a

2.. 9" ..
F: z ......r#

a. (lumbar)... J

4-

as

Maw; .
m, hwy”.
4.: 11..

sxflffififi raw...
égmwnfi

t . .. .
Nu 751W...

am
#3... 3.4mm”?
5

in? .v.

1 .l ’
3... 1.-
.35).,

r .3.

 

 

«mm?
#903 2

Sygr-Sw3 LIBRARY

Michigan State
University

 

 

 

 

This is to certify that the

thesis entitled

"Supercritical Fluid Extraction of
Quercetin from Onion Skins"

presented by

Karina Gorostiaga Martino

has been accepted towards fulfillment
of the requirements for

M.S. degree in Biosystems Engineering

J £1...

7
Major professor

 

 

 

 

Date At; 2/1300;

0-7639 MS U is an Affirmative Action/Equal Opportunity Institution

 

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 cJCIRC/DateDuepsisz

SUPERCRITICAL FLUID EXTRACTION OF QUERCETIN FROM ONION SKINS
By

Karina Gorostiaga Martino

A THESIS
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
MASTER IN SCIENCE
Department of Biosystems Engineering

2002

ABSTRACT
SUPERCRITICAL FLUID EXTRACTION OF QUERCETIN FROM ONION SKINS
By

Karina Gorostiaga Martino

Supercritical fluid extraction with ethanol-modified carbon dioxide of quercetin
from onion skins (red and yellow varieties) was studied. Static and dynamic modes were
investigated as extraction methods, with the static mode yielding the highest amount of
quercetin recovered. The total amount of 0.024 g of quercetin per kg of onion skin was
obtained for the red variety, and 0.020 g per kg for the yellow variety, at 5700 psi, 40°C,
an average of 7.6% (molar concentration) of ethanol, and in an extraction period of 2.5 h.

The extraction of pure quercetin from spiked diatomaceous earth was also
investigated, extraction conditions were the same as for the onion skins. A maximum of
0.036 g of quercetin per kg of matrix was extracted.

The modifier was found to have an important influence on quercetin recovery.
The greater the amount of ethanol collected in the trap, the greater the amount of
quercetin recovered. This trend was found for both varieties and for the pure quercetin.

Identification and quantification analysis were conducted in a High Performance

Liquid Chromatography system.

ACKNOWLEDGMENTS

I would like to give my special thanks to the Department of Food Science and
Human Nutrition for letting me use its facilities. I am Specially appreciative to Dr.
Maurice Bennink for his guidance and generosity in letting me access to his laboratory
equipment, and to Dr. Kirk Dolan and Dr. Mark Uebersax for also permitting me to use
their laboratory equipment.

I would also like to thank to the Department of Civil and Environmental
Engineering, especially to Dr. Tomas Voice, Dr. Paul Loconto and Mr. Joseph Nguyen
for their valuable contribution to my work, with their guidance and letting me use their
laboratory equipment, at the beginning of this project.

I would like to give my appreciation and respect to my committee members, Dr.
Carl Lira and Dr. Kirk Dolan for their technical support, and Dr. Daniel Guyer for giving
me the opportunity to pursue my degree, for all his support and guidance and his strong
confidence in my work.

Finally, thanks to all my fiiends from Agricultural Engineering and Food Science
departments for making the last two years very enjoyable and productive. A very Special

thanks to my lovely husband and family for their unending support and encouragement.

TABLE OF CONTENTS

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

LIST OF TABLES ‘ VI

LIST OF FIGURES VII

1. INTRODUCTION 1

1.1. FLAVONOIDS ......... -_ _- - - - . -- l

12. QUERCETIN - -- _ - _ _ _ .................. -_ 2

2. LITERATURE REVIEW 5

2.1. THEORIBS OF SUPERCRITICAL FLUID (SF) - _ - - -- - - --5

2.2.1. Analytical Method ........................................................................................... 12

2. 2. 2. Modeling the supercriticalfluids .................................................................... 13

3. MATERIALS AND METHODS 14

3.1. REAGENTS AND SAMPLES - 14

3.2. SF EXTRACTION EQUIPMENT ................................................................................... 14

3. 3. LIQUID CHROMATOGRAPHIC SYSTEMS .................................................................... 17

3. 4. SAMPLE STORAGE - ______ _ - - _ ............. 17

3. 5. SUPERCRITICAL FLUID EXTRACTION -- - - - -- - -- _ 18

3.5.1. Sample preparation ......................................................................................... 18

3.5.2. Experimental Procedure ................................................................................. 18

3.6. SAMPLE ANALYSIS -- ................................................ '. .......................... 22

3.7.SOLID-LIQUID EXTRACTION __ - _ -- _ _- _ -- 23

4. RESULTS AND DISCUSSION 25

4.1. OPTIMAL EXTRACTION CONDITIONS __ - _ -_ -_ -- - 25

4.2. PRELIMINARY EXPERIMENTS ................................................................................... 26

4.2.1. Static mode ...................................................................................................... 26

4.2.2. Dynamic mode ................................................................................................ 27

4.2.3. Comparison between modes ........................................................................... 28

4.3. STATIC MODE ......................................................................................................... 29

4.3.1. Red variety ...................................................................................................... 29

4.3.2. Yellow variety .................................................................................................. 33

4.3.3. Comparison between varieties ........................................................................ 36
4.3.4. Comparison of the preliminary and primary experiments conducted with

yellow variety ............................................................................................................ 3 7

4.3.5. Comparison of SF with solid-liquid extraction ............................................... 38

4.4. STATIC MODE FOR PURE QUERCETIN ....................................................................... 39

5. CONCLUSIONS 44

6. RECOMMENDATIONS 46

6.1. EXPERIMENTAL IMPROVEMENTS - -- - -46

6 2 FUTURE WORK ........................................................................................................ 46

iv

APPENDIX A. EXPERIMENTAL DATA

 

APPENDIX B. HPLC CHROMATOGRAMS AND OTHERS

 

 

APPENDIX C. STATISTICAL ANALYSES
REFERENCES

 

49

69

LIST OF TABLES

Table l. Quercetin content (g/kg of onion portion, fresh weight) ................................... 11
Table 2. Flavonol level and composition .......................................................................... 12
Table 3. Methanol extraction data .................................................................................... 24
Table 4. Results from static mode .................................................................................... 27
Table 5. Results from dynamic mode ............................................................................... 27
Table 6. Comparison between modes ............................................................................... 28
Table 7. Results for red variety ......................................................................................... 29
Table 8. Results for yellow variety ................................................................................... 34
Table 9. Results for the red and yellow varieties ................................ ' .............................. 3 7
Table 10. Results from preliminary and primary experiments for yellow variety ........... 38
Table 11. SF and methanol extractions results ................................................................. 39
Table 12. Results from the pure quercetin extraction ....................................................... 40
Table 13. Extraction ratios for yellow variety and pure quercetin ................................... 42
Table 14. Comparison of the results from the yellow variety and pure quercetin..... ....... 43
Table A1 . Experimental data collected during static mode .............................................. 49
Table A2. Experimental data collected during dynamic mode ......................................... 51
Table A3. Data for calibration curve ................................................................................ 52
Table A4. Experimental data and results .......................................................................... 54
Table A5. Experimental data collected and results ........................................................... 56
Table A6. Data colleCted and results from static mode extraction of pure quercetin ....... 58

vi

LIST OF FIGURES

Figure 1. Flavonoids structure. Quercetin (free aglycone): Rland R2 = Hydrogen ............ 2
Figure 2. Phase diagram of a pure component. Mukhopadhyay (2000). ............................ 6
Figure 3. SFE steps ............................................................................................................. 7
Figure 4. SF equipment ..................................................................................................... 16
Figure 5. SF equipment with recirculation. ...................................................................... 16
Figure 6. Preliminary experiments schema ....................................................................... 19
Figure 7. Primary experiments schema. ............................................................................ 22
Figure 8. Correlation between quercetin extracted and ethanol volume per vial ............. 30
Figure 9. Correlation between ethanol volume per vial and time ..................................... 31
Figure 10. Correlation between extraction ratio per vial and time ................................... 31

Figure 11. Correlation between quercetin extracted and ethanol volume collected per vial

................................................................................................................................... 34
Figure 12. Correlation between ethanol volume collected per vial and time ................... 35
Figure 13. Correlation between extraction ratio per vial and time ................................... 36

Figure 14. Correlation between quercetin extracted and ethanol volume collected per vial

................................................................................................................................... 41
Figure Al . Calibration curve ............................................................................................ 53
Figure B1. Giddings state curves. Giddings J. C. et al., 1968 .......................................... 60
Figure B2. Static mode method. Example of the chromatograms .................................... 61
Figure B3. UV spectrum of quercetin standard and extract from SFE ............................. 62
Figure B4. Dynamic mode method. Example of the chromatograms .............................. 63
Figure B5. Static mode for red onion variety. Example of the chromatograms ............... 64

vii

Figure B6. Mass Spectrometry of the standard sample ..................................................... 65

Figure B7. Mass spectrometry of the SFE extract ............................................................ 66
Figure B8. Static mode for the yellow variety. Example of the chromatograms ............. 67
Figure B9. Example of the chromatograms of the pure quercetin extract ........................ 68

viii

1. Introduction

1.1. Flavonoids

F lavonoids are plant polyphenols found frequently in fruits, vegetables, and
grains. Divided into several subclasses, they include the anthocyanidins, pigments chiefly
responsible for the red and blue colors in fruits, fruit juices, wines, and flowers; the
catechins, concentrated in tea leaves; the flavonones and flavanone glycosides, found in
citrus and honey; and the flavones, flavonols and flavonol glycosides, found in tea leaves,
fruits, vegetables, and honey. Flavonoids are known for their hydrogen-donating
antioxidant activity as well as their ability to complex divalent transition metal cations.
Moreover, they promote human health These compounds are active against allergies,
inflammation, viruses, hypertension, arthritis, and are reported to prevent mutations,
carcinogens, cancer, and AIDS (Merken et al., 2000). The supercritical fluid extraction of
quercetin (one member of the flavonol group) from onion skins, was the main focus of
this study.

Flavonoids, derived biosynthetically from phenylalanine, are pigments found
widespread in plants. Three moles of malonyl-coenzyme A (CoA) from glucose
metabolism condense to form ring A, a reaction catalyzed by chalcone synthetase (Figure
1). Rings B and C are also derived from glucose metabolism, but via the Shikimate
pathway in which phenylalanine is converted to cinnarnic acid and then to cournaric acid.
Coumaric acid CoA and three malonyl CoAs then condense in a single enzymatic step to
form naringenin chalcone. The C-ring closes and becomes hydrated to form 3-

hydroxiflavonoids (e. g. catechins), 3,4-diol flavonoids (e. g. quercetin), and procyanidins.

 

0H 0
Figure 1. F lavonoids structure. Quercetin (free aglycone): Rland R2 = Hydrogen

F lavones and flavonols are usually found in plants as O-glycosides. The flavonols
have a hydroxyl group at C3, where the flavones have a hydrogen atom. Glycosides of the
flavonol quercetin predominate in vegetables, whereas glycosides of the flavonol
kaempferol and of the flavones apigenin and luteolin also exist.

The formation of flavone and flavonol glycosides depends normally on the action
of light, so that in general the highest concentrations of these compounds are nearly
always found in the flee-standing leaves. Onions are the only exception (Herrmann,
1976). Flavones and flavonols do not contribute markedly to the coloration of the plant
except where they occur in very high concentration, as in the skins of onions, or when

they are complex with metals (Herrmann, 1976).

1.2. Quercetin

Hayashi (2000) has demonstrated that tumor growth in mice is significantly
inhibited using a combination of quercetin and chalcone (open chain flavonoid).
Furthermore, Caltagirone (2000) reported that quercetin and apigenin inhibit the growth,

the invasiveness, and metastatic potential of melanoma; therefore, flavonols may

constitute a valuable tool in combination therapy for metastatic melanoma. Finally, Xing
et al. (2001) stated that quercetin has the potential to become a chemopreventive and/or
chemotherapeutic agent for prostate cancer.

Unlike other vegetables and in contrast to white-skinned varieties, onions with
pigmented skins have exceptionally high flavonol content. Quercetin appears mainly in
the free form (3, 3’, 4’, 5, 7-Pentahydroxyflavone), i.e. as the aglycone, 67-86% of total
quercetin (aglycone and glycosides), and to a smaller extent as Spiraeoside (quercetin-4’-
glycoside). The epidermis of onion scales contains quercetin glycosides exclusively,
mainly as Spiraeoside, which is formed first, the diglycosides are formed later and
increase continuously during storage. The flavonol concentration decreases from the
outer to the inner scales, with higher levels in the outer epidermis (Herrmann, 1976).

Bilyk et al. (1984), found that the skin of the Yellow Globe hybrid variety has the
highest free quercetin (aglycone) content in the dry Skin (53%, w/w, of the total
quercetin). Other varieties studied were Sweet Spanish Utah, Early Yellow Globe,
Yellow Globe Hybrid, Sweet Spanish Hybrid, Red Hamburger, Walla Walla and
Evergreen Long White Bunching, the range of free quercetin content in these varieties
was 23 to 51% (w/w, of the total quercetin).

In 1997, Price et al. investigated different varieties, such as Red Baron,
Rijnsburger, Rose and Albion. They found that Rijnsburger variety had the highest
amount of free quercetin (0.039 g per kg fresh weight).

