Characteristics of physico-chemical properties, volatile compounds and free fatty acid profiles of commercial set-type Turkish yoghurts

doi:10.4236/ojas.2011.11001

Paper Menu >>

Journal Menu >>

Vol.1, No.1, 1-9 (2011) Open Journal of Animal Sciences

doi:10.4236/ojas.2011.11001

Characteristics of physico-chemical properties, volatile

compounds and free fatty acid profiles of commercial

set-type Turkish yoghurts

Chemical properties of Turkish yoghurts

Zehra Güler1, Young W. Park2*

1Department of Food Engineering, Mustafa Kemal University, 31034-Antakya-Hatay, TURKEY

2Agricultural Research Station, Fort Valley State University, Fort Valley, GA 31030, USA; parky@fvsu.edu

Received March 25,2011;revised April 25,2011;accepted April 30,2011

ABSTRACT

Ten most popular brands of commercial Turkish

set-type yoghurts were collected from local re-

tail outlets in Hatay, Turkey for two separate pe-

riods, and analyzed for basic nutrients, phys-

ico-chemical properties, volatile aroma com-

pounds and free fatty acid profiles to compare

their differences among the yoghurt products.

The results showed that there were significant

differences (P < 0.05, 0.01, or 0.001) and varia-

tions in physico-chemical indices, volatile aro-

ma compounds and volatile free fatty acid pro-

files among the yoghurt brands, which ulti-

mately influence the flavor quality of the pro-

duct. Acetaldehyde was predominant volatile

compound in yoghurts, which followed by

acetone, acetoin, diacetyl and ethanol. The level

of diacetly was inversely related to titratable

acidity, acetaldehyde and ethanoic acid. From

ketones with high carbones 2-undecanone and

2-pentadecanone were higher than 2-butanone,

2-nanonane and 2-tridecanone. These ketones

are related to fat content of yoghurt. Among

short chain free fatty acids, ethanoic (acetic)

acid was the most abundant in yoghurts, fol-

lowed by hexanoic, octanoic and butanoic acids.

These differences in detected chemical compo-

sitions of volatile compounds and free fatty ac-

ids would be applicable to predict flavor, nutri-

tional value, quality control or shelf-life of the

commmercial set-type Turkish yoghurts.

Keywords: Turkish Yoghurt; Free Fatty Acids;

Volatile Compounds; Physico-Chemical Properties

1. INTRODUCTION

Yoghurt has been increasingly popular in many parts

of the world, especially in Europe, North America and

the Middle East. High consumption and popularity of

yoghurt especially among women, children and teena-

gers are due to its nutritional and health benefits (Park,

1994; Hekmat and McMahon, 1997).

Although the origin of yoghurt is not definitely known,

historical records indicate that the origin of yoghurt was

in the Middle East (Tamime and Robinson, 2001) and it

was first made by Turks when they were in Middle Asia

and it was named as ‘‘yoghurt’’(Tamime and Deeth,

1980). Today, the product has gained international

recognition with this word, and many other countries use

‘‘yoghurt’’.

Streptococcus thermophilus and Lactobacillus del-

brueckii subsp.bulgaricus are the two typical strains

used for yoghurt production. The role of these two

starter bacteria in yoghurt manufacture can be summa-

rized as milk acidification and synthesis of aromatic

compounds. Flavor compounds that contribute to the

final aroma of yoghurt may be divided into four catego-

ries: non-volatile acids (lactic or pyruvic), volatile acids

(butyric or acetic), carbonyl compounds (acetaldehyde

or diacetyl) and miscellaneous compounds (amino acids

or products formed by thermal degradation) (Tamime

and Robinson, 2001). These flavor compounds of

yoghurt are influenced by the chemical composition of

milk base, processing conditions (e.g. heat treatment,

homogenization), the ratio, activity and strains of starter

culture used and incubation period (Beshkova et al.,1988;

Kneifel et al.,1992; Ulbert and Kniefel, 1992; Hassan et

al., 2003). In general, the overall properties of yoghurt,

such as acidity level, free fatty acid (FFA) content, the

production of aroma compounds (acetaldehyde, diacetly,

Z. Güler et al. / Open Journal of Animal Sciences 1 (2011) 1-9

acetoin) as well as the sensory profile, and nutritional

value, are important traits of products.

In the Turkish markets, set-style yoghurts account for

almost 100% of consumption. In set-style yoghurts, fer-

mentation takes place in the final pots, so they are sub-

jected to less industrial handling, and the possibility of

contamination is lower. In Turkey, yoghurt is one of the

greatest volumes of dairy products produced, which is

1 010 000 tons/per year (FAO, 2006). The consumers

prefer yoghurts of different brands due to their different

flavor and texture. However, few studies have been

conducted on characterization of nutritional and chemi-

cal specifications of different commercial yoghurt pro-

ducts sold in Turkey.

Therefore the objectives of this study were to: (1)

determine basic nutrient composition and some phys-

icochemical characteristics of commercially marketed

Turkish yoghurts, (2) quantify volatile compound and

free fatty acid compositions of the products, and (3)

evaluate the relationships among the tested parame-

ters.

2. MATERIAL AND METHODS

2.1. Preparation of Experimental Yoghurt

Samples

Two lots of set-type yoghurts made from cow milk

were obtained from 10 most popular commercially mar-

keted brands in Hatay, Turkey marketed during two sep-

arate periods in March and December, 2007.

2.2. Chemical Analyses

2.2.1. Analyses of Basic Nutrients And

Physicochemical Indices

Total solids, fat, protein, ash contents and titratable

acidity value of yoghurts were determined according to

the Association of Official Analytical Chemist (AOAC,

2003) methods. pH was measured using a pH meter

(Orion, Thermo, Austin, TX, USA). Lactose content was

estimated as the difference between total solids and the

sum of fat, protein and ash contents.

Color characterisitcs were measured by using a

Minolta Chromameter (model CR-400 Tokyo, Japan)

calibrated with a manufacturer-supplied white cali-

bration plate. The L (dark = 0 and light = 100), a (red

= +a and green = –a) and b (yellow = +b and blue =

–b) values were measured. The L*, a* and b* reading

was carried out in triplicate for each sample. Results

were expressed as Chroma (C* = [(a*)² + (b*)²]0.5 ),

hue angle (hab = tan–1[(a*)(b*)–1]), and whiteness in-

dex (WI = 100 – [(100 – L)2 + a2 + b2]0.5 ). Analyses

were carried out in duplicate obtaining two yoghurt

samples from each brand were collected during two ex-

perimental period (March and December, 2007), and

each sample was analyzed in duplicate.

