 

It was realized that the concentrations of benzaldehyde and benzyl alcohol were
lower when wet cherry pits were used instead of dry cherry pits, but the amount of
hydrolyzate was higher. There was a bi gger difference observed in benzaldehyde and
benzyl alcohol concentration at lower water to pit ratio, with a difference of around 12
percent and 30 percent for benzaldehyde and benzyl alcohol respectively at 3:1 ratio to a

difference of 6 percent and 5 percent for benzaldehyde and benzyl alcohol respectively at

75

 

10:1 ratio. But the total quantity of hydrolysate obtained using wet cherry pits was much
higher than amount obtained using dry cherry pits. Since the cherry pits were already wet,
the total overhead in wetting the cherry pits was minimal.

When the total amount of benzaldehyde and benzyl alcohol extracted for each
ratio were plotted for wet and dry cherry pits, the results obtained are as shown in Figure

3.2.

 

18‘

161

14~

   

 

   

+ Dry Pits-Benzaldehyde
+Wet Pits-Benzaldehyde
+ Dry pits-Benzyl Alcohol

-x—Wet Pits-Benzyl Alcohol

10‘

 

 

 

amount of adsorbate in hydrolyzate,microiiter

4+

 

 

 

\
o 1 fl 1 I r
o 2 4 e s 10 12
Water: Pit ratio

Figure 3.2 Comparison of wet cherry pits with dry cherry pits

76

It was observed that the total amount of benzaldehyde extracted from the pits with
8:1 water to pit ratio was more with wet pits than dry pits. However, there was a no
change in the amount of benzyl alcohol recovered from the cherry pits. Using wet cherry
pits also eliminated the necessity of drying the cherry pits, as the pits are shipped wet and
stored in a freezer before use. Thus hydrolysis carried out with 8:1 water to wet cherry pit
ratio at 55 C seemed to have the best yield. It was observed that the amount of
benzaldehyde and benzyl alcohol extracted in the hydrolysis changed with different pits.
There was a 20 % drop in the amount of benzaldehyde recovered by hydrolysis from
cherry pits which were dried and stored in the laboratory for more than two months, than
the newly dried cherry pits. However, for all kind of cherry pits, a 7:1 ratio using wet

cherry pits was found to be the optimum choice.

3.2 Filtration

The preliminary filtration of the hydrolyzate, necessary to get rid of the cherry pit
shells and kernels, is done by passing it through a sieve of 180 microns. This process is
very quick, and it is efficient in removing the large particles. The hydrolyzate is poured
into a NalgeneT" cylindrical tank with No 80. sieve manufactured by Dual Company,
placed on the top. The pits are emptied out of the sieve pan as necessary. The
hydrolyzate still constitutes lot of compounds besides benzaldehyde, benzyl alcohol and
benzoic acid. If the hydrolyzate is passed through the adsorption bed without further
filtration, there is a deposit of cake on the surface of the resin. As the thickness of the
cake increases, it becomes impermeable to the fluid and finally builds up backpressure. In

order to avoid this, the hydrolyzate is filtered and then passed through the adsorption bed.

77

Earlier, filtration was performed in a metal cylinder with a diameter of about 3 inches and
a length of about 15 inches. The following items composed the filter (bottom to top): 4
grams of glass wool, 32 grams of diatomaceous earth (normally used in swimming pool
filters), 130 grams of sand, and 50 grams of cherry pits for the purpose of spreading
liquid flow over the entire diameter of the filter. A pump pushed the hydrolyzate through
the filter from a feed tank. While this filter effectively removed all noticeable
particulates from the hydrolyzate, the filter cake caused it to plug very quickly. In spite
of adding diatomaceous earth to the hydrolyzate feed to make the filter cake sufficiently
porous to avoid plugging, the filtration bed had to be recharged twice with glass-wool,
sand and earth for filtering 20 liters of hydrolyzate and there was a loss of around 1 liter
of hydrolyzate in this process. And it was difficult to cepe up with the flow rate of the
fluid passing through the bed. Hence, a new filtration unit with pressure filter was set up.
The hydrolyzate is passed through a pressure filter with S-micron polypropylene filter
bag. It is then passed through the same pressure filter with l-micron polypropylene filter
bag. This filtration technique is faster and the wastage of hydrolyzate in this method is
minimal. The filtered hydrolyzate is then stored in a 20-L glass jar, which in turn was
placed in an ice bath. The chilled condition helps to slow down the bacterial growth in

the hydrolyzate.

3.3 Adsorption
PharMed tubing carries the cooled hydrolyzate from the bottom of the jar to the
Masterflex pump. The tubing then carries the hydrolyzate from the pump to the top of

the adsorption bed. The adsorption bed itself consists of a glass tube measuring 6

78

centimeters in diameter. Silicon stoppers are placed on the top and bottom openings in
order to hold the contents. The beds run completely filled with hydrolyzate. In order to
avoid air entering the glass bed, the hydrolyzate is passed through an air trap before
entering the glass adsorption bed. Experiments were done in the phase I of the USDA
grant to find the resin best suited for the cherry pit experiments. Breakthrough behavior
of benzaldehyde and benzyl alcohol were studied on XAD-4 (from Rohm and Haas) and
SP-850 (from Mitsubishi Chemical), the best known resins for benzaldehyde adsorption.
Experimental work revealed that SP-850 was the better resin for adsorbing benzaldehyde
and benzyl alcohol from the hydrolyzate.

The waste stream flowing through the bottom of the bed is analysed every hour to
determine whether benzaldehyde and benzyl alcohol are breaking through. Figure 3.3
shows the breakthrough behavior of benzaldehyde and benzyl alcohol for an experiment
done with 160 ml of SP-850 resin. The superficial velocity of the feed flowing through
the adsorption bed is 2.6 cm/min. The outlet concentration is normalized by the inlet feed
concentration. The concentration of benzaldehyde and benzyl alcohol in the feed stream
was 0.90 mol/m3 and 0.48 moi/m3 respectively. Except the diameter, all the parameters
listed in Table C-4.1 were used without any kind of change. The bed diameter was 0.0603
m and the height of the bed was 0.105 m. The breakthrough behavior is predicted by the
Fortran code given by Moon in 1987 for multicomponent liquid phase adsorption. Since
it wasn’t possible to quantify the concentration of benzoic acid in the feed, the benzoic

acid’s concentration was assumed to be 0.20 moi/m3.

79

 

1.8

1.8 .

1.4 1

1.2 r

 

 

 

 

0.8 <
0.6 .
0.4 .
0.2 -
I O
0 ..A' v A A .‘ Y ‘ f Y
0 100 200 300 400 500 800 700 800 900
Time, minutes

Figure 3.3 Breakthrough Behavior(1600 mi SP-BSO resin) at 2.6 cmlmin

80

 

0 Experimental Benzaldehyde
I Experimental Benzyl-alcohol
Calculated Benzaldehyde

 

 

l—Calcuiated Benzyl-alcohol

 

 

3.4 Regeneration

After adsorption, the adsorption bed is backwashed with water to eliminate all the
solids deposited on the top surface of the bed. Then the resin is regenerated at room
temperature with carbon dioxide pressurized to 1200 psi g in order to obtain the desired
components. Benzaldehyde is completely soluble in liquid carbon dioxide (Francis,
1954). Carbon dioxide is a relatively inexpensive gas and it fulfills the all “natural”
requirement for increased product value. Furthermore, its liquid and gaseous properties
can both be utilized at room temperature without any heating—only pressurization and
depressurization. Carbon dioxide is pressurized to its liquid state and passed through the
bed where it removes adsorbates from the pores of the resin. It is then depressurized into
its gaseous state where it relinquishes its solutes. Figure 3.4 gives a schematic diagram of
the experimental setup used for the desorption of benzaldehyde and benzyl alcohol from
the resin.

