Journal of Dairy Science
Volume 89, Issue 8 , Pages 2919-2927, August 2006

Quantification of Trace Volatile Sulfur Compounds in Milk by Solid-Phase Microextraction and Gas Chromatography–Pulsed Flame Photometric Detection

Department of Food Science and Technology, Oregon State University, Corvallis 97331

Received 1 November 2005; accepted 28 December 2005.

Article Outline

Abstract 

Volatile sulfur compounds have been reported to be responsible for the sulfurous off-flavors generated during the thermal processing of milk; however, their analysis has been a challenge due to their high reactivity, high volatility, and low sensory threshold. In this study, reactive thiols were stabilized and the volatile sulfur compounds in milk were extracted by headspace solid-phase microextraction, and analyzed by gas chromatography and pulsed-flame photometric detection. Calibration curves for 7 sulfur-containing compounds were constructed in milk by the standard addition technique. Raw, pasteurized, and UHT milk samples with various fat contents were analyzed. Compared with raw and pasteurized samples, UHT milk contained substantially higher concentrations of hydrogen sulfide, methanethiol, carbon disulfide, dimethyl trisulfide, and di-methyl sulfoxide. The high odor activity values calculated for methanethiol and dimethyl trisulfide suggested that these 2 compounds, in addition to di-methyl sulfide reported in a previous study, could be the most important contributors to the sulfurous note in UHT milk.

Key words: milk off-flavor, volatile sulfur, solid-phase microextraction, pulsed-flame photometric detection

 

Back to Article Outline

Introduction 

Thermal processing of milk can generate off-flavors. Although pasteurization of milk (typically at 72°C for 15s) can impart a slight cooked, sulfurous note, it is acceptable to most consumers. Ultrahigh temperature sterilization (135 to 150°C for 3 to 5s) can induce strong sulfurous, cooked, cabbage-like notes in milk (Shipe, 1980), thus limiting its acceptance (Steely, 1994). Researchers have identified that hydrogen sulfide (H2S), methanethiol (MeSH), carbon disulfide (CS2), dimethyl sulfide (DMS), dimethyl disulfide (DMDS), and di-methyl trisulfide (DMTS) are related to this cooked off-flavor defect (Shipe, 1980; Christensen and Reineccius, 1992; Simon and Hansen, 2001; Datta et al., 2002). Several other sulfur compounds including benzothiazole, dimethyl sulfoxide (DMSO), dimethyl sulfone (Me2SO2), carbonyl sulfide (COS), sulfur dioxide, butanethiol, and thiophene have been found in heated milk, but their roles in milk flavor have not been well elucidated (Scanlan et al., 1968; Shibamoto et al., 1980; Shipe, 1980; Moio et al., 1994).

Most sulfur compounds have distinct aroma characteristics; thus, analysis of volatile sulfur compounds is particularly important. However, reliable quantification of sulfur compounds is challenging due to their high reactivity and low concentration in foods (Chin and Lindsay, 1994; Haberhauer-Troyer et al., 2000; Murray, 2001; Nielsen and Jonsson, 2002). Numerous extraction methods, including static headspace, purge and trap, and solvent-assisted flavor evaporation (Mussinan and Keelan, 1994; Steely, 1994; Contarini et al., 1997; Bendall, 2001; Toso et al., 2002) have been evaluated to analyze sulfur compounds. These methods are either time consuming, require an exhaustive concentration procedure that often leads to the loss of the more volatile compounds such as H2S, MeSH, and DMS, or cause MeSH oxidation and other artifact formation (Steffen and Pawliszyn, 1996). Headspace solid-phase microextraction (HS-SPME) can substantially shorten analysis time, reduce sample manipulation, and minimize artifact formation (Wercinski and Pawliszyn, 1999).

The widely used flame-ionization detector is not sufficiently sensitive for the detection of most sulfur compounds present in milk (Scanlan et al., 1968; Rerkrai et al., 1987; Contarini et al., 1997; Vazquez-Landaverde et al., 2005). Although the flame photometric detector has been often utilized for detection of sulfur compounds, its application poses difficulties due to reduced response with coeluting hydrocarbons (quenching) besides its inconsistent selectivity and sensitivity (Burmeister et al., 1992; Aguerre et al., 2001; Simon and Hansen, 2001). The pulsed-flame photometric detector (PFPD) can overcome these limitations and has rapidly found application as a sulfur-specific detector (Jing and Amirav, 1997; Aguerre et al., 2001; Lestremau et al., 2003). Whereas flame photometric detection uses a continuous flame, PFPD uses a pulsed flame that ignites 2 to 4 times per second to burn compounds eluting from the gas chromatography (GC) column. By setting a timed “gate delay” for the response, the luminescence emission of hydrocarbons can be eliminated from that of sulfur compounds because their emission times are different. Thus, the PFPD produces cleaner chromatograms than the flame photometric detector, making it possible to detect sulfur compounds at low concentrations (Amirav and Jing, 1995). Solid-phase microextraction coupled to GC/PFPD has been successfully used for the analysis of volatile sulfur compounds in several foods and beverages (Pelusio et al., 1995; Hill and Smith, 2000; Fan et al., 2002; Dreher et al., 2003; Burbank and Qian, 2005b; Fang and Qian, 2005). However, few studies reported the accurate quantification of volatile sulfur compounds in foods (Fang and Qian, 2005). Although DMS in milk can be easily quantified by SPME extraction and gas chromatography-flame photometric detection (Vazquez-Landaverde et al., 2005), other trace volatile sulfur compounds cannot be quantified due to the poor sensitivity of this detector. Therefore, the objective of this study was to develop a rapid HS-SPME/GC-PFPD quantification technique for trace volatile sulfur compounds considered important to the flavor of commercially processed milk.

