WWW” .

1
f

fig WIN!INHIHIWWHHWWW

 

MICHIGMI STATE UNIVERSITY LIBRARIES

Ill l ll ll llllllllll W l l llllllllfll

3 1293 01712 737

 

 

 

This is to certify that the
dissertation entitled
OPTIMIZATION OF ONION FLAVOR RECOVERY IN THE PROCESS OF
SUPERCRITICAL CARBON DIOXIDE EXTRACTION AND DESORPTION

presented by

Chulaporn Saengcharoenrat

has been accepted towards fulfillment
of the requirements for

Ph . D - degree in Agricul tural Engineering

 

9J27W

Major profess;- I

Date 2/26/98

MSU is an Afflrmafiw Action/Equal Opportunity Institution 0-12771

 

LIBRARY

Michigan State
Unlverslty

 

 

 

PLACE IN RETURN BOX
to remove this checkout from your record.
TO AVOID FINES return on or before date due.

 

I DATE DUE DATE DUE DATE DUE

' Ii

 

HUD

2 3 2001
1,

£1 1 ‘,L
U U

if

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

use chIRC/DabOuopGS—p.“

OPTIMIZATION OF ONION FLAVOR RECOVERY IN THE PROCESS OF
SUPERCRITICAL CARBON DIOXIDE EXTRACTION AND DESORPTION

By

Chulaporn Saengcharoenrat

A DISSERTATION

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

DOCTOR OF PHILOSOPHY

Department of Agricultural Engineering

1998

ABSTRACT

OPTIMIZATION OF ONION FLAVOR RECOVERY IN THE PROCESS OF
SUPERCRITICAL CARBON DIOXIDE EXTRACTION AND DESORPTION

By

Chulaporn Saengcharoenrat

Supercritical carbon dioxide extraction of onion juice in the presence of onion
pulp at 3000 psi, 37°C, with a carbon dioxide flow rate of one liter STP per minute
significantly increased the gravimetric yield of onion oil extracted by 60 % over filtered
onion juice, however, a decrease in total peak area of identified onion flavor compounds
per 20 ug of onion oil was found.

Enhancement of supercritical carbon dioxide extraction of onion oil from onion
juice was accomplished by adsorption of onion oil onto a polymeric adsorbent prior to
supercritical carbon dioxide desorption. At 3000 psi, 37°C, with a carbon dioxide flow
rate of one liter STP per minute, an increase in amount of onion juice loaded onto the
adsorbent from one to six liters resulted in an increase in the gravimetric yield of onion
oil desorbed by weight. Additionally, the weight of onion oil desorbed per volume of
carbon dioxide used also increased. However, the gravimetric yield of onion oil by
percentage with respect to amount of onion juice loaded onto the adsorbent decreased.

The desorption of onion oil from the adsorbent loaded with one liter of onion

juice by 800 liters STP of supercritical carbon dioxide at different pressure, temperature,

and density was investigated. At constant temperature; when density was increased from
0.69 to 0.86 g/ml at 37°C, the gravimetric yield of onion oil by percentage was increased
from 0.0112 to 0.0222 %. When density was increased from 0.79 to 0.86 g/ml at 50°C,
the gravimetric yield of onion oil by percentage was increased from 0.0189 to 0.0231 %.
At constant density of 0.86 g/ml, an increase in temperature from 37°C to 50°C increased

the gravimetric yield of onion oil by percentage from 0.0222 to 0.0231 %.

ACKNOWLEGEMENTS

I would like to thank Dr. Daniel E. Guyer, my major professor, for his suggestions
and support.

I would like to thank Dr. Carl T. Lira for his idea of application of adsorption in
onion oil recovery prior to supercritical carbon dioxide desorption.

I would like to thank my committee members, Dr. Jerry N. Cash, Dr. Carl T. Lira
and Dr. Aj it K. Srivastava for their suggestions and contributions.

I would like to thank Dr. Kris A. Berglund and the Crop and Food Bioproccssing
Center for the funding of equipment and lab facilities.

I would like to thank Dr. Jerry N. Cash, Dr. Mary E. Zabik, and Dr. Gale M.
Strasburg for their permissions to use their lab facilities.

I would like to thank Dr. Stanley Ries for his suggestions in statistical analysis
and experimental design.

I would like to thank Dr. Douglas A. Gage for the Mass Spectrometry Facility and
his suggestions in analysis and an interpretation of mass spectral data.

I would like to thank the Royal Thai Government for giving me a scholarship and
an opportunity to pursue my study.

I would like to thank my family, especially my mother and my sister, for their

support and encouragement.

iv

TABLE OF CONTENTS

LIST OF TABLES ..................................................................................................... vii
LIST OF FIGURES .................................................................................................... ix
1. INTRODUCTION ................................................................................................ 1
2. LITERATURE REVIEW ..................................................................................... 3
2.1. Onion Flavor Development ........................................................................... 3
2.2. Supercritical Carbon Dioxide Extraction ...................................................... 3
2.3. Enhancement of Supercritical Carbon Dioxide Extraction ........................... 5
2.4. Adsorption ..................................................................................................... 8
2.5. Effects of Supercritical Carbon Dioxide Pressure, Temperature, and
Density ........................................................................................................... ll
3. MATERIALS AND METHODS ......................................................................... 13
3.1. Preparation of Raw Material ......................................................................... 13
3.1.1. Onion Juice Preparation ....................................................................... 13
3.1 .2. Adsorbent Preparation ......................................................................... 13
3.2. System Operation .......................................................................................... 14
3.2.1. Adsorption System ............................................................................... 14
3.2.2. Supercritical Carbon Dioxide Extraction and Desorption System ....... 14
3.2.3. GC-MS Analysis of Onion Oil and Identification of Onion
Flavor Compounds ............................................................................... 15
3 .3. Experimentation ............................................................................................ 1 7
3.3.1. Studies of Effects of Onion Pulp on Supercritical Carbon Dioxide
Extraction Yield ................................................................................... 17
3.3.2. Adsorption and Supercritical Carbon Dioxide Desorption of
Onion Oil ............................................................................................. 18
3.3.2.1. Adsorption Selection ............................................................... 18

3.3.2.2. Effects of Different Quantities of Onion Juice Loaded
onto Adsorbent on Supercritical Carbon Dioxide

Desorption of Onion Oil ......................................................... 19

3.3.2.3. Effects of Supercritical Carbon Dioxide Pressure,
Temperature, and Density on Onion Oil Desorption .............. 20
4. RESULTS AND DISCUSSION .......................................................................... 22

V

4.1. Effects of Presence of Onion Pulp on Supercritical

Carbon Dioxide Extraction Yield of Onion Oil ....................................... 22
4.2. Adsorbent Selection ...................................................................................... 25

4.3. Effects of Different Quantities of Onion Juice Loaded onto Adsorbent
on Supercritical Carbon Dioxide Desorption of Onion Oil .......................... 29

4.4. Effects of Supercritical Carbon Dioxide Pressure, Temperature,

and Density on Onion Oil Desorption ........................................................... 36
5. CONCLUSIONS .................................................................................................. 41
6. RECOMMENDATIONS ..................................................................................... 43
BIBLIOGRAPHY ...................................................................................................... 44
APPENDIX A ............................................................................................................ 48
APPENDIX B ............................................................................................................ 54
APPENDIX C ............................................................................................................ 58
APPENDIX D ............................................................................................................ 62
APPENDIX E ............................................................................................................ 63

vi

LIST OF TABLES

Table 2.1. Examples of Adsorption and Supercritical Carbon Dioxide
Regeneration Studies .............................................................................. 7

Table 3.1. Supercritical Carbon Dioxide Regeneration Condition .......................... 21

Table 4.1. Gravimetric Yield, Total Peak Area of Identified Onion Flavor
Compounds, and Composition of Onion Oil .......................................... 24

Table 4.2. GC-MS Analysis Results of Desorbed Onion Oil from XAD-l6
and XUS 40323 ....................................................................................... 28

Table 4.3. Gravimetric Yield and GC-MS Analysis Results of Desorbed
Onion Oil at Different Supercritical Carbon Dioxide Desorption
Conditions ............................................................................................... 39

Table A. 1. Mass Spectral Data of Identified Onion Flavor Compounds ................. 48

Table B. 1. Analysis of Variance of Gravimetric Yield (by Percentage) of
Onion Oil Obtained from Supercritical Carbon Dioxide Extraction
of Onion Juice in an Absence and Presence of Onion Pulp at 1500 Liters
STP of Carbon Dioxide Passed ............................................................... 54

Table 8.2. Analysis of Variance of Total Peak Area of Identified Onion
Flavor Compounds per 20 pg of Onion Oil Obtained from
Supercritical Carbon Dioxide Extraction of Onion Juice in an
Absence and Presence of Onion Pulp at 1500 Liters STP of Carbon
Dioxide Passed ........................................................................................ 55

Table B.3. Analysis of Variance of Gravimetric Yield (by Percentage) of
Onion Oil Desorbed from XAD-l6 and XUS 40323 Loaded with
One Liter of Onion Juice at 800 Liters STP of Carbon Dioxide Passed. 56

Table B.4. Analysis of Variance of Gravimetric Yield (by Percentage) of
Onion Oil Desorbed from XAD-16 at Different Supercritical
Carbon Dioxide Conditions and 800 Liters STP of Carbon Dioxide
Passed ...................................................................................................... 57

vii

Table D. 1. Correlation Equations Between Gravimetric Yield (Y) and Carbon
Dioxide Volume (X), and Their Coefficient of Determination (r2) ....... 62

viii

LIST OF FIGURES

Figure 2.1. Formation of Volatile Sulfur Compounds in Onions ............................... 4

Figure 2.2. Idealized Fixed Bed Breakthrough Curve and Mass Transfer Zone
(MTZ) ........................................................................................................ 9

Figure 3.1. Supercritical Carbon Dioxide Extraction and Desorption System ........... 16

Figure 4.1. Gravimetric Yield of Onion Oil Desorbed From XAD-l6 and
XUS 40323 as a Function of Carbon Dioxide Volume ............................ 27

Figure 4.2. Gravimetric Yield (by Weight) of Onion Oil Desorbed From XAD-16
Loaded with Different Amounts of Onion Juice as a Function of
Carbon Dioxide Volume ........................................................................... 30

Figure 4.3. Gravimetric Yield (by Percentage) of Onion Oil Desorbed From
XAD-16 Loaded with Different Amounts of Onion Juice as
a Function of Carbon Dioxide Volume ..................................................... 31

Figure 4.4. Concentration of Onion Oil in Effluent as a Function of Volmne
of Onion Juice Passed Through the Adsorbent Bed ................................. 33

Figure 4.5. Comparison of Yield of Onion Oil per Carbon Dioxide Volume
Between Direct Supercritical Carbon Dioxide Extraction and
Adsorption-Supercritical Carbon Dioxide Desorption ............................. 35