Carbon dioxide is pressurized to 1200 psi and then passed through a 500 ml
regeneration chamber. At 298 K and 1200 psi, the molar density of carbon dioxide as
reported by Angus and Armstrong is 0.01425 moi/cc. Around 5.739 L of pressurised
carbon dioxide is passed through the bed which is equivalent to 2000 L of atmospheric
carbon dioxide. The flow rate of carbon dioxide is maintained at 1Umin of atmospheric
carbon dioxide which is equivalent to 0.172 Uhr of pressurised carbon dioxide. The
superficial velocity of pressurised carbon dioxide through the bed is 54.83 cm/hr.To
analyse the concentration of benzaldehyde and benzyl alcohol in the carbon dioxide

leaving the regeneration chamber, a sampling loop is provided. The carbon dioxide can

81

be passed through the sampling loop by opening valves V1 and V2. The working of the
sampling loop is given in the next section.

Carbon dioxide is depressurised to 500 psi and then passed through the collection
chambers. Benzaldehyde and benzyl alcohol is collected in this stage. The lines carry
carbon dioxide deep into the collection chambers. This gives them enough retention time
in the collection chambers. Two chambers are provided to collect most of the
benzaldehyde and benzyl alcohol from carbon dioxide.

Care must be taken to prevent the lines from freezing during depressurization. The
resulting product from the collection chamber is an aqueous solution much more
concentrated with the desired components than the hydrolyzate. The product contains
water because of the moisture remaining in the bed after adsorption. Also, this product is
a brown color. However, the product recovered from synthetic runs is pale yellow in
color. Hence, the brown color can be attributed to oils, glycosides, proteins and other

compounds in the hydrolysate.

82

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 3.4 Schematic Layout of the Regeneration Setup.

. Carbon dioxide cylinder
. Gas booster

. Regulator

. Regeneration chamber
Sampling loop

. Collection chamber

. Meter valve

. Flow meter

ooqo‘gnewwv—

83

3.4.1 Sampling loop

The sampling section is shown in figure 3.5 and figure 3.6. From the regeneration
chamber a stainless steel tubing carries the carbon dioxide to the six port sample
switching valve manufactured by Valco Instruments Company, model number C6W.
Figure 3.6 illustrates the load position and the sample position. A valve (valve 3, Figure
3.5) and a 16 gauge female syringe fitting is connected to port number one. Port number
two is connected via 1/ 16 inch tubing to a valve (valve 4, Figure 3.5) and then to a 16
gauge needle. Port four is the inlet from the regeneration chamber and port five is the
outlet. When the sample valve is placed in the load position, Figure 3.6a, the carbon
dioxide flows into port four out of port three to the sample loop into port six and then out
of port five under operating pressure. To obtain a sample, the valve is rotated to sample
position, Figure 3.6b, the needle valve is opened at port two to allow material from the
loop flow into a 10 ml conical volumetric flask. Another tubing extends from the
evacuated test tube to a burette sealed with a silicon stopper on the top and having a
tubing running from the bottom of the burette to an overflow bottle. Port three and
six contains the 0.132 ml sample loop constructed from approximately 5.2 inches of l/l6
inch O.D.' stainless steel tubing. The method applied in measuring the volume of the

sampling loop is given in Appendix C-4 along with the raw data.

Syringe Connection

 

H

 

 

 

 

 

 

Burette F

 

 

 

 

 

 

 

 

 

 

 

.._.l
Overhead
Bottle
V 1
V2 1 S
Volumetric
From
. Flask
Regeneration
Chambers
To
Regeneration
Collection
Chambers Teflon Tubing

Figure 3.5 Schematics of the Sampling Loop

85

To regeneration collection chambers

 

From
Regeneration
Chambers Syringe

LOAD Position

 

To regeneration collection chambers Sample
Collection

   

 

From

 

Regeneration
Chambers

Syringe

 

SAMPLE Position

Sample
Collection

Figure 3.6 Schematic of Sampling Valve Operation in Sampling Section.

86

3.5 Mass balance and analysis

An attempt was made to complete the mass balance for the adsorption run illustrated
in figure 3.3. By integrating the area of the benzaldehyde and benzyl alcohol
breakthrough curves, it could be realized that 3.54 ml of benzaldehyde and 0.28 rrrl of
benzyl alcohol was loaded on the resin. The displacement of benzyl alcohol molecules
by benzaldehyde molecules towards the end of the run attributed to the lower amount of
benzyl alcohol on the resin. The amount of benzaldehyde and benzyl alcohol lost while
rinsing the resin with water after adsorption was less than 0.01 ml.

The resin was split into two halves of 80-ml resin. One half of the resin was
regenerated by solvent regeneration. The method adopted for solvent regeneration was
simple. 4 L of 100 % ethanol was pumped from the bottom of the bed at a superficial
velocity of 2.6 cm/min and flow rate of 75 ml/min. The ethanol solution was injected into
the GC after dilution. The benzaldehyde and benzyl alcohol content in the ethanol as a
function of ethanol passed through the bed is given in figure 3.7. Integration of the curves
showed that 1.60 ml of benzaldehyde and 0.12 ml of benzyl alcohol was recovered by
solvent regeneration. Thus 89 % of the benzaldehyde and 85 % of the benzyl alcohol
loaded on the resin could be recovered by solvent regeneration. However, the results
from carbon dioxide regeneration weren’t that encouraging. 600 L of atmospheric carbon
dioxide was passed through the resin bed at the flow rate of 2 IJmin. The final sample
from the collection chamber carried 0.54 ml of benzaldehyde and 0.04 ml of benzyl
alcohol. The pattern followed in the adsorbate desorption from the resin is illustrated in
Figure 3.8. The collection chambers were washed with 100 % ethanol after the

regeneration. The amount of benzaldehyde and benzyl alcohol recovered by this washing

87

Effluent concentration, ppm

 

1200

 

 

 

1ooo<

m4

800« Benzaldehyde

m.

200.

BenzylAloohoi
o ‘5 Y ‘F r+ I 1' I
0 1000 1500 2000 2500 3000 3500 4000
Voiumeoiethanolpessedmll

Figure 3.7 Solvent Regeneration Pattern

88

 

were 0.21 and 0.02 ml respectively. The amount of benzaldehyde and benzyl alcohol lost
in the waste carbon dioxide stream was continuously monitored from a gas bulb placed
before the final flow meter. GC from this bulb showed peaks of 0.3 and 0.2 area count for
benzaldehyde and benzyl alcohol respectively, for a 0.5 ml gas injection. This reading
accounted for 0.02 ml of benzaldehyde and 0.01 ml of benzyl alcohol. The resin was
soaked in 100 % ethanol after the regeneration. GC of this ethanol showed no trace of
benzaldehyde and benzyl alcohol. Hence the total benzaldehyde and benzyl alcohol
traced was 0.75 and 0.07 respectively, accounting for just 46 % of the benzaldehyde and
50 % of the benzyl alcohol loaded on the resin. Table 3.4 shows the desorption results of
benzaldehyde. Benzyl alcohol couldn’t be followed since the concentration was too low.

The overall recovery of benzyl alcohol was found by running a GC sample of the

 

 

 

 

 

 

 

collected product.
Table 3.4 Desorption Results
Time, GC reading of Amount of Total amount of
rrrinutes sample, area count benzaldehyde in the benzaldehyde carried by
total sample, mlxlO" carbon dioxide, m1/minx103
10 0.51 3.77 8.19
20 0.43 3.18 6.90
30 0.32 2.36 5.13
40 0.31 2.29 4.97
80 0.21 1.55 3.37
140 0.09 0.66 1.44

 

 

 

 

 

 

89

Rate of desorption of benzaldehyde, mllmln

 

 

 

 

 

0 l I I i f
0 20 40 60 80 1 00

Time, minutes

Figure 3.8 Carbon dioxide Regeneration Results

I

120

140

160

3.6 Distillation and resin reconditioning

The sample obtained from the collection chambers is distilled to get the final
products. Distillation is carried out under carbon dioxide to reduce the oxidation of
benzaldehyde. There is a considerable loss of benzaldehyde and benzyl alcohol in this
step. However, the product obtained after distillation are found to be more than 92%
pure.