Extraction of flavonols from onion tissue, in the above noted studies, was
accomplished by solvent extraction with methanol. Unfortunately at industrial level,

problems arise with this technique when removing organic solvent residues fiom the final

product, disposing of waste methanol, and exposing personnel to the extracting solvent.
Therefore, there is an important need for rapid and clean methods for the extraction and
determination of this highly valuable natural product, quercetin. Supercritical fluid
extraction provides several advantages over traditional liquid-solvent—based extraction
methods including improved selectivity, expeditiousness, automation and environmental
safety. The avoidance of organic solvents residues is a major goal in the isolation of
natural products which may be commercialized as food additives (Terra et al., 1998).

The present study investigated the hypothesis that quercetin could be extracted
from a natural matrix (onion skins), using odorless, tasteless, inert and nontoxic
supercritical carbon dioxide.

In order to test this hypothesis, the following objectives were proposed.

1. To develop a technique for the extraction of quercetin.

2. To identify quercetin and quantify the amount extracted, from red and yellow
onion varieties.

3. To compare the extraction of quercetin in the presence of other similar
compounds from the natural matrix (onions), with the extraction of the pure

quercetin from the spiked inert matrix (diatomaceous earth).

2. Literature Review

2.1. Theories of Supercritical Fluid (SF)

SFs utilize the ability of certain chemicals to become excellent solvents for certain
solutes under an appropriate combination of temperature and pressure. There are several
advantages of these fluids compared to liquid organic solvents. They have a higher
diffusion coeflicient and a lower viscosity. The absence of surface tension allows for
rapid penetration into the pores of heterogeneous matrices helping to enhance extraction
efficiencies. Depending on the conditions of temperature and pressure, the solubility of
the various components in the SF can be varied. Therefore, selectivity during extraction
may be manipulated. Finally, SFs do not leave a chemical residue (Rozzi et al., 2002).
2.1.1. Carbon dioxide as a supercritical solvent

In her latest work, Mukhopadhyay (2000) reviewed relevant characteristics and
properties of supercritical carbon dioxide. For example, when a gas is compressed to a
sufficiently high pressure, it becomes liquid. A gas can also be heated beyond a specific
temperature, at which point no amount of compression will cause it to become a liquid.
This temperature is called the critical temperature and the corresponding vapor pressure
is called the critical pressure. These values of temperature and pressure define a critical
point, which is unique to a given gas or liquid. Figure 2 shows a phase diagram that
describes the state of gas or liquid, called a SF, when both the temperature and pressure

exceed the critical point (CP) values.

    

solid liquid

Pressure ———-

T

 

 

Temperature —-—-

Figure 2. Phase diagram of a pure component. Mukhopadhyay (2000).

' The SF assumes many of the properties of both gas and liquid. In the supercritical
region (Figure 2), with only small changes in temperature and pressure, maximum
solvent capacity and the widest range of variations in solvent properties can be achieved
(Mukhopadhyay, 2000).

Carbon dioxide is one of the most desirable SF solvents for extraction of natural
products for use in foods and medicines. It is inert, inexpensive, easily available,
odorless, tasteless, environment-friendly, and GRAS (generally regarded as safe).
Furthermore, no additional “green house effect” results from using carbon dioxide as the
supercritical fluid solvent since it is already present in the environment. It is obtained as a

by-product from fermentation processes or fertilizer manufacture; therefore, its use as an

extractant does not cause any further increase in the amount of carbon dioxide present in
the earth’s atmosphere (Mukhopadhyay, 2000).
2.1.2. Supercritical fluid extraction (SF E) system

The four primary steps involved in SFE are compression, extraction, expansion,
and separation; the pieces of equipment commonly utilized are a compressor, a high-

pressrn'e extractor, a pressure reduction valve and a low pressure separator (Figure 3).

 

 

/\

\J
@L mpg _.'
C02

Reduction valve
compressor Separator

 

 

V

Extractor

 

 

 

Figure 3. SFE steps

2.1.3. Extraction modes

SFE can be performed in a static, dynamic or coupled static/dynamic mode. When
a fixed amount of SF interacts with the analyte and the matrix, a static extraction takes
place. The extraction vessel containing the matrix is pressurized with the chosen SF at a
certain temperature. The SF remains in the vessel, without flow, until the extraction is
completed. No SF enters or leaves the vessel during the extraction. The high diffusivity
of the fluid causes it to permeate the matrix and remove some portion or all of the
analyte. Typically, a static mode is used when modifiers and derivatizing reagents are

added (Taylor, 1995).

If fresh SF is continuously passed over or through the sample matrix, a dynamic
extraction takes place (Taylor, 1995). This mode permits the continuous flow of SF until
the extraction is completed and uses as much SF as is needed for that period of time.

The coupled mode is especially usefirl when the analyte must diffuse to the matrix
surface to be extracted (Taylor, 1995). In this mode, there is a combination of no flow
and continuous flow through the vessel for a fixed period of time. For example,
dynamic/static mode, means first there is a period of time when the SF passes through the
vessel continuously. Then, the inlet and outlet valves of the vessel are closed and the SF
remains in the vessel without flow for a fixed time. Extraction can also occur in
static/dynamic mode, in which the steps are reversed.

2.1.4. Modifier introduction

An entrainer (or modifier) is a cosolvent, such as ethanol, water or various gases,
which can be used in the food industry (Palmer, 1995). The entrainer increases the
polarity and solvent strength of the SF while it retains the sensitivity of solubility with
respect to pressure and temperature. F urtherrnore, a cosolvent can improve the selectivity
of separation by preferentially interacting with one or more components and facilitating
selective fiactional separation. For example, for systems consisting of polar solutes,
solubility is significantly enhanced with polar cosolvents due to dipole-dipole interactions

and hydrogen bonding (Mukhoapadhyay, 2000).

2.2. Related research

There has been no previous research related to the extraction of quercetin flour a
natural matrix with modified SF carbon dioxide.

However, Terra et al. (1998) studied the SF extraction of spiked diatomaceous
earth (inert matrix) with phenolics including gallic acid, (+)—catechin, (-)-epicatechin,
caffeic acid, p-coumaric acid, myricetin, t-resveratrol, quercetin and salicylic acid. Pure
carbon dioxide and methanol carbon dioxide was used. Low-molecular-weight phenolics
containing 2-3 polar groups such as t-resveratrol, p—coumaric acid and salicylic acid were
recovered with pure carbon dioxide (mean recovery 2 95%). However, the inclusion of
methanol was necessary to recover more polar compounds such as gallic acid, caffeic
acid, catechins and quercetin; the recovery rates ranged between 30% and 70%.
Myricetin was not recovered.

Palma et al. (1999), studied the extractability of eight polyphenolic compounds,
found naturally in grapes, from spiked sand samples. Such compounds included benzoic
acids, cinnamic acids, benzoic aldehydes, catechin and resveratrol. Gallic acid was the
most polar and the least polar was resveratrol. They used high density and high
percentage of modifier, due to the high polarity of phenols. The average recovery of the
eight phenols was 88%. After modified conditions of extraction, such as modifier content
and restrictor/trap temperature, recoveries for all eight compounds ranged from 92% to
100%.

Resveratrol was successfully extracted from grape skin using supercritical fluid,

and coupled static (30 seconds)/dynarnic (15 min) mode. The operating conditions were

150 bars, 40°C and 7.5% ethanol. They recovered almost 100% in 15 min (Pascual-Marti
et al., 2001 ).

Grape glycosides were recovered fiom spiked earth and real grapes (100% in both
cases) using methanol-modified supercritical carbon dioxide and coupled static/dynamic
mode as an extraction method. Extraction conditions were: gas density, 0.95 g/ml;
temperature, 40°C; supercritical fluid flow, 1.5 mein; static time, 15 min; dynamic
time, 20 min; restrictor temperature, 50°C; liquid trap temperature, 30°C; liquid trap
solvent, water; liquid trap volume, 8 m1 (Palma et al., 2000).

The last two papers mentioned above studied the extraction of compounds similar
to quercetin from natural matrices. All of them used modified-carbon dioxide as a
supercritical fluid and tried static/dynamic mode as an extraction method. For the present
study, similar extraction conditions were tested.

Bilyk et al. (1984) investigated the presence of quercetin and kaempferol of eight
onion varieties (Table 1). They separated the dry skins (portion A), outer rings (portion
B) and inner rings (portion C) and extracted, with methanol, the flavonol glycosides.
Flavonols were detected and quantified by thin-layer chromatography, high- ormance
liquid chromatography, and spectrophotometric analyses. They concluded that the skin of
all onion varieties contained quercetin in both the aglycone and the glycosides form.

Some Specific varieties also contained small amounts of kaempferol.

10

Table l. Quercetin content (g/kg of onion portion, fresh weight)

 

[ . flrcetin

I Variety Portion a fiee Total b
[Carmer hybrid 7.73 (23%) 34.15 e 0.460
0.062 i 0.006
0.027 t 0.001
3.87 (24%) 16.53 i 0.410
0.295 t 0.005
0.002 t 0.001
i618 (39%) 16.06 e 0.040
10.053 :1: 0

ND C

7.54 93%) 14.16 :t 0.380
10.055 i 0
0.010 :1: 0
4.73 (50%) 9.51 :1: 0.090
0.113 t 0.012
0.032 t 0.001
2.61 (40%) 6.6 :E 0.180
ND

ND

2.69 (51%) 5.30 :1: 0.090
0.082 :1: 0.008
0.001 :t: 0
vergreen long white bunching leaves 0.06 (5%) 1.14 :1: 0.007

bulbs ND D .
' A = dry skin; B = outer rings (scales 1-3); C = inner rings (remaining rings)
b Results are given as mean :1: standard deviation for triplicate determinations.
° ND = not detectable.

Adapted from Bylik et al., 1984.

 

 

 

 

 

Sweet Spanish Utah

 

 

 

Early yellow globe

 

 

 

Yellow Globe hybrid

 

 

 

Sweet Spanish hybrid

 

 

 

IRed hamburger

 

 

 

Walla walla

 

 

nw>nw>ow>nw>nw>ow>ow>

 

 

 

 

 

 

 

 

In a different study, extraction of flavonols from thin onion varieties was
investigated (Price et al., 1997. Table 2). The major flavonoids of mature onion bulbs

were confirmed to be the quercetin diglycosides and quercetin monoglycoside, using

11

methanol extraction and a combination of chromatographic comparisons, mass
spectrometry and nuclear magnetic resonance spectroscopy.

Table 2. F lavonol level and composition

 

Content (mg/kgfiesh weight) a % total

’lavonols
ng ng Q Total l,
Red baron 1375 394 9 1778 88.30
Rijnsburger 1117 360 39 1516 93.00
Rose 1052 302 15 1369 94.20

Albion 50 36 3 89 61.00
' Abbreviations: ng, quercetin 3,4’-O-diglycoside; ng, quercetin 4’-0-
monoglycoside; Q, quercetin. Price et al., 1997.

Variety

 

 

 

 

 

 

 

 

 

 

 

 

 

2.2.1. Analytical Method

High performance liquid chromatography (HPLC) is used as an analytical method
for quantification of flavonoids, including quercetin. Columns used. in the system are
almost exclusively reversed-phase (RP), ranging from 100 to 300 mm in length, 4.6 mm
internal diameter, and packed with C13 column material. Mobile phase systems are
usually binary, with an aqueous acidified polar solvent (solvent A) such as aqueous acetic
acid, perchloric acid, phosphoric acid, or formic acid, and a less polar organic solvent
(solvent B) such as methanol or acetonitrile. Flow rates range between 1.0 and 1.5
mL/min. Thermostatically controlled columns are normally kept at ambient or Slightly
above ambient temperatures. Injection generally ranges from 1 to 100 11L. Phenols absorb
in the ultraviolet (UV) region; flavones, flavonols and flavonol glycosides are usually

detected at wavelengths 270, 280, 350, 365, and 370 nm (Merken et al., 2000).

12

2. 2. 2. Modeling the supercritical fluids

There are many methods that describe the solubility of substances in SFS. Two of
them are the solubility parameter and process modeling. The first one describes the
solubility of the solute under laboratory conditions and the second method is to model the
process itself with experiments (Rozzi et al., 2002). In the present study the combination
of the theory (solubility parameter) and the practice was attempted. The extraction
conditions were determined with the solubility parameter, and then these extraction

conditions were adapted to the equipment limitation and operation.

13

3. Materials and Methods

The study was divided in two parts. Preliminary experiments explored methods
for extraction and analysis of quercetin, and familiarizatiOn with the supercritical fluid
(SF) and high performance liquid chromatography (HPLC) system. The preliminary
results lead to the second study phase, where primary experiments and analytical

procedures were focused and replicated.

3.1. Reagents and samples

Industrial grade carbon dioxide (C02, BOC Gases, Murray Hill, NJ.) was used as
a SF. Yellow (Spartan Banner 80 and Sweet Sandwich) and red (Mars) onion varieties
were obtained from the Muck Farm (Michigan State University, East Lansing, MI).
Methanol was purchased from Honeywell International Inc. Burdick and Jackson
(Muskegon, MI). Ethyl alcohol 190 Proof was purchased from Pharmco (Brookfield,
CT). [Phosphoric acid 86.3% was obtained from J. T. Baker (Phillisburg, NJ). Quercetin

standard and diatomaceous earth were obtained from Sigma (St. Louis, M0).

