Quality changes during storage of cooked and sliced meat products measured with PTR-MS and HS-GC–MS

Meat Science 95 (2013) 302–310

Contents lists available at SciVerse ScienceDirect

Meat Science journal homepage: www.elsevier.com/locate/meatsci

Quality changes during storage of cooked and sliced meat products measured with PTR-MS and HS-GC–MS E.S. Holm a,⁎, A.P.S. Adamsen b, A. Feilberg b, A. Schäfer c, M.M. Løkke b, M.A. Petersen a a b c

Department of Food Science, Quality & Technology, Faculty of Science, University of Copenhagen, Rolighedsvej 30, 1958 Frederiksberg C, Denmark Department of Engineering, Air Quality Engineering, Faculty of Science and Technology, Aarhus University, Blichers Allé 20, 8830 Tjele, Denmark DMRI, Danish Technological Institute, Maglegårdsvej 2, 4000 Roskilde, Denmark

a r t i c l e

i n f o

Article history: Received 26 November 2012 Received in revised form 26 March 2013 Accepted 15 April 2013 Keywords: Volatile organic compounds Proton-Transfer-Reaction Mass-Spectrometry Cooked and sliced meat products Spoilage

a b s t r a c t The changes in the VOC composition of industrially produced saveloy were measured with Proton-TransferReaction Mass-Spectrometry (PTR-MS) and HeadSpace Gas chromatography–mass spectrometry (HS-GC–MS) during a six weeks storage period. A decrease in the volatile organic compounds contributing to the fresh aroma of saveloy was the main change observed with both PTR-MS and HS-GC–MS. Samples of four other types of cooked and sliced meat product were measured with PTR-MS in the middle and at the end of the four week shelf-life period. These measurements showed an increase in m/z 69, 71, 87 and 89 for the pork loin and in m/z 61 for the herbal saveloy samples. These ions were assigned to the microbial spoilage markers: acetic acid, 2- and 3-methylbutanol, 2- and 3-methylbutanal, diacetyl and acetoin. Overall, this study shows that PTR-MS has potential for quality control of cooked and sliced meat products. © 2013 Elsevier Ltd. All rights reserved.

1. Introduction Odor and aroma are important parameters for consumer acceptability, and measurement of the composition of volatile organic compounds (VOCs) is therefore a useful tool for evaluation of the quality of cooked and sliced meat products. The aroma of fresh cooked and sliced meat products derives mainly from thermally induced lipid oxidation during cooking and from the spices added to the product (Ho, Oh, & Bae-Lee, 1994; Holm, Schäfer, Skov, Koch, & Petersen, 2012; Mottram, 1998). However, during slicing and further processing the product is subjected to the in-house microbial flora of the processing facility (Gounadaki, Skandamis, Drosinos, & Nychas, 2008; Samelis, Kakouri, & Rementzis, 2000). This flora is often dominated by Lactic Acid Bacteria (LAB) typically in combination with Brochothrix thermosphacta and Pseudomonas spp. (Bjorkroth, Vandamme, & Korkeala, 1998; Borch, KantMuermans, & Blixt, 1996; Holm et al., 2012). During storage these bacteria metabolize the nutrients present on the surface of the product and start producing unpleasant odors. Furthermore, the amount of the VOCs contributing to the aroma of the fresh product starts to decrease. These changes in the volatile profile of the product will have a negative impact on the sensory quality (Borch et al., 1996; Dainty & Mackey, 1992; Leroy, Vasilopoulos, Van Hemelryck, Falony, & De Vuyst, 2009). In a recent study the microbially produced VOCs: 2- and 3-methylbutanol, 2- and 3-methylbutanal, diacetyl and acetoin, measured by HeadSpace Gas chromatography–

⁎ Corresponding author. Tel.:+45 28726541; fax: +45 72202744. E-mail address: [email protected] (E.S. Holm). 0309-1740/$ – see front matter © 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.meatsci.2013.04.046

mass spectrometry (HS-GC–MS), have been suggested as chemical markers for the sensory spoilage of sliced saveloy (Holm et al., 2012). Measurements of these chemical markers in the fresh meat product may provide an estimate of the expected shelf-life and can also be used as an index of the current microbial state of the product. Proton-Transfer-Reaction Mass-Spectrometry (PTR-MS) may be a suitable method for rapid at- or on-line measurements of the chemical markers during industrial production of cooked and sliced meat products. The major advantages of PTR-MS are its ability to provide direct, fast and continuous measurement of the VOCs in the headspace of the product. By using the PTR-MS technique it is furthermore possible to detect VOCs present in pptv-levels (Lindinger, Hansel, & Jordan, 1998). The PTR-MS instrument is equipped with a drift tube in which the VOCs are subjected to soft chemical ionization by protonated water. Although the PTR-MS relies on soft ionization some degree of fragmentation of VOCs such as aldehydes, alcohols and carboxylic acids will occur (Brown, Watts, Märk, & Mayhew, 2010; Buhr, van Ruth, & Delahunty, 2002). Furthermore, PTR-MS does not include a GCseparation step. Therefore all VOCs obtained from the sample headspace enter the MS simultaneously (Lindinger et al., 1998). In combination with the VOC fragmentation, this presents a challenge to the data processing and identification of the VOCs present in the headspace of complex foods. Nevertheless, PTR-MS has many interesting applications in food research. These include on-line measurement of VOC development during fermentation processes and classification of various foods according to geographical origin based on their PTR-MS fingerprint (Biasioli, Yeretzian, Gasperi, & Märk, 2011). In the present study, changes in the VOC composition of industrially produced saveloy were monitored by PTR-MS and HS-GC–MS during