Figure 4.6. Gravimetric Yield (by Percentage) of Onion Oil Desorbed as
a Function of Carbon Dioxide Volume at Different Supercritical
Carbon Dioxide Conditions ...................................................................... 37

Figure A. 1. GC Chromatogram of Onion Oil Obtained From Supercritical
Carbon Dioxide Extraction of Onion Juice at 3000 psi and 37°C
Showing the Selected Peak of Onion Flavor Compounds ....................... 51

Figure C. l. Supercritical Carbon Dioxide Extraction and Desorption Diagram ........ 58

Figure E. 1. GC Chromatogram of Onion Oil Obtained From Supercritical
Carbon Dioxide Extraction of Onion Juice at 3000 psi and 37°C ........... 63

Figure E.2. GC Chromatogram of Onion Oil Obtained From Supercritical
Carbon Dioxide Extraction of the Mixture of Onion Juice and
Onion Pulp at 3000 psi and 37°C ............................................................. 63

Figure E.3. GC Chromatogram of Onion Oil Desorbed From XAD-16 Adsorbent
at 3000 psi and 37°C ................................................................................ 64

Figure E.4. GC Chromatogram of Onion Oil Desorbed From XUS 40323
Adsorbent at 3000 psi and 37°C .............................................................. 64

Figure E.5. GC Chromatogram of Onion Oil Desorbed From XAD-16 Adsorbent
at 1500 psi and 37°C ................................................................................ 65

Figure E.6. GC Chromatogram of Onion Oil Desorbed From XAD-l6 Adsorbent
at 3000 psi and 50°C ................................................................................ 65

Figure E.7. GC Chromatogram of Onion Oil Desorbed From XAD-16 Adsorbent
at 4159 psi and 50°C ................................................................................ 66

1. INTRODUCTION

Onion oil used in the food industry is produced by steam distillation. High
temperature of distillation processes yields cooked onion flavor oil. Recent studies
(Dron et al.,1997; Nuss et al.,1997; Sinha et al.,1992) showed that supercritical carbon
dioxide extraction of onions was an alternative to steam distillation and solvent
extraction in onion oil production due to the moderate processing temperature. In
addition, supercritical carbon dioxide is nontoxic, inflammable, and easily separated
from extracted products by phase separation. Solvent power of supercritical carbon
dioxide can be adjusted by changing its pressure, temperature, and density in order to
maximize extraction products. However, the process feasibility of the supercritical
carbon dioxide extraction is limited by its high pressure operation.

It was hypothesized that the efficiency of the supercritical carbon dioxide
extraction process could be increased by concentration of onion oil using an adsorption
process prior to supercritical carbon dioxide extraction at optimal processing
conditions. Therefore, optimization of supercritical carbon dioxide extraction of onion
oil including onion pulp utility, onion oil adsorption, and adjustment of pressure and
temperature were studied.

In this study, the following objectives were accomplished:

To examine supercritical carbon dioxide extraction of onion oil from onion juice
as compared to a mixture of onion juice and onion pulp.

To investigate effects of different amounts of onion juice loaded onto adsorbents
on supercritical carbon dioxide desorption of onion oil.

1

To attempt the determination of adsorption breakthrough of onion oil.
To examine effects of supercritical carbon dioxide pressure, temperature, and

density on onion oil desorption from a polymeric adsorbent.

2. LITERATURE REVIEW

2.1. Onion Flavor Development

When onions are crushed, enzyme alliinase (EC 4.4.1.4) is released to convert
flavor precursors, alk(en)yl-L-cysteine sulfoxides, to pyruvate, ammonia, and
intermediate sulfenic acids. The sulfenic acids condense to form odorous thiosulfinates
or decompose to yield aliphatic aldehydes. The thiosulfinates are unstable and break
down nonenzymatically to yield alkyl and alkenyl sulfides, thiophene derivatives, cyclic
sulfur containing compounds, or are oxidized to the corresponding thiosulfonates.
(Borman,1993; Hanum,1995; Whitfield and Last,199l; Whitaker,l976) Lancaster and
Boland (1990) summerized formation of volatile sulfur compounds in onions as shown

in Figure 2.1.

2.2. Supercritical Carbon Dioxide Extraction

Supercritical fluid extraction is a unit operation that exploits the dissolving
power of fluids at temperatures and pressures above the critical point. (Irani and
Funk,1977) In the critical region, a substance that is a gas at normal conditions exhibits
a liquid-like density and a much increased solvent capacity that is pressure-dependent.
By operating in the critical region, the pressure and temperature can be used to regulate
density, which regulates the solvent power of a supercritical fluid. Separation of
extraction products can be done by decreasing the fluid pressure. (McHugh and

Krukonis, 1 993; Rizvi et al.,1986a; Sanders, l 993; Williams, 1 98 1)

o
I all"
R-S-CHz-CH(NH2)COOH fl R-SOH + enaco coon + NH3
(1) (2) (10)
(a -CHSCH-CH)
H
> C I 8*
\
O CH3CH2 o-

(3) \

I
R-S-S-R CHacHQCHO + s
(:1) \ (1 1)

RSR + RSSR + $0;

I
I
I
I 0
* I
RSSR + R-SS-Rw
(6) l
0 (R- CH 3014- CH)

Me @Me + R339 939 + 93339 etc.
()5 (9)

Figure 2.]. Formation of Volatile Sulfur Compounds in Onions.
(1) S- alk(en)yl cysteine sulfoxide; (2) S- alk(en)yl sulfenic acid,

(3) thiopropanal S-oxide; (4) thiosulfinate;

(5) monosulfide; (6) disulfide;

(7) thiosulfonate; (8) dimethyl thiophene;
(9) trisulfide; (10) pyruvate;

(l 1) propanal

R = methyl [CH3], propyl [C3H7], or propenyl [CH3-CH=CH]
From Lancaster and Boland (1990).

5
Carbon dioxide is a particularly attractive solvent in the food industry. It is a

GRAS (generally regarded as safe) substance, incombustible, nonexplosive, and
environment—friendly. The critical point of carbon dioxide is at 31°C and 1073 psi.
Because the critical temperature of carbon dioxide is near ambient, it is an attractive
solvent for processing heat-sensitive materials. Application of supercritical carbon
dioxide extraction in the food industry includes tea and coffee decaffeination, hop
extraction, spice extraction, and flavor extraction. (McHugh and Krukonis,l993; Rizvi
et al.,l986b; Sanders,l993)

The investigation of supercritical carbon dioxide extraction of onion oil from
single strength onion juice by Sinha et al. (1992), Clavey et al. (1994), and Dron ct al.
(1997) showed that moderate extraction temperature yielded onion oil which had fresh-

like onion flavor.

2.3. Enhancement of Supercritical Carbon Dioxide Extraction

Economic feasibility to commercialize supercritical carbon dioxide extraction is
limited because of its high pressure operation which raises capital investment in the
processing plant. A fairly low extractable content in most food and flavor raw materials
necessitates the use of large volume extraction vessels, therefore, it affects processing
efficiency. (Sanders, 1993)

List et al.(1989) found that during supercritical carbon dioxide extraction of
seed oil, measured solubility was less than potential solubility of the oil in dense gas
phase because the measured solubility was a function of the ratio of available oil in

seeds to supercritical carbon dioxide present in an extraction vessel. Therefore, solute

6
availability can be considered as one of limiting factors for supercritical carbon dioxide

extraction.

Henry’s law for a dilute system can be applied to supercritical carbon dioxide
extraction of onion flavor from onion juice (Lira, C.T. :personal communication in
1995). When carbon dioxide and liquid phase are in equilibrium at a particular
temperature and pressure, the solute concentration in the carbon dioxide phase is a
ftmction of the solute concentration in the liquid phase as shown in equation (1).

y = kx (1)

y = the solute concentration in the carbon dioxide phase

k = a constant

x = the solute concentration in the liquid phase
Therefore, the concentration of solute in the liquid phase affects extraction rate.

Pretreatment of raw materials by increasing solute concentration in the liquid
phase aids a more efficient extraction. Treatment of crushed onion by adding enzyme
transpeptidase (flavor precursor formation enzyme) before supercritical carbon dioxide
extraction increased yield of onion oil by 20 to 25%. Total peak area of flavor
compounds increased by 20 to 60%. (Hanum et al.,1995)

Nuss et al.,1997 found that concentration of onion juice from 10°Brix to
18°Brix by reverse osmosis increased supercritical carbon dioxide extraction yield of
onion oil by 33 to 71% without altering its characteristic fresh aroma.

Adsorption is another method used to recover compounds from dilute streams
before supercritical carbon dioxide regeneration of adsorbed compounds. Examples of

adsorption and supercritical carbon dioxide regeneration studies are shown in Table 2.1.

7

Table 2.1. Examples of Adsorption and Supercritical Carbon Dioxide Regeneration
Studies.

 

 

Condition of
Adsorption Process Purpose Supercritical Reference
Type Carbon Dioxide
Activated To recover pesticides from 2200 to 3990 psi, deFilippi et al.,1980
carbon synthetic solutions. 120°C
Activated To recover ethyl acetate 1280 to 1900 psi, Tan and Liou,l988
carbon from synthetic solution. 27 to 65°C
Activated To recover ethyl acetate 1190 to 2396 psi, Srinivasan et al.,
carbon from synthetic solution 35 to 68°C 1990
Synthetic To recover organics from 2200 to 3000 psi, deFilippi et al.,1983
resin industrial waste water 25 to 55°C
adsorbents from manufacturing
pesticides and synthetic
solutions
Activated To recover benzaldehyde 1000 to 3000 psi, Fetzer and Lira,
carbon from synthetic solution 25 to 50°C 1991
and
synthetic
resin

adsorbents

2.4. Adsorption

Adsorption is a separation process of a substance (adsorbate) from one phase by
its accumulation or concentration at the surface of another phase (adsorbent).
(Weber,1985) Adsorption from solution onto a solid may relate to the solvophobic or
lyophobic character of the adsorbate (insolubility of the adsorbate in the solvent phase)
or to a particular affinity of the adsorbate for the surface of the adsorbent. The
adsorption can be influenced by the following factors: (1) the concentration, molecular
weight, molecular size, molecular structure, molecular polarity, and configuration of the
adsorbate; (2) the surface area, physicochemical nature of the surface, the surface
availability, and physical size and form of particles of the adsorbent; and (3) solution
temperature, pH, and the presence of other competing adsorbate compounds in the
solution. (Bemardin, 1 985; Weber, 1 985)

In a fixed bed adsorption process with downflow loading, the upper portions of
the bed are continuously contacted by fresh feed solution and become fully loaded with
the adsorbate first. Solution containing solute not adsorbed by the upper portions of the
bed contacts with the lower portions of the bed which later become fully loaded with
adsorbate. This results in the formation of an adsoption front or a mass-transfer zone
which moves downward. When the solute starts to appear in the effluent, a loading
curve or breakthrough curve rises. (Figure 2.2) Breakthrough is defined as the time or
volume processed when the effluent is no longer acceptable. Complete exhaustion is
defined as when the solute concentration in the effluent is in equilibrium with the

influent. (Bemardin, 1 985; Fox and Kennedy, 1 985)

 

 

 

 

 

 

 

EXHAUSTION

    
  

O
N

BREAKTH ROUGH

»-----<

 

CONCENTRATION
p- - - - - - - - O -

 

0 I MTZ

VOLUME TREATED

Figure 2.2. Idealized Fixed Bed Breakthrough Curve and Mass Transfer Zone (MTZ).
(Bemardin, 1 985)

Criteria of adsorbent selection are based on the characteristics of adsorbate
compounds and their solvent. The hydrophobicity of onion flavor compounds allows
them to be separated from onion juice, which contains significant amount of water, by
hydrophobic adsorbents, i.e. Amberlite XAD-16 and Dowex XUS 40323.