2.2.2. Analyses of Free Fatty Acids (Ffa) and

Benzoic Acid

Extraction and quantification of FFA and benzoic acid

were carried out according to the method of Deeth et al.

(1983) with slight modifications as reported by Güler

(2008). Heptanoic acid was added to all experimental

yoghurt samples at the time of extraction. FFAs were

analyzed by a GC-MS (Agilent 6890 gas chromatograph

and 5973 N mass selective detector; Agilent, Palo Alto,

CA, USA). Column used for FFA and benzoic acid sep-

aration was a DB-FFAP-column (30 m × 0.25 mm id ×

0.25 µm film thickness). Analyses were carried out in

triplicate.

For GC operating conditions of FFA and benzoic acid

analysis, helium was used as a carrier gas with a con-

stant flow rate of 1 mL



min–1. The GC oven tempera-

ture was set to 50˚C for 5 min then raised to 230˚C at a

rate of 5˚C min–1 and held at 230˚C for 20 min. The in-

jector temperature was 250˚C, and the run time was 58

min. The GC column was connected without splitting to

the ion source of the Agilent 5973N model quadrupole

mass selective detector which was operating in the scan

mode within a mass range 33 to 330 m z–1 at 1 scan s–1.

The interface line to MS was set at 280˚C. The MS was

ope- rated in an electron impact mode at electron energy

of 70eV and was calibrated by auto-tuning. Identifica-

tion of the compounds was performed by a computer-

matching of their mass spectral data with those of

known compounds from the Mass Spectral Database

(Wiley7n.1/Nist02.L.). To compensate the amount of

loss during the extraction and clean-up, heptanoic acid

(C7) was used as internal standard. Pre-analyses of the

milk and yoghurt had ensured that heptanoic acid was

absent.

2.2.3. Analysis of Volatile Compounds (VC)

VCs were determined by static head space technique

according to Güler (2007). VCs were analyzed using a

Agilent model 6890 gas chromatography (GC) and 5973

N mass selective detector (MS) (Agilent, Palo Alto, CA,

USA). Columns used for FFA separation HP- INNOWAX

capillary column (30 m × 0.32 mm id × 0.25 µm film

thickness). The volatile compounds were separated un-

der the following conditions: injector temperature 200˚C;

carrier gas helium at a flow rate of 1.4 mL



min–1; oven

temperature program initially held at 50˚C for 6 min and

then programmed from 50˚C to 180˚C at 8˚C



min–1

held at 180˚C for 5 min. The interface line to MS was set

at 250˚C. Identification of the compounds was also con-

ducted by a computer -matching of their mass spectral

Z. Güler et al. / Open Journal of Animal Sciences 1 (2011) 1-9

data with those of known compounds from the Nist 02.L.

Mass Spectral Database. Based on the peak resolution,

their areas were estimated from the integrations per-

formed on selected ions. The resulting peak areas were

expressed in the arbitrary area units. Quantification of

constituents was calculated by external standard tech-

nique.

2.3. Statistical Analysis

Statistical analysis was performed using the SPSS

version 9.05 for Macintosh (SPPSS Inc./Chicago, III.,

U.S.A.). Physicochemical parameters were submitted to

one-way analysis of variance (ANOVA). Duncan multi-

ple mean comparison test (P < 0.05) was used to state

the differences among the brands. Principal component

analysis (PCA), using a correlation matrix, was carried

out on mean of chemical compounds including volatiles,

free fatty acids. PCA is a statistical technique used to

identify the minimum number of latent variables, called

`principal components`, that explain the highest amount

of observed variability among studied samples. This

technique simplifies the initial multidimensional data

matrix without the loss of any important information,

thus making results easier to interpret.

To simplify the interpretation of results, flavor, physical

properties and volatile compounds were studied sepa-

rately. Varimax rotation was performed on variables.

Eigenvalues over 1.0 were considered as the selection

criterion for the minimum number of components. Cor-

relation analysis was also performed to determine rela-

tionships among individual physicochemical parameter,

FFA and volatile compounds in the yoghurt samples.

3. RESULTS AND DISCUSSION

3.1. Physicochemical Characteristics

Characteristics of basic nutrient composition and phy-

sico-chemical parameters of 10 most popular Turkish

yoghurts are summarized and compared in Table 1. The

mean values for total solids, fat, protein, ash, titratable

acidity and pH were 15.30%, 3.77%, 4.88%, 0.99%,

1.58% (lactic acid) and 4.22, respectively. The mean

total solids, fat and titratable acidity of yoghurts were

within the range of Turkish yoghurt standard, which

were > 15.5 ± 0.5%, 1.5% - 3.8% < and 0.6% - 1.6%

(lactic acid), respectively. As shown in Table 1, there

were significant differences (P < 0.05, 0.01 and 0.001) in

physicochemical properties between the 10 yoghurts

tested. Titratable acidity values of yoghurts obtained

from brands AS, SK, TV and YR were slightly high than

maximum level (1.6% lactic acid) reported in Turkish

standard, while that of the other samples was close to the

standard level. High titratable acidity is not suprising

since yoghurts are characterized primarily by its acidity.

Titratable acidity is one of the most important parame-

ters with respect to the shelf-life of fermented milk

products and also a reasonable indication of the per-

formance of the starter culture (Tamime and Robinson,

2001). High titratable acidity in yoghurts is probably due

to high total solids content, especially in protein since

titratable acidity was positively correlated to total solids

(0.632, P < 0.01) and protein (0.533, P < 0.01). Accord-

ing to Turkish yoghurt standard (TSI, 2006), yoghurts

AS, SK, TV and YR are classified as whole

milk-yoghurt (3.8% fat <) and the other samples were

classified as fat-yoghurt. The chemical characteristics of

Turkish yoghurts are different from values reported by

Tamime and Robinson (2001). This may be attributed to

the yoghurt-making technique since corn starch in most

countries is used in natural yoghurt maufacture for the

fortification of milk or sweet flavor. These yoghurts

would have a low fat and high carbohydrate contents.

For natural yoghurt-making, the use of some ingredients

is forbidden by Turkish reguations. Traditionally, it is

known that Turkish people prefer sour yoghurt flavor.

Characteristics of yoghurt color is another important

marketable (consumer acceptance) quality attribute.