3.2. SF extraction equipment

The SF equipment is shown in Figure 4.

Extractions were conducted in a 500 ml (10) stainless steel high-pressure vessel
(Autoclave Engineers, Inc., Erie, PA). The vessel was wrapped with heating tape (120
volts) and insulated with foil bubble wrap to maintain constant temperature. The
temperature was monitored and controlled with a thermocouple and temperatrue

controller (Omega Engineering, Inc., Stamford, CT).

14

Carbon dioxide from a gas cylinder (1) was compressed by an air-driven booster
compressor (3, Haskel, Inc, Burbank, CA) and stored in a 2 L reservoir (4). The gas
cylinder had a C02 regulator (2) attached to it. Pressure gauges (5, 8) were used to
monitor the pressure in the reservoir and extraction vessel, respectively. A forward
pressure regulator (7, Tescom Corporation, Elk River, MN), positioned between the
reservoir and the extraction vessel, controlled the extraction pressure. The flow of gas
was monitored using a dry test meter (14, American Dry Test Meter Model DTM-200A-
3, American Meter Co., Philadelphia, PA). Two shutoff valves (6, 11, Autoclave
Engineers, Inc., Erie, PA) were used to control the C02 stream in the system. For the
preliminary experiments, a piston pump (9, Model 305, Gilson, Middletown, WI) was
used to inject the modifier in the C02 feed line, and later, to recirculate the extract to that
same line (Figure 5). The micrometering valve (12, Autoclave Engineers, Inc., Erie, PA)
was opened, the pressure was reduced to atmospheric, the extract was then separated
from the gas and collected in a glass tube trap (13). The trap consisted of a glass vial (3.7
ml) inserted in a glass tube, surrounded by dry ice. Cartridge heaters embedded in steel
saddle blocks were mounted to the micrometering valve and set at a temperature just high

enough to keep the valve from freezing due to the instantaneous expansion of C02.

15

4

 

Figure 4. SF equipment. Carbon dioxide cylinder, 1. Carbon dioxide regulator, 2.
Booster compressor, 3. Compressed gas reservoir, 4. Pressure gauges, 5, 8. Shut-off
valves, 6, 11. Forward pressure regulator, 7. Piston pump, 9. Vessel, 10. Micrometering
valve, 12. Trap, 13. Gas meter, 14. Entrainer reservoir, 15. TC: temperatme control

4

 

Figure 5. SF equipment with recirculation. Carbon dioxide cylinder, 1. Carbon
dioxide regulator, 2. Booster compressor, 3. Compressed gas reservoir, 4. Pressure
gauges, 5, 8. Shut-ofl‘ valves, 6, 11. Forward pressure regulator, 7. Piston pump, 9.
Vessel, 10. Micrometering valve, 12. Trap, 13. Gas meter, 14. TC: temperature control

16

3.3. Liquid chromatographic systems

Preliminary analysis of the extracts was conducted using an HPLC system
consisting of Turbochrom chromatography software (version 4.1), and a Perkin Elmer
binary LC pump (Model 250, Norwalk, CT) connected to a Rheodyne (Cotati, CA)
manual injector (5 11L injection loop). The chromatographic column used for this analysis
was a BetaBasic C13, 250 x 4.6 mm, 5 pm particle size, and a 150 A pore size (Keystone
Scientific, Inc., Bellefonte, PA). A guard column (packed with C13, 1 cm x 4 mm, 5 um
particle size) was connected in front of the column. Compounds were detected and
identified using a Perkin Elmer Diode Array detector (Model 235 C).

During initial development of SF extraction and analysis methods, problems were
encountered with the above HPLC system. Therefore, a different HPLC system was
implemented at the time the primary experiments began. The system consisted of
PEAKSIMPLE for WINDOWS ’95 software. A Dionex gradient pump/QUAT (Model
AGP-l, Cotati, CA) was connected to a Rheodyne manual injector (10 ILL injection
loop). The same column and guard column as in the preliminary analysis were used. An
LC spectrophotometer was used as a detector (Waters, Lambda-Max, Model 481,

Milford, MA).
3.4. Sample storage

In order to preserve the raw material during the total research period, onions were

stored at -20°C. Extracts were stored at 6°C.

17

3.5. Supercritical fluid extraction

3. 5. 1. Sample preparation

3.5.1.a. Natural sample: the onion’s outer dry skin layers were manually separated from
the bulb and diced (broken) into small flakes approximately 0.5 x 0.5 in.

3.5.l.b. Spiked model: quercetin (z0.l g) was dissolved in 20 ml of absolute methanol.
This solution was spiked into 50 g of diatomaceous earth. The spiked solid was oven-
dried at 50°C for 2 h and then stored at room temperature for at least 12 h to evaporate
methanol. The ratio of quercetin loaded per weight of earth was chosen based on the
minimum ratio found in the methanol extraction of quercetin from the onion skins. The
minimum ratio, assumed to be the worst-case scenario (smallest amount of quercetin in
natural matrix), was used to investigate if small amounts of quercetin can be extracted in
higher amounts from the inert than fiem the natural matrix with similar quercetin
concentration. .

3.5.2. Experimental Procedure

3 .5.2.a. Preliminary experiments: static and dynamic extraction modes were conducted
separately. Figure 6 shows the schema of this study phase.

Three different static mode experiments were performed. A range of 9 to 12 g of
diced onion skins was used. The entrainer concentration was approximately 3% (molar
concentration). Both entrainer and onion skins were added directly to the vessel before it
was sealed. System pressure was set'to 7700 psi, and the heating tape was set to 40°C. To
start up the extraction, the vessel was slowly filled (to avoid onion skins plugging the
vessel’s outlet and to prevent fieezing the C02 feed line) with C02. The system reached

the desired temperature and pressure in 25 to 30 min. Once the system reached the

18

desired conditions, to collect the extract, the micrometering valve was opened, and the
desired amount of C02 (T able Al) passed the dry test meter, as fast as possible in order
to maintain the static mode method. The micrometering valve was then closed, and the
vial was changed for a new one. The extraction collection plus waiting period between
collections lasted 10 - 20 min, in order to collect higher volumes of extract for sample
analysis. Next, the micrometering valve was opened again, and the next extract was
collected as described above. The extraction period lasted 2.5 h. The extract consisted of

ethanol and polyphenols, including quercetin.

 

Preliminary experiments

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Static mode ‘
Extraction Extraction Extraction
#1 #2 #3 L
Dynamic mode
Extraction Extraction
#1 #2

 

 

 

 

 

 

Figure 6. Preliminary experiments schema

l9

Based on early extractions (before preliminary experiments) conducted with the
same SF equipment, the amount of ethanol added was first estimated, and then confirmed

(once the actual volume of C02 in the system was determined), with the following

 

 

 

equation:
% EtOH = 5'0” "'0’“ * 100 Eqn.1
C02 moles + EtOH moles
Where:
C H OH * ml C H OH
EtOHmoles=p(2 5 ). (2 5 ) Eqn.la
Molecular Weight (CZHSOH)
L ice.)
C02 moles = Eqn.lb

Molar Volume

p (CZHSOH)= 0.8 g/ml

Molecular Weight (CZHSOH) = 46 g/ mole

Molar Volume (C02 ) = 24.34 L / mole (at 298 K and 1 atrn)

EtOH = ethanol
ml (C2H50H) = ml of ethanol used during the extraction period
L (C02) = liters of carbon dioxide used to fill the system.

Two difi‘erent Mmic mode extractions were conducted. The amount of diced onion
skins on average was 11 g. Part of the total amount (2.0 to 6.6 %, molar concentration,
based on Equation 1) of ethanol was added to the vessel, the rest of it was pumped to the
C02 feed line with a piston pump during the extraction period (flow rate range: 0.1 to 0.5
ml/min). The vessel pressrn'e and temperature were set to 7700 psi and 40°C,

respectively. Once extraction conditions were reached, the micrometering valve was

20

opened, the C02 flow started, and the piston pump started to pump the modifier. To
collect the extract, C02 passed through the trap for 10 — 20 min (Table A2), then the
micrometering valve was closed, and the vial was changed for a new one. The
micrometering valve was immediately opened again for the next collection. Extraction
period was about 3 h total.

3.5.2.b. Primary experiments: static mode experiments were conducted for the yellow and
red onion varieties. The schema can be seen in figure 7.

The vessel pressure and temperature were set to 5700 psi (maximum pressure
allowed by the piston pump to recirculate) and 40°C, respectively. The 500 ml vessel was
wrapped with the heating tape and heated 24 h before the extraction, in order to have
rmiform vessel temperature by the time of extraction. Diced onion skins and ethanol (%
calculated based on Equation 1) were added to the vessel before it was sealed. The piston
pump was set to 5 ml/min for recirculation. Once the desired conditions were reached, the
micrometering valve was opened, 15 L of C02 passed through the trap (as fast as possible
to maintain the static mode), then the micrometering valve was closed, and the vial was
changed for a new one. This procedure and the waiting period between collections lasted
10 — 18 min. Then, the next sample collection started once again with the opening of the
micrometering valve (Table A4 and A5). After the collection, the extract volume was
measured with a 5 or 1 ml pipette.
3.5.2.c. Spiked procedure: the same conditions and procedure as the static mode
described above were used (Table A6). The filling procedure of the vessel was different.
First, a piece of glass wool was introduced, followed by a filter paper, then, ethanol was

added (which filtered through the paper). Finally, a mixture of the spiked earth and glass

21

wool was introduced. The filter paper and glass wool acted as a support for the spiked
earth, it also prevented the direct contact of the earth with ethanol. A piece of glass wool

was then added again to the top of the vessel before sealing.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Primary experiments
Red variety Yellow variety Spiked earth
Five Five Five
replications replications replications

 

 

 

 

 

 

 

 

 

Figure 7. Primary experiments schema

3.6. Sample analysis

Extracts collected during SF extraction were injected into an offline HPLC
system, without any further preparation. Analysis of the extracts fi'orn the preliminary and
primary experiments was conducted as follows. Mobile phase consisted of a mixture of
40% phosphoric acid (0.5%, v/v, water solution) and 60% absolute methanol (v/v). Flow
rate was maintained constant at one ml/min, and the detector was set at a wavelength of
280 nm. Period range of analysis was 10 to 15 min for each sample. Calibration curves
were developed prior to analysis with quercetin standards each time that a new mobile
phase was prepared. Five standards of known concentration were prepared and injected
into the HPLC system. Retention time and area of the peak were recorded. Simple linear

regression was applied to obtain the calibration curves (Figure A1.Table A3).

22

Before extracts were injected into the HPLC, standards were run at least twice in
order to confirm retention time and area under the curve for quercetin. Then, samples
were run once (since the volume of the extracts were not enough to rim several times).

Mass spectrometry analysis was conducted at the Department of Biochemistry,
Mass Spectrometry facility (Michigan State University) to identify and confirm the
presence of quercetin in the extract. Matrix Assisted Laser Desorption Ionization

(MALDI) Time of flight Mass Spectrometry was the method used.

3.7. Solid-liquid extraction
Along with each SF extraction, a methanol extraction was conducted. From the
skins prepared for the SF extraction, a portion of it was separated and steeped in absolute
methanol (Table 3), 1:40 (w/v) at room temperature, for 24 hours (Bilyk et al., 1984).
Then, methanol was separated from the Skins by filtering under reduced pressure
(aspirator) through filter paper, grade 601, size 7 cm (Ahlstrom, Mt. Holly Springs, PA).
Methanol extracts were analyzed with the same method and HPLC system as the

SF extracts.

23

Table 3. Methanol extraction data

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Extraction 21:3: Methanol Extractioni 21:51,: Methanol
if weight “Hf“ is weight 'z'df)“
(8) (m) (g) m
Prelimim experiments

Static Mode Dynamic Mode
1 1.103 44 1 0.464 19
2 0.475 19 2 0.502 20
3 0.247 10

Primary e riments

Red varie Yellow variey
1 0.429 17 1 0.234 11
2 0.442 18 2 0.483 19
3 0.371 1 5 3 0.465 19
4 0.794 32 4 0.238 10
5 0.825 33 5 0.221 8

 

 

24

4. Results and discussion

4.1. Optimal extraction conditions

Extraction conditions were chosen based on the solubility parameter and
equipment limitations. In theory, maximum solubility should be attained when the
solubility parameter of the extracting fluid (C02) is equivalent to that of the solute, in this
case quercetin (King, 1989). The solubility parameter of C02 can be estimated based on

the Giddings equation (Giddings et al., 1968),
6 = 1.25R"2(p, /2.66) Eqn, 2

Where: 8 = solubility parameter, Pc = critical pressure, and pt = reduced density.

From this equation, to maximize the solubility parameter, maximum reduced
density should be attained. Based on Giddings’ reduced states curve (Figure Bl), for a
given reduced pressure, lower reduced temperature gives higher reduced density.
Therefore, during extraction, the goal was to maintain higher pressure at lower
temperature. The effect was made to maximize the solubility parameter of C02, since
quercetin is a very polar compound (five hydroxyl groups), and therefore has a high
solubility parameter value. In theory, the solubility parameter is the quantitative meaning
of polarity (Schoenmakers et al., 1982).

Exact numbers could not be calculated since other variables could also affect both
quercetin and C02 solubility parameters, such as the presence of other similar compounds
in the onion matrix, in which case, extraction must be optimized for groups of
compounds rather than a single target analyte (Hawthorne, 1990). Additionally, Giddings
relationship is useful when target analytes represent a large percentage of the bulk sample

(Hawthorne, 1990), whereas quercetin is found in minor concentrations in onion skins.