E.S. Holm et al. / Meat Science 95 (2013) 302–310

a six weeks storage period. Furthermore, four other types of commercially cooked and sliced meat products were investigated. The VOC composition of these products was measured with PTR-MS in the middle and at the end of their four week shelf-life period. The overall objective of this paper was to evaluate PTR-MS as a tool for detection of quality changes in cooked and sliced meat products with focus on previously identified chemical markers. A positive result would be an important step towards a practical application of chemical markers for early spoilage detection in the meat industry. 2. Materials and methods

303

standard proton transfer reaction rate coefficient of 2 × 10−9 cm3/s (Lindinger et al., 1998) was used to convert measured signals to concentration levels in parts per billion by volume of the headspace (ppbv, nL of volatile compound/L of air). The fragmentation patterns of the following authentic standards were studied via PTR-MS measurement in order to support the assignment of PTR-MS ion fragments in the samples of cooked and sliced meat products: Hexanal (Merck KGaA, Darmstadt, Germany), acetoin (ChemService Inc., West Chester, PA, USA), 2-methylbutanal, diacetyl (Sigma-Aldrich, St. Louis, MO, USA) and α-pinene (Acros Organics, Geel, Belgium). The intensity of the protonated molecular ion (M + 1) and the most important fragments are shown in Table 1.

2.1. Experimental setup 2.3. HS-GC–MS measurement For the first part of the experiment, sliced and Modified Atmosphere Packed (MAP) saveloy samples were supplied by a commercial meat processing facility. Saveloy is a gently seasoned sausage made from minced pork meat. During a six weeks storage period the saveloy samples were kept at 5 °C and the changes in VOC composition were measured with PTR-MS and HS-GC–MS. After 3, 4 and 5 weeks of storage, a series of saveloy packages were opened and kept at 5 °C for another four or six days. This allowed oxygen to enter the packages and affect the microbial metabolism. Measurements of the VOC composition with PTR-MS were done at day 1, week 3, week 3 + 4 days (+4 days indicates 4 days after package opening), week 3 + 6 days, week 4 + 6 days, week 5, week 5 + 4 days and week 5 + 6 days. HS-GC–MS measurements were done at day 1, week 4, week 4 + 4 days, week 4 + 6 days and week 5 in order to support the assignment of the PTR-MS ion fragments. A sensory experiment was done at day 1, week 4, week 4 + 4 days and week 5 to establish whether the changes in VOC composition affected the sensory perception of the product. For the second part of the experiment packages of sandwich ham, pork loin, ‘herbal saveloy’ and ‘pork flank sausage’ were purchased in a supermarket. Herbal saveloy (In Danish: Jægerpølse) is a more seasoned version of saveloy which contains garlic and chives (Allium schoenoprasum), whereas the pork flank sausage (In Danish: Rullepølse) is made from rolled and cooked pork flank which in this case was seasoned with chives and ramson (wild garlic, Allium ursinum). The VOC composition of these four products was measured with PTR-MS in the middle of their four week shelf-life period, and again near their expiration date after being subjected to atmospheric air via package opening during the final week. 2.2. PTR-MS measurement About 25 g of sample was coarsely chopped and placed in a 500 mL closed glass flask equipped with a purge top, and conditioned in a water bath at 30 °C for 10 min. During the measurement a 150 mL/min flow of filtered air (120 cm 3 Supelpure HC Hydrocarbon Trap, Sigma-Aldrich, St. Louis, MO, USA) was let through the sample flask and then diluted with a filtered air flow of 300 mL/min. The purge and dilution air flow were controlled by mass flow controllers (Sierra Instruments Inc., CA, USA). The inlet flow of the high sensitivity PTR-Quadrupole-MS (Ionicon Analytik, Innsbruck, Austria) was ~84 mL/min and the excess air flow was discharged. The PTR-MS drift tube was operated at standard conditions with a temperature of 60 °C, a pressure of 2.14–2.20 mbar and voltage of 600 V. The E/N ratio was ~137 Td for all measurements. Here E represents the electric field strength and N is the buffer gas density (Lindinger et al., 1998). The MS scanned a m/z range from 21 to 200 with a dwell time of 200 ms. Ten cycles of each sample was recorded and the average values from the intermediate cycles of each m/z were subtracted from the background measurements made on an empty glass flask. For each treatment five independent repetitions of the PTR-MS measurements were done using a new package of meat product for each replication. A

