Document Text (Pages 101-110) Back to Document

Application of ultra high hydrostatic pressure for investigating the binding of flavor compounds to ß-lactoglobulin via headspace solid phase microextraction-gas chromatography

by 1977- Hoang, Tinyee Arden, PhD

Page 101


A headspace solid phase microextraction (HS-SPME) method was developed for
the detection of selected flavor compounds, diacetyl, δ-decalactone, 2-
methylbutyraldehyde, and ethyl lactate, in sodium phosphate buffer (0.01M, pH 9). The
method required a total analysis time of 1.7 h for both extraction (10 min) and detection.
PDMS/DVB and CAR/PDMS SPME fibers were used to examine their extraction
efficiencies for the flavor compounds tested. Extraction time was investigated to
determine the analytical performance of these fibers for the selected flavor compounds.


The calibration plots were reproducible and linear (RP


> 0.954) for the selected flavor
compounds with the PDMS/DVB fiber tested. δ-Decalactone was not detected when
using the CAR/PDMS fiber due to the long chain length and presence of a hydrocarbon
tail of δ-decalactone under the current GC conditions.


Page 102

The extraction of flavor compounds from food involves concentrating the flavor
compounds using headspace, purge and trap, liquid-liquid extraction, solid phase
extraction, or simultaneous distillation/extraction techniques (Cai and others 2001;
Cullere and others 2003; Apps and Tock 2005; Blythe and others 2006), followed by
quantitation. The challenges of flavor compound analysis include extensive preparation
time or excessive use of organic solvents. Static headspace using a gas-tight syringe and
headspace solid-phase microextraction (SPME) are simple techniques, which are portable
and inexpensive, that eliminate these drawbacks (Saraulio and others 1997). SPME
requires no solvents or complicated apparatus that may introduce errors, can be used to
concentrate volatile and nonvolatile compounds in liquid or vapor phases, provides linear
results over wide concentrations of analytes, and can be coupled with any gas
chromatograph or gas chromatograph-mass spectrometer system for the identification and
quantitation of volatile compounds (Pawliszyn 2001).
With headspace (HS)-SPME, equilibrium involves the partitioning of analytes
between the aqueous phase of sample, the vapor phase and the extraction phase of the
SPME fiber. Extraction conditions must therefore be systematically optimized to increase
the partitioning of analytes in the coated fiber. Besides extraction conditions and analyte
properties, the choice of fiber coating is one of the most important aspects of
optimization, because the coating is dependent on analyte properties. Several SPME fiber
coatings are commercially available for extraction of nonpolar and polar compounds
from liquid or gaseous samples. Pawliszyn (2001) suggested considering
polydimethylsiloxane (PDMS) first, because of its ability to withstand high injection

Page 103

temperatures. Although PDMS is a nonpolar liquid phase, it can successfully be applied
to extracting more polar compounds after optimizing extraction conditions (Pawliszyn
2001). Factors that improve the extraction of analytes by SPME are agitation, addition of
salt to the analyte solution, sample volume, headspace volume, pH or temperature of the
analyte solution, or sampling mode (immersion vs. headspace).
Gas chromatography (GC) is an important tool in analytical chemistry. Capillary
columns increase separating capabilities of GC so that complex compounds such as
flavors can be successfully resolved from solutions containing hundreds of components
(Braithwaite and Smith 1999). The factors that influence the retention of flavor
compounds are composition and properties of the mobile phase; type and properties of
the stationary phase; the intermolecular forces between the compound(s) and stationary
and mobile phases; and temperature of the column. Electrostatic interactions between
molecules include polar van der Waal’s retention forces arising from interaction between
molecules having a surface charge; and nonpolar dispersion forces between neutral
molecules or functional groups. Polar van der Waal’s retention forces are dipole-dipole
interactions and hydrogen bonding between molecules. Hydrogen bonded solvents will
attract polar solute molecules but will exhibit varying degrees of repulsion to nonpolar
molecules. Thus, solute molecules will be attracted towards the phase of similar polarity
in the GC system. In nonpolar solvents or stationary phase, London’s dispersion forces
are the main interactions between molecules. London dispersion forces are weak
intermolecular forces that arise from the attractive force between transient multipoles of
nonpolar molecules without permanent multipole moments. Multipole moments are
created when the electron densities are not evenly distributed throughout the nonpolar