Back to Article Outline

Materials and Methods 

Chemical Standards 

Dimethyl disulfide, DMTS, DMSO, Me2SO2, and iso-propyl disulfide (IPDS) were purchased from Aldrich Chemical Co. Inc. (Milwaukee, WI); DMS was purchased from TCI America Inc. (Portland, OR); CS2 was from EMD Chemicals Inc. (Gibbstown, NJ); and ethyl methyl sulfide (EtMeS) was from TCI-EP Ltd. (Tokyo, Japan).

A cylinder of gaseous MeSH was purchased from Aldrich Chemical Co. Inc. (St Louis, MO) and used to prepare a 34.1 g/kg stock solution by bubbling MeSH gas into cold methanol. The solution was kept at −17°C and used within 2 wk. All glassware used for the MeSH solution was deactivated by soaking it in a 5% dimethyl-dichlorosilane solution in toluene, and then rinsing it with toluene, methanol, and finally water.

A 1.08 g/kg stock solution of H2S was prepared by dissolving 0.077g of sodium sulfide (Na2S·9H2O, Sigma) in 10g of a 50mM phosphate buffer (pH = 8.5). The solution was then diluted with the same buffer to achieve an H2S concentration of 15.9 mg/kg and kept at 4°C before use (no more than 1 wk).

Carbonyl sulfide was prepared according to a published method (Metrohm, 2001) with some modification. Concentrated sulfuric acid was added dropwise to potassium thiocyanate salt (both from Mallinckrodt/J. T. Baker Inc., Phillipsburg, KY) in a stoppered Erlenmeyer flask. The generated COS gas was trapped by bubbling into a separate flask containing cold distilled water. The resulting solution was stored at 4°C, and used no later than 1 d after preparation.

Milk Samples 

Raw homogenized milk samples with 1 and 3.25% fat were obtained locally (Lochmead Farms, Junction City, OR). Sodium azide (0.02%) was added and the samples were stored at −17°C until analyzed (no more than 3 d). Pasteurized milk samples with 0, 1, 2, and 3.25% fat content from 2 different commercial brands (A and B) were purchased locally, stored at 4°C, and analyzed before their expiration date. Ultra-high temperature milks with 1 and 3.25% fat were purchased in Mexico (Leche Araceli, Grupo Fomento Queretano, Embotelladora la Victoria, Queretaro, Mexico), stored at 4°C, and analyzed before their expiration date.

SPME 

A Stableflex 85-μm carboxen-polydimethylsiloxane fiber (Supelco, Bellefonte, PA) was used. Before use, the fiber was conditioned in a split/splitless GC injector port at 300°C for 90min. The fiber was then placed in the SPME adapter of a CombiPAL autosampler (CTC Analytics AG, Zwingen, Switzerland) fitted with a vial heater. The same fiber was used to complete all experimental measurements. The samples were preequilibrated for 5min at 30°C under agitation at 500rpm. The extraction was performed at 30°C for 15min under agitation at 250rpm.

GC-PFPD 

A Varian CP-3800 gas chromatograph (Varian Inc., Walnut Creek, CA) equipped with a pulsed-flame photometric detector was used. Volatile compounds extracted with the SPME fiber were desorbed at 300°C for 6min in the injector. The inlet liner was previously deactivated with 15% N,O-bis(trimethylsilyl)trifluoroacetamide in hexane. The injector was maintained in splitless mode for 4min. Separation of analytes was achieved using a fused silica capillary column (DB-FFAP, 30 m×0.32mm, 1.0-μm film; Agilent Technologies, Inc., Palo Alto, CA). Nitrogen was used as the carrier gas at 2 mL/min. The oven temperature was kept at 35°C for 5min, increased to 150°C at 10°C/min, held for 1min, and then increased to 220°C at 20°C/min, with a final holding time of 5min. The PFPD was operated in the sulfur mode at 300°C and 500V with air 1 at 17 mL/min, air 2 at 10 mL/min, and hydrogen at 14 mL/min.