VOCs were extracted from the sample by dynamic headspace extraction coupled with GC–MS using the method described by Holm et al., 2012. The samples were conditioned in a closed glass container on a 30 °C water bath for 10 min, and then purged with a N2-flow of 60 mL/min for 15 min. The VOCs in the sample were retained on a sorbent trap which contained a 73 mg Tenax TA (mesh 60–80) and 100 mg carbograph 1 TD (Markes International Ltd., Llantrisant, UK). The traps were back purged with a flow of 20 mL/min for the removal of water. An ATD 400 automatic thermal desorption system (Perkin Elmer, Waltham, MA, USA) was used to desorp and cryofocus the sample before loading it on the column. For each measure point three independent replications of the dynamic headspace extraction and GC–MS analysis were performed using a new package of meat product for each measurement. Measurement of the following authentic standard compounds were done with HS-GC–MS to aid the identification of VOCs in the sample headspace: 2-methylbutanol and hexanal (Merck KGaA, Darmstadt, Germany), 3-methylbutanal and acetoin (ChemService inc., West Chester, PA, USA), 1-octen-3-ol, acetic acid, diacetyl, 2-methylpropanol, 2-heptanol and 1-hexanol (Sigma-Aldrich, St. Louis, MO, USA), octanal, 2-heptanone, heptanal, α-pinene, limonene, 3-carene, dimethyl disulfide and dimethyl trisulfide (Acros Organics, Geel, Belgium) and 2-pentylfuran (Lancaster Synthesis, Windham NH, USA). The retention times of the chromatographic peaks were standardized using the Kovats Linear Retention Index (LRI) (Kováts, 1958) calculated from HS-GC–MS runs of a C5–C15 alkane standard (Air Liquide, Paris, France). The HS-GC–MS data was processed using MSD Chemstation software (D.01.02.16, Agilent Technologies, Santa Clara, CA, USA). The chromatographic peaks were integrated based on target and qualifier ions characteristic for their mass spectra. Due to co-elution 2-methylbutanol and 3-methylbutanol were quantified together. Identification of the chromatographic peaks was done using the NIST/EPA/NIH mass spectral library (V.1.7a, Agilent Technologies, Santa Clara, CA, USA) and HS-GC–MS runs of authentic standards. The LRI of chromatographic peaks was moreover compared with the LRI of Table 1 Fragmentation patterns observed for PTR-MS measurement (E/N ~ 137 Td) of five authentic standards. The main fragments of the standards are provided along with their intensity, shown in parenthesis, relative to the most abundant fragment. Compound

Molecular mass (g/mol)

Protonated molecular ion (M + 1) and relative intensity (%)

Main fragments and relative intensities (%)

Hexanal Acetoin

100 88

101 (4) 89 (8)

55 43 55 45 69 59 43 81

2-Methylbutanal Diacetyl α-Pinene

86

87 (75)

86 136

87 (100) 137 (77)

(100), 83 (82) (17), 45 (100), (27), 87 (45) (100), 55 (11), (18) (39), 73 (11) (32), 45 (53), (100), 7 (25)