Table 1. Mean basic nutrient content (g/100 g yoghurt) and physico-chemical indices of 10 brands of commercial Turkish set-type

yoghurts1.

Brands

Properties

TS2

15.19 ± 0.62b

15.79 ± 0.26bc

13.16 ± 1.09a

15.07 ± 0.42b

14.82 ± 0.13b

15.00 ± 0.57b

15.64 ± 0.04bc

15.15 ± 0.05b

16.46 ± 1.47cd

16.72 ± 0.41d

***

Protein

5.24 ± 0.54bc

4.82 ± 0.33ab

4.64 ± 0.84ab

4.85 ± 0.41ab

4.55 ± 0.57ab

4.77 ± 0.51ab

4.36 ± 0.27ab

4.13 ± 0.47a

5.91 ± 1.21c

5.32 ± 0.44bc

Fat

3.50 ± 0.12a

5.35 ± 0.06d

3.35 ± 0.19a

3.55 ± 0.10a

3.48 ± .22a

3.65 ± 0.17ab

3.30 ± 0.23a

4.18 ± 0.17c

3.95 ± 0.50bc

Lactose

5.45 ± 0.29ab

4.66 ± 0.42ab

4.38 ± 1.80a

5.97 ± 0.45bc

5.80 ± 0.74bc

5.90 ± 0.54bc

6.75 ± 0.38cd

6.82 ± 0.71d

5.16 ± 0.49ab

6.51 ± 1.16cd

Ash

1.01 ± 0.02e

0.96 ± 0.00de

0.79 ± 0.08a

0.90 ± 0.02bc

0.91 ± 0.03cd

0.85 ± 0.06b

0.88 ± 0.01bc

0.90 ± 0.01bc

0.96 ± 0.01de

0.93 ± 0.01cd

***

4.51 ± 0.07c

4.18 ± 0.07ab

4.18 ± 0.23ab

4.18 ± 0.10ab

4.35 ± 0.19bc

4.22 ± 0.10ab

4.07 ± 0.16a

4.23 ± 0.25ab

4.05 ± 0.04a

4.25 ± 0.16ab

TA3

1.43 ± 0.03ab

1.72 ± 0.01ef

1.37 ± 0.02a

1.60 ± 0.05d

1.42 ± 0.02ab

1.57 ± 0.09cd

1.77 ± 0.05

1.49 ± 0.12bc

1.83 ± 0.02f

1.64 ± 0.03de

***

WI4

78.22 ± 0.4b

77.57 ± 0.5a

78.30 ± 0.4b

78.27 ± 0.2b

78.61 ± 0.38b

78.39 ± 0.3b

78.17 ± 0.3b

78.05 ± 0.6b

77.77 ± 0.1a

78.72 ± 0.3b

Chroma

7.38 ± 0.2ab

7.26 ± 0.1a

7.39 ± 0.1ab

7.69 ± 0.1c

7.66 ± 0.08cd

7.88 ± 0.1d

7.50 ± 0.1b

7.78 ± 0.1d

8.32 ± 0.1e

7.74 ± 0.1d

HA5

–25.25 ± 0.2c

–24.52 ± 0.2cd

–25.57 ± 0.2c

27.36 ± 0.1b

–27.13 ± 0.23b

–27.88 ± 0.1b

–25.56 ± 0.2c

–27.40 ± 0.3b

–30.20 ± 0.2a

–27.37 ± 0.3b

1Each observation is mean ± standard deviation of four yoghurt samples; 2TS: Total solids; 3TA: Titratable acidity; expressed as g lactic acid 100 g-1 of yoghurt;

4WI: Whiteness index; 5HA: Hue angle; a.b.c Means with different superscript within the same row are significantly different between yoghurt brands; P: Sig-

nificance level. NS: non significant; *P<0.05, **P<0.01, and ***P<0.001.

Z. Güler et al. / Open Journal of Animal Sciences 1 (2011) 1-9

Yoghurt color is a measure of whiteness index (WI,

black = 0 and white = 100), chroma (C*; color intensity)

and hue angle (hab; color purity). Whiteness index,

chroma and hab values of samples varied from 77.57 to

7.25 and –30.20 to 78.72, 8.32 and –24.52, respectively

(Table 1). The whiteness index (–0.594, P<0.01) and hab

(0.360, P < 0.05) values of yoghurts were correlated to

fat since higher hab is more yellow hue. Chroma (color

intensity) was correlated to total solids (0.360, P < 0.05)

and protein (0.318, P < 0.05). The color values of cow

milk yoghurts in this study were different from the re-

sults observed for goat milk-yoghurt (Guler and Park,

2009) and for nonhomogenized cow milk yoghurt (Var-

gas et al., 2008). This apparently accounts for the type of

milk used for yoghurt-making, the chemical composition

of yoghurt and yoghurt-making technique since the gel

opacity is related to the fat content, casein ratio and their

aggregation level (Hutchings, 1999).

3.2. Free Fatty Acids Profiles

Free fatty acids (FFA) are generated by both lipolytic

processes (C4-C20) and bacterial fermentation (C2-C4).

Quantification of the levels of short-chain FFAs would

be important since their concentration can cause flavor

changes and defects in milk based foods. The results of

our study (Table 2) showed that hexadecanoic (C16),

octadecanoic (C18), cis 9-Octadodecanoic (C18:1), tet-

radecanoic (C14), ethanoic acid (C2), dodecanoic (C12)

and hexanoic (C6) acids were the major FFAs found in

the Turkish yoghurts. This outcome was similar to the

previous report on cow milk yoghurts made with differ-

ent types of cultures (Guler, 2008). There were signifi-

cant differences (P < 0.05, 0.01 and 0.001) in C2 to C12

and C18 free fatty acids concentrations among different

brands of the yoghurts.

Principal component analyses (PCA) applied to free

fatty acids is shown in Table 3. Varimax rotation was

applied to more clearly for grouping individual com-

pounds. The compounds demonstrating the similar

chemical properties and/or origin were grouped into the

one component and were related to each other in propor-

tion to the magnitude of their coefficients. The first

component is the combination of variables that explains

the greatest amount of variation. The second compound

defines the next largest amount of variation. According

to the PCA, the first thr ee components explained 77.42%

of the total variation. The first component, accounting

for 33.39% of total variance, was an intense free fatty

acid, with high loadings for butanoic, hexanoic, octanoic,

decanoic and dodecanoic acids with positive coefficient,

and octadecanoic and cis, 9-octadecenoic acid with neg-

ative coefficient. In other words, the C4 to C12 free fatty

acids were negatively related to the different types of C18

fatty acids. This result was in consistent with the

previous reports (Calderon et al.,1984; Lee et al., 2004).