Amberlite XAD-16 and Dowex XUS 40323, styrene-divinyl-benzene resins, are
nonionic and hydrophobic adsorbents. They are used to adsorb hydrophobic molecules
from polar solvents and volatile organic compounds from vapor streams.

Applications of Amberlite XAD-l6 polymeric adsorbent include recovery and

purification of antibiotics, water soluble steroids, enzymes, amino acids, and protein;

10

enzyme immobilization; fruit juice debittering and upgrading; and removal of nonpolar
compounds from polar solvents. (Rohm and Hass Company,1993)

Applications of Dowex XUS 40323 polymeric adsorbent are purification of
antibiotics, vitamins, and amino acids; and removal of phenol, chlorinated
hydrocarbons, aromatic hydrocarbons, pesticides, and surfactants from water. (Dow
Chemical Company, 1 990)

Krings (1993) investigated the adsorption and desorption properties of several
adsorbents for the recovery of aroma compounds in fermentation processes using a
model solution of aroma compounds. The most effective adsorbent was activated
carbon. However, the disadvantage of the activated carbon was poor desorption
capacity when organic solvents were used. This was due to very strong solute-adsorbent
interactions, which could not be overcome by the commonly used desorption solvents.
The second best materials were styrene-divinyl-benzene resins and zeolite. They had
adsorption rates similar to the activated carbon, but also good desorption rate. However,
further studies by Krings (1994) found that the zeolite was not applicable in low
pressure adsorption columns, because its small particle size caused a significant
pressure drop across the fixed bed packings. Therefore, the applicability of the styrene-
divinyl-bezene resins XAD-16 and Lewatit 1064 for low pressure downflow adsorption
of aroma compounds in a fixed bed was investigated by Krings and Berger (1995).

F etzer and Lira (1991) studied adsorption for the recovery of benzaldehyde by
using non-functional polymeric resins, a slightly functionalized weakly basic anion
exchange resin, and a granular activated charcoal. In batch adsorption, the slightly

functionalized weakly basic anion exchange resin and the granular activated charcoal

11
had higher adsorption rate than the non-functional polymeric resins. In comparison

among the non-functional polymeric resins, XUS 40323 had the highest adsorption rate.
When supercritical carbon dioxide was used as a regenerating solvent in benzaldehyde
recovery from the slightly functionalized weakly basic anion exchange resin and the
granular activated charcoal, the amount of benzaldehyde recovered from the granular
activated charcoal was extremely less than from the slightly functionalized weakly basic
anion exchange resin. A higher amount of benzaldehyde was recovered from the
granular activated charcoal by using vacuum steam regeneration.

These studies of using adsorbents in product recovery indicated that both
adsorption and desorption capacity should be taken into the consideration of adsorbent

selection.

2.5. Effects of Supercritical Carbon Dioxide Pressure, Temperature, and Density
In the critical region, carbon dioxide density is strongly pressure and
temperature dependent. Increasing its density at a given temperature or increasing its
temperature at a given density can increase the solvent power of carbon dioxide.
(Sanders,l993) Favati et al.,1991 studied the influence of pressure and temperature on
the supercritical carbon dioxide extraction of evening primrose oil. At constant
temperature, oil solubility in supercritical carbon dioxide and percentage of oil
recovered increased when the pressure was increased. At pressure above 30 MPa,
increasing the extraction temperature resulted in a higher oil solubility, while the
opposite effect was observed at lower pressures. These solubility trends were explained

by recognizing that an increase in extraction temperature affected both the solute and

12

the solvent; the vapor pressure of the solute was increased, while the density of the
solvent was decreased. However, only at low pressure was the density of the gas
notably affected by a change in the temperature. Similar results of the effects of
pressure and temperature on solute solubility in supercritical carbon dioxide and
extraction yields were found in supercritical carbon dioxide extraction of terpenes from
orange essential oil by Temelli et al.(1988) and supercritical carbon dioxide extraction
of volatile constituents from a model plant matrix by Smith et al.(1992).

Desorption of ethyl acetate from activated carbon by supercritical carbon
dioxide was more favorable at higher pressures due to the increase of density, but

optimal temperatures were found to depend on pressure. (Tan and Liou,l988)

3. MATERIALS AND METHODS

3.1. Preparation of Raw Material
3.1.1. Onion Juice Preparation

Two onion varieties were used in experiments: the cultivar Norstar was used in
the studies of effects of onion pulp on supercritical carbon dioxide extraction yield in
1996, and the cultivar Commanche was used in adsorption and supercritical carbon
dioxide desorption of onion oil in 1997. The onions were stored at 28°C.

Onions were peeled, cut, and juiced by an ACME juicerator Model 6001
(ACME Juicer Manufacturing, Co., New Hartford, CT) with a filter which separated the
pulp from the juice. The onion juice was held at room temperature for 45 minutes. In
the studies of effects of onion pulp on extraction yield, onion pulp and juice were mixed
in a certain ratio before holding. In adsorption and desorption studies, the held onion
juice was further filtered through G8 Fisher glass fiber filters (Fisher Scientific,

Pittsburgh, PA) before adsorption to prevent clogging of the adsorbent bed.

3.1.2. Adsorbent Preparation

Amberlite XAD-16 polymeric adsorbent (Rohm and Haas Co., Philadelphia,
PA) and Dowex XUS 40323 (Dow Chemical Co., Midland, M1) were preconditioned
before use in adsorption. 200 ml of methanol were added to approximately 40 g of
adsorbent in a 250 ml beaker. The materials were allowed to stand at room temperature
at least 12 hours before the methanol was decanted and replaced with distilled water.

The adsorbents were rinsed with distilled water three times.

13

14

3.2. System Operation
3.2.1. Adsorption System

Onion juice was gravitationally fed to a 75 ml-preconditioned adsorbent bed by
a 250 ml-separatory funnel attached at the top of the column. Feed flow rate was
approximately 10 ml per minute. The adsorbent bed was backwashed with distilled

water and was drained before supercritical carbon dioxide regeneration.

3.2.2. Supercritical Carbon Dioxide Extraction and Desorption System

The extraction and desorption were performed with the apparatus shown in
Figure 3.1. Industrial grade carbon dioxide (AGA Gas, Inc., Ann Arbor, MI) was
compressed by an air-driven booster compressor (Haskel, Inc., Burbank, CA) and stored
in a two-liter reservoir (Autoclave Engineers, Inc., Erie, PA). Carbon dioxide flowed
through a shut-off valve (Autoclave Engineers, Inc., Erie, PA) and a pressure regulator
(Tescom Corporation, Elk River, MN) was used to set the pressure to the desired value.
Two pressure gauges (range : 0 to 10,000 psi) were used to observe the pressure in the
reservoir and extraction vessel. A two-liter bolted closure stirred autoclave Model
BCOZOOSSOSAG (Autoclave Engineers, Inc., Erie, PA) with a temperature controller
was used in the studies of effects of pulp. In adsorption and desorption studies, a high
pressure stainless steel tube with 2.12 cm id. and 43 cm length (High Pressure
Equipment, Co., Erie, PA) was used as an adsorption column and a desorption vessel.
Wire mesh screens, were placed at both ends of this vessel to retain adsorbents. This

tubular vessel was wrapped with heating tape and insulated. The temperature was

15
monitored by a thermocouple and temperature controller (Omega Engineering, Inc.,
Stamford, CT).

The carbon dioxide flowed through the extraction vessel in the upflow direction
and was passed to a separator through a shut-off valve (Autoclave Engineers, Inc., Erie,
PA) and a heated micrometering valve (Autoclave Engineers, Inc., Erie, PA) used for
adjusting the exit flow to the desired value. The heat probe which was used to heat this
micrometering valve was controlled at 40°C in order to prevent freezing and to
minimize thermal decomposition of onion flavor. However, actual temperature within
the valve and around the flow was not measured. At this valve, the pressure was
reduced to atmospheric pressure to separate extraction products from the carbon
dioxide. The separated products were collected in a separator which consisted of a 3.7
ml-glass vial placed in a glass side arm test tube with a rubber stopper attached to a
stainless steel tube connected to the heated micrometering valve. The depressurized
carbon dioxide was passed through a flow meter and totalizer (American Dry Test

Meter Model DTM-200A-3, American Meter Co., Philadelphia, PA) and vented.

3.2.3. GC-MS Analysis of Onion Oil and Identification of Onion Flavor
Compounds
Onion oil solution was prepared by dissolving 0.02 g of onion oil in 2 m1 of
methylene chloride. 2 ul of the solution was injected into the GC-MS. The GC-MS
system consisted of a JEOL AX-505H double-focusing mass spectrometer connected to

a Hewlett-packard HP5890J gas chromatograph via heated interface. GC separations

l6

3&2 .GEQV .Eoumxm 558809 can :ouoabxm 2585 5530 Boutoceasm ._.m “mama

.3258». 95a 8
0.5.00.5 3025 31....

Con..> guaguxmv
saga-a‘ thwum Id
.5 Av

9C.) Smugddou

.= 2...: KW.
O :u _ LHI .2.... s... 2..»

Econ-:33! faucet 50:35.1
«£83:

 

 

 

 

 

 

 

 

  
  

 

 

 

3.3.: 8::
3.... 3....

 

 

 

 

 

17
were accomplished on a DB-l fused silica capillary column (30 mm x 0.25 mm id. x

0.25 mm film coating) with helium as a carrier gas at a flow rate of 1 ml per minute.
The GC temperature program was as follows: holding at initial temperature (35°C) for 5
minutes, increasing temperature from 35°C to 200°C at 5°C per minute, and holding
temperature at 200°C for 20 minutes. The mass spectrometer was operated at the
following condition: interface temperature was at 280°C, ion source temperature was at
220°C, and the scan rate was 1 second per scan over the m/z range from 35 to 500.
Mass spectra were obtained by electron ionization at 70 eV.