25

The solubility parameter of C02 is assumed to be affected also by the solubility parameter
of the entrainer, as is reported with liquid mixtures (Karger et al., 1973).

However, this theory was still used in this project as an acceptable estimation of
optimal operation conditions, based on the results of caffeine extraction predictions done

by King et al. (1990).

4.2. Preliminary experiments
4. 2. I . Static mode

The total ratio of quercetin extracted per liter of C02 (extraction ratio) was found
to be in the range of 0.19 to 0.57 (Table 4). Experimental data collected during each
extraction can be found in Table A1. Variation in this ratio is assumed to be the result of
non-uniform distribution of solute concentration through the vessel, and it might be
possible that gas channeling inside the vessel had been formed. The collected data (Table
A1) and the results (Table 4) suggest that there is a relationship between volume of C02
passed through the trap, and total quercetin extracted. At the highest volume of C02
passed, the highest amount of quercetin was extracted. However, this proportional
relationship is not found between extraction duration and total quercetin extracted.

An example of the chromatograms and the UV absorbance chart that confirmed

the presence ofquercetin can be seen in Figures B2 and B3.

26

Table 4. Results from static mode

 

 

 

 

 

 

 

 

 

 

. Extraction g of
ExtractioniEm'ac.tloll Conc.‘ Total. ratio (pg oflquercetinlkgTouu
duration uercetrn , - , (:02
ff (min ) (pg/III (pg) iquercetmlL of onron (L) c
‘ of C02) " skins
1 132 0.017 71.57 0.57 0.0062 129
2 136 0.004 11.45 0.19 0.0010 60
3 192 0.012 49.16 0.41 0.0054 121

 

 

 

" b Values shown are average in time
‘ Concentration of quercetin per ethanol volume collected in the trap
° Carbon dioxide passed through trap during extract collection

4. 2. 2. Dynamic mode

Table 5 shows the results from the dynamic mode. The experimental data
collected during each extraction can be seen in Table A2. It can be observed from the

results that at highest volume of C02 passed through the trap, the highest total quercetin

is extracted. However, total quercetin appears to be independent from extraction ratio and

 

 

 

 

extraction duration.
Table 5. Results from dynamic mode
. Extraction g of
Extractionidemcimn Conc.' Flow rate Total. ratio (pg oflquercetin/kgToull

uratron . I, tquercetrn , , C02
# (min) (Org/pl» (L/mm) (pg) quercetin/L of onron (L) a

' of C02) ‘ skins
l 174 0.005 1.11 9.9 0.0481 9.80E-04 207
2 169 0.005 2.03 12.81 0.0365 1.03E-03 403

 

 

 

 

 

 

 

 

 

“’1” ° Values shown are average in time
' Concentration of quercetin per ethanol volume collected in the trap
d Carbon dioxide passed through trap during extract collection

An example chromatograrn can be seen in Figure B4.

27

 

4. 2. 3. Comparison between modes

Static and dynamic modes were conducted separately in order to compare their
performance (total quercetin extracted). As seen in Table 6 the static mode had a higher
total quercetin and extraction ratio than the dynamic mode. This result can be due to
higher contact time between the solute and the dense gas in the static mode. Continuous
flow during dynamic mode may not provide sufficient time for interaction between the
solute and the gas

Table 6. Comparison between modes

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Emcfi lC . Total Extractionf
, on onc. .
Mm“ # (pg/pl) “mmgflfé’iiin
(143) of €02) h
Static 1 0.017 71.57 0.569
2 0.004 11.45 0.187
3 0.012 49.16 0.408
Dynamic 1 0.005 9.90 0.048
2 0.005 12.81 0.037
' Concentration of quercetin per ethanol volume collected in the trap
" b Average values in time

Based on these preliminary results, Static mode was the extraction method chosen
to compare the SF extraction of quercetin between yellow and red onion varieties.
Moreover, to improve reproducibility and to achieve uniform distribution of solute

through the vessel, a recirculation pump was added to the SF equipment.

28

4.3. Static Mode
4.3.1. Red variety

In five replicate extractions (Table 7) on average, 287.86 pg of quercetin could be
extracted (in 150 min), with a ratio of 1.91 ug of quercetin per liter of C02; this generates
a total amount of quercetin of 0.024 g per kg of onion skins. Even though there is a small
variation in the extraction druation (143 to 157 min), weight of onion Skins (10.50 to
13.90 g), and percentage of ethanol (6.80 to 7.78%), the amount of quercetin extracted

had no relationship with any of these factors. Specific details of each extraction can be

 

 

 

 

 

 

 

 

 

 

seen in Table A4.
Table 7. Results for red variety
. Extraction g of Onion %
Extraction $323,? Conc. Lqutowl' ratio (rig oflquercetinlkg skins Ethanofi
#, (date) (min) (pg/III ) ' quercetin/L of onion weight (molar
(pg) of C02) " skins (g) cone.)
1, (1 1/08/01) 145 0.037 313.22 2.09 0.027 1 1.80 7.09
2, (1 1/12/01) 143 0.037 322.64 2.09 0.027 13.90 7.78
3, (1 1/26/01 152 0.035 290.91 1.94 0.023 12.80 6.80
4, (01/22/02) 152 0.029 286.57 1.91 0.024 10.50 7.78
5, (01/25/02) 157 0.029 225.98 1.51 0.019 12.00 7.78
150 0. 034 287.86 1.91 0. 024 12.20 7.45

 

 

 

 

 

 

 

 

 

' Concentration of quercetin per ethanol volume collected in the trap.

 

" b Values shown are average in time

An example of the calibration curve for analysis is shown in Figure A1 (Table
A3). An example of the chromatograms from HPLC analysis and results from the Mass
Spectrometry analysis are shown in Figures B5, B6, and B7.

Since the results suggest no effect of time, onion weight and ethanol

concentration (within the range that these parameters varied) on total quercetin extracted,

29

the influence of ethanol volume collected in the trap was investigated. In order to
evaluate any significant correlation between the quercetin extracted and the ethanol
volume collected in the trap, a statistical analysis was performed. To conduct this
analysis, the average of the five replications for each of the 10 vials collected, was
calculated (Appendix C, analysis C1). The regression analysis showed that, at 95%
confidence level, a significant linear relationship exists between these two variables
(R2=O.82, p< 0.001). This result suggests that there is a limiting factor in the collection of
the solute of interest, with less ethanol volume less quercetin is collected per vial (Figure

8).

 

 

R2 = 0.3255 .

L- - .,_. L_¢—»—— ——1

 

 

 

 

Quercetin (pg)
t» «h
o o

 

 

-—-N
CO
I

 

O
_ Rm . -....+...

 

0 0.2 0.4 0.6 0.8 1 1 .2
Ethanol (ml)

l
i
l
. . . . r , I
l
1
Figure 8. Correlation between quercetin extracted and ethanol volume per vial
Furthermore, regression analysis of ethanol volume (average of five replications
for each of the ten vials) versus time, and extraction ratio (average of five replications for
each of the ten vials) versus time were done in order to find any possible trend (Appendix

C, analyses C2 and C3). Both analyses suggest the same trend, as time increases, volume

of ethanol and extraction ratio decrease. At 95% confidence level, a Significant linear

30

relationship was found for both regressions (Figures 9 and 10), at p=0.001 for ethanol

versus time, and p< 0.001 for extraction ratio versus time.

 

 

1.8 T; R2 = 0.7592

1.6 T
1.4 i
1.2

 

 

 

 

 

 

0.8 T N'L‘xm i

0.6
0.4
0.2

 

Ethanol collected (ml)
I
i

 

 

 

O 50 100 150 200

Time (min)

 

Figure 9. Correlation between ethanol volume per vial and time

 

 

 

 

 

 

 

 

 

 

 

 

 

, i
l 7 i ‘
A R2=0.7819 l
8 6 4% --~ I
l U 1
E i, 5 j i
‘ 3- 4 l
i .2 4 i — 1
i a s - i - 1
H 2 h \ -_\\ 5
1 E T ' rr t i
t5 ‘7 La
o T . -—-——~~ . ~ . g
f o 50 100 150 200
i Time(min) E

 

Figure 10. Correlation between extraction ratio per vial and time

31

From these results, the influence that the entrainer volume collected in the trap
has on the amormt of quercetin extracted can be observed. The reasons for the trends
shown in Figures 8, 9, and 10, could be the following. First, that ethanol was saturated
with quercetin therefore, no more quercetin could be trapped in the collection trap (direct
relationship shown in Figure 8); second, not all the ethanol that was coming out with the
gas stream was trapped (trend shown in Figure 9); third, within the vessel, ethanol has
better affinity with the onion skins than with the gas stream, therefore, afier certain time
no more ethanol was pulled out of the vessel (Figure 9); and finally, quercetin has better
affinity with ethanol than with C02, so the dense gas cannot compete with the entrainer
and quercetin stays where ethanol stays (Figure 10).

The above reasoning also suggests that saturation of the C02 with quercetin could
not be reached or detected during the extraction, since constant concentration of quercetin
(pg/L of C02) in time was not achieved. Other similar compounds, such as resveratrol,
gallic acid, catechins, p-coumarin, grape glycosides, etc., reached saturation fast (15
min), as it is reported by Palma et al. (2000), Pascual-Marti et al. (2001), and Tena et al.
(1998). These studies were conducted in a much smaller vessel (2:7 ml) than the present
study (500 ml). This fact is assumed to be one of the causes of a better understanding of
the early stages of the extraction process, since they had a better control, for example, of
the filling time (seconds), pressure and temperature, less void volume, less potential for
channeling formation inside the vessel.

There are also other factors that may influence the results, for instance, the
presence of other similar compounds in the onion skins, such as kaempferol and

myricetin. These compounds may be competing with quercetin for a place in the gas

32

stream and ethanol. On the other hand, it might be possible, that forces in the onion skin ,
due to bonds with these similar compounds, are stronger than the solvent power of C02,
so extraction of quercetin becomes difficult.

The results of this study show that it is possible to overcome the difficulties
mentioned above, and the experimental limitations (vessel volume, trap collection,
ethanol introduction, etc.), since the extraction of quercetin from the natural matrix was

successfully accomplished.

4. 3. 2. Yellow variety

In five different extractions (Table 8) on average, a total of 173.87 pg of quercetin
could be extracted, with an extraction ratio of 1.142 pg of quercetin per liter of C02 (in
156 min); this generates 0.02 g of quercetin per kg of onion peels. Again, as in the red
variety case, small variations in extraction durations (151 to 164 min),- weight of onion
skins (8.10 to 8.60 g) and ethanol concentration (7.69 to 8.17%) have no relationship with
the extraction ratio of solute per gas (within the range that these parameters varied).

Experimental data collected of each extraction can be found in Table A5.

33

Table 8. Results for yellow variety

 

Total

. Extraction g of Onion %
Extraction Extractiont Cone. '

. . ratio (pg oflquercetin/kg skins Ethanoll
#, (date) duration (pg/pl) Iquemnntquercetin/L of onion weight (molar
(mum) (pg) of C02) " skins (g) cone.)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

l , (02/07/02) 156 0.015 192.523 1.268 0.022 8.6 7.69
2, (02/13/02) 151 0.016 201.448 1.324 0.025 8.1 7.69
3 , (02/20/02) 159 0.020 225.076 1.461 0.023 9.6 8.17
4, (02/28/02) 164 0.020 149.193 0.989 0.018 5.8 7.69
5, (03/05/02) 150 0.014 101.098 0.668 0.012 8.2 7.69

156 0. 01 7 1 73.868 1.142 0. 020 8.06 7. 79

 

‘ Concentration of quercetin per ethanol volume collected in the trap
" b Values shown are average on time

An example of the chromatograms can be found in Figure BS.
Again, the relationship between ethanol collected and quercetin extracted per vial
was tested (regression can be seen in Figure 11). At 95% confidence level, a significant

linear relationship was found, with a p< 0.001 and r2=0.877 (Appendix C, analysis C4).

 

'7 —1

A
O

l
E

R2 = 0.8772

Mb.)
CU!

 

N
M

 
 

 

 

Quercetin (pg)
N
C

 

 

15 ._~

. 10 E

l 5 E '

é o . » - . - E
E o 0.5 1 1.5 2 E
E Ethanol(ml) E

 

Figure 11. Correlation between quercetin extracted and ethanol volume collected
mt vial

34

Figures 12 and 13 show the same trend as in the red variety, as time increases,
ethanol volume collected and the extraction ratio per vial decrease. The same
assumptions as those for the red variety were made in this case. Statistical analysis
(Appendix C, analyses C5 and C6) showed a significant linear relationship at 95%
confidence level, with p=0.004 for ethanol versus time, and p=0.001 for extraction ratio

versus time.