The second component, explaining 27.77% of total va-

riance, was defined by tetradecanoic and hexadecanoic

acids with positive correlations. The last component,

accounting for 16.26% of total variance, was related to

ethanoic acid with high negative coefficient. Ethanoic

and benzoic acid too was seperately grouped from other

free fatty acids. This result implies that acetic acid was

not produced from only lipolysis by lipases, but from

other biochemical pathways.

Among short chain free fatty acids, ethanoic (acetic)

acid was the most abundant in yoghurts, followed by

hexanoic, octanoic and butanoic acids. This result was

Table 2. Profiles of mean concentrations (µg 100g-1 of yoghurt) of free fatty acids and benzoic acid in 10 brands of Turkish commer-

cial set-type yoghurts1.

Brands

Free Fatty Acids

Ethanoic (C2)

12.4 ± 4.0ab

23.8 ± 8.1bc

5.3 ± 2.7a

5.4 ± 4.0a

15.1 ± 7.2ab

3.3 ± 2.8a

8.5 ± 5.3ab

12.1 ± 11.6ab

13.9 ± 2.3ab

34.6 ± 23.7c

Butanoic (C4)

6.3 ± 3.8c

9.1 ± 6.1cd

5.4 ± 5.0b

12.1 ± 5.4e

3.6 ± 2.2ab

4.5 ± 2.1d

8.6 ± 9.3cd

14.7 ± 10.3f

1.3 ± 0.6a

9.8 ± 3.2cd

***

Hexanoic ( C6)

9.2 ± 5.3bc

10.2 ± 3.2cd

3.9 ± 3.2a

16.2 ± 12.2e

9.5 ± 7.4b

5.9 ± 2.9a

9.0 ± 11.2bc

14.4 ± 11.5e

5.1 ± 0.7a

14.3 ± 6.4d

***

Octanoic (C8)

2.2 ± 0.4a

2.1 ± 0.4a

4.5 ± 2.0f

3.4 ± 6.1g

6.4 ± 5.0cd

3.4 ± 0.8ab

4.6 ± 2.9bc

7.6 ± 4.2e

3.2 ± 1.0ab

5.8 ± 1.3

***

Total (C2-C8)

30.1 ± 13.6

45.3 ± 17.7

19.1 ± 13.0

37.4 ± 27.9

34.6 ±22.0

17.1 ±8.7

30.7 ± 28.7

48.9 ± 37.7

23.59 ± 4.74

64.6 ± 34.7

Decanoic (C10)

4.4 ± 1.5a

3.8 ± 0.9bc

10.4 ± 4.1f

6.4 ± 5.2g

8.4 ± 3.6d

5.9 ± 0.9bc

6.9 ± 3.2cd

11.2 ± 5.2f

6.9 ± 2.6d

8.8 ± 3.2e

***

Dodecanoic (C12)

7.2 ± 2.6a

10.1 ± 5.4bc

15.7 ± 4.8e

10.4 ± 17.2f

10.7 ± 4.0bc

13.2 ± 3.1e

9.7 ± 2.7bc

11.6 ± 4.3d

8.8 ± 3.8c

8.9 ± 1.1b

***

Tetradecanoic (C14)

35.7 ± 12.3

34.7 ± 15.5

48.8 ± 14.1

38.4 ± 17.1

48.1 ± 8.6

36.0 ± 9.8

37.7 ± 5.2

39.5 ± 10.5

39.8 ± 12.1

32.3 ± 7.9

Total (C10-C14)

47.3 ± 16.4

48.7 ± 21.8

74.9 ± 23.1

55.2 ± 39.5

67.2 ± 16.2

55.2 ± 13.9

54.3± 11.3

62.4 ± 20.1

55.6 ± 18.6

50.0 ± 12.1

Pentadecanoic (C15)

3.0 ± 0.7 b

3.3 ± 1.5 b

6.2 ± 3.6 c

1.6 ± 1.9 a

5.8 ± 1.7 c

3.6 ± 1.7 b

2.4 ± 1.8 a

2.53 ± 1.81 b

2.9 ± 1.1 b

Hexadecanoic (C16)

186.5 ± 53.5

158.2 ± 68.1

209.2 ± 60.7

170.5 ± 81.0

217.9 ± 33.2

183.9 ± 41.3

200.3 ± 3.4

198.7 ± 21.1

227.9 ± 25.3

153.3 ± 51.5

9-Hexadecenoic (C16:1)

2.9 ± 1.9 c

0.2 ± 0.1 a

3.5 ± 1.9 c

2.9 ± 2.7 c

4.5 ± 4.3 d

1.7 ± 1.2 b

7.3 ± 4.5 e

Heptadecanoic (C17)

4.4 ± 1.8 c

2.6 ± 0.4 b

6.1 ± 1.7 d

3.1 ± 2.3 b

1.9 ± 1.0 a

1.8 ± 0.5 a

4.7 ± 0.8 c

6.8 ± 3.2 d

Octadecanoic (C18)

120.5 ± 40.2bc

88.6 ± 40.9ab

98.8 ± 26.6abc

78.5 ± 42.6a

111.9 ± 29.8ab

75.3 ± 16.6

141.2 ± 31.0c

144.5 ± 44.1bc

131.4 ± 35.0bc

75.5 ± 28.1ab

9-Octadecenoic (C18:1)

88.6 ± 48.1

59.8 ± 31.4

73.1 ± 26.0

56.9 ± 16.2

80.7 ± 31.7

111.8 ± 27.0

64.9 ± 7.4

43.0 ± 19.0

99.3 ± 18.0

40.2 ± 25.5

Total (C15- C18:1)

405.9 ± 146.9

309.8 ± 142.1

392.8 ± 119.4

307.7 ± 141.8

425.8 ± 100.0

380.6 ± 1.9

415.1 ± 49.4

388.5 ± 86.5

473.1 ± 85.5

278.6 ± 109.5

Benzoic acid

9.08 ± 0.2e

4.37 ± 0.3a

4.84 ± 0.37ab

7.2 ± 0.3cd

7.00 ± 0.82c

5.53 ± 0.41b

7.37 ± 0.48cd

7.94 ± 0.35d

10.92 ± 0.76f

7.37 ± 0.48cd

1Each observation is mean ± standard deviation of four yoghurt samples; a.b.cMeans with different superscript within the same row were significantly different

between yoghurt brands; P: significance level. NS: non significant; *P<0.05, **P<0.01, and ***P<0.001.