Mass spectra and elution order obtained from GC-MS analysis were compared
to those from references (Dron,1994; Hanum,1995; Nuss,1994) in order to identify
onion flavor compounds. Total peak area of identified onion flavor compounds per
weight of onion oil injected into the GC-MS was reported. Percentage peak area of each
identified onion flavor compound was calculated based on the total peak area of

identified onion flavor compounds.

3.3. Experimentation

Orders of experiments were randomized within each sub-study in order to avoid
variation in onions due to storage time.
3.3.1. Studies of Effects of Onion Pulp on Supercritical Carbon Dioxide Extraction
Yield

Two weight ratios of onion juice to onion pulp, 1:0 and 2:1, were tested. 800 g
of the material was used in supercritical carbon dioxide extraction at 3000 psi and

37°C. Carbon dioxide flow rate was one liter STP per minute. The autoclave vessel was

18

operated at the stirring speed of 500 rpm. After 1500 liters STP of carbon dioxide were
passed, extraction products were weighed and analyzed by GC-MS.
Gravimetric yield (by percentage) was calculated by using the following

equation:

gravimetric yield, % = weight of onion oil extracted @1 gram) X 100 (2)
weight of onion juice (or pulp) used (in gram)

This experiment was designed as a randomized complete block design with

replications as blocks. Three replications of the experiment were performed.

3.3.2. Adsorption and Supercritical Carbon Dioxide Desorption of Onion Oil
3.3.2.1. Adsorbent Selection

Two types of adsorbents, Amberlite XAD-l6 and Dowex XUS 40323, were
used. 75 ml of each adsorbent were loaded with one liter of onion juice. Adsorbents
were regenerated by supercritical carbon dioxide at 3000 psi, 37°C, and one liter STP
per minute. Desorbed oil was weighed at discrete intervals, afier 200, 400, 600, and 800
liters STP of carbon dioxide were passed. Gravimetric yield (by percentage) was
calculated by using equation (2).This experiment was triplicated.

Gravimetric yield (by percentage) calculated from the final weight of desorbed
onion oil at 800 liters STP of carbon dioxide passed was used in statistical analysis.
Therefore, this experiment was considered as a randomized complete block design with
replications as blocks. The adsorbent yielding the greatest quantity of desorbed oil was

chosen for further studies.

19
GC-MS analysis was performed on the desorbed oil from the first replication of

desorption for each adsorbent.

3.3.2.2. Effects of Different Quantities of Onion Juice Loaded onto Adsorbents on
Supercritical Carbon Dioxide Desorption of Onion Oil

Three different amounts of onion juice, two, three, and six liters, were loaded
onto 75 ml of selected adsorbent. Supercritical carbon dioxide regeneration of the
adsorbent was run at 3000 psi, 37°C, and one liter STP of carbon dioxide per minute.
After 200, 400, 600, 800, and 1500 liters STP of carbon dioxide were passed, desorbed
oil was weighed. Gravimetric yield (by percentage) was calculated by using equation
(2).

Three replications of the two-liter and three-liter loading of onion juice onto the
adsorbent and the following desorption were performed.

However, six-liter loading of onion juice onto the adsorbent was performed
once in order to attempt to determine a breakthrough curve. During preparation of onion
juice, 20 ml of onion juice was collected from every one liter of onion juice prepared.
These onion juice samples were mixed together and 60 ml of the mixed onion juice was
extracted by 500 liters STP of supercritical carbon dioxide extraction at 3000 psi, 37°C,
and one liter STP of carbon dioxide per minute to determine the concentration of onion
oil in the onion juice fed onto the adsorbent column. During the adsorption, 80 ml of
effluent was collected at different volumes of onion juice passed. 60 ml of each
collected sample was extracted by supercritical carbon dioxide at 3000 psi, 37°C, and

one liter STP of carbon dioxide per minute. Onion oil extracted from each effluent was

20

weighed after 500 liters STP of carbon dioxide were passed. Concentration of onion oil
in each effluent was calculated by using equation (3).
Concentration of onion oil (ppm) = weight of onion oil extracted (in milligram) (3)
weight of effluent sample used (in kilogram)
The adsorbent loaded with six liters of onion juice was regenerated by
supercritical carbon dioxide at 3000 psi, 37°C, and one liter STP per minute. After 200,
400, 600, 800, and 1500 liters STP of carbon dioxide were passed, desorbed onion oil

was weighed. Gravimetric yield (by percentage) was calculated by using equation (2).

3.3.2.3. Effects of Supercritical Carbon Dioxide Pressure, Temperature, and
Density on Onion Oil Desorption

One liter of onion juice was loaded onto selected adsorbent. Supercritical
carbon dioxide regeneration of adsorbent was carried out at different conditions as
shown in Table 3.1. Desorbed oil was weighed when 200, 400, 600, and 800 liters STP
of carbon dioxide were passed. Oil obtained at 800 liters STP of carbon dioxide passed
was analyzed by GC-MS. Weight of this oil was used in statistical analysis. The

experiment was triplicated and considered as a randomized complete block design.

21

Table 3.1. Supercritical Carbon Dioxide Regeneration Condition.

 

 

Temperature, °C Pressure, psi Density, g/rnl
37 1500 0.69
3000 0.86
50 3000 0.79

4159 0.86

 

4. RESULTS AND DISCUSSION

4.1. Effects of Presence of Onion Pulp on Supercritical Carbon Dioxide Extraction
Yield of Onion Oil

Supercritical carbon dioxide extraction of the onion juice and the mixture of the
onion juice and onion pulp was performed at 3000 psi, 37°C, one liter STP of carbon
dioxide per minute, and 1500 liters STP of carbon dioxide passed. Table 4.1 shows
onion oil gravimetric yields and results from the GC-MS analysis of 20 ug of onion oil
injected. Onion oil gravimetric yields obtained from the onion juice and the mixture of
the onion juice and the onion pulp were significantly different at the five percent
significance level. (Table 3.1)

Identification of onion flavor compounds was done by comparison of the mass
spectra and the order of elution from the GC-MS analysis (Table Al) to references
(Dron,1994; Hanum,1995; Nuss,1994). Peak area of each identified onion flavor
compound was summed. This total peak area was used in calculation of percentage
peak area for each identified compound. Larger total peak area of identified onion
flavor compounds per 20 ug of onion oil injected into the GC-MS was found in the
analysis of the onion oil obtained from onion juice extraction. However, there was no
significant difference in the total peak area of identified onion flavor compounds per 20
ug of onion oil injected into the GC-MS between the onion oil obtained from the onion
juice extraction and the extraction of the mixture of the onion juice and the onion pulp.
(Table B.2) Similar composition of these two onion oils is also shown in Table 4.1. GC

chromatograms of both onion oils are shown in Figure El and E2. Rising of the

22

23
baseline of these chromatograms during the run probably resulted from the oven
temperature programming. A large peak was found at the end of the run because the
column was heated in order to drive off unwanted compounds at the final stage of the
oven temperature programming.

The higher gravimetric yield but lower total peak area of identified onion flavor
compounds per 20 pg of onion oil injected into the GC-MS of the mixture of the onion
juice and the onion pulp indicated that some unidentified compounds were extracted.
However, the nonsignificant difference in the total peak area of identified onion flavor
compounds per 20 ug of onion oil injected into the GC-MS between these two onion oil
indicated that the onion pulp contained a significant amount of onion flavor.
Suspension of the onion pulp in the mixture might increase contact of carbon dioxide

and raw materials during the extraction.

24

Table 4.1. Gravimetric Yield, Total Peak Area of Identified Onion Flavor Compounds,
and Composition of Onion Oil.

 

 

 

Onion Oil Obtained from
Mixture of Onion
Onion Juice Juice and Onion Pulp
Gravimetric Yield, %by weight 0.02809‘ 0.04552‘
Total peak area of identified flavor 21551 .12NS 14503.58NS
compounds per 20 ug of onion oil
Percentage peak area:
thiopropanal S-oxide 5.68 4.75
3,4-dimethylthiophene 3.98 3.00
methyl l-propenyl disulfide 5.92 6.47
l-propenyl propyl disulfide 2.80 . 3.78
3-ethyl-1,2-dithi-(4/5)-ene 26.00 28.09
methyl propyl trisulfide 1.46 2.63
methyl l-propenyl trisulfide 9.27 7.89
dirnethyl tetrasulfide 3.87 3 .95
diallyl thiosulfinate 2.53 3 .58
l-propenyl propyl trisulfide 1 1 .72 12.61
diallyl trisulfide 16.00 13 .90
diallyl tetrasulfide 8.72 7.72
methyl 3,4-dimethyl-2-thienyl 2.05 1.62

disulfide

. Significant difference at the five percent level.

NS Nonsignificant difference at the five percent level.

25
4.2. Adsorbent Selection

Polymeric adsorbents, XAD-16 and XU S 40323, loaded with one liter of onion
juice were regenerated by supercritical carbon dioxide at 3000 psi and 37°C. The
carbon dioxide flow rate was fixed at one liter STP per minute. Gravimetric yield by
percentage was plotted as a function of carbon dioxide volume in Figure 4.1. Non-linear
regression analysis using a tentative model equation (18) in Appendix C was performed
in order to find a functional relationship between these two variables. Non-linear
regression equations and their coefficient of determination, r2, are shown in Table D. 1.

Onion oil desorption curves shown in Figure 4.1 were inverse exponential. At
the beginning of desorption, the high concentration of onion oil in the adsorbent phase
resulted in a high desorption rate. A decrease in the onion oil concentration in the
adsorbent phase during desorption resulted in a lower desorption rate. The steeper slope
of the desorption curve for XAD-16 indicated a more rapid desorption of onion oil. This
could partially result from the higher concentration of onion oil on XAD-16 as
implicitly shown in a higher constant b1 in the non-linear regression equation in Table
D.1. The higher constant b1 of XAD-16 also implicitly showed the higher adsorption
capacity. The higher constant b2 (the higher constant k) of XUSe40323 showed that
onion oil could be easier desorbed from XUS 40323 than fi'om XAD-16. However,
onion oil yields obtained from both adsorbents after 800 liters STP of carbon dioxide
passed were not significantly different at the five percent significance level. (Table 8.3)

GC chromatograms of the onion oil are shown in Figure E3 and E4. Table 4.2
shows results from the GC-MS analysis. The number of identified onion flavor

compounds in the onion oil obtained from adsorption and supercritical carbon dioxide

26

desorption were less than from supercritical carbon dioxide extraction. This indicated
that some identified onion flavor compounds in onion juice were not adsorbed onto the
adsorbent. However, the composition of the onion oil desorbed from XAD-l6 and XUS
40323 were slightly different. According to desorption results, XAD-16 was selected
for further studies.