 

 

 

 

 

 

 

 

 

 

 

3 E
=7 2.5
5 R2=0.6711
B 2 .
3 l
-8- l 5 N
"o' 1 ' h~
g mxtkl
m 0.5
E. 0 , ; . T l
I o 50 100 150 200
l
‘ Time (min)

 

 

Figure 12. Correlation between ethanol volume collected per vial and time

35

 

 

 

 

 

 

 

E a: E R2=0.7363

E O 1

E g 4 +

E 3

E .2 3 E

E E" 'E'

.5 2 ~-~\\- ‘"

‘ 8 E I \“~‘

E g IE I I \-\'\I\'\‘ E

as E \-

: 04 ~ . . z

E 0 50 100 150 200 E
E

E Time (min) E

 

Figure 13. Correlation between extraction ratio per vial and time

4. 3. 3. Comparison between varieties

A statistical analysis was performed (Appendix C, analysis C7) in order to test
any significant difference between g of quercetin extracted per kg of onion skins from
each variety. At 95% confidence level there is no significant difference (p=0.182)
between them, with less standard deviation (0.003) for the red than the yellow variety
(0.005). Results from the yellow and red onion varieties extractions (average of five

replications) are shown in Table 9.

36

Table 9. Results for the red and yellow varieties

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

. g of
”Eim‘ll'lqmm
(min) of onron

skins

Red
145 0.027
143 0.027
152 0.023
152 0.024
157 0.019
Yellow

156 0.022
151 0.025
159 0.023
164 0.018
150 0.012

 

 

 

 

4. 3. 4. Comparison of the preliminary and primary experiments conducted with yellow
variety

Statistical analysis was conducted to the values shown in Table 10 (Appendix C,
analyses C8, C9 and C10). It was found that at 95% confidence level the total quercetin
extracted (p=0.007), g of quercetin per kg of onion skins (p=0.003), and ethanol volume
collected per vial (p=0.004) had a significant difference between the preliminary and the
primary experiments conducted. This means that primary experiments have significantly
higher values than preliminary ones. However, the extraction ratio did not show a
significant difference between the preliminary and primary experiments (p=0.01,

appendix C, analysis C11). From these results it can be observed that, although the

37

extraction ratio did not present a statistical difference, the ethanol collected in the primary
experiments was higher, which gave higher amount of quercetin extracted and g of
quercetin per kg of onion skins. These results also suggest that the recirculation pump
and dry ice added to the collection trap may have improved the amount of quercetin
extracted.

Table 10. Results from preliminary and primary experiments for yellow variety

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Ethanol
. volume Extraction of
ExtractionEEJmcanE Cone. ' EcollectedEq Total. ratio (pg of Equergefin/kg
uratron . uercetm , ,

# (min.) (pg/pl) m the (P8) quercetméL of omen

trap of C02) skins
(ml)

Treliminary 1 132 0.017 0.468 71.57 0.57 0.006
2 136 0.004 0.234 1 1.45 0.19 0.001
3 192 0.012 0.206 49.16 0.41 0.005
Primary 1 156 0.015 1.107 192.52 1.27 0.022
2 151 0.016 1.180 201.45 1.32 0.025
3 159 0.020 1.090 225.08 1.46 0.023
4 164 0.020 0.733 149.19 0.99 0.018
5 150 0.014 0.706 101.10 0.67 0.012

 

 

 

' Concentration of quercetin per volume of ethanol collected in the trap
" b Values shown are average in time

4. 3. 5. Comparison of SF with solid-liquid extraction

Since SF and methanol extractions were not performed in the same time basis (3

and 24 h, respectively), it is not possible to directly compare the amount of quercetin

extracted in each extraction. However, it was found (Appendix C, analyses C12 and C13)

that standard deviation is higher for methanol extraction (3.034 g of quercetin per kg of

onion skin, red variety; 3.214 g of quercetin per kg of onion skin, yellow variety) than for

38

SF extraction (0.003 g of quercetin per kg of onion skin, red variety; 0.005 g of quercetin

per kg of onion skin, yellow variety). Results fiom both extractions can be seen in Table

11.

Table 11. SF and methanol extractions results

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

. SF (90‘ Methanol (g of
Variety Extrzctron “Tm tall/1kg quercetin/kg of
skins) onron skins
Red 1 0.027 10.14
2 0.027 15.90
3 0.023 8.73
4 0.024 10.82
5 0.019 8.33
Average = 0. 024 I0. 78
Yellow 1 0.022 3.80
2 0.025 10.41
3 0.023 3.07
4 0.018 3.48
5 0.012 7.62
Average = 0. 020 5.68

 

 

 

4.4. Static mode for pure quercetin

 

Five replications were done using the same extraction conditions and procedure
(Table A6) as in the previous red and yellow variety cases. A maximum amount of 1800
pg was extracted from the spiked inert matrix with a maximum extraction ratio of 12 pg

of quercetin per liter of C02 (Table 12).

39

Table 12. Results from the pure quercetin extraction

 

. Total Extraction
Extraction Cone. EquercetinEraho (pg of

#. (date) Etna/11L) (pg) «macaw
0 2

 

1,(03/26/02) 0.015 107.79 0.70
2, (04/01/02) 0.104 650.43 4.34
3, (04/03/02) 0.040 247.06 1.65
4, (04/10/02) 0.158 1832.50 12.02

5, (04/17/02) 0.087 706.02 4.69

" b Values shown are average in time
" Concentration of quercetin per volume of ethanol collected in the trap

 

 

 

 

 

 

 

 

 

 

Even though the results were not as reproducible as with the natural onion matrix,
similar trends were found. After a regression analysis (Appendix C, analysis C14) was
applied to quercetin extracted versus ethanol volume collected per vial (average values of

five replications for each vial), at 95% confidence level, a significant linear relationship

was found (p< 0.001, Figure 14).

40

 

 

R2 = 0.9177

 

 

 

Quercetin (pg)
8
O
E

 

 

0 a. 1 . . w E
0 0.2 0.4 0.6 0.8 1 f
Ethanol (ml) E

 

Figure 14. Correlation between quercetin extracted and ethanol volume collected
per vial

A decreasing trend was also found between ethanol collected versus time and
extraction ratio versus time (Appendix C, analyses C15 and C16). However, they did not
have a significant linear relationship as they had with the natural matrix. These results
suggest that the collection trap was not efficient enough to collect all the ethanol and
solute that were coming in the gas stream. No ethanol-matrix and/or solute-matrix
interaction is supposed to occur, since the matrix is a spiked earth.

Extraction ratios were also compared (Appendix C, analysis C17). Extraction
ratios from the yellow variety presented less standard deviation (0.314 pg/L) than the
pure quercetin (4.444 pg/L). The difference between the first three extraction ratios fi'om
the pure quercetin was assumed to be due to varying degrees of channeling formation
inside the vessel. To avoid the potential of this channeling glass wool was mixed with the

earth for the last two extractions. Extraction ratios values can be seen in Table 13.

41

Table 13. Extraction ratios for yellow variety and pure quercetin

 

Extraction

ratio (pg of

. quercetin/L of
Extraction €02) a

#

 

Pure quercetin
0.7
4.34
1 .65
12.02
4.69

 

 

 

 

 

M-FUJNH

 

Yellow varieg
l .27
1.32
1 .46
0.99

5 0.67
' Average values in time

 

 

 

 

hUJN—e

 

 

 

 

 

Analysis of variance was performed (Appendix C, analysis C18) to the values
presented in Table 14. The amount of quercetin extracted per kg of onion skins showed
less standard deviation (0.005 g of quercetin per kg of onion skin) than the amount of
quercetin extracted per kg of earth (0.013 g of quercetin per kg of earth). However, the
values shown (g ofquercetin extracted/kg of matrix) in Table 14 did not show significant
difference at 95% confidence level (p=0.3 75). These results suggested that there is no
effect due to matrix, quercetin collected is the same for both, natural and inert matrices
(for this study and under the conditions that it was conducted).

An example of- the chromatograrns for pure quercetin can be seen in Figure B9.

42

Table 14. Comparison of the results fi'om the yellow variety and pure quercetin

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Quercetin
of present
EExtractio quegrcetin before
# xtrac t % extraztfron (g
0‘ matrrx quercetin/kg
of matrix)
Onion skins
1 0.022 3.80 '
2 0.025 “MI
3 0.023 3-07
4 0.018 3-43
5 0.012 8.13
Earth
1 0.002 2.50
2 0.013 2.78
3 0.005 2.56
4 0.036 2.42
5 0.014 2.47
' Based on methanol extraction

43

5. Conclusions

This study confirms the potential of supercritical fluid extraction as an alternative
method to extract quercetin from a natural matrix, in this case, onion skins. Essential data
on the extraction process have been provided for the first time. After comparing the
amount of quercetin that can be extracted with dynamic (11 pg total, average of two
replications) and static (44 pg total, average of three replications) modes, the static mode
was found to be the best method.

There is no statistical difference between the amount of quercetin recovered fi'om I
the red (0.024 :t 0.003 g of quercetin/kg of onion skins, average of five replications) and
yellow (0.020 i 0.005 g of quercetin/kg of onion skins, average of five replications)
onion varieties. The overall amount and reproducibility (g of quercetin per kg of onion
skins), for both varieties, has been improved by adding a recirculation system to the SF
equipment.

The modifier has an important influence on the amount of quercetin extracted. At
the highest ethanol volume collected per extract, the highest amount of quercetin was
extracted. This trend is found in both varieties and the pure solute.

Even though the extraction ratio (pg of quercetin/L of C02) from the natural
matrix has less standard deviation (1.142 t 0.314 pg/L) than the extraction ratio from the
spiked inert matrix (4.680 i 4.444 pg/L), higher values were obtained from the spiked
inert matrix (as much as 8.23 times higher, comparison between maximum values, 12.02

pg/L on earth and 1.46 pg/L on onion skins).

When the amount of quercetin extracted per kg of onion skins is compared to the
amount extracted per kg of earth, there is no statistical difference between them, meaning

that there is no effect due to matrix (under the extraction conditions of this study).

45

6. Recommendations

6.1. Experimental improvements
It might be helpful for future experiments:

0 F reeze-dry the onion skins before the extraction to ensure no water is present in
the sample.

0 The current method of recirculation can be improved. For this study the flow rate
of this recirculation was not known exactly. An online flow meter would be
helpful to ensure what portion of the gas stream and solute are been recirculated.

o The current collection trap could be improved in order to ensure the total
collection of ethanol and solute. A solid trap is a possibility, such as C13 column.
A liquid trap can be also implemented such as specific volume of ethanol or
water. Furthermore, a combination of both solid and liquid might also improved

the collection system.

6.2. Future work
It would be beneficial to further investigate:
o If storage period and temperature had any influence on quercetin content or
extraction.
0 The saturation point (when concentration of quercetin is constant in time) of the
pure quercetin, in order to estimate the equilibrium constant, and develop a mass
transfer model.

0 Application of the above model to the natural matrix.

46

How to extract the maximum ethanol volume and quercetin from the vessel if
they tend to interact with the onion skins.

Identify and quantify the other similar compounds that are extracted along with
quercetin, in order to find any possible pattern that they are following.

Once the other compounds are identified, how to improve selectivity of SFE by

further separation of these compounds (by different traps).

47

APPENDICES

48

APPENDIX A. Experimental Data

Preliminary results

Table A1. Experimental data collected during static mode

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Volume: . Total Volume Extraction g of
Extraction # Ti!” Vi“ per Cone. quercetin C02 ratio (pg of quercetin/kg
(mm.)1 # extract (pg/pl) (pg) (L) quercetin/L of ornon
(ml) of C02) skins
5 1 0.62 0.015 9.36 11.91 0.79 8.10E-04
20 2 0.48 0.014 6.86 9.57 0.72 5.90E-04
35 3 0.37 0.012 4.44 10.24 0.43 3.80E-04
53 4 0.46 0.022 10.26 25.27 0.41 8.80E-04
1 70 5 0.27 0.016 4.29 10.60 0.41 3.70E-04
85 6 0.44 0.012 5.41 15.54 0.35 4.70E-04
100 7 0.53 0.013 7.10 15.63 0.45 6.10E-04
1 17 8 0.54 0.021 1 1.34 15.32 0.74 9.80E-04
132 9 0.50 0.025 12.50 15.08 0.83 1.08E-03
T oral/Average 4.21 0.017 71.57 129.16 0.57 8 6.17E-03
3 1 0.32 0.004 1.18 6.00 0.20 1.00E-04
20 2 0.18 0.000 0.00 4.00 0.00 0.00E+00
35 3 0.28 0.001 0.39 4.00 0.10 3.40E-05
55 4 0.21 0.000 0.00 5.00 0.00 0.00E+00
69 5 0.29 0.003 0.84 6.00 0.14 7.20E-05
2 81 6 0.26 0.005 1.25 5.00 0.25 1.10E-04
93 7 0.20 0.005 1 .02 5.00 0.20 8.70E-05
102 8 0.26 0.006 1.48 5.00 0.30 1.30E-04
1 1 1 9 0.20 0.006 1.28 5.00 0.26 1.10E-04
1 19 10 0.20 0.006 1.18 5.00 0.24 1.00E-04
128 l l 0.21 0.010 2.08 5.00 0.42 1.80E-04
136 12 0.20 0.004 0.74 5.00 0.15 6.30E-05
Total/Average 2.81 0.004 1 1.45 60.00 0.19 9.80E-04

 

' Concentration of quercetin per ethanol volume per extract

49

 