Z. Güler et al. / Open Journal of Animal Sciences 1 (2011) 1-9

Table 3. Principle component loadings of free fatty acids in

yoghurts following varimax rotation.

Principal Component

Free Fatty Acids

PC1

PC2

PC3

Ethanoic acid (C2)

–0.14

–0.38

0.72

Butanoic acid (C4)

0.76

0.08

0.32

Hexanoic acid ( C6)

0.81

0.08

0.48

Octanoic acid (C8)

0.83

0.48

0.09

Decanoic acid (C10)

0.74

0.51

–0.05

Dodecanoic acid (C12)

0.69

0.41

–0.42

Tetradecanoic acid (C14)

–0.32

0.73

0.05

Pentadecanoic (C15)

0.66

0.09

–0.50

Hexadecanoic (C16)

–0.51

0.81

0.19

9-Hexadecenoic (C16:1)

0.61

0.14

0.04

Heptadecanoic (C17)

0.55

0.03

0.11

Octadecanoic (C18)

–0.51

0.40

0.38

9-Octadecenoic (C18:1)

–0.68

0.49

–0.04

Benzoic acid

0.14

0.10

0.83

% explained variance

27.48

26.75

19.77

Numbers in bold are primary importance due to the magnitude of coeffi-

cient.

consistent with that reported by the other authors (Rasic

and Kurman, 1978; Warsy, 1983). Beshkova et al. (1998)

found that the formation of volatile free fatty acids

(C2-C10) was more active in the mixed yoghurt cultures

than in the pure ones owing to the stimulating effect of

protocol-operation between the two thermophillic

species on the metabolic activities, which are responsible

for the formation of free fatty acids. In fact, volatile ac-

ids is not only produced from lipolysis by lipases but

also from several biochemical pathways including the

fermentation of lactose or citrate and the degradation

(oxidative deamination or decarboxylation) of amino

acids (alanine and serine) which are the most important

precursor of most volatile fatty acids (Kneifel et al.,

1992; Beshkova et al.,1998). The present study showed

that butanoic (0.396, P < 0.05), hexanoic (0.423, P <

0.01) and octanoic (0.320, P < 0.05) acids were corre-

lated to lactose content. Acetic acid was also positively

correlated to total solids (0.373, P < 0.05) and fat (0.345,

P < 0.05), where similar results were found by Kamin-

erides et al. (2007).

The short-, medium- and long-chain FFAs represented

approximately 6.8 - 7.8, 10.8 - 13.0 and 82.2% - 79.2%,

respectively, in the experimental yoghurts. Despite the

quantitative importance of medium- and long-chain

FFAs, they are not the main contributors to flavor of

dairy products (Molimard and Spinnler, 1996).

The content of benzoic acid of yoghurts showed sig-

nificant differences between the brands (P < 0.05). The

levels of benzoic acid were higher than value reported

earlier for goat milk yoghurt (Guler, 2007), whereas

were lower than levels reported by Tamime and Robin-

son (2001). This may be due to the type of milk and

starter culture used for yoghurt. As known, benzoic acid

formed from breakdown of hippuric acid by yoghurt

organisms acts as a natural preservative. Benzoic acid

and ethanoic acid were grouped in the same principal

component (PC3) (Table 3). These acids may occur

from the similar sources.

3.3. Profiles of Volatile Compounds

In volatile compounds analysis, the headspace

analysis method is more convenient than the ether

extract one (Kang et al., 1988). Of headspace methods,

static headspace is the most satisfactory for quantifi-

cation of acetaldehyde, acetoin, 2-butanone and

2-pentanone (Ott et al., 1999). The main aroma com-

pounds identified in typical Turkish yoghurts are

shown in Table 4. Except for acetaldehyde, all the

other volatiles were not detected in some yoghurt

samples. Quantitatively, the major volatile compound

in the headspace and contributing to the flavor of

set-type yoghurt appeared to be acetaldehyde, which

was reported by other researchers (Kneifel et al.,1992;

Ott et al., 1997). Significant (P < 0.001) variations

were observed in acetaldehyde concentrations of yo-

ghurts among the brands. The mean concentration of

acetaldehyde was 50.74 ppm. Acetaldehyde was posi-

tively correlated to protein (0.314, P < 0.05) and ti-

tratable acidity (0.350, P < 0.05) since the fortification

and heat treatment of milk have been reported to in-

crease the production of acetaldehyde (Tamime and

Robinson, 2001). This increase in acetaldehyde may

subsequently elevate free amino acids and growth

factors for lactobacilli. Ketones are common con-

stituents of yoghurt as other volatile compounds.

However, some methyl ketones were not detected in

all the yoghurts (Table 4). Even though diketone di-

acetyl (2,3 butanedione) was significantly varied

from sample to sample (P < 0.001), it was detected in

yoghurt samples of brands AS, IC, PN and TV. Di-

acetyl content showed significant correlations to lev-

els of titratable acidity (–0.539, P < 0.01), acetalde-

hyde (–0.365, P < 0.05), ethanoic acid (–0.351, P <

0.05), octanoic (0.493, P < 0.01) and decanoic (0.513,

P < 0.01) acids. The similar results were observed by

Kneifel et al. (1992), Ulbert and Kneifel (1992) and

Serra et al. (2009). Monnet and Corrieu, (2007) indi-

cated that diketones in yoghurt come only from py-

ruvate, since thermophilic starter cultures are not

able to metabolize citrate. S. thermophilus strains

possess an α-acetolactate synthase and an acetohy-

droxy acid synthase, which produce α-acetolactate

Z. Güler et al. / Open Journal of Animal Sciences 1 (2011) 1-9

Table 4. Profiles of concentrations (µg 100g–1 of yoghurt) of mean volatile compounds of 10 brands of commercial set-type Turkish

yoghurts1.