Page 104

molecule. Therefore, multipoles of one nonpolar molecule may interact with multipoles
of other nonpolar molecules (French 2000).
Diacetyl, 2-methylbutyraldehyde, ethyl lactate, and δ-decalactone were selected
based on the flavor profiles of foods perceived to contain or are rich, such as dairy
products (Badings and Neeter 1980; Milo and Reineccius 1997), chocolate, or creambased
Diacetyl, or 2,3-butanedione, is one of the major flavor compounds in dairy
products and wine. Diacetyl contributes a characteristic buttery aroma and complexity to
the final sensory impact of food. Diacetyl is naturally produced upon the conversion of
lactose to lactic acid by lactic acid bacteria in cultured milks and fresh cheeses (Urbach
1995). After alcoholic fermentation ceases when fermentative yeasts die, malolactic
fermentation continues, either naturally or by inoculation, in wines and produces diacetyl
among other aroma compounds (Hayasaka and Bartowsky 1999). Diacetyl is an
important flavor compound in many dairy products and wine, yet is not studied or
extensively subjected to quantitative analyses due to analytical difficulties in the
quantitation from its high volatility. Diaz and others (2004) evaluated fiber selection,
extraction time, temperature and ionic strength for headspace SPME analysis of diacetyl
from water samples of the Llobregat River in Spain. They determined that extraction of
diacetyl for 30 min with a CAR/PDMS fiber and polar column at 60 °C resulted good


linearity (rP


> 0.999) and good reproducibility (R.S.D. < 20%).
2-Methylbutyraldehyde (2MB) is a Strecker aldehyde that contributes a
characteristic cocoa aroma. 2MB is formed from the deamination and decarboxylation of
branched chain amino acid isoleucine by brewing yeast, and can be reduced

Page 105

enzymatically to ‘fusel’ alcohols (Ford and Ellis 2002). The ‘fusel’ alcohols and their
subsequent esters are important in the development of flavor during the brewing process.
Foese and others (1999) were able to detec 2-methylbutyraldehyde quantitatively in
disinfected water by purging water samples with N2 and trapping on Tenax-TA ™
adsorbent. The adsorbent was thermally desorbed by heating to 200 C, transferring the
analyte into a nonpolar GC column. Guillen and Ibargoitia (1999) used dichloromethane,
an organic solvent, to extract liquid smoke components, such as 2-methylbutyraldehyde,
and concentrated the components by evaporation prior to direct injection into a GC. Little
research is available on 2-methybutyraldehyde as a flavor ingredient in foods.
Ethyl lactate (EL) contributes a characteristic fruity, buttery and butterscotch
aroma. EL is used in yeast fermentations to produce roasted aromas that contribute to the
desirable flavor of thermally processed foods such as meat and bread (Rhlid and others
2002). Ethyl lactate was identified as a flavor compound in fresh coconut sap via direct
injection of a solvent concentrate into a GC-MS (Borse and others 2006). According to
the official method of the Office International de la Vigne et du Vin (OIV) (1994), ethyl
lactate is quantitatively determined using gas chromatography with a polar capillary
column. Soufleros and others (1994) modified the mthod by OIV in which analytes were
extracted by a mixture of solvents, mixed with an internal standard, and injected directly
into a GC with a polar capillary column.
Although fruits such as peaches, apricots and nectarines contain the pleasant
aromas of δ-lactones (Tamura and others 2005), δ-decalactone (dDL) imparts a creamy
and coconut note to foods. Guillot and others (2006) used headspace-SPME-GC to
quantitate various flavor compounds, including γ-decalactone, from apricot puree. They