The resin is stored under brine solution to prevent growth of mould. Resin is washed
with ethanol and water before reuse. Around 3 its of 20 % ethanol solution is passed
through the bottom of the bed. Then, the resin bed is washed with 3 Its of water till the

fluid sitting on the top of the bed becomes clear.

91

REFERENCES

Angus, B., and Armstrong, K.M., Thermodynamic Tables Project, International Union of
Pure and Applied Chemistry, Commission on Thermodynamics and Thermochemistry,

Gribb, A., and Lira, C.T., “Benzaldehyde recovery process design”, Report to the Cherry
Marketing Institute, (1992)

Hagedom, M.L.; “Differentiation of Natural and Synthetic Benzaldehydes by H Nuclear
Magnetic Resonance”; J. Agric. Food Chem, 1992, 40, 634.

Kawabe, T.; Morita, H.; “Production of Benzaldehyde and Benzyl Alcohol by the
Mushroom Polyporus Tuberaster K2606”; J. Agric. Food Chem, 1994, 42, 2556.

Krueger, H.W. and Reesman, R.H. Carbon Isotope Analyses in Food Technology. Mass
Spectrometry Reviews 1 (205-236).(1983)

Krueger, DA. and Krueger, H.W. Carbon Isotopes in Vanillin and the Detection of
Falsified "Natural" Vanillin. Agricultural and Food Chemistry 31 (1265-1268).

Lamer, T.; Spinnler, H.E.; Souchon, 1.; Voilley, A.; “Extraction of Benzaldehyde from
Fermentation Broth by Pervaporation”; Process Biochemistry, 1996, 31, 533.

Li, C.P.; Swain, E.; Poulton, J .; “Prunus Serotina Amygdalin Hydrolase and Prunasin
Hydrolase”; Plant Physiology, 1992, 100, 282.

Soule, A., Studies of Gas Adsorption on Activated Carbon and Cherry Flavor Recovery
from Cherry Pits, Master’s Thesis, 1999

Swain, E.; Poulton, J.; “Utilization of Amygdalin during Seedling Development of
Prunus Serotina”; Plant Physiology, 1994, 106, 437.

Walther, D., and Maurer, G., High Pressure Vapour-Liquid Equilibria in Binary Mixtures
of Carbon Dioxide and Benzaldehyde, Bromobenzene, Chlorobenzene, 1,2-
Dichlorobenzene and 2-Chloro-1-Methylbenzene at Temperatures between 313 and 393
K and Pressures up to 22 Mpa, Phys Chem 96 (1992).

Zheng, L.; Poulton, J.E.; “Temporal and Spatial Expression of Amygdalin Hydrolase and
Mandelonitrile Lyase in Black Cherry Seeds”; Plant Physiology, 1995, 109, 31.

92

APPENDIX A

PHYSICAL PROPERTIES

Table A-1. Physical Properties of Benzaldehyde

 

 

 

 

 

 

 

Formula C7H60

Molecular weight 106.12 g/gmol

Boiling point 179.0° C @ 760 mm Hg
112.5° C @100 mm Hg

Melting point -26° C

Auto ignition temperature 190° C

Flash point 62° C

Vapor pressure 10mm Hg @ 62.0° C

60mm Hg @ 99.6° C
100mm Hg @ 112.5° C

400mm Hg @ 154.1° C

 

Solubility of benzaldehyde in water

Solubility of water in benzaldehyde

0.6 % @ 20° C

1.5 % @ 20°C

 

 

 

 

Viscosity 1.4 ycm sec @ 25° C
Relative density(water=l) 1.05
Relative vapor density(watem1) 3.65

 

 

93

 

Table A-2. Physical Properties of Benzyl Alcohol

 

 

 

 

 

 

 

Formula C7H30

Molecular weight 108.14g/gmol

Boiling point 205.0° C @ 760 mm Hg
142.0° C @ 100 mm Hg

Melting point -15.3° C

Auto ignition temperature 436° C

Flash point 93° C

Vapor pressure 10mm Hg @ 92.6° C

60mm Hg @ 129.3° C
100mm Hg @ l4l.7° C

400mm Hg @ 183.0° C

 

Solubility of benzyl alcohol in water

Solubility of water in benzyl alcohol

4.4 % @ 25°C

8.0 % @ 20° C

 

 

 

 

Viscosity 8.0 g/cm sec @ 20° C
Relative density(water=1) 1.04
Relative vapor density(water=1) 3.70

 

 

94

 

Table A-3. Physical Properties of Benzoic Acid

 

 

 

 

 

 

 

Formula C711602

Molecular weight 122.12 g/gmol

Boiling point 249.0° C @ 760 mm Hg
186.0° C @ 100 mm Hg

Melting point 122.4° C

Auto ignition temperature 573° C

Flash point 12l.0° C

Vapor pressure 10mm Hg @ 132.1° C

60mm Hg @ 172.8° C
100mm Hg @ 186.2° C

400mm Hg @ 227.0° C

 

Solubility of benzoic acid in water

0.35 % @ 25° C

 

 

 

 

Viscosity 1.2 g/cm sec @ 130° C
Relative density(water=1) 1.32
Relative vapor density(water=1) 4.2

 

 

 

APPENDIX B

GC CALIBRATION

The analysis of the liquid and gas phases of the experimental samples was
performed using a gas chromatograph. A Perkin-Elmer model 8500 equipped with a FID
detector was used for the gas chromatography. The column was an Alltech Associates,
Inc. Econo-capTM capillary column, catalog number 19653, 15 m long x 1.2 mm 111 with a
EC-wax stationary phase.

Calibration standards are prepared by weighing out each component for the
sample solution in the volumetric flask used as a sample container and filling the flask to
the volume line with water. A clean stir bar is added to each sample and they are placed
on a magnetic stirrer for at least an hour to assure that each solution will be
homogeneous. The compositions of the sample solutions are given in table BI and table
3.2. 1 u] of the sample solution was injected in the Gas Chromatograph for three to four
times and a mean value of the area peak was noted. Sample solutions spanning a wide
concentration range were prepared. The results are presented in table B3 and B.4.
Calibration curves were prepared as shown in figure B.1 and B2, and a response factor
for benzaldehyde and benzyl alcohol was calculated.

The column operating conditions are given in figure B.3 while a sample

chromatograph is given in figure B.4.

96

Table B-1. Sample Solutions of Benzaldehyde/Water

 

 

 

 

 

 

 

 

 

Sample Description Water m1 Benzaldehyde g Benzaldehyde ppm

1 100 0.0068 68

2 100 0.0084 84

3 100 0.0126 126
4 100 p 0.0189 189
5 200 0.0462 231
6 200 0.0630 315
7 200 0.0840 420

 

 

 

 

Table B-2. Sample Solutions of Benzyl Alcohol [Water

 

 

 

 

 

 

 

 

 

Sample Description Water ml Benzyl alcohol ,g Benzyl alcohol, ppm

1 100 0.003 1 3 1

2 100 0.0073 73

3 100 0.0135 135
4 100 0.0166 166
5 200 0.0456 228
6 200 0.0706 353
7 200 0.0894 447

 

 

 

 

97

 

 

Table B-3. Benzaldehyde/Water GC Calibration Data

 

 

 

 

 

 

 

 

 

Sample Description Sample Volume (111) Benzaldehyde (Mean Area count/til)
1 1 0.921
2 1 1.123
3 1 1.702
4 1 2.558
5 1 3.090
6 1 4.189
7 1 5.679

 

 

 

