???????????????????????—Volume 64, No. 3, 1999
Volatilities of Short-Chain Fatty Acids
in a Fermented Milk Model System
as Affected by Stabilizers
L. Chen, E. Boyle-Roden, and S. A. Rankin
Authors Chen and Boyle-Roden are affiliated with the Dept. of Nutrition & Food
Science, Univ. of Maryland, College Park, MD 20742. Author Rankin is affili-
ated with the Dept. of Animal & Avian Sciences, Univ. of Maryland, College Park
MD 20742-2311. Direct inquiries to Dr. S. A. Rankin.
A fermented milk model system was employed to investi-
gate the volatilities of butyric, caprylic, and caproic acids as
affected by low, intermediate, and high levels of carrageenan,
guar and xanthan gums. Carrageenan exhibited no effect on
fatty acid volatility. Intermediate levels of guar gum increased
the volatility of caprylic acid. Low and high levels of guar
gum had no effect on free fatty acid volatility. At low and
high levels, xanthan gum had no effect on volatilities of fatty
acids. The intermediate level of xanthan gum increased vola-
tilities of caproic and caprylic acids. The effects of viscosity
were potentiated with increased fatty acid chain length.
Key Words: aroma, dairy, fatty acids, chromatography,
THE DEVELOPMENT OF REDUCED-FAT PRODUCTS FREQUENTLY
involves the replacement of fat with hydrocolloid stabilizers to
achieve desirable sensory characteristics. Although such stabilizers
may mimic certain characteristics of lipids, studies have shown that
hydrocolloid stabilizers can bind or inhibit the release of aroma-ac-
tive compounds (Voilley et al., 1990; Rankin and Bodyfelt, 1996).
Nonvolatile food components such as proteins and carbohydrates
may alter the volatility of flavor compounds. Casein was observed
to strongly bind esters (Voilley et al, 1990), acetate, and acetone
(Thanh et al., 1992). Hansen and Heinis (1991) reported that the
presence of casein decreased the sensory perception of vanilla. Cer-
tain carbohydrates were found to bind alcohols, aldehydes, esters
(Thanh et al., 1992) and diacetyl (Rankin and Bodyfelt, 1996). Inter-
actions between volatiles and carbohydrates may occur via ionic in-
teractions, hydrogen bonding, hydrophobic effects, formation of in-
clusion complexes, and impedance of molecular mobility due to in-
creased viscosity (Voilley et al., 1990).
Desirable aroma in fermented dairy products is directly related to
the presence and availability of aroma compounds including volatile
free fatty acids (VFA). Butyric, caproic and capric acids were identi-
fied as potent aroma volatiles in 3-yr-old Cheddar cheese (Chris-
tensen and Reineccius, 1995). VFA’s also impart the rancid flavor
characteristic of dry hard Italian cheeses (Eaton, 1994). Although
many compounds and classes of compounds have been evaluated
for interaction and binding to stabilizer gums, no research has been
published on the potential for VFA’s to be affected by such interac-
tions. Our preliminary research demonstrated that hydrocolloid sta-
bilizers can influence the volatility of VFA’s in aqueous media (Chen,
1997). The objective of this study was to further characterize the
effects of hydrocolloid type, level, and viscosity on volatilities of
VFA’s in a fermented milk model system under dynamic headspace
MATERIALS & METHODS
Butyric, caproic, caprylic, and valeric acids, carrageenan (type 1
commercial grade), xanthan (practical grade) and guar gums were
purchased from Sigma Chemical Co. (St. Louis, MO). GC/MS-grade
eluting solvent, CS2, was obtained from Sigma-Aldrich (St. Louis,
Dynamic headspace analysis
Model systems sampled under dynamic conditions were prepared
using carrageenan, and xanthan and guar gums at concentrations of
0.025%, 0.1% and 0.4% (wt/wt) by dispersing dry hydrocolloid into
commercially pasteurized and homogenized skim milk under con-
stant agitation using a magnetic stir bar in 250-mL glass vials with
teflon-lined screw-top caps. Skim milk with no added stabilizer was
used as a control. All experimental units were vat-pasteurized at 70?C
for 10 min in a circulating water bath. Samples were allowed to cool
at room temperature to ~50?C at which point a 300-?g aliquot con-
taining butyric, caproic, and capric acids was added. The final con-
centration of each VFA in the model system solution was 1.00 mg/g.
Each experimental unit was stored overnight at 4?C to complete sta-
bilizer hydration and to allow VFA binding to reach equilibrium.
The final pH was 5.2 ?0.1.