Page 106

reported that γ-decalactone was extracted in more quantities by a PDMS fiber than by
Carboxen/PDMS fiber when using a polar (DB-WAX) capillary column. Tamura and
others (2005) assessed the authenticity of δ-decalactone from peach, apricot, and
nectarine fruits by direct injection of analytes containing the flavor compound into a DB-
WAX column using a temperature program of 5-200 °C at a rate of 5 °C/min.
The objective of this research was to develop one method that can accommodate
the extraction and quantitation of diacetyl, ethyl lactate, δ-decalactone, and 2-
methylbutyraldehyde, separately and together, using HS-SPME coupled with gas
Diacetyl, 2-methylbutyraldehyde, δ-decalactone, and ethyl lactate were purchased
from Sigma-Aldrich (St. Louis, MO, USA ). The SPME fibers were purchased from
Supelco (Bellefonte, PA, USA). Polydimethylsiloxane-divinylbenzene (PDMS/DVB) of
65 µm thickness and carboxen- polydimethylsiloxane (CAR/PDMS) of 75 µm thickness
SPME fibers were selected based upon their affinity for the flavor. Fibers were
conditioned prior to use according to the manufacturer’s instructions: PDMS/DVB was
inserted into the GC injector at 250 °C for 0.5 h and CAR/PDMS at 300°C for 1-1.5 h.
SPME Procedure
Individual standard stock solutions of selected flavor compounds were prepared at
a concentration of 1 µg/mL in 10% ethanol and water to solubilize the flavor compounds.
An aliquot of each standard stock solution was mixed with 2 mL of sodium phosphate
buffer (0.01 M, pH 9.0) and 0.65 g NaCl to selected concentrations (0.001-0.01 µg/mL)
in a 4 mL amber vial sealed with a silicone septum.

Page 107

An amber vial containing buffer, individual or combined flavor compounds, and
salt was placed on a magnetic hot stir plate (model PC-220, Corning, NY) and stirred at
400 rpm for 5 min to pre-equilibrate volatiles for each headspace (HS)-SPME analysis. A
pre-equilibration temperature of 37°C was chosen to mimic the temperature of the mouth
when flavor compounds are released during eating. The buffer containing the flavor
compounds and salt continue to stir during adsorption. The SPME needle pierced the
septum and the fiber was extended through the needle to bring the stationary phase in
contact with the assay headspace above selected concentrations of flavor solutions. The
fiber was withdrawn into the needle after an exposure time from 1 to 60 min. The
volatiles adsorbed to the SPME fiber were thermally desorbed into the injection port of a
Hewlett-Packard HP 5890A (Agilent Technologies, Palo Alto, CA, USA) gas
chromatograph connected to a FID detector in splitless mode. A DB-WAX column (60 m
x 0.32 mm i.d., 0.5 um film thickness, J&W Scientific, Folsom, CA) was chosen
according to methods by Diaz and others (2004), OIV (1994), Guillot and others (2006),
and Tamura and others (2005). Helium was used as the carrier gas. The injector
temperature was 250°C, according to the method of Adahchour and others (2005). The
oven temperature program was adopted from Tamura (2005) with a change at which the
oven was maintained at 33 °C, for 5 min before increasing to 50 °C, at a rate of 2 °C/min,
then to 200 °C, at a rate of 5 °C/min, and held for 37 min.
A standard curve, to assure a linear response of HS-SPME, was prepared by
adding selected flavor compounds in 10% ethanol to amber vials containing 2 mL sodium
phosphate buffer (0.01M, pH 9.0) and 0.65 g NaCl to yield solution concentrations from
4.02 µg/mL to 20.62 µg/mL of the flavor compounds. Volatile compounds in the vial

Page 108

headspace were extracted using HS-SPME from 1 to 60 min at 37°C, and a calibration
curve was prepared by plotting peak area against solution concentration of specific flavor
The characterization of chemical compounds using absolute gas chromatographic
retention times, or even specific retention volumes, is problematic, due to changes in
most of the critical gas chromatographic parameters for a given column that can occur
gradually with time. Thus, marked differences often exist between one chromatographic
system and the next (Peppard and Ramus 1988). Consequently the Kovats’ retention
index (KI) overcomes such problems. The KI relates the retention of compounds during
isothermal GC to the retention of a series of n-alkanes analyzed under identical
conditions. Each compound is thus bracketed by two hydrocarbons, both of which are
assigned a retention index value corresponding to the number of carbon atoms in the
molecule multiplied by 100. The KI for each flavor compound for the selected
temperature GC program was based on the following equation:

KI t
= 100 *



R( y)
R( y)


where tBRB is retention time, x is the solute of interest (i.e. flavor compound), y is the
normal alkane with y number of carbon atoms eluting before x, and z is the normal alkane
with z number of carbon atoms eluting after x.
Statistical Analysis
Data were analyzed by using one-way analysis of variance (ANOVA) procedure
using the General Linear Model, with further analysis using Tukey’s pairwise comparison
test to determine significance at p < 0.05 (SAS for Windows, version 9.1 TS Level, 1M2,
SAS Institute., Cary, NC, USA). 93

Page 109

Extraction Time
The extraction time profiles of the PDMS/DVB and CAR/PDMS SPME fibers
were established by plotting detector response against extraction time. The equilibration
time is reached when a further increase in extraction time does not result in a significant
increase in detector response. Figures 1 and 2 present the time profiles for the
PDMS/DVB and CAR/PDMS fibers, respectively. Different extraction time profiles were
observed for the analytes on the tested coatings. The signals obtained increased with
increasing extraction time for all analytes; equilibration time was, however, different for
each coating. The PDMS/DVB fiber is a made of a blend of porous divinylbenzene
polymer particles with a liquid polydimethylsiloxane polymer. The PDMS polymer has a
high degree of porosity (1.5 mL/g) that can physically retain analytes, producing a strong
retention of analytes that fit tightly into the pores. DVB is mesoporous with micropores
that are fairly larger than Carboxen particles, and is ideal for trapping CB6B-CB15B analytes
(Pawliszyn 1999). The combination of PDMS and DVB has better affinity for polar
analytes. CAR/PDMS, on the other hand, is ideal for the analyses of molecules in the CB2B-

CB12B range, and has unique pores that pass completely through the particles so that small

analytes can be rapidly desorbed (Pawliszyn 1999).
The equilibration time for diacetyl, δ-decalactone, and ethyl lactate was 10 min,
and for 2-methylbutyraldehyde was 20 min for the PDMS/DVB fiber. The semipolar
nature of PDMS/DVB allows for the extraction of the selected flavor compounds,
including dDL, which has a polar head and a 5-carbon tail. Higher uptake of diacetyl,
ethyl lactate, and 2-methylbutyraldehyde was observed for the CAR/PDMS fiber,

Page 110

because the polar fiber has a higher affinity for the low molecular mass compounds, and
Diaz and others (2004) have reported that the fiber presented greater efficiency than other
polar fibers, such as CAR-DVB-PDMS. This observation indicated that, though dDL is
within the detectable limits of CAR/PDMS, dDL is not detectable under the GC
conditions set in the method, because of its chain length and polarity (Figure 2).
For quantitative analysis, it is not necessary for the analytes to reach equilibrium,
as long as the extractions are carefully timed and the mixing conditions and volumes
remain constant (Ai 1997). An extraction time of 10 min was selected to keep the total
extraction time comparable with the chromatographic run time. The equilibrium
conditions for the selected flavor compounds reached after about 20 min, and the slope of
the curve is small. Therefore, a 1-min deviation in the exposure of the fiber results in
about the same absolute error independent of the extraction time (Pawliszyn 1997). The
PDMS/DVB fiber was chosen over the CAR/PDMS fiber, because of higher recovery of
diacetyl, ethyl lactate, 2-methylbutyraldehyde, and δ-decalactone and decreased
extraction time.
Analytical Characteristics
To check the performance of the procedure, SPME was applied to the analysis of
sodium phosphate buffer solutions containing known concentrations of flavor compounds
using the PDMS/DVB fiber. The linearity of the method was investigated by determining
calibration plots over the concentration range 5-20 µg/mL. Series of three concentrations
were obtained by spiking sodium phosphate buffer with the flavor compounds to generate
calibration plots. Each solution was extracted with the PDMS/DVB fiber, in replicates of
two, and analysis was performed by GC-FID. The line of best fit for the relationship

© 2009 All Rights Reserved.