304

E.S. Holm et al. / Meat Science 95 (2013) 302–310

potential matching compounds using the C20M column from internet database www.flavornet.org (Acree & Arn, 2004). The maximum difference in LRI allowed for a potential match was 50 units in the LRI. However, not all the three identification methods were used for each compound. The following denotation will be used: N (NIST mass spectral library), S (compound standards) and L (comparison of LRI). 2.4. Sensory test A sensory test was included in the storage experiment on the saveloy samples. The test was performed with a panel consisting of ten people experienced in working with meat and meat spoilage. Four times during the six weeks storage experiment the sensory quality of the saveloy samples was compared to a reference sample stored at −1 °C. In previous experiments in our sensory laboratory sliced saveloy stored at −1 °C has been shown to be stable from a sensory perspective during a four weeks storage period (data not shown). The difference in odor and taste between the saveloy sample and the reference was evaluated on a scale from 0 to 5, where 0 was no deviation and 5 was strong deviation. The panel furthermore had the opportunity to comment on the observed changes. The sensory test was done on day 1, week 4, week 4 + 4 and week 5 of the saveloy storage experiment. 2.5. Data processing Chemometric analysis was done using the PLS Toolbox (version 5.2.2, Eigenvector Research Inc., Wenatchee, WA, USA). The PLS toolbox runs in the MATLAB environment (version 7.6.0.324, The Mathworks Inc., Natick, MA, USA). For the Principal Component Analysis (PCA) different preprocessing techniques were tested to optimize the visualization of the data. Auto-scaling proved to be the best preprocessing technique for the PTR-MS data as well as the HS-GC–MS data. Statistical analysis was performed in JMP v. 8.0.1 (SAS institute, Cary, NC, USA). For the HS-GC–MS data and the PTR-MS data ANOVA models were built describing the relevant VOC variables using ‘storage time’ as fixed effect. Based on the ANOVA models Tukey Honest Significant Different (HSD) tests were done to locate significant differences (on a 95% level) between the levels of the relevant variables. 3. Results and discussion 3.1. Changes in VOC composition of sliced saveloy measured with HS-GC–MS A total of 46 VOCs were isolated from the sliced saveloy samples with HS-GC–MS during the storage experiment. Twenty one of these were terpenes, however, other compound classes such as aldehydes, ketones, alcohols, and sulfur containing compounds were also observed. The complete list of VOCs extracted from sliced saveloy and their peak areas are shown in Table 2. In Table 2 it is seen that the aroma of the fresh saveloy samples is roughly composed of VOCs formed by thermal degradation of lipids during cooking and VOCs from the flavoring ingredients added to the product. Several of the straight chain aldehydes, ketones and alcohols found in the product headspace are well-known lipid degradation products (Ho et al., 1994; Mottram, 1998). Table 2 furthermore shows that the peak area of some of these lipid oxidation products, including hexanal and 2-butanone, decreased significantly with storage time. Similar decreases were observed for several of the terpenes and the sulfur containing compounds shown in Table 2. The terpenes originate from the spices added to the product whereas several of the sulfur containing compounds could derive from onion which was also added to the product (Järvenpää et al., 1998; Løkke et al., 2012; Meynier, Novelli, Chizzolini, Zanardi, & Gandemer, 1999). Overall, this means that the VOCs which are assumed to contribute to the fresh aroma of saveloy decrease with storage time.

Previous studies have shown that the shelf-life of cooked and sliced meat products is determined mainly by microbial reactions (Holm et al., 2012; Leroy et al., 2009; Samelis et al., 2000). However, in Table 2 it is seen that none of the VOCs extracted from the saveloy samples and measured with HS-GC–MS increased significantly with storage time in this study. This implies that the level of microbially produced VOCs was very limited. An overview of the distribution of the saveloy samples based on their VOC composition measured with HS-GC–MS is given in Fig. 1. This figure shows the bi-plot from a PCA model including the saveloy samples and all 46 VOCs. This bi-plot shows Principal Component (PC) 1 vs. PC2 which accounts for 66% of the variation in the VOC composition of the saveloy samples. The first PC accounts for the main variation in the dataset. On this PC the saveloy samples measured at day 1 and week 4 have positive scores whereas the vast majority of the samples from week 4 + 4 days, week 4 + 6 days and week 5 have negative scores. The majority of the VOCs included in the PCA model are associated with the saveloy samples from day 1 and week 4. This indicates that these VOCs generally were present in the highest level at day 1 and week 4. However, the peak areas of acetic acid (Ac1) and 2,2-dimethylpropanoic acid (Ac2) were associated with the samples from week 4 + 4 days, week 4 + 6 days and week 5. Acetic acid is a known metabolite of LAB, which has been related to spoilage of meat products during storage (Dainty & Mackey, 1992; Laursen, Byrne, Kirkegaard, & Leisner, 2009). However, as seen in Table 2, this increase in acetic acid and 2,2dimethyl-propanoic acid was not significant in this experiment. Acetoin (K3), which is a known B. thermosphacta metabolite (Dainty & Hibbard, 1983), was located right between the samples from day 1 and the samples from week 4 + 4 days and 4 + 6 days in the bi-plot in Fig. 1. Overall, these observations suggest that the microbial contribution to the VOC composition is beginning to increase, though not significantly, at week 4 + 4 days and week 4 + 6 days. 3.2. Changes in the VOC composition of sliced saveloy measured with PTR-MS The results of the PTR-MS measurements made on the sliced saveloy samples during the storage experiment are shown in Table 3. The list of VOCs contributing to specific ions shown in Table 3 was made based on fragmentation patterns reported in the literature (Brown et al., 2010; Buhr et al., 2002; Feilberg, Liu, Adamsen, Hansen, & Jonassen, 2010; Lasekan & Otto, 2009; Løkke et al., 2012; Maleknia, Bell, & Adams, 2007), along with fragmentation patterns recorded in this study, and comparison with GC–MS measurements. The HS-GC–MS measurements of the VOC composition showed a decrease in the level of the terpenes, the sulfur compounds and the lipid oxidation products with storage time. This pattern was also observed in the PTR-MS measurements. A significant decrease in m/z 81 and 137 was observed with storage time. These are the major ion fragments of monoterpenes, which have a molecular mass of 136 g/mol. The protonated molecular ion (M + 1) of the monoterpenes is m/z 137 whereas m/z 81 corresponds to the loss of a C4H8 fragment (Maleknia et al., 2007). The m/z 49, 63 and 95 which were assigned to the M + 1 of methanethiol, dimethyl sulfide (DMS) and dimethyl disulfide (DMDS) respectively also decreased significantly with storage time. As seen in Table 3, m/z 49 was found in high amounts, 600 ppb at day 1, compared to m/z 63 and 95. Methanethiol, found at m/z 49, is a highly volatile compound which is easily transformed to DMDS during HSGC–MS analysis. There is furthermore a high risk of breakthrough of methanethiol on the sorbent traps during the headspace sampling procedure (Andersen, Hansen, & Feilberg, 2012). Methanethiol is therefore a good example of a VOC, which is poorly quantified by HSGC–MS but quantified well by PTR-MS. The fragmentation patterns of aldehydes and alcohols, which are predominant lipid oxidation products, were studied by Buhr et al. (2002). Upon protonation aldehydes and alcohols undergo fragmentation by