It was found that both polymeric adsorbents, XAD-l6 and XUS 40323, changed
their color from white to pink after onion oil adsorption and supercritical carbon
dioxide desorption cycle. The pink color was possibly developed by the reaction
between carbonyl compounds and free amino acids in the onion juice. (Lancaster and

Boland, 1990)

27

 

0.03-
ale
ale
a ................................. §
2. 002~ 9% ..................
“03 ------------------
>- n’."’. ”””” 0
g j: l,/’% O
16 fix” 0
.§ S
> ,z',/
E 0.01-r ,-’// O
0 .’.'//
.’/
.’/
.’/
- .’/
.’/
.’l
.’l
I
I,
0.00% I

 

I 1 I I I I l
0 1 00 200 300 400 500 600 700 800
Volume of Carbon Dioxide, Liters STP

*6 XAD-16

<> XUS 40323
-- ------ XAD-16
----- XUS 40323

Figure 4.1. Gravimetric Yield of Onion Oil Desorbed From XAD-16 and XU S 40323 as
a Function of Carbon Dioxide Volume.

28

Table 4.2. GC-MS Analysis Results of Desorbed Onion Oil from XAD-16 and
XUS 40323.

 

 

 

Onion Oil Desorbed from
XAD-16 XUS 40323
Gravimetric yield, % by weight 0.02216NS 0.01915NS
Total peak area of identified flavor 12494.31 13063.08
compounds per 20 pg of onion oil
Percentage peak area:
3 ,4-dimethylthiophene 6.45 3. l 0
methyl 1-propenyldisulfide 3.27 5.24
1-propenyl propyl disulfide 1 .91 3.70
3-ethyl-1 ,2-dithi-(4/5)-ene 28 .62 27.50
methyl propyl trisulfide l .01 1.09
methyl l-propenyl tn'sulfide 9.55 6.62
dimethyl tetrasulfide l .62 5.19
diallyl thiosulfinate 1.78 3.16
l-propenyl propyl trisulfide 40.83 39.69
diallyl trisulfide 4.95 4.73

 

NS Nonsignificant difference at the five percent level.

29
4.3. Effects of Different Quantities of Onion Juice Loaded onto Adsorbent on
Supercritical Carbon Dioxide Desorption of Onion Oil.

Different amounts of onion juice were loaded onto XAD-l6 before supercritical
desorption at 3000 psi, 37°C, and one liter STP of carbon dioxide per minute. Figure
4.2 and 4.3 show the gravimetric yield by weight and by percentage, respectively, of
onion oil desorbed from XAD-16 loaded with different amounts of onion juice as a
function of the volume of carbon dioxide passed through the vessel. Non-linear
regression analysis using a tentative model equation (1 8) in Appendix C was performed
in order to find a functional relationship between the gravimetric yield and the volume
of carbon dioxide passed. The correlation equations and their coefficient of
determination, r2, are shown in Table D. 1.

Desorption curves shown in Figure 4.2 and 4.3 were inverse exponential.
Increasing the amount of onion juice loaded from one liter up to six liters resulted in an
increase in the amount of onion oil desorbed with respect to total carbon dioxide
volume passed. This indicated that the complete XAD-l6 exhaustion was not reached.
However, initial desorption rate increased when the amount of onion juice loaded
increased from one to three liters. Similar initial desorption rate was found when three
and six liters of onion juice were loaded. This indicated that carbon dioxide was
probably saturated by the onion oil, therefore, the amount of onion oil desorbed was
limited.

A decrease in percentage yield of onion oil desorbed when the amount of onion

juice loaded was increased (Figure 4.3) indicated that the onion oil started to leak from

30

0.91

Gravimetric Yield, 9

 

 

 

V I l I l I l l
0 200 400 600 800 1000 1200 1400 1600
Volume of Carbon Dioxide, Liters STP

1 liter of onion juice loading
2 liters of onion juice loading
3 liters of onion juice loading
6 liters of onion juice loading
- ------- 1 liter of onion juice loading
............. 2 liters of onion juice loading
----- 3 liters of onion juice loading
— 6 liters of onion juice loading

ODDS

Figure 4.2. Gravimetric Yield (by Weight) of Onion Oil Desorbed From XAD-16
Loaded with Different Amounts of Onion Juice as a Function of Carbon Dioxide
Volume.

31

0.03-

a
.-".o
’
"

O
O
N
l
\
amen.

0.01—

Gravimetric Yield, °/o

 

 

0.00

 

l I l l I l l l
0 200 400 600 800 1000 1200 1400 1600
Volume of Carbon Dioxide, Liters STP

ale 1 liter of onion juice loading
[:1 2 liters of onion juice loading
A 3 liters of onion juice loading
0 6 liters of onion juice loading
-- ------ 1 liter of onion juice loading
2 liters of onion juice loading
----- 3 liters of onion juice loading
— 6 liters of onion juice loading

Figure 4.3. Gravimetric Yield (by Percentage) of Onion Oil'Desorbed From XAD-16
Loaded with Different Amounts of Onion Juice as a Function of Carbon Dioxide
Volume.

32
the adsorption column during the adsorption which increased the onion oil

concentration in the effluent during the adsorption as shown in Figure 4.4.

Results of breakthrough determination described in section 3.3.2.2 are shown in
Figure 4.4. The breakthrough or loading curve was a plot of column effluent onion oil
concentration (or concentration of onion oil remaining in the effluent) versus the
volume of onion juice processed. During the adsorption process, onion juice was fed
onto the fixed-bed column of the XAD-16 adsorbent. At the beginning of the adsorption
process, the onion oil concentration in the effluent was low. This indicated that onion
oil probably was rapidly adsorbed by the adsorbent. After onion juice was continuously
fed onto the adsorbent column, the onion oil concentration in the effluent started to
increase and approached the onion oil concentration of the onion juice (feed) which was
335 ppm. It was likely that onion oil could start to leak from the column. This possibly
resulted from the saturation of onion oil in the adsorbent phase which could decrease
the adsorption ability of the adsorbent. However, Lira, C.T. (personal communication in
1997) suggested that the competitive adsorption probably occurred. The competitive
adsorption was described by Bemardin (1985) as the chromatographic effect which
resulted from the presence of competing adsorbent compounds. Since physical
adsorption is a reversible phenomenon, the presence of materials with a particularly
high affinity for the adsorbent can tend to displace materials of lesser affinity from the
adsorbent under continued application. (Bemardin,]985)

The leakage of onion oil from the adsorption column due to the exhaustion of
the adsorbent and the competitive adsorption possibly affected the percentage yield of

onion oil desorbed when the amount of onion juice loaded was increased.

33

 

 

400—
E “““““““““““““““““““““““““ .- ------- fu“.
0- o
0- 3004
c.
.9
E
‘5
o 200—
c
O
U
5
t:
.9 100—
c
O

O O
O ' I ' l ' I ' I ‘ I ' l
O 1 000 2000 3000 4000 5000 6000

Volume of Onion Juice Passed, ml

. effluent
----- onion juice (feed)

Figure 4.4. Onion Oil Concentration in Effluent as a Function of Volume of Onion
Juice Passed Through the Adsorbent Bed.

From mass balance of onion oil during the adsorption, total mass of onion oil in
the onion juice passed through the adsorbent column equaled to the sum of mass of
onion oil adsorbed onto the adsorbent column and mass of onion oil remaining in the
effluent. Total mass of onion oil in the onion juice passed through the adsorbent column
was calculated by multiplying the onion oil concentration of the onion juice by the
volume of onion juice passed through the adsorbent column. By integrating the area
under the breakthrough curve using the trapezoidal rule, the mass of onion oil
remaining in the effluent was calculated. Then, the mass of onion oil adsorbed onto the

adsorbent was calculated.

34

It was assumed that onion oil concentration of onion juice in each experiment
was constant. Therefore, the breakthrough curve from loading of six liters of onion juice
onto the adsorbent could be used in the calculation of the mass of onion oil adsorbed for
two-liter loading and three-liter loading.

The calculated mass of onion oil adsorbed onto the adsorbent by loading two,
three, and six liters of onion juice was 0.613, 0.800, and 0.868 g, respectively. After
regeneration of this adsorbent by 1500 liters STP of supercritical carbon dioxide at 3000
psi, 37°C, and one liter STP per minute, 0.529, 0.663, and 0.837 g of onion oil was
recovered from the adsorbent loaded with two, three, and six liters of onion juice,
respectively. The recovery percentage of onion oil was 86.3, 82.9, and 96.4% for two,
three, and six-liter loading of onion juice, respectively.

When the amount of onion juice loaded was increased from two to three liters,
the recovery percentage of onion oil decreased. This probably resulted from the
incomplete desorption of onion oil from the adsorbent. However, when the amount of
onion juice loaded was increased from three to six liters, the recovery percentage
increased. This indicated that other compounds were likely to be desorbed from the
adsorbent loaded with six liters of onion juice. (Lira, C.T. :personal communication in
1998)

Comparison of yield of onion oil per carbon dioxide volume between the direct
supercritical carbon dioxide extraction and the combination process of the adsorption

and the supercritical carbon dioxide desorption is shown in Figure 4.5.

35

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0.0006
0.00056
0)
.12
X
.9
0 0.00044
8
*5 0.0004 ..
0 0.00035
‘6
52
:1
as
Q
’5 0.0002 +
5 0.00015
'E
O
“6
O)
0
Direct Desorption Desorption Desorption
Extraction From 2 From 3 From 6
Liters Liters Liters
Loaded loaded Loaded

Figure 4.5. Comparison of Yield of Onion Oil per Carbon Dioxide Volume Between
Direct Supercritical Carbon Dioxide Extraction and Adsorption-Supercritical Carbon
Dioxide Desorption.

The weight of onion oil desorbed after 1500 liters STP of carbon dioxide passed
per the volume of carbon dioxide used was 3.50x10'4, 4.42x10'4, and 5.58x104 g of oil
per one liter STP of carbon dioxide when two, three, and six liters of onion juice were
loaded onto the adsorption column respectively. However, only 1.53x10'4 g of onion oil
per one liter STP of carbon dioxide was found when onion juice was directly extracted
by supercritical carbon dioxide at the same condition. Adsorption with high loading of
onion juice increased the effectiveness of supercritical carbon dioxide because of the
increase in available amount of onion oil in the adsorbent phase. However, yield per

unit of raw material decreased.

36
4.4 Effects of Supercritical Carbon Dioxide Pressure, Temperature, and Density
on Onion Oil Desorption.

XAD-l6 polymeric adsorbent loaded with one liter of onion juice was
regenerated by supercritical carbon dioxide at different conditions. Onion oil desorbed
was weighed after 200, 400, 600, and 800 liters STP of carbon dioxide were passed.
Figure 4.6 shows gravimetric yield of desorbed onion oil by percentage as a function of
carbon dioxide volume passed at different supercritical carbon dioxide conditions.
Correlation between the gravimetric yield and the volume of carbon dioxide passed was
determined by non-linear regression analysis using a tentative model equation (18) in
Appendix C. Results of the non-linear regression analysis including correlation
equations and their coefficient of determination, r2, are shown in Table D].