Table B-4. Benzyl Alcohol/Water GC Calibration Data

 

 

 

 

 

 

 

 

 

Sample Description Sample Volume (111) Benzyl alcohol(Mean Area count/v.1)
l 1 0.573
2 1 1.360
3 1 2.510
4 1 3.121
5 1 4.271
6 1 6.564
7 1 8.301

 

 

 

98

 

 

 

rowemaru

GC response AClmicroilter

 

 

 

0 l r . .
0 100 200 300 400 500

Benzaldehyde concentration, ppm

Figure 8.1. Benzaldehyde Concentration Versus GC Response

 

 

 

 

o l T i T
0 1 00 200 300 400

Benzyl alcohol concentration, ppm

GC response AClmlcro liter

Figure 8.2. Benzyl Alcohol Concentration Versus GC Response

SECTION 1 GC CONTROL

1 2
OVEN TEMP (DEG C) 60 210
ISO TINIE (NHN) 0.0 7.0
RAMP RATE (DEG C/MIN) 30.0
HWD l RANGE OFF HWD l POLARITY A-B
FID 2 SENS LOW
DET ZERO ON
INITIAL DET 2
INJ 1 TEMP OFF
IN] 2 TEMP 240
DET l TEMP OFF
DET 2 TEMP 240
FLOW l 5 MUMIN
CARRIER GAS 1 HE
FLOW 2 5 MIJMIN
CARRIER GAS 2 HE
PRESSURE 3 5.0 PSIG
EQUILIB TIME 0.5 MIN
TOTAL RUN TIME 12.0 MIN

Figure B.3 Column Operating Conditions

100

4.62

3.11

 

 

“m
~—

 

 

Figure B.4. A Sample Chromatograph

101

It was observed from Figures BI and B2 that the GC response was a linear function of
the concentration for benzaldehyde and benzyl alcohol. The GC response factors for a 1
ul of liquid sample were found to be 73.97E-6 ill/area count and 53.91E-6 til/area count
for benzaldehyde and benzyl alcohol respectively. Hence the area count obtained for a
particular injection multiplied by their response factor would give the ppm concentration

of that compound in the sample.

102

APPENDIX C
PROCEDURES AND RAW DATA
C-l. Void fraction calculations
The skeletal density and the bulk density were calculated to find out the void ratio
of SP-850 resin. The procedure adopted and the raw data for finding them is summarized.
1. Resin was thoroughly washed with water to get rid of the salts.
2. A particular volume of washed resin (V1 ml) was measured using a 10 ml graduated
cylinder.
3. The resin was transferred using a pipette to decant the liquid and then dried in a
vacuum oven for 24 hours at a temperature of 60°C.
4. The resin was then weighed to give the final mass (m gm) of the resin.
5. The dried resin was then added to a finite volume (V2 ml) of 40 % ethanol solution.
6. The total volume (V3 ml) of the ethanol solution and the resin was measured after 30
minutes, thus giving the change in volume (AV: V3-V2 ml).
7. pb and ,0, were calculated using the formulas :

pb = m/V, and

p,=m/AV.

103

Table C-1.1 pb and p, Calculations

 

 

 

 

 

 

 

 

V1 ml m gm V2 ml V3 ml AV ml p, gm/ml p, gm/ml
10 3.40 40 43.2 3.2 0.340 1.062
12 4.10 40 43.9 3.9 0.341 1.051
18 6.25 80 85.9 5.9 0.347 1.059
20 6.80 80 86.5 6.5 0.340 1.046
24 8.20 80 87.8 7.8 0.341 1.051
25 8.60 80 88.2 8.2 0.344 1.048
30 10.20 100 109.7 9.7 0.340 1.051

 

 

 

 

 

 

 

 

 

The same experiments were repeated by using a different ethanol solution and the

results found were exactly same. Hence, pb and p, used in all the calculations were 0.34

gm/ml and 1.051 gm/ml respectively. The void fraction of the resin is given by

8 =[ -£1], and its value is found to be 0.677.
P,

C-2. Equilibrium adsorption behavior of benzaldehyde, benzyl-alcohol and benzoic
acid solution.
Equilibrium adsorption behavior of benzaldehyde, benzyl-alcohol and benzoic

acid was studied in a liquid phase adsorption at room temperature. The following

104

procedure was used to get a single point on the isotherm for benzaldehyde and benzyl-
alcohol.
1. Water was sparged with carbon dioxide for 15 minutes.

2. 250 ml of the sparged water was put in a 400 ml PYREXT" beaker.

3. A specific volume of benzaldehyde or benzyl alcohol was measured and
added to the beaker.
4. The solution was stirred on a low setting for 10 minutes to let the
concentration even out.
5. A G.C of the solution was taken to measure the initial concentration.
6. A specific volume of resin was added to the flask.
To measure the resin:
A. Salt water was rinsed out of resin and it was put in fresh water.
B. A pipette was used to suck some of the water/resin.
C. All the water was drained carefully so that only resin remained in the
auto pipette.
D. A small portion of the beaker solution was put in a small Erlenmeyer
flask onto the scale.
E. The resin was carefully added, by shaking it out of the tube into the
solution until desired mass.
F. The small portion of resin/solution was added to the beaker. The resin
was rinsed into the beaker using the solution.
7. The beaker was covered with parafilm and stirred for two hours.

8. The final concentration was measured after two hours.

105

Here, the adsorbent loading is given by the grams of adsorbate adsorbed onto the resin

per gram of wet resin. The resin is washed prior to weighing to get rid of the salts.

Experiments were done to calculate the ratio of dry resin weight to wet resin weight. A

particular volume of resin was taken and weighed. It was then washed with water to

remove the salts. A pipette was used to remove the drain the water. The resin was then

weighed again. The results are tabulated in Table C.2.1.A.

Table C.2.1.A. Results of Experiments to Detect the Ratio of Dry Resin Weight / Wet

 

 

 

 

 

 

Resin Weight
Weight of Dry Resin, Weight of Wet Resin, Ratio of Dry Resin Weight to

gm gm Wet Resin Weight

5 7.50 0.666

8 12.50 0.640

10 15.65 0.638

15 23.00 0.652

20 31.25 0.64

 

 

 

 

Hence, it was found that the ratio of dry resin weight to wet resin weight is 0.64.

The same procedure was observed for collecting the equilibrium adsorption data

for benzoic acid, except the method to detect the concentration. The Gas Chromatograph

couldn’t accurately sense the concentration of benzoic acid, so the concentration was

detected using a pH test-meter. Calibration of the pH meter is given in Table C.2.1.B.

106

 

The results are given in table C22, C23 and C24. The values of equilibrium adsorbate
concentration in the adsorbent phase are tabulated in gm adsorbate/gm wet adsorbent. For
model fitting, these values were converted to gm adsorbate/gm dry adsorbent by diving

them by 0.64.