E.S. Holm et al. / Meat Science 95 (2013) 302–310

305

Table 2 Overview of the volatile organic compounds (VOCs) extracted from the saveloy samples with HS-GC–MS. The table includes the peak area/1000 of the 46 VOCs for each of the five days of measurement, the linear retention index (LRI) and the target ion used for quantification. HS-GC–MS measurements were done at day 1 (D1), week 4 (W4), week 4 + 4 days (W4 + 4), week 4 + 6 days (W4 + 6) and week 5 (W5). The VOCs were identified using the NIST mass spectral database (N), comparisons of the observed LRI with the LRI in internet database flavornet.com (L) or HS-GC–MS runs of authentic standard compounds (S). Letters a to d are used to indicate significant differences in the peak area of the VOCs between measure points based on a Tukey HSD test. Nr.

Code

Compound name

LRI

Id-quality

T-ion

D1

W4

W5

W4 + 4

W4 + 6

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46

A1 K1 E1 K2 S1 Ah1 S2 T1 T2 T3 S3 A2 T4 T5 T6 T7 T8 T9 Ah2 T10 T11 T12 S4 Ah3 F1 T13 Ah4 T14 Ah5 T15 K3 Ah6 S5 S6 A3 S7 T16 E2 Ah7 Ac1 T17 T18 T19 Ac2 T20 T21

Propanal Acetone Ethyl acetate 2-Butanone 1-(Methylthio)-propane Ethanol 1-(Methylthio)-(Z)-1-propene α-Pinene α-Thujene Camphene Dimethyl-disulfide Hexanal β-Pinene Sabinene 3-Carene α-Phellandrene β-Myrcene (+)-4-Carene 1-Ethoxy-2-propanol Limonene β-Phellandrene Eucalyptol Methyl-propyl-disulfide 2- and 3-methyl-1-butanol 2-Pentylfuran γ-Terpinene 3-Methyl-3-buten-1-ol ρ-Cymene 1-Pentanol δ-Terpinene Acetoin 1-Hexanol Dimethyl-trisulfide Dipropyl-disulfide Nonanal 1,2-Dithiolane p-α-Dimethylstyrene Ethyl octanoate 1-Octen-3-ol Acetic acid Cis-β-terpineol Camphor Linalool 2,2-Dimethyl-propanoic acid Terpinen-4-ol Safrole

765 802 865 884 903 941 999 1004 1009 1040 1050 1068 1078 1094 1116 1133 1140 1148 1164 1166 1174 1182 1199 1200 1203 1212 1234 1238 1242 1248 1265 1331 1341 1343 1368 1400 1407 1409 1427 1429 1452 1488 >1488 >1488 >1488 >1488

N N N–L N N N–L–S N N–L–S N–L L–N N–L–S N–L–S N–L N–L N–L–S N–L N–L N N N–L–S N–L N–L N N–L–S N–L–S N–L N N–L N–L N–L N–L–S N–L–S N–L–S N N–L N N–L N–L N–L–S N–L–S N N–L N N N N

58 43 43 43 61 45 88 93 93 93 94 56 93 93 93 93 93 121 45 68 93 43 122 55 81 93 41 119 55 93 45 56 126 43 57 148 132 88 57 43 71 95 71 57 71 162

8000 28000a 93000a 4800a 1500a 500000a 4200a 20000a 52,000 1200a 30000a 900a 12,200 17000a 2300a 1200ab 3200a 260 950a 5000a 5900 1100a 9000a 150a 160a 6000 110a 11,000 200 1400 600 290a 20000a 4000a 110 310a 700 200 130 100 300a 80 550a 0 960a 800