Brands

Volatiles

Acetaldehyde¹

50.23

± 7.78a

53.55

± 5.14a

47.25

± 3.34a

56.03

± 5.56ab

55.40

± 15.76ab

40.50

± 16.25a

50.35

± 24.47a

39.60

± 12.13a

72.73

± 2.05b

41.65

± 3.00a

Diacetyl¹

3.33

± 0.35b

10.81

± 0.14e

3.37

± 0.51b

4.34

± 0.27c

7.90

± 0.09d

1.10 a

Acetoin¹

8.75

± 2.03cd

6.26

± 0.22bc

3.63

± 0.30a

5.58 ± 0.88b

4.20

± 0.22a

5.75

± 1.55b

7.53

± 0.78c

12.79

± 0.57d

Acetone¹

5.13

± 0.22ab

5.35

± 0.23ab

5.01

± 0.74ab

9.32

± 0.74d

14.10

± 1.64e

6.25

± 1.30bc

4.40

± 0.64a

5.63

± 0.35ab

6.73

± 0.77c

Ethanol¹

9.15

± 0.19e

5.14

± 0.20c

3.90

± 0.29b

5.20

± 0.16c

4.23

± 0.27b

8.12

± 0.52d

2.86

± 0.38a

2-Butanone²

4.45

± 1.05

2-Nanonane²

2.73

± 0.47

2-Undecanone²

9.50

± 0.91e

3.00

± 0.41b

1.71

± 0.19

5.46

± 0.35c

0.59

± 0.08a

10.60

± 0.49f

6.56

± 0.90d

2-Tridecanone²

6.15

± 0.13e

2.56

± 0.13c

1.41

± 0.31b

5.20

± 0.28

3.50

± 0.20d

0.65

± 0.21a

2-Pentadecanone²

10.71

± 0.63f

5.23

± 0.19d

0.90

± 0.29a

2.13

± 0.10b

2.65

± 0.12c

5.50

± 0.37d

6.87

± 0.38e

1.95

± 0.37b

Ethylacetate¹

8.67

±1.03d

5.95

± 0.66c

2.13

± 0.25a

2.35

± 0.30a

3.57

± 0.22b

8.55

± 0.36d

3.13

± 0.07b

2.24

± 0.57a

1Each observation is mean ±standard deviation of four yoghurt samples; a.b.cMeans with different superscript within the same row were significantly different

between yoghurt brands; P: significance level; NS: non-significant; *P < 0.05, **P < 0.01, and ***P < 0.001.

and 2-hydroxyacetolactate, respectively from py-

ruvate. These two α-aceto acids are generally metabo-

lized into more neutral compounds to maintain pH ho-

meostasis (Tsau et al., 1992) by spontaneous decarboxy-

lation. They can be converted into diacetyl, or by means

of enzymatic mechanisms into branched-chain amino

acids such as valine, leucine or isoleucine (Monnet and

Corrieu, 2007). The former amino acids have been de-

scribed as essential for L. delbrueckii subsp. bulgaricus

responsible for development of acidity (Ott et al., 2000).

These observations could explain why diacetyl was neg-

atively correlated to titratable acidity, ethanoic acid and

acetaldehyde as mentioned above. On the other hand,

diacetyl was negatively correlated (–0.322, P < 0.05) to

acetoin, as reported earlier (Warsy, 1983). This may be

due to the reduction of diacetyl to acetoin. Urbach (1995)

reported that diacetyl can be reduced to acetoin by the

starter bacteria, but its subsequent reduction to bu-

tan-2-one is brought about by the non starter lactic acid

bacteria (NSLAB). Therefore, the high concentrations of

ethanol and 2-butanone found in yoghurt from brand TV

could be attributed to the presence of the non starter lac-

tic acid bacteria (NSLAB), although microbial analyses

were not done in yoghurts. This observation confirmed

that 2-butanone was positively and negatively correlated

to ethanol (0.342, P<0.05) and butanoic acid (–0.497, P

< 0.01), respectively.

There were significant differences in acetone (2-pro-

panone) concentrations in yoghurts (P < 0.001). This

methyl ketone is derived from β-oxidation of saturated

free fatty acids depending on the lipolytic activity of

yoghurt strains (Tsau et al., 1992). These observations

could be explained why the saturated free fatty acids

such as butanoic (–0.493, P < 0.01) and hexanoic

(–0.396, P < 0.05) were inversely correlated to acetone.

With respect to the other ketones, the amounts of

2-nonanone (0.836, P < 0.01) 2-undecanone (0.328, P <

0.05) 2-tridecanone (0.460, P < 0.01) and 2-pentade-

canone (0.460, P < 0.01) showed a positive correlation

with fat contents of yoghurts. Stelios et al. (2007) found

that these ketones increased in yoghurts depending on

the increase in fat content and storage time.

The other volatile compounds such as ethanol and

ethyl acetate showed significant differences among yo-

ghurts (P < 0.01). Ethanol was correlated to acetoin

(–0.382, P < 0.05) and acetone (–0.322, P < 0.05) and

ethyl acetate (–0.562, P<0.01), lactose (–0.409, P < 0.01)

pH (–0.383, P < 0.05), fat (0.581, P < 0.01). These find-

ings were similar to the results reported by Kaminarides

et al. (2007). Ethyl acetate was also correlated to acetoin

(0.469, P < 0.01), 2-nanonane (–0.412, P < 0.01),

2-undecanone (–0.722, P < 0.01), and 2-pentadecanone

(–0.390, P < 0.05). This observation confirmed that these

ketones decreased in products with progressive storage

Z. Güler et al. / Open Journal of Animal Sciences 1 (2011) 1-9

since ethyl acetate is increased in yoghurts depending to

time of storage (Güler, 2007). Ethyl acetate was nega-

tively correlated to ethanoic acid (0.305, P < 0.05), bu-

tanoic (0.361, P < 0.05), hexanoic (0.441, P < 0.01), oc-

tanoic (0.358, P < 0.05) and decanoic (0.341, P < 0.05)

acids. As has been known, ethyl acetate comes from es-

terification of ethanol with acetyl-coenzyme A.