Quantification 

A standard stock solution containing 20.4 mg/kg of MeSH, 60μg/kg of CS2, 12μg/kg of DMDS, 20μg/kg of DMTS, and 600 mg/kg of DMSO was prepared in methanol. The standard stock solution was then diluted with methanol to 1:2, 1:3, 1:4, 1:5, and 1:6 strengths. The H2S stock solution (15.9 mg/kg) was diluted with 50mM phosphate buffer (pH = 8.5) to 1:2, 1:3, 1:4, 1:5, and 1:6 strength. A separate aqueous solution containing 2.6 g/kg of Me2SO2 was prepared using distilled water and then diluted to 2.2, 1.8, 1.4, 1, and 0.6 g/kg concentrations. To construct the calibration curves, 10-g aliquots of raw milk (3.25% fat) adjusted to pH 2.6 with tartaric acid were spiked with 0.01g of each of the solutions to give the targeted final added concentration for each compound (Table 1). Finally, 0.01g of the internal standard solution containing 1.63 mg/kg of EtMeS and 0.61 mg/kg of IPDS in methanol was added to the spiked milk samples. The volatile sulfur compounds were extracted by HS-SPME and analyzed using GC-PFPD. Triplicate analysis at each concentration level and linear regression plots were used to construct calibration curves of the concentration ratio [(compound concentration)/(internal standard concentration)] vs. the response ratio [squared root of the (area of compound)/(area of internal standard)].

Table 1. Calibration curves for trace volatile sulfur compounds1 in milk
CompoundInternal standardCalibration concentrationsRegression equation2R2Quantification limit3
H2SEtMeS2.6, 5.3, 7.9, 10.6, 13.3, and 15.9μg/kgy = 0.084x0.9550.005μg/kg
MeSH 3.4, 6.8, 10.2, 13.6, 17, and 20.4μg/kgy = 0.101x0.9930.004μg/kg
CS2 10, 20, 30, 40, 50, and 60 ng/kgy = 11.383x0.9740.033 ng/kg
DMDSIPDS2, 4, 6, 8, 10, and 12 ng/kgy = 48.544x0.9950.021 ng/kg
DMTS 3.3, 6.6, 10, 13.3, 16.6, and 20 ng/kgy = 11.655x0.9470.088 ng/kg
DMSO 100, 200, 300, 400, 500, and 600μg/kgy = 0.0009x0.9761.145μg/kg
Me2SO2 0.6, 1, 1.4, 1.8, 2.2, and 2.6 mg/kgy = 0.0003x0.9660.003 mg/kg

1H2S = Hydrogen sulfide; EtMeS = ethyl methyl sulfide; MeSH = methanethiol; CS2 = carbon disulfide; DMDS = dimethyl disulfide; IPDS = isopropyl disulfide; DMTS = dimethyl trisulfide; DMSO = dimethyl sulfoxide; and Me2SO4 = dimethyl sulfone.

2y = Squared root (compound area/internal standard area); x = (compound)/(internal standard).

3Quantification limit calculated as the concentration that gives a signal 10 times that of noise.

Ten grams of milk samples (raw with 1 and 3.25% fat; pasteurized brands A and B with 0, 1, 2, and 3.25% fat; and UHT with 1 and 3.25% fat) were adjusted to pH 2.6 with tartaric acid. Internal standard solution (0.02g) diluted to half strength (0.81 mg/kg of EtMeS and 0.30 mg/kg of IPDS) was added and the volatile sulfur compounds were analyzed using the HS-SPME/GC-PFPD procedure described previously. The concentration of volatile compounds was calculated based on the squared root of the peak area ratio of the compound to the internal standard. All samples were analyzed in triplicate.

Statistical Analysis 

Statistical evaluations including ANOVA, linear regression, and Tukey's honestly significant difference (α = 0.05) were conducted using Statgraphics Plus 5.0 (Manugistics Inc., Rockville, MD).

Back to Article Outline

Results and Discussion 

Stabilization of Volatile Sulfur Compounds in Milk Matrix 

It has been reported that sulfur compounds are extremely reactive and artifacts can be formed during analysis (Block and Calvey, 1994; Haberhauer-Troyer et al., 2000). In this work, all standards were tested individually for stability and artifact formation during sample extraction and analysis. Most of the compounds were stable and only gave a single chromatographic peak. However, MeSH showed significant DMDS artifact formation (Figure 1A). This observation was consistent with previous reports that MeSH in aqueous solutions can be readily oxidized to form DMDS and DMTS probably caused by the presence of oxygen, transition metals, or free radicals (Chin and Lindsay, 1994; Haberhauer-Troyer et al., 2000). This artifact formation makes the quantitative analysis of volatile sulfur compounds in food unreliable. Recently, methanethiol has been stabilized in cheese and wine matrices (Burbank and Qian, 2005a,b; Fang and Qian, 2005) and quantitative analysis of volatile sulfur compounds in wine was achieved. Similar approaches were attempted in this study to stabilize MeSH in milk. The stability of the MeSH standard was tested by spiking it separately into pure water, ethanol, and a pH 2.6 tartaric acid aqueous solution. Minimum DMDS formation was achieved for the tartaric acid solution, suggesting that acidity could stabilize MeSH in milk. This is consistent with the work reported by Kastner et al. (2003), who found that the oxidation rate of ethanethiol, another highly reactive mercaptan, decreased exponentially when pH was lowered from 5.1 to 3.6. Therefore, an acidified matrix was considered desirable to prevent MeSH oxidation in milk. Decreasing the sample pH from 6.6 to 2.6 minimized the DMDS formation from MeSH (Figure 1B) allowing the construction of a calibration curve by the standard addition technique. Acidic conditions were also necessary to achieve total conversion of Na2S standard to H2S, another important sulfur compound to be quantified in this work.

  • View full-size image.
  • Figure 1. 