Table C-2.1.B Calibration of pH-meter

 

 

 

 

 

 

 

 

 

 

pH reading Concentration of benzoic acid in moi/m3
3.25 0.562
3.50 0.316
3.75 0.177
4.00 0.100
4.50 0.031
5.00 0.010
5.50 0.003
6.00 0.001

 

 

 

107

Table C-2.2 Equilibrium Adsorption Data for Benzaldehyde Solution

 

y- equilibrium adsorbate

q-equilibrium adsorbate concentration in

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

concentration in the fluid phase the adsorbent phase,
moi/m3 gm adsorbate/gm wet adsorbent
0.0470 0.0070
0.0943 0.0090
0.1475 0.01 15
0.1577 0.0135
0.0169 0.0185
0.2357 0.0209
0.3397 0.0265
0.3960 0.0276
0.6338 0.0407
0.6458 0.0430
0.7304 0.0449
0.8215 0.0435
1.0842. 0.0547
1.3397 0.0630
1.7730 0.0676
1.8868 0.0694
2.1048 0.0709
2.6418 0.0739
2.9614 0.0800
3.4131 0.0822
3.8679 0.0815
4.6870 0.0852

 

108

 

Table C-2.3 Equilibrium Adsorption Data for Benzyl Alcohol Solution

 

y- equilibrium adsorbate

q-equilibrium adsorbate concentration in

 

 

 

 

 

 

 

 

 

 

 

 

 

 

concentration in the fluid phase the adsorbent phase,
moi/m3 gm adsorbate! gm wet adsorbent
0.1853 0.0062
0.2263 0.0051
0.2768 0.0076
0.3943 0.0072
0.6079 0.0132
0.8647 0.0163
0.9979 0.0194
1.4614 0.0243
1.7022 0.0266
2.0675 0.0312
2.8888 0.0370
4.0581 0.0468

 

109

 

Table C-2.4 Equilibrium Adsorption Data for Benzoic Acid Solution

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

y- equilibrium adsorbate q-equilibrium adsorbate concentration in
concentration in the fluid phase, the adsorbent phase,
moi/m3 gm adsorbate! gm wet adsorbent
0.0508 0.0036
0.1239 0.0070
0.1639 0.0088
0.1803 0.0094
0.2377 0.0123
0.2704 0.0134
0.3704 0.0166
0.3998 0.0177
0.5002 0.0202
0.6557 0.0240
0.7868 0.0265
0.9098 0.0292

 

110

 

C-3. Breakthrough behavior of benzaldehyde, benzyl-alcohol and benzoic acid in a

fixed bed adsorption of single component.

Experiments were performed to study the breakthrough pattern of benzaldehyde,

benzyl alcohol or benzoic acid in a fixed bed adsorption of single component. The

procedure employed was as follows.

1.

15 ml of resin was washed with water to get rid of the salts. The height of the
resin bed was measured.

The resin was loaded on the glass bed and backwashed with 20 weight percent
ethanol in water solution and then back-washed with water till the fluid on the
top of the bed looked clear.

A feed solution was prepared with benzaldehyde, benzyl alcohol or benzoic
acid. A GC of the feed was taken to quantify the concentration.

The feed was passed downward through the bed at a superficial velocity of 2.7
cm/min.

GC of the waste stream was taken at a regular interval to analyze the

concentration of the adsorbate in the outlet.

Experiments were performed with 20ml and 25 ml of resin. The diameter of the bed was

also varied. The concentration for benzoic acid was again measured by following the pH.

The results of the experiments are tabulated in table C.3.1 to C.3.12

lll

Table C-3.1 Benzaldehyde Breakthrough Behavior, Experiment-1

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Adsorbate Benzaldehyde
Height of bed 6.35 cm
Superficial velocity 2.7 cm/min
Diameter of bed 2.22 cm
Adsorbate concentration in the feed 0.83 mol/m3
Time, minutes C/Co

60 0.000

180 0.015

300 0.061

360 0.050

420 0.080

480 0.280

540 0.690

600 0.886

660 I 0.920

 

 

112

 

Table C.3.2 Benzaldehyde Breakthrough Behavior, Experiment -2

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Adsorbate Benzaldehyde
Height of bed 8.53 cm
Superficial velocity 2.7 cm/min
Diameter of bed 2.22 cm
Adsorbate concentration in the feed 1.13 mol/mj
Time, minutes C/Co

30 0.000

150 0.010

270 0.021

330 0.009

390 0.007

450 0.054

510 0.060

570 0.250

630 0.720

690 0.920

 

 

113

 

Table C-3.3 Benzaldehyde Breakthrough Behavior, Experiment -3

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Adsorbate Benzaldehyde
Height of bed 12.19 cm
Superficial velocity 2.7 cm/min
Diameter of bed 6.03 cm
Adsorbate concentration in the feed 1.05 moi/m3
Time, minutes C/C0

60 0.013

180 0.010

300 0.015

360 0.020

420 0.022

480 0.024

600 0.026

660 0.030

720 0.070

780 0.090

840 0.210

900 0.540

960 0.870

1020 0.920

 

 

114

 

Table C-3 .4 Benzaldehyde Breakthrough Behavior, Experiment -4

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Adsorbate Benzaldehyde
Height of bed 11.17 cm
Superficial velocity 2.7 cm/min
Diameter of bed 6.03 cm
Adsorbate concentration in the feed 1.09mol/m3
Time, rrrinutes C/C0

60 0.013

180 0.019

300 0.020

360 0.045

420 0.039

480 0.045

540 0.039

600 0.040

660 0.040

720 0.080

780 0.057

840 0.092

1020 0.920

 

 

115

 

Table C-3.5 Benzyl Alcohol Breakthrough Behavior, Experiment-5

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Adsorbate Benzyl alcohol
Height of bed 5.08 cm
Superficial velocity 2.7 cm/min
Diameter of bed 2.22 cm
Adsorbate concentration in the feed 0.63 moi/m3
Time, minutes C/C0

30 0.000

60 0.001

90 0.002

120 0.160

155 0.530

190 0.860

210 0.920

240 0.924

 

 

116

 

Table C.3.6 Benzyl Alcohol Breakthrough Behavior, Experiment-6

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Adsorbate Benzyl alcohol
Height of bed 6.60 cm
Superficial velocity 2.7 cm/min
Diameter of bed 2.22 cm
Adsorbate concentration in the feed 0.48 mol/m3
Time, minutes C/CO

30 0.000

60 0.000

90 0.000

150 0.050

180 0.201

210 0.450

227 0.663

256 0.812

273 0.839

 

 

117

 

Table 03.7 Benzyl Alcohol Breakthrough Behavior, Experiment-7

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Adsorbate Benzyl alcohol
Height of bed 8.12 cm
Superficial velocity 2.7 cm/min
Diameter of bed 6.03 cm
Adsorbate concentration in the feed 0.48 mol/m3
Time, rrrinutes C/Co

40 0.000

70 0.000

100 0.001

130 0.001

160 0.002

190 0.1 10

220 0.221

246 0.371

268 0.538

310 0.831

340 0.839

 

 

118

 

Table C-3.8 Benzyl Alcohol Breakthrough Behavior, Experiment-8

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Adsorbate Benzyl alcohol
Height of bed I 8.38 cm
Superficial velocity 2.7 cm/min
Diameter of bed 6.03 cm
Adsorbate concentration in the feed 0.48 moi/m3
Time, rrrinutes C/Co

35 0.000

105 0.002

140 0.002

175 0.003

210 0.101

268 0.458

306 0.821

335 i 0.882

360 0.901

 

 

119

 

Table 03.9 Benzoic Acid Breakthrough Behavior, Experiment-9

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Adsorbate Benzoic acid
Height of bed 5.08 cm
Superficial velocity 2.7 cm/min
Diameter of bed 2.22 cm
Adsorbate concentration in the feed 0.26 mol/m3
Time, minutes C/Co

10 0.000

25 0.001

55 0.007

80 0.009

210 0.180

270 0.371

280 0.470

313 0.532

365 0.691

395 0.734

440 0.890

470 0.945

510 0.961

 

 

120

 

Table C-3. 10 Benzoic Acid Breakthrough Behavior, Experiment-10

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Adsorbate Benzoic acid
Height of bed 8.08 cm
Superficial velocity 2.7 cm/min
Diameter of bed 2.22 cm
Adsorbate concentration in the feed 0.26 mol/m3
Time, rrrinutes C/C0

30 0.000

150 0.000

270 0.000

330 0.1 1 1

390 0.231

450 0.320

510 0.450

570 0.709

630 0.795

690 0.921

718 0.942

 

 

121

 

Table C-3.11 Benzoic Acid Breakthrough Behavior, Experiment-11

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Adsorbate Benzoic acid
Height of bed 12.19 cm
Superficial velocity 2.7 cm/min
Diameter of bed 6.03 cm
Adsorbate concentration in the feed 0.19 mol/m3
Time, minutes C/C0

30 0.000

70 0.000

210 0.001

420 0.021

600 0.080

660 0.140

720 0.241

828 0.550

853 0.601

929 0.821

1035 0.939

 

 

122

 

Table C-3.12 Benzoic Acid Breakthrough Behavior, Experiment-12

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Adsorbate Benzoic acid
Height of bed 7.50 cm
Superficial velocity 2.7 cm/min
Diameter of bed 6.03 cm
Adsorbate concentration in the feed 0.19 mol/m3
Time, minutes C/Co

40 0.000

170 0.002

300 0.020

360 0.150

420 0.271

480 0.451

521 0.561

600 0.830

660 0.898

 

 

123

 

C-4. Breakthrough behavior of mixtures of benzaldehyde, benzyl-alcohol and
benzoic acid in a fixed bed adsorption.