Principal component analyses (PCA) applied to volatile

compounds is shown in Table 5. Four components ex-

plained 82.02% of total variation with 15.58% loss of

variance (Table 5). The first component, accounting for

29.05% of total variance, was defined by acetoin and

acetone in negative coeffıcient, as opposed to 2-nanonane,

2-undecanone, 2-tridecanone, 2-pentadecanone with posi-

tive coefficients that are ketones made up one group. In

other words, acetoin and acetone were inversely related

to the ketones mentioned. The second component (PC2),

accounting for 22.99% of total variance, was mainly

related to acetaldehyde, ethanol and 2-butanone with

positive coefficient. This result confirmed that acetalde-

hyde, ethanol and 2-butanone are produced by lactic acid

bacteria from lactose fermentation or amino acid catabo-

lism (Molimard and Spinnler, 1996). The similar obser-

vations were made by Ulbert and Kneifel (1992). On the

other hand, alcohol dehydrogenase does not seem to be

involved in the reduction of acetaldehyde concentration

as reported previously (Ott et al., 1999). The third com-

ponent, explaining 18.57% of total variance, by ethyl

acetate. The last component, responsible for 11.5% of

variance, was related to was characterized by ethyl ace-

tate. The last component, responsible for 11.5% of vari-

ance, was related to diacetyl. These results suggest that

the occurrence of diacetyl and ethyl acetate may be dif-

ferent from the other volatiles as mentioned above.

Table 5. Principle component loadings of each volatile com-

pound in yoghurts following varimax rotation.

Principal Component

Volatiles

PC1

PC2

PC3

PC4

Acetaldehyde

–0.12

0.74

0.40

–0.03

Acetoin

–0.72

–0.10

0.09

0.36

Acetone

–0.58

–0.13

–0.56

–0.39

Diacety

0.00

–0.66

0.12

0.69

Ethanol

0.57

0.59

–0.23

0.32

2-Butanone

–0.23

0.81

0.03

0.32

2-Nanonane

0.83

0.24

0.16

–0.16

2-Undecanone

0.78

–0.24

–0.44

0.17

2-Tridecanone

0.79

–0.23

0.44

–0.13

2-Pentadecanone

0.93

–0.23

0.15

–0.07

Ethyl acetate

–0.59

–0.29

0.61

–0.15

% explained variance

29.05

22.99

18.57

11.50

Numbers in bold are primary importance due to the magnitude of coeffi-

cient.

In overall, the concentrations of volatile compounds

such as acetaldehyde, acetone and acetoin in the yo-

ghurts of this study were similar to the mean values re-

ported by Ozer et al. (2007) and Güler et al. (2009),

whereas the amount of ethanol was low. These authors

did not find diacetyl in yoghurts. The levels of diacetyl

ranged from less than 1 ppm (Ulbert and Kneifel, 1992;

Ligor et al., 2008) up to more than 40 ppm (Macciola et

al., 2008). The amount of 2-butanone was similar to that

obtained by Stelios et al., (2007) for yoghurt made from

ovine milk. Kneifel et al.(1992) and Xanthopoulos et al.

(1994) showed that the levels of acetaldehyde (2.0 to

41.0 ppm), diacetyl (0.2 to 2.3 ppm), acetoin (2.2 to

28.2), ethanol (0.2 to 9.9 ppm), acetone (1.8 to 3.4 ppm),

and butanone-2 (0.1 - 0.6 ppm) as their samples had a

wide variation in yoghurts. Discrepancies between stud-

ies might have been resulted from the factors including

the use of different starters, synergistic effects of the

microflora itself, fermentation conditions, protein used

for increasing the dry matter or fat content of yoghurt,

and milks from different mammalian species. Further-

more, the analytical method applied may also be a

source of divergent volatile compounds concentrations.

4. CONCLUSIONS

This study confirmed that significant differences exist

in physcochemical indices, volatile compounds and free

fatty acids composition among different brands of

Turkish commercial set-type yoghurt. Gas chromato-

graphic analysis of the yoghurts can also easily be

adapted to quality control or shelf-life studies of yoghurt.

Acetaldehyde was predominant volatile compound in

yoghurts, followed by acetone, acetoin, diacetyl and

ethanol. Diacetly was negatively correlated to titratable

acidity, acetaldehyde and ethanoic acid. Acetone was

inversely correlated to the saturated free fatty acids such

as butanoic and hexanoic. With respect to the other ke-

tones, the amounts of 2-nonanone, 2-undecanone,

2-tridecanone and 2-pentadecanone showed positive

and negative correlations with fat and ethyl acetate con-

tents of yoghurts. It was found that ethanoic (acetic) acid

was the most abundant in yoghurts among short chain

free fatty acids. In addition, hexanoic, octanoic and bu-

tanoic acids showed the next high levels of short chain

free fatty acids in the popular 10 brands of the Turkish

commercial yoghurt products. These differences in vola-

tile compounds and free fatty acids compositions as well

as physico-chemical properties characterized in this

study may applicable for estimation of flavor, nutritional

value, quality control or shelf-life of the commmercial

set-type Turkish yoghurts, and other yoghurt products.

5. ACKNOWLEDGEMENTS

This work was supported by the T.R. Prime Ministry State Planning

Z. Güler et al. / Open Journal of Animal Sciences 1 (2011) 1-9

Organization (DPT) project (Project no: 02 K 120860), and the authors

gratefully acknowledge the GC-MS provided by the DPT project. The

authors are also thankful to Ersin Gök and Gökan Diler for their

assistance in conducting the experiments at the Food Engineering

Department, Mustafa Kemal University, Hatay, Turkey.

REFERENCES

[1] AOAC. (2003) Official Methods of Analysis. Vol.I.17th

ed. Association of Analytical Washington, DC, USA

[2] Beshkova, D., Simova, E., Frengova, G. and Simov, Z.

(1998) Production of flavour compounds by yogurt start-

er cultures. J. Ind. Micro. Biotechn. 20: 180-186.

[3] Calderon, I., De Peters, E. J., Smith, N. E. and Franke,

A. A. (1984) Composition of goat’s milk: Changes within

milking and effects of a high concentrate diet. J. Dairy

Sci. 67: 1905-1911.

[4] Deeth, H. C., Fitz-Gerald, C. H. and Snow, A. J. (1983) A

gas chromatographic method for the quantitative deter-

mination of free fatty acids in milk and milk products.

New Zealand J. Dairy Sci. and Techn. 18: 13-20.

[5] FAO (2006) Food Agricultural Organization.

www.faostat.fao.org

[6] Güler, Z. (2007) Changes in salted yoghurt during stor-

age. Int. J. Food Sci. and Techn. 42: 235-237.

[7] Güler Z. (2008) Evaluation of lipolysis in set-type

fermented milks made with different commercial yogurt

starter cultures during storage. Milchwissenchaft. 33:

73-77.

[8] Güler,Z. and Park, Y.W. (2009) Evaluation of chemical

and color index characteristics of goat milk, its yoghurt

and salted yoghurt. Trop. and Subt. Agroecosystems, 11:

37-39.