7000 23000b 73000b 4000ab 1300ab 400000b 3600ab 16000ab 5000 700ab 23000ab 800ab 10,000 8000ab 1800ab 1500ab 2300ab 2800 510b 4000ab 6000 900ab 6000ab 60ab 126a 6000 92ab 8000 90 1600 480 250ab 4000b 2500b 90 120b 500 190 90 600 0b 63 290bc 7 670ab 600

3900 20000b 64000b 3200bc 1150b 380000bc 3110b 10000ab 2000 600ab 19000ab 720b 6000 4000b 1300b 1800a 1700b 2400 440bc 3300ab 5000 680b 2900bc 20b 112ab 6000 73ab 7000 50 1500 430 220b 1500b 540c 100 10c 300 173 84 400 0b 58 200c 300 500b 550

20,000 14000c 39000c 2300cd 600c 350000bc 1950c 8000ab 3000 500ab 2400b 790ab 7000 8000ab 1300b 700b 1700b 1500 460bc 2800b 4100 660b 1000c 104ab 30b 4000 90ab 4200 130 1000 1200 240ab 590b 400c 90 10c 170 170 70 1200 30b 68 463ab 0 810a 500

2000 10000c 26000d 1500d 360c 320000c 1320c 6400b 2000 250b 1100b 640b 5200 6000b 1000b 730b 1400b 1400 420c 2500b 3800 610b 430c 40b 101ab 3700 60b 5000 140 900 420 220b 260b 170c 90 0c 280 160 92 200 7b 57 370abc 400 690ab 510

water elimination, in some cases followed by further reactions, e.g. elimination of H2. The high fragmentation levels of these compounds, make it difficult to relate them to specific PTR-MS ions. The characteristic fragment ions of aldehydes were reported to be m/z 55 and 69, whereas fragment ions of alcohols often contribute to m/z 43 and 57 (Brown et al., 2010; Buhr et al., 2002). In this study no clear tendency was observed for the changes in m/z 57 and m/z 69 was observed, and no significant changes were observed for m/z 43 and m/z 55. The PTR-MS measurements also showed that ethanol was by far the most abundant compound in the headspace of sliced saveloy. The protonated molecular ion of ethanol (m/z 47) was found in levels corresponding to between 1900 ppb and 7000 ppb and the fragment ions of ethanol at m/z 29 (C2H5+) and m/z 45 were also found in high amounts. It should be noted that the PTR-MS measurements of ethanol to some extent will underestimate the concentration, due to significant fragmentation of protonated ethanol. In addition, the C2H5+ fragment is likely to undergo proton transfer to water, since the proton affinity of ethylene is lower than that of water. Ethanol was also found to be an abundant peak in the HS-GC–MS measurements.

An overview of the development in the VOC composition measured with PTR-MS is provided in Fig. 2. This figure shows the bi-plot (PC1 vs. PC2) from a PCA model including the PTR-MS ions shown in Table 3. PC1 and PC2, which are shown in the bi-plot, account for 66% of the variation in the data. The variation between the five different packages of saveloy measured on the same day is outlined in the figure. This enables an evaluation of the variation in the VOC composition of the saveloy packages stored at the same conditions. The samples from day 1, week 3 and week 5, which were kept in closed packages until measurement, were generally characterized by positive scores on PC1. This was most pronounced for the samples from day 1 which had the highest scores on PC1. The majority of the PTR-MS ions had the highest intensity in the samples stored in closed packages from day 1, week 3 and week 5. This includes m/z 81 and 137 from the terpenes along with m/z 49, 63 and 95 from methanethiol, DMS and DMDS respectively. In Fig. 2 it is seen that the samples subjected to package opening four or six days before measurements generally had negative scores on PC1. These samples contained higher levels of m/z 31, 45 and 46. For the samples measured in week 5 + 4 days and week 5 + 6 days one of the five

306

E.S. Holm et al. / Meat Science 95 (2013) 302–310

0.3

T19

0.2 T20 Ah4

Ah2 T17

PC 2 (13%)

0.1

Week4+4days Week4+6days

T5 A3 Ah3 Ah5 S7 Day 1 T21 T18S6 A2 S5 T12 Ah1 Ah6 T4 T8 T6 S4 A4 T3 K1 T16 T1 T14 T10 T1E1 K2 F1 S3 E2 S2 S1

K3 A1

Ac1

0 Ac2

Week4

-0.1 Week 5

T11 T9 T13 T15

-0.2 T7

-0.3 -0.3

-0.2

-0.1

0

0.1

0.2

0.3

0.4

0.5

PC 1 (53%) Fig. 1. Bi-plot (PC1 vs. PC2) from the PCA model including the 46 volatile organic compounds (VOCs) found with HS-GC–MS measurements of the saveloy samples at day 1, week 4, week 4 + 4 days, week 4 + 6 days and week 5. The data was auto-scaled for before calculating the PCA model. The VOCs are denoted by the compound code given in Table 2.