Table A1 . Continuation

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

3 l 0.20 0.000 0.00 6 0.00 0.00E+00
12 2 0.23 0.012 2.69 6 0.45 3.00E-04
23 3 0.22 0.013 2.79 ' 6 0.47 3.10E-04
32 4 0.21 0.012 2.46 5 0.49 2.70E-04
41 5 0.21 0.014 2.92 6 0.49 3.20E-04
52 6 0.27 0.014 3.78 7 0.54 4.20E-04
62 7 0.23 0.015 3.36 6 0.56 3.70E-04
71 8 0.22 0.014 3.15 6 0.52 3.50E-04
81 9 0.22 0.009 1.94 6 0.32 2.10E-04
3 92 10 0.20 0.018 3.60 6 0.60 4.00E-04
101 1 1 0.20 0.013 2.56 5 0.51 2.80E-04
109 12 0.23 0.012 2.69 7 0.38 3.00E-04
1 18 13 0.18 0.012 2.20 6 0.37 2.40E-04
131 14 0.20 0.01 1 2.20 6 0.37 2.40E-04
142 15 0.19 0.012 2.32 7 0.33 2.60E-04
153 16 0.23 0.01 1 2.53 6 0.42 2.80E-04
164 17 0.17 0.013 2.18 6 0.36 2.40E-04
174 18 0.20 0.012 2.30 6 0.38 2.50E—04
181 19 0.15 0.01 1 1.67 6 0.28 1.80E-04
192 20 0.16 0.012 1.84 6 0.31 2.00E—04
T otaI/A verage 4.12 0.012 49.16 121 0.41 5.40E-03
Calculations for all tables
0 Concentration in ppm: from HPLC analysis
0 Concentration of quercetin per volume of ethanol:
pg/yl=ppm/1000 Eqn.A1
- Total quercetin:
pg = ethanol(ml) “ conc.(pg / pl) * 1000 Eqn. A2

50

 

0 Total ratio:

,ug / L = quercetinUrgfl C02 (L)

o Quercetin per weight of onion peel:

g / kg = [(quercetingug) /(1000 * g of onion peel used )]

Eqn. A4

Table A2. Experimental data collected during dynamic mode

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

VolumeE Total Flow Volum Exam“ quegrgtin/
Em?“ _('fni:°)v;"EexP£a 1&3; uereetin rate or C0; or kg or
' (ml) (pg) (Ia/min» (L) quercetin! onion
Lof C01) skrns

0 l 0.10 0.006 0.64 1.25 16 0.0402 6.37E-05

14 2 0.10 0.005 0.51 1.00 14 0.0363 5.04E-05

28 3 0.10 0.005 0.50 1.00 16 0.0315 5.00E-05

42 4 0.08 0.007 0.54 1.00 13 0.0415 5.35E-05

55 5 0.10 0.007 0.72 1.00 16 0.0448 7.10E-05

69 6 0.14 0.006 0.84 1.13 15 0.0563 8.37E-05

1 83 7 0.15 0.005 0.81 1.15 15 0.0538 7.98E-05

96 8 0.17 0.006 1.09 1.13 15 0.0728 1.08E-04

l 10 9 0.19 0.003 0.48 1.14 16 0.0301 4.76E-05

124 10 0.18 0.005 0.82 1.18 14 0.0589 8.1’6E-05

137 11 0.23 0.003 0.75 1.13 13 0.0580 7.47E-05

149 12 0.17 0.004 0.62 1.16 15 0.0415 6.16E-05

162 13 0.22 0.003 0.68 l .15 14 0.0489 6.78E-05

174 14 0.28 0.003 0.87 1.13 15 0.0581 8.64E-05

otal/AverageE 0.16 0.005 9.90 1.11 207 0.0481 9.80E-04

 

 

‘ Concentration of quercetin per ethanol volume per extract

51

 

Table A2. Continuation

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0 1 1.100 0.000 0.000 1.580 43 0.0000 0.00E+00

24 2 0.300 0.001 0.260 2.050 27 0.0095 2.06E-05

38 3 0.530 0.007 3.960 2.220 30 0.1322 3.20E-04

50 4 0.500 0.004 2.060 2.400 27 0.0762 1.66E-04

61 5 0.090 0.007 0.660 1.760 24 0.0273 5.29E-05

2 72 6 0.100 0.009 0.900 3.000 27 0.0333 7.26E-05

85 7 0.350 0.005 1 .5 80 1.870 42 0.0376 1 .27E-04

107 8 0.200 0.01 1 2.140 1.870 45 0.0475 1.73E-04

129 9 0.100 0.010 1.020 1.870 30 0.0341 8.24E-05

148 10 0.090 0.001 0.120 1.870 53 0.0023 9.96E-06

169 l 1 0.070 0.002 0.1 10 1.870 55 0.0019 8.64E-06

Total/A verage 0.310 0.005 12.810 2.030 403 0.0365 1.03E-03
HPLC analysis

Table A3. Data for calibration curve

 

 

 

 

 

 

 

Standard
tconcentrationE
Area (ppm)
2135.08 178.72
1097.10 89.36
573.55 44.68
188.55 26.81

 

 

52

 

E Calibration curve 5
y = 0.0833x E
/- R2 = 0.99 E

E
E
E
I

§

 

Conc (ppm)
‘6 8 8

 

 

O

 

l l 1 TE 1

0.00 500.00 1000.00 1500.00 2000.00 2500.00 E

Figure A1. Calibration curve

 

 

Calculation of quercetin concentration

y = 0.0833 " x Eqn. A5
Where:

y = concentration of quercetin in ppm

x=areaunderthepeak

53

Primary results

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Red variety
Table A4. Experimental data and results
Volume Extraction g of .
Extraction Time Vial per Cone. quiltctflin Volume ratrzfotg (13:20)?"
#, (Date) (mm.) # extract (pg/p1) " (pg) CO; (L) quercetin/L onion
1"“) or C02) skins
0 1 0.92 0.057 52.376 15 3.49 0.0044
20 2 0.91 0.008 6.857 15 0.46 0.0006
38 3 0.91 0.064 58.283 15 3.89 0.0049
53 4 0.78 0.056 44.004 15 2.93 0.0037
1, 70 5 1.08 0.066 71.324 15 4.75 0.0060
(1 1/08/01) 89 6 0.59 0.026 15.053 15 1.00 0.0013
104 7 0.89 0.051 45.772 15 3.05 0.0039
115 8 0.49 0.022 10.863 15 0.72 0.0009
130 9 0.43 0.015 6.497 15 0.43 0.0006
145 10 0.32 0.007 2.187 15 0.15 0.0002
Total
/A verage 7.32 0.037 313.216 150 2.09 0.0265
0 1 1.10 0.078 85.383 15 5.69 0.0061
17 2 0.79 0.045 35.378 15 2.36 0.0025
35 3 1.21 0.019 22.796 15 1.52 0.0016
46 4 0.89 0.021 18.827 15 1.26 0.0014
2, 63 5 1.02 0.040 41.153 15 2.74 0.0030
(1 1/12/01) 82 6 1.00 0.033 33.078 15 2.21 0.0024
96 7 0.78 0.028 21.505 15 1.43 0.0015
1 12 8 0.90 0.028 25.374 15 1.69 0.0018
127 9 0.87 0.026 22.193 15 1.48 0.0016
143 10 0.83 0.020 16.947 15 1.13 0.0012
Total
/Average 9.39 0.034 322.636 150 2.15 0.0232

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

' Concentration of quercetin per ethanol volume per extract

54

 

Table A4. Continuation

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0 1 0.85 0.053 44.938 15 3.00 0.0035
17 2 0.94 0.049 45.715 15 3.05 0.0036
35 3 0.92 0.050 46.431 15 3.10 0.0036
51 4 0.90 0.043 38.469 15 2.56 0.0030
3,01,26,01) 70 5 0.84 0.039 32.740 15 2.18 0.0026
89 6 0.76 0.034 25.953 15 1.73 0.0020
103 7 0.68 0.025 17.155 15 1.14' 0.0013
120 8 0.74 0.020 15.109 15 1.01 0.0012
136 9 0.71 0.020 14.260 15 0.95 0.0011
152 10 0.60 0.017 10.140 15 0.68 0.0008
Total/Average 7.94 0.035 290.910 150 1.94 0.0227
0 1 1.04 0.053 54.845 15 3.66 0.0046
16 2 1.20 0.069 82.386 15 5.49 0.0069
35 3 1.10 0.031 33.965 15 2.26 0.0029
50 4 0.94 0.005 5.161 15 0.34 0.0004
4, (01/22/02) 72 5 1.09 0.045 48.513 15 3.23 0.0041
90 6 0.61 0.019 11.785 15 0.79 0.0010
103 7 0.68 0.014 9.493 15 0.63 0.0008
120 8 0.69 0.025 17.352 15 1.16 0.0015
137 9 0.64 0.016 10.110 15 0.67 0.0008
152 10 0.74 0.018 12.959 15 0.86 0.0011
Total/Average 8.73 0.029 286.569 150 1.91 0.0241
0 1 0.76 0.039 29.695 15 1.98 0.0025
17 2 0.75 0.032 23.920 15 1.59 0.0020
39 3 0.91 0.042 38.045 15 2.54 0.0032
55 4 0.87 0.020 17.667 15 1.18 0.0015
5,(01/25/02) 74 5 0.90 0.040 35.556 15 2.37 0.0030
98 6 0.73 0.029 21.138 15 1.41 0.0018
109 7 0.67 0.027 17.845 15 1.19 0.0015
125 8 0.77 0.020 15.530 15 1.04 0.0013
142 9 0.68 0.013 9.099 15 0.61 0.0008
157 10 0.69 0.025 17.479 15 1.17 0.0015
Total/Average 7.73 0.029 225.976 150 1.51 0.0188

 

55

 

Yellow variety

Table A5. Experimental data collected and results

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Volume Total Extraction g of
Extraction Time Vial per Cone. . Volume ratio (pg of quercetin!
4, (date) E(min.)1 # extract (pg/pl) ' “mm“ (C02) quercetin/L kg or onion
(ml) ("9 0fC02) peels
0 1 3.10 0.023 70.085 15 4.67 0.0080
18 2 1.10 0.018 20.277 15 1.35 0.0020
37 3 1.20 0.020 23.886 15 1.59 0.0030
52 4 1.15 0.017 19.251 17 1.13 0.0020
1, 68 5 0.78 0.031 24.166 15 1.61 0.0030
(02/07/02) 92 6 0.87 0.013 11.011 15 0.73 0.0010
106 7 0.83 0.010 8.207 15 0.55 0.0010
123 8 0.72 0.009 6.502 15 0.43 0.0010
140 9 0.63 0.007 4.516 15 0.30 0.0010
156 10 0.69 0.007 4.622 15 0.31 0.0010
Total/
Average 11.07 0.015 192.523 152 1.27 0.0220
0 1 1.60 0.025 40.647 15 2.71 0.0050
17 2 1.60 0.022 34.671 15 2.31 0.0040
35 3 1.30 0.015 19.688 15 1.31 0.0020
63 4 1.30 0.019 24.162 17 1.42 0.0030
2, 71 5 1.10 0.017 18.273 15 1.22 0.0020
(02/13/02) 90 6 1.00 0.019 18.663 15 1.24 0.0020
103 7 1.10 0.014 15.582 15 1.04 0.0020
119 8 1.00 0.013 13.356 15 0.89 0.0020
135 9 0.90 0.010 9.203 15 0.61 0.0010
151 10 0.90 0.008 7.203 15 0.48 0.0010
Total
/Average 11.80 0.016 201.448 152 1.32 0.0250

 

' Concentration of quercetin per ethanol volume per extract

56

 

Table A5. Continuation

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0 1 1.55 0.024 37.899 16 2.37 0.004

18 2 1.40 0.024 33.806 15 2.25 0.004

36 3 1.25 0.022 28.020 15 1.87 0.003

52 4 1.20 0.026 30.754 17 1.81 0.003

3’ (02/20/02) 70 5 1.10 0.020 21.621 15 1.44 0.002
89 6 1.00 0.021 21.336 15 1.42 0.002

108 7 0.90 0.016 14.344 15 0.96 0.001

127 8 0.80 0.014 1 1.473 15 0.76 0.001

143 9 0.90 0.018 16.643 15 1.1 1 0.002

159 10 0.80 0.01 1 9.180 15 0.61 0.001

T otal/A verage 10.90 0.020 225.076 153 1.46 0.023
0 1 1.00 0.020 20.084 15 1.34 0.002

20 2 0.77 0.009 6.657 15 0.44 0.001

41 3 0.68 0.026 17.382 15 1.16 0.002

57 4 0.77 0.010 7.338 17 0.43 0.001

4, (02/28/02) 81 5 0.61 0.018 1 1.118 15 0.74 0.001
99 6 0.73 0.043 31.367 15 2.09 0.004

1 14 7 0.72 0.030 21.841 15 E 1.46 0.003

132 8 0.68 0.023 15.735 15 1.05 0.002

148 9 0.68 0.018 12.084 15 0.81 0.001

164 10 0.69 0.008 5.586 15 0.37 0.001

Total/A verage 7.33 0.020 149.193 152 0.99 0.018
0 1 1.00 0.018 17.772 15 1.18 0.002

18 2 0.61 0.007 4.258 15 0.28 0.001

35 3 0.44 0.010 4.620 15 0.31 0.001

51 4 0.64 0.012 7.617 17 0.45 0.001

5, (03/05/02) 69 5 0.69 0.036 24.828 15 1.66 0.003
87 6 0.73 0.013 9.701 15 0.65 0.001

101 7 0.84 0.014 1 1.362 15 0.76 0.001

1 18 8 0.78 0.010 7.426 15 0.50 0.001

134 9 0.60 0.010 5.910 15 0.39 0.001

150 10 0.73 0.010 7.606 15 0.51 0.001

T otal/A verage 7.06 0.014 101.098 152 0.67 0.012

 