At constant temperature, an increase in density by increasing pressure resulted in
an increase in onion oil desorption. This indicated that the solvent power of
supercritical carbon dioxide probably increased. However, the extraction ability of
supercritical carbon dioxide for onion flavor compounds was decreased because of a
decrease in total peak area of identified onion flavor compounds per 20 pg of onion oil
injected into the GC-MS. (Table 4.3) A significant increase of onion oil desorbed due to
the increase of supercritical carbon dioxide density was observed at 37°C. At the higher
temperature (50°C), the increase of onion oil desorbed due to the increase of density
resulting from increased pressure was not significant. This possibly resulted from less

difference in density changes at 50°C than at 37°C.

37

 

 

0.03-
“ ()1?
Q

°\° O 8 .................. g
D. 0.02-4 ......
E’ O 3, ... M, we
> "M Cl
.g D
Iii .....
.§ . 8
S O. 01 "" I.’ _________ a _____________
o —

 

 

0.00 I I I I l I I I
0 1 00 200 300 400 500 600 700 800
Volume of Carbon Dioxide, Liters STP

3000 psi, 37 C, 0.86 g / ml
3000 psi, 50 C, 0.79 g I ml
1500 psi, 37 C, 0.69 g I ml
4159 psi, 50 C, 0.86 g / ml
- ------- 3000 psi, 37 C, 0.86 g / ml
~ -------------- 3000 psi, 50 C, 0.79 g I ml
----- 1500 psi, 37 C, 0.69 g I ml
— 4159 psi, 50 C, 0.86 g / ml

ODD->16

Figure 4.6. Gravimetric Yield (by Percentage) of Onion Oil Desorbed as a Function of
Carbon Dioxide Volume at Different Supercritical Carbon Dioxide Conditions.

38

At constant density, onion oil desorbed increased when the temperature of
supercritical carbon dioxide increased. This could result from higher solubility of onion
oil in supercritical carbon dioxide at higher temperatures which is affected by an
increase in the vapor pressure of onion oil.

From statistical analysis of percentage yield of onion oil desorbed after 800
liters STP of carbon dioxide passed (Table 8.4), there was no significant difference in
onion oil yield desorbed at 3000 psi, 37°C; 3000 psi, 50°C; and 4159 psi, 50°C at the
five percent significance level. However, the onion oil yields obtained at these three
conditions were significantly different from one obtained at 1500 psi, 37°C at the five
percent significance level.

GC chromatograms of the onion oil desorbed at different conditions are shown
in Figure E3, E5, E6, and E7 Composition of the onion oil (Table 4.3) was similar.
However, the difference in the number of the identified onion flavor compounds in each
onion oil obtained from different conditions possibly resulted from the different solvent
power of supercritical carbon dioxide at different conditions and the thermal
decomposition of the onion flavor compounds at different conditions.

Effects of supercritical carbon dioxide desorption temperature and pressure on
the composition of onion oil were probably diminished by long desorption time and the
high injection temperature of the GC-MS analysis which induced decomposition of the

onion flavor compounds as documented by Clavey et al. (1994).

39

Table 4.3. Gravimetric Yield and GC-MS Analysis Results of Desorbed Onion Oil at
Different Supercritical Carbon Dioxide Desorption Conditions.

 

Supercritical Carbon Dioxide Condition

 

 

 

 

Temperature,°C 50
Pressure, psi 1500 3000 3000 4159
Density, g/ml 0.69 0.86 0.79 0.86
Gravimetric yield, % by weight 0.0112c 0.0222ab 0.01898” 0.0231“
Total peak area of identified 25556.08 12494.31 24240.91 17393.63
flavor compounds per 20 pg of
onion oil
Percentage peak area:

3 ,4-dimethylthiophene 5. 19 6.45 4.25 4.75
methyl propyl disulfide 0.62 - 1 .1 8 1 .44
metyl l-propenyl disulfide 2.43 3.27 2.14 0.94
dipropyl disulfide - - 0.35 1.89
dimethyl trisulfide - - 0.34 -
l-propenyl propyl disulfide 5.58 1 .91 5.13 l .68
3-ethyl-l,2-dithi-(4/5)-ene 28.52 28.62 23.54 19.19
methyl propyl trisulfide 0.42 1.01 0.97 1.69
methyl l-propenyl trisulfide 6.01 9.55 6.93 7.19
dimethyl tetrasulfide 0.79 1 .62 2. 16 2.51
diallyl thiosulfinate 1.48 1.78 1.87 2.1 1

 

40
Table 4.3. (cont’d)

 

Supercritical Carbon Dioxide Condition

 

 

 

 

Temperature,°C 37 50
Pressure, psi 1500 3000 3000 4159
Density, g/ml 0.69 0.86 0.79 0.86
Percentage peak area:

1-propenyl propyl trisulfide 36.79 40.83 33.10 36.76
diallyl trisulfide 1 1.15 4.95 16.27 14.81
methyl 3,4-dimethyl-2-thioenyl 1.03 - 1.78 5.04
disulfide

 

2‘ b’ c Significant difference at the five percent level.

5. CONCLUSIONS

1. Supercritical carbon dioxide extraction of the onion oil at 3000 psi, 37°C, one liter
STP of carbon dioxide per minute, and 1500 liters STP of carbon dioxide passed from
the onion juice in the presence of the onion pulp resulted in 60 % higher gravimetric
yield of onion oil but lower in total peak area of identified onion flavor compounds per
20 pg of onion oil compared to tests with an absence of the onion pulp.

2. In the regeneration of onion-oil-loaded polymeric adsorbents by supercritical carbon
dioxide at 3000 psi, 37°C, one liter STP of carbon dioxide per minute, and 800 liters
STP of carbon dioxide passed, percentage gravimetric yield of onion oil desorbed from
XAD-16 was slightly higher than from XUS 40323.

3. Increasing onion juice loading onto XAD-16 adsorbent ( based on two, three, and
six liters of the onion juice) before supercritical carbon dioxide desorption at 3000 psi,
37°C, one liter STP of carbon dioxide per minute, and 1500 liters STP of carbon
dioxide passed resulted in an increase in gravimetric yield of onion oil by weight.
However, percentage gravimetric yield of onion oil was decreased.

4. Onion oil desorption rate was increased by increasing supercritical carbon dioxide
density at constant temperature (from 0.69 to 0.86 g/ml at 37°C and from 0.79 to 0.86
g/ml at 50°C) or increasing temperature at constant density (from 37 to 50°C at 0.86
g/ml).

5. Adsorption of onion juice onto adsorbent prior to supercritical carbon dioxide

desorption increased the efficiency of supercritical carbon dioxide extraction of onion

41

42

juice by increasing gravimetric yield per carbon dioxide volume used and making it

possible to use a smaller extraction vessel.

6. RECOMMENDATIONS

1. Analysis techniques for analyzing all components in onion juice i.e. onion flavor

compounds and competitive adsorbates should be developed in order to determine the

capacity and selectivity of adsorbents.

2. Adsorption capacity of the regenerated adsorbents after each adsorption-

regeneration cycle should be studied in order to determine the reusability of the

adsorbents.

43

BIBLIOGRAPHY

BIBLIOGRAPHY

Bemadin, PE. 1985. Experimental design and testing of adsorption and adsorbates. In
Adsorption Technology: A Step-by- Step Approach to Process Evaluation and
Application. (F .L. Slejko, ed.). Marcel Dekker, Inc., New York, New York.

Borman, S. 1993. Component profiles of onion, garlic obtained. Chem. & Eng. News
70, 31-32.

Calvey, E.M., Matusik, J .E., White, K.D., Betz, J .M., Block, E., Littlejohn, M.H.,
Naganathan, S. and Putman, D. 1994. Off-line supercritical fluid extraction of
thiosulfinates from garlic and onion. J. Agric. Food Chem. 42, 1335-1341.

deFilippi, R.P., Krukonis, V.J., Robey, R.J. and Modell, M. 1980. Supercritical fliud
regeneration of activated carbon for adsorption of pesticides. Report EPA-600/2-
80-054, March.

deFilippi, RP. and Robey, R.J. 1983. Supercritical fluid regeneration of adsorbents.
Report EPA-600/S2-83-03 8, June.

Dow Chemical Company. 1990. XUS 40323 sorptive polymer. Application
information.

Dron, A. 1994. Yield and quality of onion flavor oil obtained by supercritical fluid
extraction and other methods. Master’s thesis. Michigan State University,
East Lansing, Michigan.

Dron, A., Guyer, D.E., Gage, D.A., and Lira, CT. 1997. Yield and quality of onion
flavor oil obtained by supercritical fluid extraction and other methods. J. Food
Proc. Eng. 20, 107-123.

F avati, F ., King, J .W. and Mazzanti, M. 1991. Supercritical carbon dioxide extraction
of evening primrose oil. JAOCS 68(6), 422-427.

44

45

F etzer, F. and Lira, CT. 1991. Adsorption for the recovery of benzaldehyde. Progress
report. Michigan State University, East Lansing, Michigan.

Fox, CR. and Kennedy, DC. 1985. Conceptual design of adsorption systems. In
Adsorption Technology: A Step-by- Step Approach to Process Evaluation and
Application. (F.L. Slejko, ed.). Marcel Dekker, Inc., New York, New York.

Hanum, T. 1995. The role of allinase and y-glutamyl transpeptidase on flavor
enhancement in onion (Allium cepa L.) Ph.D. dissertation. Michigan State
University, East Lansing, Michigan.

Hanum, T., Sinha, N.K., Guyer, DE. and Cash, J .N. 1995. Pyruvate and flavor
development in macerated onion (A Ilium cepa L.) by y-glutarnyl transpeptidase
and exogenous C-S lyase. Food Chem. 54, 183-188.

Irani, CA. and Funk, E.W. 1977. Separation using supercritical gases. In Recent
Developments in Separation Science, Vol. III, Part A. CRC Press, Inc., West
Palm Bleach, Florida.

Krings, U. 1994. Thesis. University of Hannover, Hannover, Germany.

Krings, U., and Berger, R.G. 1995. Porous polymers for fixed bed adsorption of aroma
compounds in fermentation processes. Biotechol. Techniques. 9(1), 19—24.

Krings, U., Klech, M., and Berger, R.G. 1993. Adsorption for the recovery of aroma
compounds in fermentation process. J. Chem. Tech. Biotechnol. 58, 293-299.

Lancaster, J .E. and Boland, M.J. 1990. Flavor biochemistry. In Onion and Allied Crops.
(H.D. Robinowitch and J .L. Brewster, eds.). CRC Press, Inc., Boca Raton,
Florida.

List, J .R., Friedrich, J .P. and King, J .W. 1989. Supercritical C02 extraction and
processing of oilseeds. Oil Mill Gazetteer December, 28-34.

McHugh, R.L. and Krukonis, V.J. 1994. Supercritical Fluid Extraction: Principles and

46

Practice. Butterworth-Heinemann, Stoneham, Massachusetts.