A 10-g sample was transferred to a 20-mL round-bottom flask with
a universal inlet adapter. Each flask was immersed in a 50?C circulat-
ing water bath and purged for 15 min with nitrogen at 800 mL/min.
Effluent analytes were trapped using a Carbotrap solvent desorption
tube (20/40 mesh, ORBO 101, Supelco, Inc., Bellefonte, PA). A sub-
mersible stirrer with teflon starburst stirring head was used to provide
sample agitation. Teflon tubing was used for all connections follow-
ing the gas filters. A new trap was used for each analysis.
Following sample purge, traps were immediately desorbed by
placing the trap adsorbent bed into a 2-mL vial and eluting with 150
?L HPLC grade CS2 containing 40 ?g/g valeric acid.
Gas chromatographic analysis was conducted on a Hewlett-Pack-
ard 6890 GC with a flame ionization detector. A relatively short fused
silica capillary column (Rtx-Wax, 10m, 0.53 mm i.d., 1.0 µm film
thickness, Restek Corporation, Bellefonte, PA) was employed to ef-
fect adequate resolution of the underivatized fatty acids in ~15 min.
Chromatographic parameters were as follows: injector and detector
temperatures, 220?C; oven temperatures: initial 100?C, hold 3 min;
ramp 10?C/min; final 180?C; hold 2 min. Helium carrier gas was
used with a column flow rate of 2.2 mL/min. A 2-?L sample was
injected in splitless mode. Relative analyte quantities were deter-
mined using valeric acid as an external standard.
Viscosity was determined using a Brookfield Model RV rotary
Volume 64, No. 3, 1999—???????????????????????
viscometer (Brookfield Engineering Laboratories, Inc. Stoughton,
MA). Randomly selected samples were analyzed at 4?C. For statisti-
cal data evaluation, viscosity data were normalized using log10 trans-
Experimental design and statistical analysis
Analysis of variance was used with a 33 factorial design to eval-
uate the influences of fatty acid chain length, gum type, and gum
level on volatilities of the fatty acids (Minitab Vs. 8, Minitab, Inc.,
State College, PA). The complete data set was analyzed using the
Peak Area standardized ? T ? L ? F ? TL ? TF ? LF ? TLF
where T?gum type, L?gum level, F?fatty acid type, followed by
interaction terms; each of the main effects was treated as a fixed
term. Significance was defined at P?0.05. Fisher mean compari-
sons were conducted where appropriate. Relationships between vol-
atility and viscosity data were evaluated using regression techniques.
Each experiment was completely replicated on three different days.
RESULTS & DISCUSSION
Fatty acid volatility
Fatty acid type affected VFA volatility (P?0.05). Overall recov-
eries were 1.86, 6.75, and 10.6% for butyric, caproic, and caprylic
acid, respectively. Our results confirmed previous work that has
shown an increase in volatility for higher molecular weight com-
pounds of a homologous series in aqueous systems (Buttery et al.,
Hydrocolloid type had no effect on VFA volatility. However, a
gum type ? gum level interaction warranted further evaluation. An
analysis of values from carrageenan gum-containing samples across
gum level demonstrated that carrageenan had no effect on VFA vol-
atility at any of the three levels within fatty acid type (Fig. 1a) rela-
tive to the control. This was contrary to much of research on polysac-
charide/volatile compound interactions (Solms, 1986). However, our
data demonstrated that an anionic hydrocolloid such as carrageenan
and the resulting viscosities may have no effect on VFA volatility.
The source of carrageenan we used contained lesser amounts of ?-
carrageenan, but was predominantly comprised of ?-carrageenan.
Carrageenan has been reported to exhibit strong reactivity with milk
proteins (Schmidt and Smith, 1992). This suggests that much of the
intermolecular binding potential of carrageenan is complexed with
casein proteins, and thus becomes less available as a binding ligand
for aroma compounds.
Evaluation of data from guar-containing samples demonstrated
that the presence of 0.025 to 0.4% guar gum had no effect on volatil-
ities of butyric or caproic acid. For caprylic acid, there was a slight
increase in VFA volatility at the intermediate level of guar gum (Fig
1b). This novel finding suggests that hydrocolloid gums have the
potential to de-solubilize volatile compounds from aqueous media.
The uncharged galactomannan structure of guar gum and tolerance
of ionic species such as salts and acids (Whistler and Daniel, 1985),
may relate to its ability to alter the volatility of VFA’s.