Experiments were performed to study the breakthrough pattern of benzaldehyde,
benzyl alcohol and benzoic acid in a liquid phase fixed bed adsorption. The procedure
employed was as follows.

1. 15 ml of resin was washed with water to get rid of the salts. The height of the
resin bed was measured.

2. The resin was loaded on the glass bed and backwashed with 20 percent ethanol
solution and then washed with water till the fluid on the top of the bed looked
clear.

3. A feed solution was prepared with benzaldehyde, benzyl alcohol or benzoic
acid. A GC of the feed was taken to quantify the concentration.

4. The feed was passed through the bed at a superficial velocity of 2.52 cm/min.

5. GC of the waste stream was taken at a regular interval to analyze the
concentration of the adsorbate in the outlet.

Experiments were performed with 20ml and 25 ml of resin. The superficial velocity of
the bed was also varied to 3.78 and 5.04 cm/min. The concentration of benzoic acid could
be followed for only one run. The experimental results were fitted with the predicted
results from Moon’s program for multicomponent liquid adsorption. The Freundlich

constants obtained from equilibrium adsorption data were used for fitting. Parameters like

temperature, ,0, , p, Freundlich constants, diameter of adsorption bed and particle

diameter were constant for all the experiments. Table C.4.1 enlists values of all the

parameters that were constant for all the experiments. Their values are given in units

124

compatible with the program. The Freundlich constant n corresponding to equation 2.9,

was taken from Table 2.3. The value of 11 didn’t undergo any change since it was

dimensionless. Since Moon’s program required q in krnol adsorbate/m3 of bed, K from

Table 2.3 was converted.

Table C-4.1 Constant Parameters in Multicomponent Adsorption Experiments.

 

 

 

 

 

 

 

 

 

Temperature 25°C

Diameter of bed 0.0222 m

Pb 340 kg/m3

p3 1051 kg/m3

Particle diameter 0.000425 m

Adsorbate Benzaldehyde Benzyl alcohol Benzoic acid
Freundlich constant K 0.509 0.194 0.254
Freundlich constant n 2.43 2.2 1.558

Mass transfer coefficient .2E-4 m/sec .2E-4 m/sec .2E-4 m/sec

 

 

Surface diffusion coefficient

 

.26E-ll mzlsec

 

.26E-11 mzlsec

 

.26E-1 1 m2/sec

 

The raw data from the experiments is given in table C.4.2 to €410.

125

 

Table C-4.2 Multicomponent Breakthrough Behavior, Experiment—1

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Height of bed 0.075 m

Superficial velocity 0.00042 m/sec

Adsorbate Benzaldehyde Benzyl alcohol Benzoic acid
Adsorbate concentration in feed 0.90 moi/m3 0.48 moi/m3 0.20 moi/m3
Time, minutes C/Co C/Co C/Co

120 0.020 0.004 -

157 0.020 0.120 —

171 0.010 0.240 -

185 0.020 0.430 -

248 0.010 1.183 -

324 0.028 1.560 -

374 0.077 1.526 -

424 0.172 1.380 -

 

 

 

 

126

 

Table C-4.3 Multicomponent Breakthrough Behavior, Experiment-2

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Height of bed 0.075 m

Superficial velocity 0.00063 m/sec

Adsorbate Benzaldehyde Benzyl alcohol Benzoic acid
Adsorbate concentration in feed 0.88 moi/m3 0.46 moi/m3 0.20 moi/my
Time, minutes C/Co C/Co C/C0

75 0.010 0.067 -

1 10 0.010 0.403 -

136 0.010 0.640 -

187 0.020 1.276 -

230 0.050 1.453 -

274 0.194 1.372 -

315 0.395 1.240 -

375 0.68 1.055 -

 

 

 

 

127

 

Table C-4.4 Multicomponent Breakthrough Behavior, Experiment-3

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Height of bed 0.075 m

Superficial velocity 0.00084 m/sec

Adsorbate Benzaldehyde Benzyl alcohol Benzoic acid
Adsorbate concentration in feed 0.88 moi/m3 0.46 moi/m3 0.20 moi/m3
Time, minutes C/Co CIC0 C/Co

38 0.004 0.047 -

75 0.01 1 0.309 -

1 10 0.025 0.793 -

136 0.043 1.195 -

187 0.194 1.360 -

210 0.289 1.295 -

230 0.355 1.240 -

274 0.61 1.184 -

 

 

 

 

128

 

Table C-4.5 Multicomponent Breakthrough Behavior, Experiment-4

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Height of bed 0.106 m

Superficial velocity 0.00042 m/sec

Adsorbate Benzaldehyde Benzyl alcohol Benzoic acid
Adsorbate concentration in feed 0.90 mol/m3 0.50 moi/r33 0.20 moi/m3
Time, minutes C/Co C/Co C/Co

38 0.000 0.000

75 0.000 0.000

1 10 0.000 0.000

136 0.000 0.000

187 0.000 0.000 -

210 0.000 0.000 -

230 0.000 0.010 -

252 0.000 0.011 -

266 0.000 0.323 -

312 0.010 0.850 -

460 0.020 1.610 -

540 0.030 1.574 -

688 0.476 1.240 —

735 0.593 1.187 -

775 0.819 1.098 -

 

 

 

 

129

 

Table C-4.6 Multicomponent Breakthrough Behavior, Experiment-5

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Height of bed 0.106 m

Superficial velocity 0.00063 m/sec

Adsorbate Benzaldehyde Benzyl alcohol Benzoic acid
Adsorbate concentration in feed 0.90 moi/m3 0.50 mol/mj 0.20 moi/m3
Time, minutes C/C0 C/Co C/Co

150 0.010 0.130 -

175 0.010 0.388 -

223 0.010 1.006 -

303 0.030 1.547 -

396 0.1 18 1.410 -

446 0.341 1.310 -

498 0.640 1.156 -

 

 

 

 

130

 

Table C-4.7 Multicomponent Breakthrough Behavior, Experiment-6

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Height of bed 0.106 m

Superficial velocity 0.00084 m/sec

Adsorbate Benzaldehyde Benzyl alcohol Benzoic acid
Adsorbate concentration in feed 0.90 moi/m3 0.50 moi/m3 0.20 moi/m3
Time, minutes C/Co C/Co C/Co

35 0.000 0.010 -

90 0.000 0.100 -

1 13 0.004 0.271 -

131 0.007 0.480 -

175 0.037 1.127 -

212 0.059 1.446 -

250 0.1 17 1.480 -

310 0.275 1.347 -

340 0.41 1 1.200 -

380 0.631 1.134 —

400 0.716 1.090 -

420 0.763 1.086 -

 

 

 

 

131

 

Table C-4.8 Multicomponent Breakthrough Behavior, Experiment-7

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Height of bed 0.121 m

Superficial velocity 0.00042 m/sec

Adsorbate Benzaldehyde Benzyl alcohol Benzoic acid
Adsorbate concentration in feed 0.90 mol/m3 0.46 moi/my 0.20 mol/m3
Time, minutes C/Co C/Co C/C0