    Methanethiol (MeSH) stability and dimethyl disulfide formation measured by pulsed-flame photometric detector (PFPD) in 3.25% fat raw milk spiked with 7μg/kg of MeSH; A) pH = 6.6; B) pH = 2.6. COS = Carbonyl sulfide; H2S = hydrogen sulfide; CS2 = carbon disulfide; DMS = dimethyl sulfide; DMDS = dimethyl disulfide; DMTS = dimethyl trisulfide; DMSO = dimethyl sulfoxide; and Me2SO4 = dimethyl sulfone.

Quantification of Sulfur Compounds in Commercial Milk Samples 

The volatile sulfur compounds identified in milk samples were COS, H2S, MeSH, CS2, DMS, DMDS, DMTS, DMSO, and Me2SO2. When the extraction conditions were optimized for the analysis of other trace sulfur compounds by PFPD, DMS reached the saturation limit of the PFPD and was therefore not quantified in this study. Quantification of DMS by GC and flame-ionization detection was previously reported (Vazquez-Landaverde et al., 2005) because it is the most abundant sulfur compound in milk and its quantification can be easily achieved. Quantification of COS was not performed because it is an odorless sulfur compound and it is not important to the flavor of heated milk (Simon and Hansen, 2001).

Calibration curves with high correlation coefficients (R2>0.94) and low quantification limits (signal equal to 10 times the noise) for 7 sulfur compounds were obtained (Table 1). The quantification limits were below their sensory thresholds for all compounds included in this study. The more volatile H2S and MeSH could be easily extracted and had low quantification limits whereas the less volatile DMSO and Me2SO2 were harder to extract. Carbon disulfide, DMDS, and DMTS had the highest quantification sensitivity possibly due to their high affinity to the carboxen/polydimethylsiloxane SPME fiber. Adequate repeatability of the measurements (coefficient of variation<13.6%) was achieved in this study considering the high volatility and reactivity of some of the compounds analyzed.

Commercial milk samples processed with various heat treatments were analyzed using this technique and the results showed that the different concentration of volatile sulfur compounds could be related to the difference in heat treatment (Table 2 and Figure 2). At the same fat level, UHT milk contained significantly higher concentrations of H2S, MeSH, CS2, DMTS, and DMSO than did raw and pasteurized milk. Hydrogen sulfide was the sulfur compound with the highest increase (up to 7-fold) in concentration. Methanethiol increased about 5-fold, and CS2 almost doubled. There was a significant difference between some pasteurized milk samples when compared with raw milk at the same fat level. Pasteurized sample B 1% fat showed higher concentrations of MeSH, CS2, and DMDS compared with raw and pasteurized brand A milk with 1% fat. Contrary to expectations, pasteurized samples A and B with 3.25% fat had lower concentrations of DMDS, DMTS, and DMSO than raw samples with 3.25% fat, whereas other sulfur compounds were not different. The “cooked” note perceived in pasteurized sample B with 3.25% fat was stronger compared with brand A and raw milks with 3.25% fat. In a previous work, we found that B 3.25% had a significantly higher concentration of DMS than raw and pasteurized brand A milk samples (Vazquez-Landaverde et al., 2005), whereas the thermally formed ketones and aldehydes were not significantly different.

Table 2. Concentration of volatile sulfur compounds in various milk samples
Compound2Milk samples1CV3
Raw 1%Raw 3.25%UHT 1%UHT 3.25%A 0%A 1%A 2%A 3.25%B 0%B 1%B 2%B 3.25%
H2S (μg/kg)1.78b1.21ab3.49c12.0d0.63a0.94ab0.59a0.87ab0.62a0.51a0.48a1.04ab<13.9
MeSH (μg/kg)1.95abc4.80d16.1e23.9f1.65ab1.92abc3.56bcd3.57bcd5.97d4.65d0.41a4.32cd<12.6
CS2 (ng/kg)17.1a22.1ab56.4e58.9e16.3a17.8ab18.0ab24.0b32.3c40.3d31.3c32.8c<10.5
DMDS (ng/kg)10.3b33.4ef30.3de32.8def2.24a10.21b17.1c28.3d9.37b35.8f10.1b17.2c<11.7
DMTS (ng/kg)10.7cd36.7h32.9h47.3i1.18a7.27bc13.8def16.6fg3.92ab12.2de15.7efg18.8g<8.3
DMSO (mg/kg)0.45d1.26g0.70e1.46h0.05a0.22b0.30bc0.68e0.08a0.41cd0.82f0.79ef<8.9
Me2SO2 (mg/kg)1.14cd2.64e0.78c1.26cd0.13a0.77c0.74bc1.39d0.23ab1.03cd4.72f2.21e<13.6

a–iDifferent letters for the same compound indicate significant difference between samples (Tukey HSD 95%).

1Raw 1% and Raw 3.25% = raw milk with 1 and 3.25% fat, respectively; UHT 1% and UHT 3.25% = UHT processed milk with 1 and 3.25% fat, respectively; A 0%, A 1%, A 2%, and A 3.25% = pasteurized milk of brand A with 0, 1, 2, and 3.25% fat, respectively; B 0%, B 1%, B 2%, and B 3.25% = pasteurized milk of brand B with 0, 1, 2, and 3.25% fat, respectively.