Table 3 The concentration in ppb of selected m/z-ratios found in the saveloy samples measured with PTR-MS at day 1 (D1), week 3 (W3), week 3 + 4 days (W3 + 4), week 3 + 6 days (W3 + 6), week 4 + 6 days (W4 + 6), week 5 (W5), week 5 + 4 days (W5 + 4) and week 5 + 6 days (W5 + 6). Letters a to e are used to indicate significant differences in the concentration of PTR-MS ion fragments between measure points based on a Tukey HSD test. Assignments of the ion fragments are provided based on compound standards and the existing literature. m/z

I

D1

W3 a

W5 ab

W3 + 4 bc

bc

W3 + 6 d

W4 + 6

W5 + 4

cd

bcd

W5 + 6 bc

29 31 33 35 39 41 43 45 46 47 48 49 51 55 57

180 33c 700a 30a 20 67 190 370c 10d 7000a 200a 600a 40a 50 12a

150 30c 670ab 4b 16 48 120 300c 8d 3400ab 84ab 180b 11b 9 8b

130 110abc 650ab 2b 30 35 140 800b 24b 5000ab 120ab 100bc 14b 60 10ab

130 91abc 640ab 2b 16 38 80 770b 20bc 3100ab 78ab 28c 4b 10 6b

70 90abc 420d 1b 20 80 100 670b 18c 1900b 47b 13c 3b 10 7b

90 160a 510c 1b 20 25 56 1200a 33a 2600b 70b 17c 7b 40 6b

100 70bc 620b 2b 50 100 100 400c 11d 5000ab 120ab 25c 11b 60 7b

120 140ab 610b 1b 50 100 100 690b 19c 5000ab 100ab 16c 5b 9 7b

59 61 62 63 69

330 250a 7a 10a 7a

220 150bc 4bc 5b 5b

210 190b 5ab 9a 4b

160 70de 2cd 2b 4b

200 110cd 3cd 2b 3c

100 40e 1e 3b 5b

600 60de 2cd 2b 7a

600 60de 2cd 1b 5b

71 73 75 78 81 83 87 89 95 137 138

4 16a 12b 18bc 100a 6a 3 70a 7a 74a 8a

3 11b 7b 32b 70b 4bc 2 40ab 5b 60b 5b

3 11b 20a 60a 56bc 4bc 3 60a bc 5 49b 5b

2 6de 4b 5c 34cd 6ab 1 15bc 3d 27c 3c

3 3e 5b 5c 33d 3c 2 7c 3d 26c 3c

2 7cd 4b 3c 38cd 4c 3 8c 3d 27c 3c

4 10bc ab 13 7c 37cd 7a 6 19bc cd 3 30c 4c

9 7cd 4b 3c 30d 4bc 8 15bc 3d 26c 3c

PTR-MS measurement of compound standard, IIMaleknia et al. (2007), IIIBuhr et al. (2002),

IV

Assignments EthanolIV, acetoinV MethanalII MethanolII,IV HydrogensulfideVI Various compoundsIV,V Various compoundsIII Various compoundsIII EthanolIII, ethanal Ethanol, C13-isotope EthanolIII Ethanol, C13-isotope MethanethiolVI ButanalIII, hexanalI,III, H3O+(H2O)2-cluster 2-Methylpropanol, 1-pentanol, 1-butanol, 1-hexanol, 1-octanolIII,IV AcetoneVI, propanalIII, diacetylI,V Acetic acidV, ethyl acetateIII Dimethyl sulfideVI 2- and 3-methylbutanalI,V, 1-octen-3-olIII, pentanal, octanal, nonanalIII 3-MethylbutanolIII, 1-pentanolIV 2-ButanoneIII,VI Propanoic acidVI MonoterpenesI,II HexanalI,III DiacetylI,V,VI, 2- and 3-methylbutanalI,V, acetoinI Ethyl acetateIII, acetoinI Dimethyl disulfideVI, monoterpenesII phenolVI MonoterpenesI,II Monoterpenes, C13-isotopeI,III

Brown et al. (2010), VLasekan & Otto (2009),

VI

Feilberg et al. (2010).