35 0.000 0.000 -

70 0.000 0.000 -

140 0.030 0.012 -

325 0.020 0.150 -

389 0.020 1.028 -

489 0.010 1.532 -

549 0.010 1.647 -

609 0.020 1.594 -

670 0.030 1.507 -

760 0.010 1.380 -

812 0.028 1.280 -

870 0.077 1.190 -

 

 

 

 

132

 

Table C-.9 Multicomponent Breakthrough Behavior, Experiment-8

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Height of bed 0.121 m

Superficial velocity 0.00063 m/sec

Adsorbate Benzaldehyde Benzyl alcohol Benzoic acid
Adsorbate concentration in feed 0.90 mol/m3 0.46 mol/m3 0.20 moi/m3
Time, minutes C/Co C/Co C/Co

97 0.015 0.037 -

189 0.010 0.360 -

247 0.030 0.980 -

299 0.010 1.401 -

41 1 0.103 1.521 -

520 0.371 1.326 -

 

 

 

 

133

 

Table C-4.10 Multicomponent Breakthrough Behavior, Experiment-9

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Height of bed 0.121 m

Superficial velocity 0.00084 m/sec

Adsorbate Benzaldehyde Benzyl alcohol Benzoic acid
Adsorbate concentration in feed 0.90 mol/m3 0.46 mol/m3 0.20 moi/m3
Time, minutes C/Co C/Co C/Co

30 0.000 0.020 -

75 0.000 0.045 -

105 0.010 0.070 -

140 0.010 0.250 -

187 0.026 0.928 -

220 0.022 1.3 14 -

265 0.051 1.527 -

315 0.150 1.466 -

360 0.257 1.313 -

390 0.382 1.21 1 -

415 0.554 1.170 -

 

 

 

 

134

 

C-S. Sampling loop procedure and calculations.

The procedure for sampling the liquid phase by the sampling loop is as follows:

1. Flush the line and loop

A.

B
C.
D

The sampling loop is set to sample position.

. Valve V1 and V2 are closed.

Valve V3 and V4 are opened.

. A syringe filled with ethanol is connected to Knurl-Lok finger tightened

HPLC union (a).

The sample loop is cleaned. Approximately 10 cc’s of ethanol is flushed
through the line and sample loop and into a beaker.

The syringe is removed and then air is flushed through the loop in the same

manner.

2. Load the sample in the loop

A.

B.

Valve V3 and V4 are closed.

The sampling loop is set to load position.

Valve V1 and V2 are opened to allow the carbon dioxide to flow through the
loop.

After 15-20 minutes, the loop is brought back to sample position. This
transition is done very fast, so that almost all the sample is retained in the
loop.

Valve V1 and V2 are closed.

135

3. Obtain sample and measure included gas

A.

B.

P190

1'11

The initial burette reading is noted.

Valve V4 is opened slowly to allow the sample to flow out of the loop to the
10 ml volumetric flask.

A syringe with 10 ml ethanol is connected to Valve V3.

Valve V3 is opened.

Ethanol is pushed through the sampling loop to the volumetric flask.

The overflow bottle is lowered until the water level in the bottle matches that

of the burette.

. The liquid level of the burette is recorded.

. Ethanol is added to the volumetric flask till the liquid level in the flask reaches

10 ml line.

Valve V4 is closed and the sample collected in the volumetric flask is
analysed.

The sample is then diluted to 20 % by adding 40 ml of water. The diluted

sample is then injected into the gas chromatograph.

Experiments were done to accurately measure the volume of the sampling loop.

The equipment configuration used to determine the volume of the sample loop was

identical to that used in the composition measurements except that the sample loop was

connected directly to the high pressure tank of carbon dioxide. Sample loop volumes

were calculated using the equations:

n

P V

_ mbr’enrfinal

136

Where,

V = nv(Bwp,T)

n: number of moles.

v =molar volume, mol/cm3 .

Experiments were done at room temperature.The molar volume of carbon

dioxide at elevated pressure were found by interpolating values from International

thermodynamic tables of the fluid state ; v. 3.

Table 05.1 Data for Volume Determination of Sample Loop

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Temp. Loop Final Moles carbon v carbon dioxide, Loop
K pressure, volume, dioxide x 104 cc/gmol volume, (cc)
bar cc
17.0 6.95 0.132
298.15 63.44 17.1 6.99 190.47 0.133
17.0 6.95 0.132
13.8 5.64 0.133
298.15 58.62 13.7 5.60 235.84 0.132
13.7 5.60 0.132
12.1 4.95 0.132
298.15 55.86 123 499 266.67 0.133
12.2 5.03 0.134

 

137

 

It can be realized from the table that the volume of the sampling loop is 0.132 ml. So for

all the composition measurements, 0.132 ml was used as the volume of the sampling

100p.

Experiments were done to check the precision of the sampling loop. Walther and Maurer

reported that the solubility of benzaldehyde in carbon dioxide at 313.2 K and 8.11 MPa is

0.0023 moi benzaldehyde/mo] carbon dioxide. Except the regeneration chamber, The

experimental setup was same as Figure 3.4 . A smaller regeneration chamber was used

and it was placed in a temperature bath to elevate the temperature to 40°C. The procedure

followed in operating the sampling loop was the same. However, the sample wasn’t

diluted to 20%.

Table C-5.2 Results for Checking precision of Sampling loop

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Temp, Pressure GC. Reading Moles of Molar density Moles of Mole
K MPa of sample, benzalde- of CO2 CO2 x104 fraction,
area count hydex 107 moi/cc mol/mol
0.14 9.76 0.00141
313.2 7.61 0.00525 6.93
0.14 9.76 0.00141
0.31 21.62 0.00242
313.2 8.11 0.00675 8.91
0.30 20.93 0.00234

 

138

 

The solubility obtained from experimental results is quite close to the reported
value. Hence, the sampling, loop can be used to quantify the concentration of

benzaldehyde in the carbon dioxide stream.

139

APPENEHXJJ

COMPUTER PROGRAM

/**********************************************************

PROGRAM : ADSORPTION MODEL

**************************1%******************************

The program fits the experimental data to Adsorption
models and find the constants.

***‘k******************************************************/

#include <ctype.h>

#include <stdio.h>

#include <fstream.h>

#include <iostream.h>

#include <strings.h>

#include <iomanip.h>

#include <std1ib.h>

#include <math.h>

void linear(double *xl, double *x2);
void freundlich(double *xl, double *x2);
void langmuir(double *xl, double *x2);
void lang_freu(double *xl, double *x2);

double error(double *al, double *a2);
int datapoints,response1;

struct functionresults
{ .
double a,b;
} linearresults, langresults, freuresults, lfresults,
functionresultsl;

functionresults function(double *xl, double *x2);
int main()

{
char *filepl;
FILE *filep;
int numbers, i, j,k,z;
ifstream fin;
ofstream. fout;
char filename[11], line[28];

140

double values[l99], totnumber,y[50],q[50];

int r, t, b, response, responseZ;
char data[100];

char number[120][10];

int m, n, state;

cout << "the program has been prepared for one component
"<<endl;

cout << "please prepare indata.dat with your adsorption
data," << endl;

cout << " or you can specify your own input file names”
<< endl;

continuel:
cout << " press 0 to quit or 1 to continue" << endl;
cin >>response;

if (response == 0)
{
exit(l);
}
cout << ” enter 1 for linear adsorption model. " << endl;
cout << " enter 2 for langmuir adsorption model. " <<
endl;

cout << " enter 3 for freundlich adsorption model. " <<
endl;

cout << " enter 4 for langmuir-freundlich adsorption
model. " << endl;

cin>> responsel;