2H2S = Hydrogen sulfide; MeSH = methanethiol; CS2 = carbon disulfide, DMDS = dimethyl disulfide; DMTS = dimethyl trisulfide; DMSO = dimethyl sulfoxide; and Me2SO4 = dimethyl sulfone.

3CV = % Coefficient of variation.

  • View full-size image.
  • Figure 2. 

    Concentration of volatile sulfur compounds in commercial milk samples. a) Hydrogen sulfide, b) methanethiol, c) carbon disulfide, and d) dimethyl trisulfide. Raw 1% and Raw 3.25% = raw milk with 1 and 3.25% fat, respectively; UHT 1% and UHT 3.25% = UHT processed milk with 1 and 3.25% fat, respectively; A 0%, A 1%, A 2%, and A 3.25% = pasteurized milk of brand A with 0, 1, 2, and 3.25% fat, respectively; B 0%, B 1%, B 2%, and B 3.25% = pasteurized milk of brand B with 0, 1, 2, and 3.25% fat, respectively.

Because aroma impact is dependent not only on concentration, but also on sensory threshold, the odor activity value (OAV = concentration/sensory threshold) was calculated (Table 3). In this study, OAV values for MeSH and DMTS in UHT milk samples were much greater than 1, indicating that their concentrations were much higher than their sensory thresholds, thus contributing to the aroma. According to the magnitude of its OAV, MeSH could be the most important sulfur-containing contributor to the aroma of UHT milk, at a concentration 80 to 119 times higher than its reported threshold. The second most important contributor could be DMTS, with a concentration 4 to 6 times higher than its reported threshold. The concentration of H2S was slightly higher than its reported threshold only for one UHT sample, indicating that this compound could be a contributor to the sulfurous aroma in heated milk, albeit to a lesser degree than MeSH and DMTS. A previous study (Vazquez-Landaverde et al., 2005) suggested that DMS could be an important contributor to the aroma of both heated and fresh milk with an OAV of 3 to 11. The concentration of DMS was almost 3 times higher in UHT milk than in raw milk, whereas pasteurized brand B milk samples (2 and 3.25% fat) had significantly higher concentrations (P<0.05) of DMS than raw and pasteurized brand A milk. This trend with the severity of the heat treatment was in agreement with the concentration of the trace volatile sulfur compounds observed in this study.

Table 3. Odor activity values (concentration/reported threshold) of some volatile sulfur compounds in milk samples
Compound2Aroma threshold3Milk sample1
Raw 1%Raw 3.25%UHT 1%UHT 3.25%A 0%A 1%A 2%A 3.25%B 0%B 1%B 2%B 3.25%
H2S10μg/kg40.10.10.31.2<0.1<0.1<0.1<0.1<0.1<0.1<0.10.1
MeSH0.2μg/kg49.724801198.29.6171729232.021
CS25μg/kg5<0.1<0.1<0.1<0.1<0.1<0.1<0.1<0.1<0.1<0.1<0.1<0.1
DMDS0.16μg/kg4<0.10.20.10.2<0.1<0.10.10.1<0.10.2<0.10.1
DMTS8 ng/kg41.34.54.115.90.10.91.72.00.41.51.92.3

1Raw 1% and Raw 3.25% = raw milk with 1 and 3.25% fat, respectively; UHT 1% and UHT 3.25% = UHT processed milk with 1 and 3.25% fat, respectively; A 0%, A 1%, A 2%, and A 3.25% = pasteurized milk of brand A with 0, 1, 2, and 3.25% fat, respectively; B 0%, B 1%, B 2%, and B 3.25% = pasteurized milk of brand B with 0, 1, 2, and 3.25% fat, respectively.

2H2S = Hydrogen sulfide; MeSH = methanethiol; CS2 = carbon disulfide, DMDS = dimethyl disulfide; and DMTS = dimethyl trisulfide.

3Values measured in water.

4Rychilk et al., 1998.

5ETS Laboratories, 2002.

Volatile sulfur compounds may be present originally in milk at low concentrations because of the cow's metabolism (Toso et al., 2002). They can also be generated during the thermal processing of milk (Shibamoto et al., 1980; Steely, 1994). It has been reported that the volatile sulfur compounds formed during the high-temperature heating process are mainly responsible for the development of the cooked flavor defect (Christensen and Reineccius, 1992; Steely, 1994; Simon and Hansen, 2001; Datta et al., 2002). Upon heating, sulfur-containing amino acids from whey proteins, mainly β-LG, as well as proteins associated with the fat globule membrane (Parry, 1974; de Koning et al., 1990; Chan and Reineccius, 1994; Tressl et al., 1994; Zheng and Ho, 1994) can be released from the proteins and form sulfides, thiols, and other minor sulfur-containing compounds.

Methanethiol is important to the flavor of many foods due to its low detection threshold. It has a strong and unpleasant cabbage, sulfur-like aroma (Rychlik et al., 1998). A correlation between the increase in concentration of this compound and the increase in the “cooked” flavor defect has been reported previously (Badings et al., 1981; Christensen and Reineccius, 1992; Simon and Hansen, 2001). Methanethiol is thought to be generated from methionine during heat treatment by breakdown of the sulfur-bearing side chain (Damodaran, 1996) but the actual pathway has not been elucidated. Despite the importance of MeSH to the flavor of milk and dairy products, its study has been limited due to its high reactivity and volatility. Only few studies have reported reliable quantification techniques for this compound (Burbank and Qian, 2005a; Fang and Qian, 2005).