Nuss, J .S. 1994. Concentration of onion juice volatiles by reverse osmosis and its
effects on supercritical C02 extraction. Master’s thesis. Michigan State
University, East Lansing, Michigan.

Nuss, J .S., Guyer, DE. and Gage, DA. 1997. Concentration of onion juice volatiles by
reverse osmosis and its effects on supercritical C02 extraction. J. Food Proc. Eng.
20, 125-139.

Rizvi, S.S.H., Benado, A.L., Zollweg, J .A. and Daniels, J .A. 1986a. Supercritical fluid
extraction: fundamental principles and modeling methods. Food Tech. June, 55-
65.

Rizvi, S.S.H., Daniels, J .A., Bemado, AL. and Zollweg, J .A. 1986b. Supercritical fluid
extraction: operating principles and food applications. Food Tech. July, 57-64.

Rohm and Haas Company. 1993. Amberlite XAD-16 polymeric adsorbent. Product
description.

Sanders, N. 1993. Food legislation and the scope for increased use of near-critical fluid
extraction operations in the food, flavoring and pharmaceutical industries. In
Extraction of Natural Products Using Near-critical Solvent. (M.B. King and TR.
Bott, eds.) Blackie Academic & Professional, an imprint of Chapman & Hall,
Glasglow, UK.

Sinha, N.K., Guyer, D.E., Gage, DA. and Lira, CT. 1992. Supercritical carbon dioxide
extraction of onion flavors and their analysis by gas chromatrography-mass

spectrometry. J. Agric. Food Chem. 40, 842-845.

Smith, RM. and Burford, MD. 1992. Optimization of supercritical fluid extraction of
volatile constituents from a model plant matrix. J. Chromatogr. 600, 175-181.

Srinivasan, M.P., Smith, J .M. and McCoy, B.J. 1990. Supercritical fluid desorption
from activated carbon. Chem. Eng. Sci. 45(7), 1885-1895.

47

Tan, C .S. and Liou, DC. 1988. Desorption of ethyl acetate from activated carbon by
supercritical carbon dioxide. Ind. Eng. Chem. Res. 27, 988-991.

Temelli, F., Chen, CS. and Braddock, R.J. 1988. Supercritical fluid extraction in citrus
oil processing. Food Tech. June, 145-150.

Weber, W.J.1985. Adsorption theory, concepts, and models. In Adsorption Technology:
A Step-by- Step Approach to Process Evaluation and Application. (F.L. Slejko,
ed.). Marcel Dekker, Inc., New York, New York.

Whitaker, J .R. 1976. Development of flavor, odor, and pungency in onion and garlic.
Adv. Food Res. 22, 78-133.

Whitfield, F .B. and Last, J .H. 1991 Vegetables. In Volatile Compounds in Foods and
Beverages. (H. Maarse,ed.) Marcel Dekker, Inc., New York, New York.

Williams, OF. 1981. Extraction with supercritical gases. Chem. Eng. Sci. 36(11), 1769-
1788.

APPENDIX

APPENDIX A

Table A. 1. Mass Spectral Data of Identified Onion Flavor Compounds.

 

No.

Compound

Mal

Mass Spectral Data

 

thiopropanal S-oxide

90

920.97), 910.91), 9007.04), 750.61),
740.05), 73(15.16), 720.77), 63(5.84),
620.81), 48(8.23), 4505.23), 440.17),
4205.94), 41000.00)

 

3,4-dimethylthiophene

112

114(5.62), 113(11.84), 11205.25),
111000.00), 9706.06), 770.42),
710.93),690.82), 670.72), 4503.04)

 

methyl 1-propenyl disulfide

120

12203.75), 121(9.14),1200oo.0o),
87(9.67), 8007.75), 7506.37), 7406.17),
7303.43), 7202.18), 7102.57), 4707.97),
460.58), 45 (46.83), 4109.43), 4001.71),
3903.05)

 

methyl l-propenyl disulfide

120

122(9.69), 121(9.07),120000.00),
1050.45), 8009.55), 790.85), 7505.45),
7403.20), 7308.29), 7205.07), 7107.02),
61(9.37), 4703.25), 460.90), 4503.02),
4109.11), 3908.76)

 

l-propenyl propyl disulfide

148

150(35.78),149(7.92), 148(54.96),
10801.08), 10606.85), 78(9.89), 7403.04)
7307.10), 7207.43), 64(11.87), 4504.92),
43000.00), 4104.02), 3903.72)

w

 

3-ethyl-1 ,2-dithi-(4/5)-ene

146

1470.95), 14607.83), 113000.00),
11200.85), 11101.20), 98(9.51), 9707.98)
8605.50), 7902.67), 7401.20), 7306.82),
7103.19), 45(53.08), 4106.80), 4001.19),
39(1 1.84)

‘V

 

 

 

3-ethyl-1 ,2-dithi-(4/ 5 )-ene

 

146

 

14804.38), 147(13.37),146(100.00),
11300.83), 1110.52),10104.34),
8206.07), 7401.36), 7308.07), 7200.51),
7102.16), 6904.05), 6707.06), 61(11.57),
59(19.82), 4701.55), 4503.27) 4106.36),

 

3905.92)

 

48

Table A]. (cont’d)

49

 

Compound

Massl

Mass Spectral Data

 

methyl propyl trisulfide

154

15602.55), 1550.03),154(100.00),
11203.02), 7906.44), 64(7.01), 4705.33),
4504.14), 4303.97), 4109.77)

 

methyl 1-propenyl trisulfide

152

15405.37), 1530.45),152000.00),
10500.65), 1030.48), 8806.00),
7908.90), 7300.08), 7207.02), 7107.97),
640.34), 4702.92), 4500.38), 4101.33),
3902.53)

 

10

methyl l-propenyl trisulfide

152

15404.54), 153(7.46),152(98.47),
10502.74), 88(84.40), 800.34), 7904.22),
7303.59), 72(21.63), 7105.63), 6403.04),
6104.58), 4706.92), 4600.17),
45000.00), 4101.77), 3903.43)

 

11

dimethyl tetrasulfide

158

16006.30),1590.41), 158000.00),
1110.01), 9407.77), 8002.53), 7906.57),
6407.24), 610.12), 48(3.40), 4706.76),
4606.19), 4501.48)

 

12

diallyl thiosulfinate

162

1640.49), 1630.95), 16203.37),
12909.03), 9901.29), 9800.08),
97(66.11), 870.13), 850.54), 8307.72),
69000.00), 5900.83), 5704.56),
5500.86), 4506.09), 4305.04), 4101.20),
3901.79)

 

l3

l-propenyl propyl trisulfide

180

1820.39), 1810.45),18006.33),
11604.59), 11505.18), 11405.14),
11303.61), 10600.84), 8302.28),
7503.70), 74(61.03), 7300.95), 64(11.42),
5909.42), 470.44), 4500.38), 4303.54),
41000.00), 3901.07)

 

l4

 

 

l-propenyl propyl trisulfide

 

180

 

18205.03),18101.98), 180000.00),
11603.64), 11504.75), 10601.15),
8702.06), 8309.04), 7508.64), 7401.54),
7307.52), 6406.01), 5902.68), 4703.96),
4502.13), 4307.79), 4101.72), 3906.19)

 

 

Table A. l. (cont’d)

50

 

Nol

Compound

M...|

Mass Spectral Data

 

15

diallyl trisulfide

178

1820.00), 181(6.36),180(46.31),179(6.68),
17800.15), 1 1601.86), 1 1408.22),
10602.89), 10500.35), 9903.15),
7403.74), 7304.47), 7201.59), 7100.55),
6405.23), 6106.19), 5903.18), 580.05),
4703.13), 45000.00), 4306.19),
4109.41), 3905.62)

 

l6

 

 

methyl 3,4-dimethyl-2-
thienyl disulfide

 

190

19204.81), 1910.71),19008.13),
14402.25),143(100.00), 11102.77),
9907.26), 980.11), 590.58), 4502.53),

 

410.29), 390.45)

 

 

51

02:59:00 SEE 22:0 mo xmom vogue—om 2t wEBocm 065m
was an ooom 8 81:. :oEO mo cocombxm 0265 5230 Eoutoconsm 88m 35030 :0 880 .«o EEwofiEoEU DO ._.< ocswi

 

 

 

 

 

 

 

cmom
oom oov com com OCH 0
mm.HmmH O.H«. . . b . . . L p . . . . _ p . . . p t t . L _ t . . r
.. o
m ION
e m .
fi
Tow
. 0
10m 0
. c
. m
. U
100 e
. a
. n
_ o: - 4
. I _ . . _ 1 . . . . OOH
.chnm .006 m w v N

.e.m mm.HmmH am:

52

8.383 ._.< 2&5

 

 

 

 

 

 

 

 

 

€QDG'UIUGUQ)

cmom
OOHH OOOH com com 00m. 00m
mm.HmmH O.H.¢ r . _ 1 p t p r . r _ . . . p . r . . . _ t
b
row
a m .
: .
0 ice
o_ .
low
5 .
Tom
0 UHH. .
. . . J . . . . . i . . . . . 2:
.chnm .008 ma «a NH 0H
.8.& mm.HmmH X62

53

23:53 ._.< unswE

 

 

 

 

 

 

 

 

 

 

cmom
00m.” oomH oova coma OONH OOHH
mm.HmmH O.H« . _ p . L . t _ . . . _ p t . . r
1 b
ION
N_ .
C— 1

C 10v
3 . 0
loo 0
v G
i m
f e
2 Tom a
.. 5
. Q
UHB . <

q _ 4 a . q < a 4 q 4 4 . — a OOH

.Ucfiflm .OME vN NN ON mH

.B.m mm.HmmH xmz

APPENDIX B

STATISTICAL ANALYSIS

Table B.1. Analysis of Variance of Gravimetric Yield (by Percentage) of Onion Oil
Obtained from Supercritical Carbon Dioxide Extraction of Onion Juice in an Absence
and Presence of Onion Pulp at 1500 liters STP of Carbon Dioxide Passed.

 

Treatment Block Mean Standard
I II 111 Deviation

Absence of Pulp 0.03335 0.02612 0.02479 0.02809 4.6O6x10'3

Presence of Pulp 0.04504 0.04225 0.04928 0.04552 3.540x10'3

 

 

 

Source of Degree of Sum of Mean Square F Fa=o_05
Variance Freedom Square
Total 5 5.236x10T
Blocks 2 2.526x10’5
Treatments 1 4.561x10'4 4.56lx10'4 21.59 18.51
Error 2 4.224xl05 2.112x10'5

 

54

55

Table B.2. Analysis of Variance of Total Peak Area of Identified Onion Flavor
Compounds per 20 pg of Onion Oil Obtained from Supercritical Carbon Dioxide
Extraction in an Absence and Presence of Onion Pulp at 1500 liters STP of Carbon
Dioxide Passed.