/’1'*********************************************************

Section 2 : Data Reading

**********************************************************/

cout << “The data is in indata.dat(1) or some other
file(0)' << endl;
cin >> responsez;
switch (responseZ) {
case 1:
{
filename = "Indata.dat”;
break;
}

141

case 0:

{
cout << " what is the filename
cin >> filename;
break;

}

}

for (j = 0; j <= 120; j++)
{
for (n = O; n <= 10; n++)
{
number[j][n]='\0’;
}
}
state = l;
m=-l;
n=1;
cout<<endl;
filep = fopen(filename, "r");
if (filep != NULL)
{
for (j = 0; j <= 200; j++)
{
fgets(data, 29, filep);

" << endl;

n=1;
for (i = 0; i <= 29; i++)
{
if (isdigit(data[i]) || (data[i] ==
if (state == 1)
{
m = m + 1,
n=0;
}

number[m][n] = data[i];
n = n + 1;
state = 999;

}
else

{

state = 1;
}
}

142

.'))

for(i=0;i<=29;i++)
data[i]=’\0’;

}
m = l; .
for (j = 0; j <= 120; j++)
{
if (isdigit(number[j][0]) || number[j][0] == ’.’)
{
values[m] = atof(number[j]);
m = m + l,
}
}
totnumber = m - 1; /* one already added to m */
}
else
{
cout << "it doesnt open" << endl;
exit(l);
}

/**********************************************************

Section 3 : Call to corresponding Function

*‘k********************************************************/

datapoints= int(values[1]);
m=2;
for(j=1;j<=datapoints;j++)
{
y[j]=values[m];
q[j]=values[m+1];
m:m+2;
}
switch (responsel)
{
case 1:
{
linear(y,q);
break;

}

case 2:

143

{
langmuir(y,q);
break;

}

case 3:

{
freundlich(y,q):
break;

}

case 4:

{
1ang_freu(y,q);
break;

}
}

goto continuel:

/*1%********************************************************

Linear Adsorption Model Q: k*Y

***‘k******************************************************/

void linear(double *xl, double *x2)
{

int i;

double a,lincalc[50],linerror;

cout<<endl;

linearresults= function(xl,x2);

a=(linearresults.a);

cout<< " the equation for linear adsorption model is q =
"<<a<<' * y"<<endl;

for(i=1;i<=datapoints;i++)
{
lincalc[i]=(a*x1[i]);
cout<<x1[i]<<" "<<x2[i]<<" "<<1inca1c[i]<<endl;
}
linerror=error(x2,lincalc);
cout<<'I the total error is "<<1inerror<<endl;

144

/*********************************************‘k'k***********

Freundlich Adsorption Model Q: k*Y“n

**********************************************************/

void freundlich(double *xl, double *x2)
{
static int m;
int i;
double a,n;
double y1[50],y2[50],freerror,frecalc[50];
m=0;
for(i=1;i<=datapoints;i++)
{
y1[i1=log(x1[i]);
y2li1=log(x2[i]);
if(m==0)
{
cout<<endl;
m=m+1;
}
}

freuresults= function(y1,y2);
n=freuresults.a;
a=exp(freuresults.b);
cout<< " the equation for Freundlich adsorption model is
q = "<<a<<" * y“'<<n<<endl;
for(i=l;i<=datapoints;i++)
{
frecalc[i]=(a*pow(x1[i],n));
cout<<x1[i]<<" "<<x2[i]<<" '<<frecalc[i]<<endl;
}
freerror=error(x2,frecalc);
cout<<"the total error is "<< freerror<<endl;

/**********************************************************

Langmuir Adsorption Model Q= a*Y/(l+b*Y)

'k*********************************************************/

void langmuir(double *xl, double *x2)
{

int i,m;

145

=(

double a,b;
double yl[50],y2[50],lanerror, lancalc[501;
m=0;
for(i=l;i<=datapoints;i++)
{
yl[i]=(xl[i]);
y2[i]=(xl[i])/(x2[i]);
if(m==0)
{
cout<<endl;
m=m+1;
}
}

langresults= function(yl,y2);
a=l/langresults.b;
b=langresults.a*a;

cout<<endl<<" "<<b<<" "<<a<<endl;

cout<< "the equation for Langmuir adsorption model is q

0' o
’

cout<<a<<"*y)/(l+("<<b<<"*y))"<<endl;
for(i=1;i<=datapoints;i++)
{
lancalc[i]=((a*xl[i])/(l+b*x1[i]));
cout<<x1[i]<<" "<<x2[i]<<” '<<lancalc[i]<<endl;
}
lanerror=error(x2,lancalc);
cout<<" the total error is "<<lanerror<<endl;

/**********************************************************

Langmuir-Freundlich Adsorption Model

Q=(a*Y“n)/(l+(b*Y“n))

****************************~k*****************************/

void lang_freu(double *xl, double *x2)

{

int i,j,m,k,1;

double a,b,step;

double y1[50],y2[50],lferror, lfcachSO],errorlast;
double minn,maxn,n,ntrial,finala,finalb,finaln;

cout<<” Please specify the range of n '<<endl;
cout<<" possible minimum of n'<<endl;

146

cin >>minn;

cout<<" possible maximum of n"<<endl;
cin >>maxn;
m=0;
errorlast= 10000000;
for(l=1;l<=3;l++)

{

for(j=0;j<=100;j++)
{
step=(maxn-minn)/100;
ntrial=(minn+((maxn—minn)*j/100));
for(i=1;i<=datapoints;i++)
{
yl[i]=pow((x1[i]),ntrial);
y2[i]=pow((x1[i]),ntrial)/(x2[i]);
if(m==0)
{
cout<<endl;
m=m+l;
}
}

lfresults= function(yl,y2);
a=1/lfresults.b;
b=lfresults.a*a;

for(i=l;i<=datapoints;i++)
lfcalc[i]=((a*(y1[i]))/(1+b*(yl[i])));
lferror=error(lfcalc,x2);
cout<<endl;
if(lferror<errorlast)
{
finala=a;
finalb=b;
finaln=ntrial;
errorlast=lferror;

}
}

minn=finaln—step;
maxn=finaln+step;

}

cout<< "the equation for Langmuir-Freundlich adsorption
model is“;

147

cout<<" q =( '<<finala<<" *
y“"<<finaln<<”)/(1+(~<<finalb<<"*y“"<<finaln<<"))"<<endl;

for(i=1;i<=datapoints;i++)
cout<<xl[i]<<" "<<x2[i]<<"
”<<((finala*pow(x1[i],finaln)/(l+finalb*(pow(x1[i],finaln))
)))<<endl;
cout<<" the total error is "<<errorlast<<endl;

}

/****************‘k'k'k'k'k‘k************************************

Function to calculate the best fit

***************************‘k‘k‘k'k'k'k‘k************************/

functionresults function(double *xl, double *x2)
{
functionresults functionresultsl;
int 1;
double sigmax, sigmay, sigmaxy,sigmax2;
sigmax=sigmay=sigmaxy=0;
for(i=1;i<=datapoints;i++)
{
sigmax=sigmax+x1[i];
sigmay=sigmay+x2[i]:
sigmaxy=sigmaxy+(x1[i]*x2[i]);
sigmax2=sigmax2+(x1[i]*xl[i]);
}
if(response1==l)
functionresultsl.a=sigmay/sigmax;
else
functionresultsl.a=(((double)datapoints*sigmaxy)-
(sigmax*sigmay))/(((double)datapoints*sigmax2)-
(sigmax*sigmaX));
functionresultsl.b=((sigmay-
(functionresultsl.a*sigmax))/(double)datapoints);
return(functionresultsl);

}

/**********************************************************

Function to calculate the error

*********************************************************/

148

double error(double *a1,double *a2)
{

int i;

double totalerror;

totalerror=0;

for(i=1;i<=datapoints;i++)
totalerror=totalerror+ pow((a2[i]-al[i]).2);

return( totalerror);

149