The formation of DMDS is probably due to the oxidation of MeSH (Ferreti, 1973; Chin and Lindsay, 1994). Jaddou et al. (1978) reported that DMDS concentration increased in UHT milk, but decreased in sterilized samples. Dimethyl disulfide and DMTS have a sulfur- and cabbage-like aroma (Rychlik et al., 1998) with low sensory threshold values.

The concentration of H2S in milk increases linearly with heating temperature (Hutton and Patton, 1952; Christensen and Reineccius, 1992). Also, the log concentration of H2S has a strong linear relationship with the heated flavor intensity of milk (Badings, 1978). This compound is also indirectly responsible for the formation of other sulfur compounds (Zheng and Ho, 1994). Hydrogen sulfide is produced mainly from sulphydryl groups of sulfur-containing amino acids (cysteine) in β-LG (Badings et al., 1981; Damodaran, 1996). Some authors suggested that this could be the most important contributor to the “cooked” flavor of milk because it is the major sulfur compound formed in heated milk (Badings et al., 1978; Jaddou et al., 1978; Badings et al., 1981; Rerkrai et al., 1987; Christensen and Reineccius, 1992) and it has a characteristic sulfur-like aroma (Rychlik et al., 1998). However, the OAV values calculated in this study indicate that H2S could be less important to the aroma of heated milk than previously thought. The OAV suggested that MSH and DMS were the 2 most important sulfur compounds in heated milk.

Although its origin has not been well elucidated, CS2 has been identified as a breakdown product of other sulfur compounds (Urbach, 1993). However, there are no previous reports of a correlation between CS2 and heat treatment, but our results showed that this compound could be a good indicator of heat treatment. It has a sweet, ethereal, slightly green, sulfur-like aroma (Rychlik et al., 1998) but its sensory threshold is very high and the calculated OAV was less than 1, so it most likely does not contribute to the aroma.

It has been proposed that Me2SO2 is produced in milk by the heat-induced oxidation of DMS via DMSO as the intermediate (Shibamoto et al., 1980). Under oxidative conditions, methionine is easily oxidized to methionine sulfoxide and methionine sulfone, which eventually breakdown to yield DMSO and Me2SO2, respectively (Damodaran, 1996). Shibamoto et al. (1980) found that Me2SO2 concentration decreases when milk is subjected to treatments between 60 and 90°C, but starts to increase considerably above 90°C. Moio et al. (1994) found that the Me2SO2 concentration was lower for UHT milk than for raw and pasteurized samples. Dimethyl sulfone has an aroma defined as hot milk, leather, and bovine sweat-like (Rychlik et al., 1998), but its sensory threshold is too high to be considered an important contributor to the aroma of milk.

A general trend was observed that the concentrations of H2S, MeSH, and DMTS increased with the fat level in milk (Table 2 and Figure 2). The results suggested that the concentration of sulfur compounds in milk was also affected by fat level. Although it is possible that the fat level may not have the same affect on the volatility of sulfur compounds and internal standard, it was more likely that the effects of heating on denaturation of the fat globule membrane protein were greater for samples with higher fat content, and that membrane proteins contributed to the formation of sulfides (de Koning et al., 1990).

Back to Article Outline

Conclusions 

The high reactivity and volatility of sulfur compounds makes their detection and quantification in foods a challenge. Proper sample manipulation and protection eliminated artifact formation from highly reactive sulfur compounds, allowing for the accurate quantification of volatile sulfur compounds in milk. The SPME/GC-PFPD technique developed in this study was effective in detecting a wide range of sulfur-containing compounds in milk at low concentrations, and was used successfully to study the sulfur compounds in milk subjected to different thermal processes.

Back to Article Outline

Acknowledgments 

The authors would like to thank Dairy Management Inc. (Rosemont, IL) for their financial support. P. Vazquez-Landaverde gratefully appreciates the support given by CONACYT (Mexico) for scholarship #142410.

Back to Article Outline

Supplementary data 

Download file

download text

Interpretive summary.