57

 

Pure quercetin

Table A6. Data collected and results fi'om static mode extraction of pure quercetin

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Volume . Total Extraction
Extraction #, Time Vial # per Couc. quercetin Volume ratio (pg of
(date) (mm.) extract (pg/pl) C02 (L)Equereetin/L
(m0 0‘" 0fC02)
0 1 0.50 0.015 7.37 15 0.49
17 2 0.90 0.025 22.79 15 1.52
35 3 1.10 0.019 21.05 18 1.17
50 4 0.75 0.020 14.95 15 1.00
1, (03/26/02) 64 5 0.48 0.020 9.68 15 0.65
. 82 6 0.67 0.013 8.49 15 0.57
96 7 0.57 0.012 6.85 15 0.46
110 8 0.62 0.011 6.64 15 0.44
125 9 0.68 0.009 5.87 15 0.39
139 10 0.60 0.007 4.11 15 0.27
Total/A verage 0.69 0.015 107.79 153 ‘ 0.70
0 1 0.50 0.046 22.92 15 1.53
16 2 0.76 0.093 70.85 15 4.72
35 3 0.56 0.077 43.05 15 2.87
51 4 0.60 0.122 73.14 15 4.88
2, (04/01/02) 70 5 0.71 0.110 78.41 15 5.23
88 6 0.58 0.114 65.96 15 4.40
102 7 0.65 0.118 76.92 15 5.13 .
118 8 0.60 0.139 83.16 15 5.54
135 9 0.60 0.101 60.64 15 4.04
152 10 0.62 0.122 75.38 15 5.03
Total/Average 0.62 0.104 650.43 150 4.34

 

' Concentration of quercetin per ethanol volume per extract

58

 

Table A6. Continuation

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0 1 0.86 0.028 24.35 15 1.62
19 2 0.95 0.031 29.55 15 1.97
37 3 0.44 0.034 14.99 15 1.00
53 4 0.44 0.035 15.20 15 1.01
3’ (04/03/02) 72 5 0.70 0.054 37.84 15 2.52
90 6 0.63 0.040 24.90 15 1.66
104 7 0.62 0.073 45.37 15 3.02
121 8 0.50 0.036 17.92 15 1.19
136 9 0.59 0.027 16.17 15 1.08
151 10 0.53 0.039 20.75 15 1.38
T otal/A verage 0.63 0.04 247.06 150 1.65
0 l 0.75 0.133 99.72 15 6.65
17 2 1.50 0.394 591.68 15 39.45
33 3 1.90 0.253 481.64 16 30.10
48 4 1.00 0.166 166.15 15 11.08
4,(04/10/02) 67 5 1.10 0.138 151.96 15 10.13
85 6 0.69 0.1 1 1 76.76 15 5.12
.99 7 0.75 0.202 151.39 15 10.09
120 8 0.65 0.087 56.48 . 15 3.77
130 9 0.57 0.05 28.68 15 1.91
146 10 0.65 0.043 28.03 15 1.87
T otal/Averget 0.96 0.158 1832.50 151 12.02
0 1 0.53 0.018 9.66 15 0.64
17 2 0.61 0.037 22.76 15 1.52
36 3 0.57 0.084 47.81 16 2.99
48 4 0.65 0.096 62.44 15 4.16
5, (04/17/02) 67 5 0.79 0.104 81.95 15 5.46
85 6 0.90 0.135 121.35 15 8.09
E 98 7 1.00 0.112 112.33 15 7.49
115 8 0.76 0.117 88.77 15 5.92
131 9 1.00 0.083 83.17 15 5.54
146 10 0.93 0.082 75.80 15 5.05
T otal/A verage 0.7 7 0.087 706.02 151 4.69

 

 

59

 

APPENDIX B. HPLC Chromatograms and others

Reduced Pressure

30
2

O

0.5

OJ ,

 

' I v '

 

 

 

 

 

Reduced Density

Figure B1. Giddings state curves. Giddings J. C. et al., 1968

60

 

 

Figure 32. Static mode method. Example of the chromatograms

61

 

 

‘ dorm-novne -v— v
r

Absorbance
/

”TE \\ ///\ EEK Standard /’/
; \\_/ \_. " J Eat: art E

,1

L...“

Vi , , t; v. ;VT-fif ' f. E4
200 220 240 200 280 3“ 320 340 360

Nanometers

w

Figure B3. UV spectrum of quercetin standard and extract from SF E

62

 

11750 IV

 

 

 

 

 

 

 

 

0000 IN
Ci
{4"

 

 

 

UIE

Figure B4. Dynamic mode method. Example of the chromatograms

63

 

151]] all
.__....... “E

 

 

E
1 E
It

 

“#6
o “—0
_- __ ,--_. 0'-

4......
' - _"_T"'__.m a .

 

 

 

 

 

 

_- Q...“ --

00111:?
1‘
E L
01R

Figure B5. Static mode for red onion variety. Example of the chromatograms

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

WE ”lElithetin
11
E
m 1
.. 1
3 é.
a. E' 1
. 1
E‘- i E
a. .4 EE. 1
E E '1
1t ‘ W--~E E1 01.....E EELWEE MEL 0.110 EEEEEEEEE‘EttwwE‘WW
281 310 310 alo
”(N/ll

Figure B6. Mass spectrometry of the standard sample

65

 

 

 

FF'H—Nn

ONO. 0N”

 

 

 

 

a? I:

 

8? 000

 

fl Po.‘0°

 

 

3500

3550

280

260

Mass (m/z)

Figure B7. Mass spectrometry of the SFE extract

 

 

37,00 IV

 

 

 

 

 

 

 

 

 

0,000 1111

Ann-AflA‘

 

 

 

 

5 6 -

Figure B8. Static mode for the yellow variety. Example of the chromatograms

67

 

15,125 IV

Quercet'n‘l. E

 

 

 

Eh..—
__ - O... r‘v—

 

 

 

 

Figure B9. Example of the chromatograms of the pure quercetin extract

68

APPENDIX C. Statistical Analyses

RED VARIETY

ANALYSIS C1: quercetin (1’) versus volume of ethanol (A? collected

 

Total
Volume quercetin
(ml) (pg) FROM MINITAB

 

0.934 53.447 The regression equation is
0.918 38.851 Y = - 50.7 + 96.7 X

1.010 39.904
0.876 24.826 Predictor Coef SE Coef T P
0.986 45.857 Constant -50.70 13.09 -3.87 0.005
0.738 21.401 X 96.67 15.72 6.15 0.000
0.740 22.354
0.718 16.846 8 = 6.490 R-Sq = 82.5% R-Sq(adj) = 80.4%
0.666 12.432
0.636 11.942 Analysis of Variance

 

 

 

 

 

 

 

 

 

 

 

 

 

Source DF SS MS F P
Regression 1 1592.9 1592.9 37.82 0.000
Residual Error 8 336.9 42.1

Total 9 1929.9

69

ANALYSIS C2: ethanol collected ()7 versus time (A?

 

 

 

 

Time
(min) Volume (ml)E FROM MINTTAB

0.00 0.93 The regression equation is
17.40 0.92 Y = 1.00 - 0.00237 X
36.40 1.01

 

51.00 0.88 Predictor Coef SE Coef T P
69.80 0.99 Constant 1 .00479 0.04284 23.46 0.000
89.60 0.74 X -0.0023719 0.0004723 -5.02 0.001

 

 

 

 

 

103.00 0.74
118.40 0.72 S = 0.07163 R-Sq = 75.9% R-Sq(adj) = 72.9%
134.40 0.67

 

 

 

 

149.80 0.64 Analysis of Variance

 

 

Somee DF SS MS F P
Regression 1 0.12940 0.12940 25.22 0.001
Residual Error 8 0.04105 0.00513

Total 9 0.17044

70

ANALYSIS C3: extraction ratio (Y) versus time (A?

 

Extraction
ratio (pg of
Time quercetin/L
(min) of C02) FROM MIN ITAB

 

 

 

0.00 3.56 The regression equation is
17.40 2.59 Y = 3.23 - 0.0171 X
36.40 2.66

 

51.00 1.65 Predictor Coef SE Coef T P

69.80 3.05 Constant 3.2331 0.2891 11.18 0.000
89.60 1.43 X -0.01 7071 0.003188 -5.36 0.001
103.00 1.49
118.40 1.12 S = 0.4835 R-Sq = 78.2% R-Sq(adj) = 75.5%
134.40 0.83
149.80 0.80 Analysis of Variance

 

 

 

 

 

 

 

 

 

 

Source DF SS MS F P
Regression 1 6.7028 6.7028 28.68 0.001
Residual Error 8 1.8698 0.2337

Total 9 8.5726

71

YELLOW VARIETY

ANALYSIS C4: quercetin (1’) versus ethanol volume collected (X)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Total
Volume quercetin
(ml) (pg) , FROM MINIT AB
1.650 37.297 The regression equation is
1.096 19.934 Y=-11.1+29.6X
0.974 18.719
1.012 17.824 Predictor Coef SE Coef T P
0.856 20.001 Constant -11.095 3.896 -2.85 0.022
0.866 18.416 X 29.570 3.913 7.56 0.000
0.878 14.267
0.796 10.898 S = 3.124 R-Sq = 87.7% R-Sq(adj) = 86.2%
0.742 9.671
0.762 6.839 Analysis of Variance
Source DF SS MS F P
Regression 557.32 557.32 57.12 0.000

1
Residual Error 8 78.06 9.76
Total 9 635.37

72

 

ANALYSIS C5: ethanol collected (1? versus time ()0

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Time
(min) Volume (ml) W

0.00 1.65 The regression equation is

18.20 1.10 Y=1.29-0.00414X

36.80 0.97

55.00 1.01 Predictor Coef SE Coef T P
71.80 0.86 Constant 1.29454 0.09667 13.39 0.000
91.40 0.87 X -0.004145 0.001026 -4.04 0.004
106.40 0.88
123.80 0.80 S = 0.1619 R-Sq = 67.1% R-Sq(adj) = 63.0%
140.00 0.74
156.00 0.76 Analysis of Variance

 

Source DF SS MS F P
Regression 1 0.42776 0.42776 16.32 0.004
Residual Error 8 0.20964 0.02620

Total 9 0.63739

73

 

ANALYSIS C6: extraction ratio (19 versus time (A?

 

 

 

 

 

 

 

 

 

 

 

 

Total ratio
(pg of

Time quechL of
(min.) C02)
0.0 2.454
18.2 1.326
36.8 1.248
55.0 1.048
71.8 1.334
91.4 1.226
106.4 0.954
123.8 0.726
140.0 0.644
156.0 0.456

 

 

 

FROM MINITAB

The regression equation is
Y = 1.86 - 0.00901 X

Predictor Coef SE Coef T P

Constant 1.8620 0.1797 10.36 0.000

X -0.009011 0.001907 -4.73 0.001

S = 0.3009 R-Sq = 73.6% R-Sq(adj) = 70.3%
Analysis of Variance

Source DF SS MS F P
Regression 1 2.0220 2.0220 22.34 0.001
Residual Error 8 0.7242 0.0905

Total 9 2.7462

74

COMPARISON BETWEEN VARIETIES

ANALYSIS C7: g of quercetin extracted/ kg of onion skins for both varieties

 

 

 

 

 

 

g of
quercetin] Eg of quercetin] kg
kg of onion of onion skins
skins LR“) (Yellow) W
0.027 0022 Analysis of Variance
0027 0-025 Source DF SS MS F P
0023 0-023 C2 1 0.0000400 0.0000400 2.13 0.182
0024 0-018 Error 8 0.0001500 0.0000188
0.019 0.012 Total 9 0.0001900

 

 

 

 

Level N Mean StDev
0 5 0.024000 0.003317
1 5 0.020000 0.005148

0 = red
1 = yellow

75

COMPARISON BETWEEN PRELIMINARY AND PRIMARY EXPERIMENTS

ANALYSIS C8: total quercetin obtained

 

 

 

 

 

 

 

 

 

Total quercetin (pg) ,
Preliminary Primary
71 .57 192.52
1 1.45 201.45
49.16 225.08
149.19
101 .1

 

 

FROM MINITAB
Analysis of Variance
Source DF SS MS F P

C5 1 31594 31594 16.51 0.007
Error 6 1 1482 1914
Total 7 43076

N Mean StDev
0 3 44.06 30.38
1 5 173.87 49.08

0=Pre1iminary
1=Primary

ANALYSIS C9; of quercetin per kg of onion skins obtained

 

 

 

 

 

 

 

 

 

g of quercetin/kg of onion skins
Preliminary Primary
0.006 0.022
0.001 0.025
0.005 0.023
0.01 8
0.012

 

 

 

FROM MINITAB
Analysis of Variance
Source DF SS MS F P
C8 1 0.00048 0.00048 24.00 0.003
Error 6 0.00012 0.00002
Total 7 0.0006
Level N Mean StDev
0 3 0.004 0.002646
1 5 0.02 0.005148
0 = Preliminary
1= Primary

76

ANALYSIS C10: volume of ethanol collected

 

 

 

 

 

 

 

 

 

Volume (m1) FROM MIN ITAB
Preliminary Primary Analysis of Variance
0.468 1.107 Source DF SS MS F P
0.234 1.180 C11 1 0.8181 0.8181 20.10 0.004
0.206 1.090 Error 6 0.2442 0.0407
0.733 Total 7 1.0623
0.706