 

 

 

Treatment Block Mean Standard
I II 111 Deviation
Absence of 16835.92 14988.60 32828.84 21551.12 9810.37
Pulp

Presence of 13827.63 10126.58 19556.54 14503.58 4751.18
pulp

 

 

Source of Degree of Sum of Mean Square F Fa=005
Variance Freedom Square

Total 5 312135850

Blocks 2 207714353

Treatments 1 74501659 74501659 4.98 18.51

Error 2 29919838 14959919

 

56

Table B. 3. Analysis of Variance of Gravimetric Yield (by Percentage) of Onion Oil
Desorbed from XAD- 16 and XUS 40323 Loaded with One Liter of Onion Juice at 800
liters STP of Carbon Dioxide Passed.

 

 

 

 

 

 

Treatment Block Mean Standard
I II 111 Deviation
XAD-16 0.02029 0.02559 0.02059 0.02216 2.977x10‘3
XUS 40323 0.01930 0.01677 0.02139 0.01915 2.313x10'3
Source of Degree of Sum of Mean Square F Fa=o,05
Variance Freedom Square
Total 5 4.196x10'5
Blocks 2 2.2551006
Treatments 1 1.353x10'5 1.353100“5 1.034 18.51
Error 2 2.618x10'5 1.309x10‘5

 

Table 8.4. Analysis of Variance of Gravimetric Yield (by Percentage) of Onion Oil
Desorbed from XAD-16 at Different Supercritical Carbon Dioxide Conditions and 800
liters STP of Carbon Dioxide Passed.

 

 

 

 

 

 

Treatment Block Mean Standard
I 11 HI Deviation
3000psi,37°C 0.0203 0.0256 0.0206 0.0222 2.98x10T
3000psi,50°C 0.0188 0.0207 0.0171 0.0189 1.80x10'3
1500psi,37°C 0.0109 0.0115 0.0111 0.0112 3.06x10'3
4150psi,50°C 0.0260 0.0219 0.0210 0.0230 2.67x10'3
Source of Degree of Sum of Mean Square F Fa=005
Variance Freedom Square
Total 1 1 2.995x10“
Blocks 2 1.251x10'5
Treatments 3 2.609x10'4 8.697x10’5
Error 6 2.6091005 4.349x10'6 20.00 4.76

 

Least Significant Different (LSD) at P S 0.05 = 8.674x10'4

APPENDIX C

DEVELOPMENT OF A TENTATIVE MODEL EQUATION
FOR SUPERCRITICAL CARBON DIOXIDE EXTRACTION AND DESORPTION

CO2 + solute

l

 

 

C02 phase

 

-——extracted or desorbed phase

 

 

 

T CO2

Figure C. 1. Supercritical Carbon Dioxide Extraction and Desorption Diagram.

A vessel is filled with a certain amount of an extracted or desorbed phase (L).
Carbon dioxide is passed through the extracted or desorbed phase at constant flow rate
(V). The solute in the extracted or desorbed phase is extracted into the carbon dioxide
phase and is carried out from the vessel by the carbon dioxide.

The system is assumed to be at equilibrium. The solubility of carbon dioxide in
the extracted or desorbed phase is assumed to be negligible.

At the equilibrium, the equation (1) is used to describe the correlation of the

solute concentration in both phases.

58

59
y = kX (1)
y = the solute concentration in the carbon dioxide phase
k = a constant
x = the solute concentration in the extracted or desorbed phase
A mass balance of the solute in term of a differential equation is developed by
differentiatng the weight of solute in the extracted or desorbed phase (xL) with respect

to the extraction or desorption time (t).

SIX—1.: -yV (4)
dt
L d_x + x L = -yV (5)
dt dt
Substitute equation (1);
L d_x + x d_L = -ka (6)
(It (11
L is constant, therefore, dL = 0;
dt
d_x = -k_v dt (7)
x L

Integrate equation (7) from the maximum concentration of the solute in the extracted or
desorbed phase at the beginning (x = x0 at t = t0 =0) to the concentration of the solute in

the extracted or desorbed phase at any time during the extraction or desorption (x = x,

flit-=10;
1 ti
fax = -k_V.l dt (8)
x0 x L 0
In at = -_l<_\l t1 (9)

x0 L

60

_X1 = EXPCKXII) (10)

x0 L

x, = x0 (EXP(-_ng t,)) (11)
L

Difference between the weight of the solute in the extracted or desorbed phase at the
beginning (xoL at t = to =0) and at any time during the extraction or desorption (XL

at t =t,) equals to the weight of the solute extracted or desorbed by carbon dioxide.

(x. - x.) L = Ix. - x. (EXP(-k_\l t. 1)] L (12)
L
(xo - x,) L = [1 - (EXP(-L\_’ t, ))] xoL (13)
L
Let Y = (x0 - x,) L (14)

the weight of extracted or desorbed solute
at time t,

bl = xoL = a constant (15)
The constant b) is related to the maximum weight of onion oil extracted or desorbed.

b2 = k = aconstant (16)
L

The constant b2 is related to the constant k showing the ratio of the solute concentration
in the carbon dioxide phase and the solute concentration in the extracted or desorbed
phase. The higher the constant k is, the easier the solute can be extracted or desorbed
from the extracted or desorbed phase into the carbon dioxide phase.

X = Vt, = the carbon dioxide volume passed (17)

61

A tentative model equation is:
Y = b) [1 - (EXP(-b2 X)] (18)
This tentative model was used in curve fitting of the data obtained from the supercritical

carbon dioxide desorption of onion oil from polymeric adsorbents.

 

62
APPENDIX D

RESULTS OF NON-LINEAR REGRESSION ANALYSIS
BY PlotIT PROGRAM

Table D.l. Correlation Equations Between Gravimetric Yield (Y) and Carbon Dioxide
Volume (X), and Their Coefficient of Determination (r2).

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Tentative Model Equation: Y = b1 [1 - (EXP(-b2 X)]

Figure Treatment b1 b2 r2
4.1 XAD-1 6 0.02207 0.004475 0.9462
XUS 40323 0.01907 0.004679 0.9453
4.2 1 liter 0.2208 0.004476 0.9462
2 liters 0.5070 0.003256 0.9509
3 liters 0.6353 0.003029 0.9711
6 liters 0.8563 0.002044 0.9912
4.3 1 liter 0.02207 0.004475 0.9462
2 liters 0.02549 0.003211 0.9510
3 liters 0.02118 0.003028 0.9711
6 liters 0.01356 0.002042 0.9912
4.5 1500psi, 37°C 0.01125 0.004051 0.9896
3000 psi, 37°C 0.02207 0.004476 0.9462
3000 psi, 50°C 0.01906 0.004442 0.9612
4159J)si, 50°C 0.02271 0.005910 0.9648

 

 

 

 

 

 

 

APPENDIX E

 

 

 

 

 

 

 

 

 

 

 

GC CHROMATOGRAMS
Max 1891.83 R.T.
10 20 3o 40 50 mag. abund.
A 100 ‘ ‘ ‘ ‘ ‘ ‘ ‘ ‘ ‘ ‘ ‘ ‘ “““““““ 1 ‘ ‘ ‘ ‘
< TIC
b .
u .
n 8 o -l
d .
a .
n 1
c 60‘ l
e . ,
40‘
20'
0“ .
' r ' - 1 ' ' ' 1 T ' r ' I ' ' ' ‘1.0 1891.83
1000 2000 3000 4000

Scan

Figure E.l. GC Chromatogram of Onion Oil Obtained From Supercritical Carbon
Dioxide Extraction of Onion Juice at 3000 psi and 37°C.

 

 

 

 

 

 

 

 

 

 

flax 1411.33 R.T.
110 210 310 ‘10 540 mag . abund .
10° ............ E ........ . . -
A TIC
b 1
U 1
n 80‘
d t
a 1
n 4
C 60'
e I
40‘
20‘
*_l
01 b
- r w 9979* r r7 T T ' I * fi' ' ‘1.0 1411.33
0 1000 2000 3000 4000
Sean

Figure E.2. GC Chromatogram of Onion Oil Obtained From Supercritical Carbon
Dioxide Extraction of the Mixture of Onion Juice and Onion Pulp at 3000 psi and 37°C.

63

64

 

 

 

 

 

 

Max 902.807 R.T
10 zp 30 40 so m39- ab““d-
100 x . a x . x a . A r l x a h x L x A 4 4 L A
A rrc
b .
u 4
n 80‘
d .
a i
n u
C 60‘
e 4
40‘
20‘
I
l
01 l-
‘ I ' * ' I ”f ' I ' * ’1.0 902.80?
0 1000 2000 3000 4000
Sean

Figure E.3. GC Chromatogram of Onion 0il Desorbed From XAD-16 Adsorbent at
3000 psi and 37°C.

 

 

 

 

 

 

 

Max 829.425 3.2
10 20 337 4p 50 “‘9 ab““d-
A100 ‘ ‘ * * ‘ c ‘ ‘ ‘ ‘ "4
rrc
b
11
n 801
d .
a
n
c 60‘
e 4
405
l
20-
04 h
4 -4- t . e7 - -e- s, r T . *1.o 829.425
1000 2000 3000 4000
Sean

Figure E.4. GC Chromatogram of Onion Oil Desorbed From XUS 40323 Adsorbent at
3000 psi and 37°C.

65

 

 

 

 

 

 

 

 

Max 979.386 R-T- b
10 20 39 4o 50 mag- a ““d-
100 c ‘ ‘ t
A . TIC
b .
u l
n 80.4
d 4
a l
n -4
c 601
e 'l
.l
40~
20-
0 r - . , - - - . ~ - ~ ‘1.0 979.386
1000 2000 3000 4000
Scan

Figure E.5. GC Chromatogram of Onion Oil Desorbed From XAD-16 Adsorbent at
1500 psi and 37°C.

Max 1160.96 R.T.

 

 

 

 

 

 

 

 

ip 20 30 4o 50 ”39- “bun“:
100 A A a A A A a x a 1 a a x l #4. a I a x a A
A . TIC
b .
U
n 80-1
d t
a
n 1
c 60‘
e 4
.1
40‘
207
d
0.4 i-
, _ I , - - . - ~ t - - - - 31.0 1160.96
0 1000 2000 3000 4000

Sean

Figure E.6. GC Chromatogram of Onion Oil Desorbed From XAD-16 Adsorbent at
3000 psi and 50°C.

66

 

 

 

 

 

 

 

 

Max 1241.71 R.T.
10 20 30 4o 50 M9- abund-
100 A A A A A A +4 A4 4 A I A A A A J A A A A 1 . A A A
A TIC
b
u ‘
n 80-
d 4
a <
n 4
c 60-
e .
40*
2o-
0‘ n
- . r . T . - r - r- - - 91.0 1241.71
.1000 2000 3000 4000

Sean

Figure E.7. GC Chromatogram of Onion Oil Desorbed From XAD-16 Adsorbent at
4159 psi and 50°C.