Back to Article Outline

References 

  1. Aguerre S, Lespes G, Desauziers V, Potin-Gautier M. Speciation of organotins in environmental samples by SPME-GC: Comparison of four specific detectors: FPD, PFPD, MIP-AES and ICP-MS. J. Anal. Atom. Spectrom. 2001;16:263–269
  2. Amirav A, Jing H. Pulsed flame photometer detector for gas chromatography. Anal. Chem. 1995;67:3305–3318
  3. Badings HT. Reduction of cooked flavor in heated milk and milk products. In:  Land DG,  Nursten HE editor. Progress in Flavor Research. London, UK: Applied Science Publishers; 1978;p. 263–265
  4. Badings HT, Neeter R, Van der Pol JJG. Reduction of heated flavor in heated milk and milk products by l-cystine. Lebensm. Wiss. Technol. 1978;11:237–242
  5. Badings HT, Van der Pol JJG, Neeter R. Aroma compounds which contribute to the difference in flavor between pasteurized milk and UHT milk. In:  Schreider P editors. Flavor ‘81. Berlin, Germany: Walter de Gruyter & Co.; 1981;p. 683–692
  6. Bendall JG. Aroma compounds of fresh milk from New Zealand cows fed different diets. J. Agric. Food Chem. 2001;49:4825–4832
  7. Block E, Calvey EM. Facts and artifacts in Allium chemistry. In:  Mussinan CJ,  Keelan ME editor. Sulfur Compounds in Foods. Washington, DC: American Chemical Society; 1994;p. 63–79
  8. Burbank HM, Qian MC. Quantification of volatile sulfur compounds in Cheddar cheeses of various age. Abstract 10, session 30. In: Proc. IFT Annual Meeting. New Orleans, LA. IFT, Chicago, IL. 2005;
  9. Burbank HM, Qian MC. Volatile sulfur compounds in Cheddar cheese determined by headspace solid-phase microextraction and gas chromatograph-pulsed flame photometric detection. J. Chromatogr. A. 2005;1066:149–157
  10. Burmeister MS, Drummond CJ, Pfisterer EA, Hysert DW, Sin YO, Sime KJ, et al. Measurement of volatile sulfur compounds in beer using gas chromatography with a sulfur chemiluminescence detector. J. Am. Soc. Brewing Chem. 1992;50:52–58
  11. Chan F, Reineccius GA. Kinetics o the formation of methional, dimethyl disulfide, and 2-acethylthiophene via the Maillard reaction. In:  Mussinan CJ,  Keelan ME editor. Sulfur Compounds in Foods. Washington, DC: American Chemical Society; 1994;p. 127–137
  12. Chin H-W, Lindsay RC. Ascorbate and transition-metal mediation of methanethiol oxidation to dimethyl disulfide and dimethyl trisulfide. Food Chem. 1994;49:387–392
  13. Christensen KR, Reineccius GA. Gas chromatographic analysis of volatile sulfur compounds from heated milk using static headspace sampling. J. Dairy Sci. 1992;75:2098–2104
  14. Contarini G, Povolo M, Leardi R, Toppino PM. Influence of heat treatment on the volatile compounds of milk. J. Agric. Food Chem. 1997;45:3171–3177
  15. Damodaran S. Amino acids, peptides, and proteins. In:  Fennema OR editors. Food Chemistry. 3rd. ed.. New York, NY: Marcel Dekker, Inc.; 1996;p. 412–413
  16. Datta N, Elliot AJ, Perkins ML, Deeth HC. Ultra-high-temperature (UHT) treatment of milk: Comparison of direct and indirect modes of heating. Aust. J. Dairy Technol. 2002;57:211–227
  17. de Koning PJ, Badings HT, Van der Pol JJG, Kaper J, Vos-Klompmacher EAJ. Effect of heat treatment and fat content on UHT milk. Voedingsmiddelentechnologie. 1990;23:11–14
  18. Dreher JG, Rousef RL, Naim M. GC-olfactometric characterization of aroma volatiles from the thermal degradation of thiamin in model orange juice. J. Agric. Food Chem. 2003;51:3097–3102
  19. ETS Laboratories. 2002. Subject: Sulphides in wine. http://www.etslabs.com/scripts/ets/pagetemplate/blank.asp?pageid=350 Accessed Feb. 4, 2005.
  20. Fan X, Sommers CH, Thayer DW, Lehotay SJ. Volatile sulfur compounds in irradiated precooked turkey breast analyzed with pulsed flame photometric detection. J. Agric. Food Chem. 2002;50:4257–4261
  21. Fang Y, Qian MC. Sensitive quantification of sulfur compounds in wine by headspace solid-phase microextraction technique. J. Chromatogr. A. 2005;1080:177–185
  22. Ferreti A. Inhibition of cooked flavor in heated milk by use of additives. J. Agric. Food Chem. 1973;21:939–942
  23. Haberhauer-Troyer C, Crnoja M, Rosenberg E, Grasserbauer M. Surface characterization of commercial fibers for solid-phase microextraction and related problems in their application. J. Anal. Chem. 2000;366:329–331
  24. Hill PG, Smith RM. Determination of sulfur compounds in beer using headspace solid-phase microextraction and gas chromatographic analysis with pulsed flame photometric detection. J. Chromatogr. A. 2000;872:203–213
  25. Hutton JT, Patton S. The origin of sulfhydryl groups in milk proteins and their contributions to “cooked” flavor. J. Dairy Sci. 1952;35:699–705
  26. Jaddou HA, Pavey JA, Manning DJ. Chemical analysis of flavor volatiles in heat-treated milks. J. Dairy Res. 1978;45:391–403