[9] Güler, Z.,Taşdelen, A., Şenol, H., Kerimoğlu, N., Temel,

U. (2009) The determination of volatile compounds in

set-type yoghurts by using static headspace gas chroma-

tographic method, Turkish J. Food. 34 (3): 137-142.

[10] Hekmat, S. and McMahon, D. (1997) Manufacture and

quality of iron fortified yoghurt. J. Dairy Sci. 80: 3114.

[11] Hassan, A.N., Ipsen, R., Janzen, T. and Qvist, K.B. (2003)

Microstructure and rheology of yoghurt made with cul-

tures different only in their ability to procedure exopoly-

saccharides. J. Dairy Sci. 86: 1632-1638.

[12] Hutchings, J.B. 1999. In: Food Color Appearance. 2nd Ed.

Springer, 610 pages.

[13] Kaminerides, S., Stamou, P. and Massouras, T. (2007)

Comparison of the characteristics of set type yoghurt

made from ovine milk of different fat content. Int. J.

Food Sci. and Techn. 42:1019-1028.

[14] Kang, Y., Frank, J.F. and Lillard, D.A. (1988) Gas

chromatographic detection of yoghurt flavour com-

pounds and changes during refrigerated storage. Cul.

Dairy Prod. J. 11: 6-9.

[15] Kneifel, W., Ulbert, F., Erhard, F. and Jaros, D. (1992)

Aroma profiles and sensory properties of yoghurt and

yoghurt related products. I. Screening of commercially

available starter cultures. Milchwissenchaft. 47: 362-365.

[16] Lee, J.H., Melton, S. L., Waller, J. C. and Saxton, A. M.

(2004) Modification of physicochemical characteristics

of goat milk fat by feeding protected high oleic sun-

flower oil supplements. J. Food Sci. 69: 280-286.

[17] Ligor, M., Jarmalaviciene, R., Szumski, M., Marugka, A.

and Buszewski, B. (2008) Determination of volatile and

non-volatile products of milk fermentation processes us-

ing capillary zone electrophoresis and solid phase mi-

croextraction coupled to gas chromatography. J. Sept. Sci.

31: 2707-2713

[18] Macciola, V., Candela, G. and De Leonardis, A. (2008)

Rapid gas-chromatographic method for the determination

of diacetyl in milk, fermented milk and butter. Food

Control. 19: 873-878.

[19] Molimard, P. and Spinnler, H.E. (1996) Compound

Involved in the Flavor of Surface Mold-Ripened Cheeses:

Origins and Properties. J. Dairy Sci.79: 169-184.

[20] Monnet, C. and Corrieu, G. (2007) Selection and proper-

ties of [alpha]-acetolactate decarboxylase-deficient spon-

taneous mutants of Streptococcus thermophilus. Food

Microbiology. 24 : 601-606.

[21] Ott, A. and Fay, L.B., Chaintreau, A.(1997) Deter-

mination and origin of the aroma impact compounds of

flavour. J. Agric. and Food Chem. 45: 850-858.

[22] Ott, A., Germond, J.E., Baumgartner, M. and Chaintreau,

A. (1999) Aroma comparisons of traditional and mild

yogurts: headspacegas chromatography quantification of

volatiles and origin of α-diketons. J. Agric. Food Chem.

47: 2379-2385.

[23] Ott, A., Germond, J.E. and Chaintreau, A., (2000) Vicinal

diketone dormation in yoghurt: 13C precursors and effect

of branched-chain amino acids. J. Agric. and Food Chem.

48: 724-731.

[24] Park, Y.W. (1994) Basic nutrient and mineral compo-

sition of commercial goat milk

[25] yogurt produced in the U.S. Small Rumin. Res. 13: 63-70

[26] Ozer, B., Kirmaci, H.A., Oztekin, Ş., Hayaloglu,A. and

Atamer, M. (2007) Incorporation of microbial transglu-

taminase into non-fat yogurt production, Int. Dairy J. 17:

199-207.

[27] Rasic, J.L. and Kurman, J.A. (1978) Yoghurt. Volume I,

Technical Dairy Publishing House. Copenhagen, Den-

mark.

[28] Serra, M., Trujillo, A.J., Guamis, B. and Ferragut, V.

(2009) Flavour profiles and survival of starter cultures of

yoghurt produced from high-pressure homogenized milk.

Int. Dairy J. 19: 100-106.

[29] Stelios, K., Stamou, P. and Massouras, T. (2007) Com-

parision of The Characterstics of Set Type Yohurt Made

From Ovine Milk of Different Fat Content. International

J. Food Sci. and Techn. 42: 1019-1028.

[30] Tamime, A. and Deeth, H. (1980) Yogurt: Technology

and Biochemistry. J. Food Protect. 43: 939-977.

[31] Tamime, A.Y., Robinson, R.K. (2001) Yoghurt Science

and Technology. CRC Pres. New York, US.

[32] TSI (2006) Yogurt. TS 1330. Ankara : Turkish Standar

Institution.

[33] Tsau, J.L., Guffanti, A.A. and Montville, T.J. (1992)

Convertion of pyruvate to acetoin helps to mantain pH

homeostasis in Lactobacillus plantarum. Appl. and Envi-

rom. Microbi. 58: 991-994.

[34] Ulbert, F. and Kneifel, W. (1992) Aroma profiles and

sensory properties of yoghurt and yoghurt-related prod-

ucts. II. Classification of starter cultures by means of

cluster analysis. Milchwissenchaft. 47: 432-434.

[35] Urbach, G. (1995) Contribution of lactic acid bacteria to

Z. Güler et al. / Open Journal of Animal Sciences 1 (2011) 1-9

flavour compound formation of dairy products. Interna-

tional Dairy Journal, 5, 877-903.

[36] Vargas, M., Chafer, M., Albors, A., Chiral, A. and Gon-

zalez-Martınez, C. (2008) Physicochemical and sensory

characteristics of yoghurt produced from mixtures of

cows’ and goats’ milk. Int. Dairy J. 18: 1146-1152.

[37] Warsy, J.D. (1983) Production of volatile aroma com-

pounds in Dahi. J. Agric. Res. 21 (1): 31-36

[38] Xanthopoulos, V., Picque, D., Basit, N., Boquien, C.Y.

and Corrieu, G. (1994) Methods for the determination of

aroma compounds in dairy products: a comparative study.

J. Dairy Res. 61: 289-297.