Xanthan gum-containing samples demonstrated that at low and
high levels no changes in volatilities were detected. Similar to the
other gums, xanthan had no effect on volatility of butyric acid at any
concentration. The intermediate level of xanthan gum increased the
volatility of caproic and caprylic acids as compared to the control
(Fig 1c). The increase in volatility was least for caproic acid (?28%)
and greatest for caprylic acid (?36%) suggesting that the increase in
volatility was a function of hydrophobic interactions between the
gum and the nonpolar regions of the fatty acid molecule.
Viscosity data were sigmoidal (Fig. 2); best fits were obtained
with the following polynomial models with corresponding correla-
Fig. 1—Effects of chain length and level of carrageenan (a), guar
(b), and xanthan (c) gums. Error bars represent one standard mean
error. Different letters within each gum represent significantly dif-
ferent means at the P?0.05 level.
caprylic f(x) ? 116x3 ? 713x2 ? 1390x ? 738
f(x) ? 15.4x3 ? 94.5x2 ? 184x ? 92.5
f(x) ? 58.2x3 ? 354x2 ? 700x ? 354
r2 ? 0.68
r2 ? 0.67
r2 ? 0.60
From viscosities of 1.1 to 1.6 log cps, the volatilities increased,
then decreased from 1.6 to 2.5 log cps, and then increased again.
392 Download full-text
???????????????????????—Volume 64, No. 3, 1999
Free Fatty Acid Volatility . . .
Several mechanisms of binding ligand-VFA interaction appearred to
influence VFA volatility, including the reduction of mass transfer
rates as a function of viscosity and potentially competing VFA des-
Fig. 2—Effects of viscosity on butyric (?), caproic (?), and caprylic
(?) acid volatility as determined by dynamic headspace sampling of
aqueous solution containing gum and nonfat dry milk powder.
olubilizing effects. Additionally, the ability of this model matrix to
both increase and decrease the volatility of VFA’s was (P?0.05) po-
tentiated with increasing chain length suggesting that viscosity-in-
fluenced volatilities were affected by hydrophobic interactions.
GUM HYDROCOLLOIDS MAY HAVE NO EFFECT, INCREASE OR
decrease the volatilities of VFA’s. Formulations designed to replace
lipid with fat mimetics should consider the effects that gums may
have on the volatilities of aroma-active compounds. Further studies
are needed to evaluate the effects of re-formulated food matrices on
sensory properties of food systems.
Buttery, R.G., Ling, L.C., and Guadagni, D.G. 1969. Volatilities of aldehydes, ketones,
and esters in dilute water solution. J. Agric. Food Chem. 17: 385-389.
Chen, L. 1997. The interaction between short chain free fatty acids and gums - static
and dynamic headspace analysis of model dairy systems. M.S. thesis, Univ. of Mary-
land, College Park.
Christensen, K.R. and Reineccius, G.A. 1995. Aroma extract dilution analysis of aged
Cheddar cheese. J. Food Sci. 60: 218-220.
Eaton C.D. 1994. Dairy flavors. In Bioprocess Production of Flavor, Fragrance and
Color Ingredients, A.Gabelman (Ed.), p. 169-203. Wiley, Inc., New York.
Hansen, A.P. and Heinis, J.J. 1991. Decrease of vanillin flavor perception in the pres-
ence of casein and whey proteins. J. Dairy Sci. 74: 2936-2940.
Rankin, S.A. and Bodyfelt, F.W. 1996. Headspace diacetyl as affected by stabilizers
and emulsifiers in a model dairy system. J. Food Sci. 61: 921-923.
Schmidt, K.A. and Smith, D.E. 1992. Milk reactivity of gum and milk protein solu-
tions. J. Dairy Sci. 75: 3290-3295.
Solms, J. 1986. Interactions of nonvolatile and volatile substances in foods. In Interac-
tions of Food Components. G.G. Birch and M.G. Lindley (Ed.), p.189-210. Elsevier,
Thanh, M.L., Thibeaudeau, P., Thibaut, M.A., and Voilley, A. 1992. Interactions be-
tween volatile and non-volatile compounds in the presence of water. J. Food Chem.
Voilley, A., Lamer, C., Dubois, P., and Feuillat, M. 1990. Influence of macromolecules
and treatments on the behavior of aroma compounds in a model wine. J. Agric. Food
Chem. 38: 248-251.
Whistler, R.L. and Daniel, J.R. 1985. Carbohydrates. In Food Chemistry, O.R. Fenne-
ma (Ed.), p. 121-122. Marcel Dekker, New York.
Ms received 9/3/98; revised 12/21/98; accepted 1/10/99.
Reprinted from J. Food Sci. 64(3)
©1999 Institute of Food Technologists