  27. Jing H, Amirav A. Pesticide analysis with the pulsed-flame photometer detector and direct sample introduction device. Anal. Chem. 1997;69:1426–1435
  28. Kastner JR, Das KC, Hu C, McClendon R. Effect of pH and temperature on the kinetics of odor oxidation using chlorine dioxide. J. Air Waste Manag. Assoc. 2003;53:1218–1224
  29. Lestremau F, Desauziers V, Roux J-C, Fanlo J-L. Development of a quantification method for the analysis of malodorous sulphur compounds in gaseous industrial effluents by solid-phase microextraction and gas chromatography-pulsed flame photometric detection. J. Chromatogr. A. 2003;999:71–80
  30. Metrohm. 2001. Subject: Titrimetric/potentiometric determination of hydrogen sulfide, mercaptans and carbonyl sulfide in natural gas. http://www.metrohm.co.uk Accessed January 7, 2005.
  31. Moio L, Etievant PX, Langlois D, Dekimpe J, Addeo F. Detection of powerful odorants in heated milk by use of extract dilution sniffing analysis. J. Dairy Res. 1994;61:385–394
  32. Murray RA. Limitations to the use of solid-phase microextraction for quantitation of mixtures of volatile sulfur compounds. Anal. Chem. 2001;73:1646–1649
  33. Mussinan CJ, Keelan ME. Sulfur Compounds in Foods. An Overview. In:  Mussinan CJ,  Keelan ME editor. Sulfur Compounds in Foods. Washington, DC: American Chemical Society; 1994;p. 1–6
  34. Nielsen AT, Jonsson S. Trace determination of volatile sulfur compounds by solid-phase microextraction and GC-MS. Analyst. 2002;127:1045–1049
  35. Parry RM. Milk coagulation and protein denaturation. In:  Webb BH,  Johnson AH,  Alford JH editor. Fundamentals of Dairy Chemistry. Westport, CT: AVI Publishing; 1974;p. 127–128
  36. Pelusio F, Nilsson T, Montanarella L, Tilio R, Larsen B, Facchetti S, et al. Headspace solid-phase microextraction analysis of volatile organic sulfur compounds in black and white truffle aroma. J. Agric. Food Chem. 1995;43:2138–2143
  37. Rerkrai S, Jeon IJ, Bassette R. Effect of various direct ultra-high temperature heat treatments on flavor of commercially prepared milks. J. Dairy Sci. 1987;70:2046–2054
  38. Rychlik M, Schieberle P, Grosch W. Compilation of Odor Thresholds, Odor Qualities and Retention Indices of Key Food Odorants. Garching, Germany: Deutsche Forschungsanstalt fur Lebensmittelchemie and Institut fur Lebensmittelchemie der Technischen Universitat Munchen; 1998;
  39. Scanlan RA, Lindsay R, Libbey LM, Day EA. Heat-induced volatile compounds in milk. J. Dairy Sci. 1968;51:1001–1007
  40. Shibamoto T, Mihara S, Nishimura O, Kamiya Y, Aitoku A, Hayashi J. Flavor volatiles formed by heated milk. In:  Charalambous G editors. The Analysis and Control of less Desirable Flavors in Foods and Beverages. New York, NY: Academic Press, Inc.; 1980;p. 260–263
  41. Shipe WF. Analysis and control of milk flavor. In:  Charalambous G editors. The Analysis and Control of Less Desirable Flavors in Foods and Beverages. New York, NY: Academic Press, Inc.; 1980;p. 201
  42. Simon M, Hansen AP. Effect of various dairy packaging materials on the shelf life and flavor of ultrapasteurized milk. J. Dairy Sci. 2001;84:784–791
  43. Steely JS. Chemiluminescence detection of sulfur compounds in cooked milk. In:  Mussinan CJ,  Keelan ME editor. Sulfur Compounds in Foods. Washington, DC: American Chemical Society; 1994;p. 22–35
  44. Steffen A, Pawliszyn J. Analysis of flavor volatiles using headspace solid-phase microextraction. J. Agric. Food Chem. 1996;44:2187–2193
  45. Toso B, Procida G, Stefanon B. Determination of volatile compounds in cows’ milk using headspace GC-MS. J. Dairy Res. 2002;69:569–577
  46. Tressl R, Kersten E, Nittka C, Rewicki D. Formation of sulfur containing flavor compounds from [13C]-labeled sugars, cysteine, and methionine. In:  Mussinan CJ,  Keelan ME editor. Sulfur Compounds in Foods. Washington, DC: American Chemical Society; 1994;p. 224–235
  47. Urbach G. Relations between cheese flavor and chemical composition. Int. Dairy J. 1993;3:389–422
  48. Vazquez-Landaverde PA, Velazquez G, Torres JA, Qian MC. Quantitative determination of thermally derived volatile compounds in milk using solid-phase microextraction and gas chromatography. J. Dairy Sci. 2005;88:3764–3773
  49. Wercinski SAS, Pawliszyn J. SPME Theory. In:  Wercinski SAS editors. Solid phase microextraction: A practical guide. New York, NY: Marcel Dekker, Inc.; 1999;p. 1–24
  50. Zheng Y, Ho C. Kinetics of the release of hydrogen sulfide from cysteine and glutathione during thermal treatment. In:  Mussinan CJ,  Keelan ME editor. Sulfur Compounds in Foods. Washington, DC: American Chemical Society; 1994;p. 138–146

PII: S0022-0302(06)72564-4

doi:10.3168/jds.S0022-0302(06)72564-4

Journal of Dairy Science
Volume 89, Issue 8 , Pages 2919-2927, August 2006

Article Tools

* Image Usage & Resolution