 

 

 

Level N Mean StDev
0 3 0.3026 0.1437
1 5 0.9632 0.2252

0 = Preliminary experiments
1 = Experiments

ANALYSIS C11: extraction ratio

 

 

 

 

 

 

 

 

 

Extraction ratio (pg of
quercetin/L of C02) FROM MIN [TAB
Preliminary Primary Analysis of Variance '

0.57 1.27 Source DF SS MS F P

0.19 1.32 C2 1 1.0603 1.0603 13.60 0.010

0.41 1.46 Error 6 0.4679 0.0780
0.99 Total 7 1.5282
0.67

 

 

 

Level N Mean StDev
0 3 0.3900 0.1908
1 5 1.1420 0.3 143

' 0=Pre1iminary
1=Primary

77

 

COMPARISON BETWEEN SF AND SOLID-LIQUID EXTRACTION

ANALYSIS C12: SF E and methanol for red variety

 

 

 

 

 

 

 

 

g of quercetin/kg of onion
skins

SFE Methanol
0.027 10.14
0.027 15.90
0.023 8.73
0.024 10.82
0.019 8.33

 

 

 

EROM MINITAB

Level N Mean StDev
0 5 0.024 0.003
1 5 10.784 3.034
0 = SFE

1= Methanol

ANALYSIS C13: SFE and methanol for yellow variety

 

 

 

 

 

 

 

 

g of quercetin/kg of
onion skins
SFE Methanol
0.022 3.80
0.025 10.41
0.023 3.07
0.018 3.48
0.012 7.62

 

 

 

FROM MINITAB

Level N Mean StDev
0 5 0.020 0.005
1 5 5.676 3.214
0 = SFE

1= Methanol

78

PURE QUERCETIN

 

ANALYSIS C14:guercetin extractedm versus ethanol collected (X)

 

 

 

 

 

 

 

 

 

 

 

Total
Volume quercetin

(ml) (its)
0.628 32.804
0.944 147.526
0.914 121.708
0.688 66.376
0.756 71.968
0.694 59.492
0.718 78.572
0.626 50.594
0.688 38.906
0.666 40.814

 

 

 

FROM MINITAB

The regression equation is

Y = - 165 + 322 X
Predictor Coef SE Coef T P
Constant -164.55 25.18 -6.53 0.000
X 321.54 34.04 9.44 0.000
S = 11.32 R-Sq = 91.8% R-Sq(adj) = 90.7%
Analysis of Variance
Source DF SS MS F P

Regression 1 11431 1 1431
Residual Error 8 1025 128 .
Total 9 12456

89.20 0.000

79

 

ANALYSIS C15: ethanol collected (1? versus time ()0

 

 

 

 

 

 

 

 

 

 

 

 

Time Volume
(min-l (ml)
0.0 0.628
17.2 0.944
35.2 0.914
50.0 0.688
68.0 0.756
86.0 0.694
99.8 0.71 8
1 16.8 0.626
131.4 0.688
146.8 0.666

 

 

 

FROM MINITAB
The regression equation is
Y = 0.811 - 0.00105 X

Predictor Coef SE Coef T P
Constant 0.81 120 0.06193 13.10 0.000
X -0.0010517 0.0006991 -1.50 0.171

S = 0.1038 R-Sq = 22.1% R—Sq(adj) = 12.3%
Analysis of Variance
Source DF SS MS F P

Regression 1 0.02438 0.02438 2.26 0.171
Residual Error 8 0.08618 0.01077

Total 9 0.11056

80

 

ANALYSIS C16: extraction ratio (1’) versus time ()0

 

 

 

 

 

 

 

 

 

 

 

 

Extraction
ratio (pg ofE
Time quercetin/L
Quin.) of C02)
0.0 2.186
17.2 9.836
35.2 7.626
50.0 4.426
68.0 4.798
86.0 3.968
99.8 5.238
116.8 3.372
131.4 2.592
146.8 2.720

 

 

 

FROM MINITAB
The regression equation is
Y = 6.48 - 0.0240 X

Predictor Coef SE Coef T P

Constant 6.482 1.327 4.89 0.001

X -0.02404 0.01498 -1.60 0.147

S = 2.224 R-Sq = 24.4% R-Sq(adj) = 14.9%
Analysis of Variance

Source DF SS MS F P
Regression 1 12.736 12.736 2.58 0.147
Residual Error 8 39.556 4.945

Total 9 52.292

ANALYSIS C I 7: comparison of extraction ratios fiom inert and natural matrices

 

 

 

 

 

 

 

lExtraction ratio (pg
of quercetin/L of
C02

Pure Yellow

quercetin variety
0.7 1.27
4.34 1.32
1.65 1.46
12.02 0.99
4.69 0.67

 

 

 

 

FROM MINITAB

Analysis of Variance

Source DF SS MS F P
C14 1 31.29 31.29 3.15 0.114
Error 8 79.41 9.93

Total 9 110.70

Level N Mean StDev
0 5 4.680 4.444

1 5 1.142 0.314

0 = Pure quercetin

1 = Yellow variety

81

ANALYSIS C18: quercetin collected fi'om the natural matrix versus inert matrix

 

 

 

 

 

 

 

 

g of quercetin/kg of
matrix
Natural Inert
0.022 0.002
0.025 0.013
0.023 0.005
0.018 0.036
0.012 0.014

 

 

 

W

Analysis of Variance

Source DF SS MS F P
C2 1 0.000090 0.000090 0.88 0.375

Error 8 0.000816 0.000102
Total 9 0.000906

Level N Mean StDev
0 5 0.02000 0.00515
1 5 0.01400 0.01332

 

0=natural ..
1=inert

82

REFERENCES

Bilyk A., Cooper P. L., Sapers G. M. (1984). Varietal Diflerences in Distribution
of Quercetin and Kaempferol in Onion (Allium cepa L.) Tissue. J. Agric. Food Chem.,
32, 274-276.

Caltagirone S. (2000). Flavonoids apigenin and quercetin inhibit melanoma
growth and metastatic potential. Alternative Medicine Review.
http://www.findgrticles.com/cf10/m0FDN/6 5/68727260/p1/article.jhtml?termfluercetin

Giddings J. C., Myers M. N., McLaren L., Keller R. A. (1968). High Pressure
Gas Chromatography on Nonvolatile Species. Science, 162 (67), 67-73.

Hawthorne S. B. (1990). Analytical-Scale Supercritical Fluid Extraction.
Analytical Chemistry, 62 (11), 633-642.

Hayashi A. (2000). Efl'ects of Daily Oral Administration of Quercetin Chalcone
and Modified Citrus Pectin on Implanted C olon-25 Tumor Growth in Balb-c Mice.
Alternative Medicine Review.
http://www.i'lndgrticlescom/cf 0/m0FDN/6i5/68727252/print.ihtml

Herrrnann K. (1976). Flavonols and flavones in food plants: a review. J. Food
Technology, 11, 433-448. E

Karger B. L., Snyder L. R., Horvath C. (1973). Separation Equilibria. An
Introduction to Separation Science. John Wiley & Sons. New York, NY.

King J. W. (1989). Fundamentals and Applications of Supercritical Fluid
Extraction in Chromatographic Science. Journal of Chromatographic Science, 27, 355-

364.

83

 

King J. W. (1990). Quantitative Correlations Between Solute Molecular Structure
and Solubility in Supercritical Fluids. Journal of Chromatography, 517, 449-458.

Mukhopadhyay M. (2000). Natural Extracts using Supercritical Carbon Dioxide.
CRC Press LLC. Boca Raton, Florida.

Merken H. M. (2000). Measurement of Food F lavonoids by High-Performance
Liquid Chromatography: A Review. J. Agric. Food Chem, 48 (3), 577-599.

Palma M., Taylor L. T. (1999). Statistical Design for Optimization of Extraction
of Polyphenols fiom an Inert Matrix using Carbon Dioxide-based Fluids. Analytica
Chimica Acta, 391, 321-329.

Palma M., Taylor L. T., Zoecklein B. W., Douglas L. S. (2000). Supercritical
Fluid Extraction of Grape Glycosides. J. Agric. Food Chem, 48, 775-779.

Palmer M. V., Ting S. S. T. (1995). Applications of supercritical fluid technology
in food processing. Food Chemistry, 52, 345-352.

Pascual-Marti M. C., Salvador A., Chafer A., Berna A. (2001). Supercritical
Fluid Extraction of Resveratrol fiom Grape Skin of Vitis Vinifiera and Determination by
HPLC. Talanta, 54, 735-740.

Price K. R., Rhodes M. J. C. (1997). Analysis of the Major Flavonol Glycosides
Present in Four Varieties of Onion (Allium cepa) and Changes in Composition Resulting
fiom Autolysis. J. Sci. Food Agric., 74, 331-339.

Ro- N. L., Singh R. K. (2002). Supercritical Fluids and the Food Industry.

Comprehensive Reviews in Food Science and Food Safety, 1, 33-44.

84

 

Schoenmakers P. J., Billiet H. A. H., Galan L. de. (1982). The Solubility
Parameter as a Tool in Understanding Liquid Chromatography. Chromatographia, 15,
205-214.

Taylor L. T. (1995). Strategies for Analytical SFE. Analytical Chemistry, A, 364-
370.

Tena M. T., Rios A., Valcarcel M. (1998). Supercritical Fluid Extraction of T-
Resveratrol and Other Phenolicsfiom a Spiked Solid. Frenesius J. Anal. Chem, 361 , 143-
148.

Xing N., Chen Y., Mitchell S. H., Young C. Y. F. (2001). Quercetin Inhibits the
Expression and Function of the Androgen Receptor in LNCaP Prostate Cancer Cells.

Carcinogenesis, 22 (3), 409-414.

General references

Barton A. F. M. (1975). Solubility Parameters. Chemical Reviews, 75 (6), 731-
753.

Dron A., Guyer D. E., Gage D. A., Lira C. T. (1997). Yield and Quality of Onion
Flavor Oil Obtained by Supercritical Fluid Extraction and Other Methods. Journal of
Food Process Engineering, 20, 107-124.

Fedors R. F. (1974). A Method for Estimating Both the Solubility Parameters and
Molar Volumes of Liquids. Polymer Engineering and Science, 14 (2), 147-154.

Giddings J. C., Myers M. N., King J. W. (1969). Dense Gas Chromatography at

Pressures to 2000 Atmospheres. Journal of Chromatographic Science, 7, 276-283.

85

Ems

 

[ 3"
.1 i

 

 

 

Halddnen S. H., Karenlampi S. 0., Heinonen I. M., Mykkanen H. M., Torronen
A. R. (1998). HPLC Method for Screening of Flavonoids and Phenolic Acids in Berries.
J. Sci. Food Agric., 77, 543-551. ‘

Hirota S., Shimoda T., Takahama U. (1998). Tissue and Spatial Distribution of
F lavonol and Peroxidase in Onion Bulbs and Stability of Flavonol Glucosides during
Boiling of the Scales. J. Agric. Food Chem, 46, 3497-3502.

Karastogiannidou C. (1999). Efi'ects of Onion Quercetin on Oxidative Stability of
Cook-chill Chicken in Vacuum-sealed Containers. Journal of Food Science, 64 (6), 978-
981.

McDonald M. S., Hughes M., Burns J., Lean M. E. J., Matthews D., Crozier A.
(1998). Survey of the Free and Conjugated Myricetin and Quercetin Content of Red
Wines of Dtflerent Geographical Origins. J. Agric. Food Chem., 46, 368-3 75.

McHugh M., Krukovis V. (1994). Supercritical Fluid Extraction: Principles and
Pratices. Second Edition. Butterworth-Heinemann. Stoneham, MA.

Nuss J ., Guyer D. E., Gage D. A. (1997). Concentration of Onion Juice Volatiles
by Reverse Osmosis and its Eflects on Supercritical CO; Extraction. Journal of Food
Process Engineering, 20, 125-139.

Ollanketo M., Hartonen K., Riekkola M-L., Holm Y., Hiltunen R. (2001).
Supercritical Carbon Dioxide Extraction of Lycopene in tomato skins. Eur. Food Res.
Technol., 212, 561-565.

Reid R. C., Prausnitz J. M., Poling B. E. (1987). The Properties of Gases and

Liquids. Fourth Edition. McGraw-Hill. New York, NY.

86

 

Saengcharoenrat C. (1998). Optimization of Onion Flavor Recovery in the
Process of Supercritical Carbon Dioxide Extraction and Desorption. Doctor of
Philosophy’s thesis. Michigan State Universtiy. East Lansing. MI.

Simandi B., Sass-Kiss A., Czukor B., Deak A., Prechl A., Csordas A., Sawinsky
J. (2000). Pilot-scale Extraction and Fractional Separation of Onion Oleoresin using
Supercritical Carbon Dioxide. Journal of Food Engineering, 46, 183-188.

Taylor L. T. (1996). Supercritical Fluid Extraction. John Wiley & Sons, Inc. New
York, NY.

Westwood S. A. (1993). Supercritical Fluid Extraction and its use in

Chromatographic Sample Preparation. CRC Press, Inc. Boca Raton, FL.

87

MICHDGAN STATE UNIVERSITY LIBRARIES

1 1‘ ’1 ‘ "‘ll‘ lllll

,lllwhiltllllllllMl JIM ,x, mi! N
3 1293 024 053

 

7O 0