E.S. Holm et al. / Meat Science 95 (2013) 302–310

307

0.4

0.3

39 41 59 71 87

PC 2 (20%)

0.2 Week5+6days 31

43

0.1 Week5+4days

35 49 61 62 63 73

57 48

Week3+6days

55

46 45

0

75

83

Week4+6days

Week3+4days

-0.2

89

Day1

Week5 78

-0.1

47 69 33 29 51

-0.1

95 138

137

81

Week3

0

0.1

0.2

0.3

0.4

PC 1 (46%) Fig. 2. Bi-plot (PC1 vs. PC2) from the PCA model including the PTR-MS ion fragments and the saveloy samples measured at day 1, week 3, week 3 + 4 days, week 3 + 6 days, week 4 + 6 days, week 5, week 5 + 4 days and week 5 + 6 days. The PCA model includes all the PTR-MS ions shown in Table 3. The data was auto-scaled for before calculating the PCA model.

replications was separated from the remaining samples with high scores on PC2. These samples contained increased levels of a group of ions including m/z 59, 71 and 87. As seen in Table 1 and Table 3 m/z 87 is the M + 1-ion of diacetyl and 2- and 3-methylbutanal, whereas a fragment ion of acetoin also contributes to this ion. Moreover, a fragment ion of 3-methylbutanol contributes to m/z 71 and a fragment ion of diacetyl contributes to m/z 59. As mentioned, these VOCs have previously been identified as chemical markers for the sensory shelf-life of saveloy (Holm et al., 2012). Though the increase in m/z 59, 71 and 87 is not significant according to the Tukey HSD test, the PCA model indicates that the microbial spoilage processes start to have an impact on the VOC composition of the saveloy samples at week 5 + 4 days and week 5 + 6 days. The five replicate samples from week 5 + 4 days and week 5 + 6 days seen in Fig. 2 illustrate the degree of variation between different packages of cooked and sliced meat product, which can be found in practice. It should moreover be noted that the expected shelf-life of the saveloy samples was four weeks according to the label. The observed changes in VOC composition beyond this point are therefore mainly of theoretical interest.

to the odor of the product at this time point. No significant effect of storage time on the taste descriptor was observed in the saveloy samples. 3.4. PTR-MS measurements of the four types of cooked and sliced meat product In order to study the extent of spoilage in other types of cooked and sliced meat products, PTR-MS measurements of herbal saveloy, pork flank sausage, sandwich ham and pork loin were also included in the experiment. The saveloy samples used in the first part of the experiment were obtained directly from the production facility and afterwards kept at 5 °C until measurement. In contrast the four other product types were purchased in local super markets. The risk that these samples had been subjected to temperature fluctuations during storage and distribution is therefore increased compared to the saveloy 3

a ab

ab

2

3.3. Sensory evaluation of the quality changes in the sliced saveloy samples The results of the sensory evaluation are seen in Fig. 3. This figure shows the perceived quality changes in odor and taste relative to a reference sample measured on a scale from 0 to 5. As seen in Fig. 3 there was a significant increase in the odor difference between the sample and reference from day 1 to week 4 + 4 days. At week 4 + 4 days seven of the ten assessors moreover commented that they noted a decrease in the spicy aroma of the samples compared to the reference. This corresponds well with the observed decrease in terpene intensity observed in HS-GC–MS and PTR-MS measurements. Furthermore, two of the assessors noted a slightly sour odor at week 4 + 4 days. This indicates that the microbial spoilage processes are starting to contribute

-

-

1

Odor

Taste

b -

0 Day 1

Week 4

Week 4+4days

Week 5

Fig. 3. The results of the sensory evaluation of the sliced saveloy samples made on day 1, week 4, week 4 + 4 days and week 5. A 10 person sensory panel evaluated the saveloy samples in relation to a reference stored at −1 °C. The difference in odor and taste was assessed on a scale from 0 to 5, where 0 was no difference and 5 was large difference. Letters a and b are used to indicate significant differences between sensory evaluations within the same attribute.

308

E.S. Holm et al. / Meat Science 95 (2013) 302–310

samples. The VOC composition of the four types of cooked and sliced meat product was measured with PTR-MS in the middle of their shelf-life period and again close to their expiration date, after being subjected to package opening for one week. In the final PTR-MS measurement made near the expiration date the samples of especially pork loin and herbal saveloy appeared spoiled and clear sour and butter-like off-odors were noted when preparing these samples for measurement. The changes in VOC composition determined by PTR-MS were accounted for in a PCA model from which the bi-plot (PC1 vs. PC2) is shown in Fig. 4. In this bi-plot the herbal saveloy samples measured in the middle of the shelf-life period were separated from the remaining samples on PC1. This is partly due to the content of m/z 81, 137 and 138 corresponding to the monoterpenes. This illustrates that herbal saveloy contained a higher level of spices than the other product types tested. In the PCA model shown in Fig. 4, PC2 mainly describes the difference between the pork loin and herbal saveloy samples measured at the end of the shelf-life period and the remaining samples. The samples of pork loin and herbal saveloy measured at the end of the shelf-life period are associated with several fragment ions including m/z 59, 69, 71, 87 and 89. In Tables 1 and 3 it is seen that fragments of 2and 3-methylbutanal contribute to m/z 69 whereas M + 1 of acetoin is m/z 89. As previously highlighted 2- and 3-methylbutanol, diacetyl and acetoin contribute to m/z 59, 71 and 87. These VOCs were all among the previously suggested chemical markers of spoilage in sliced saveloy. The changes in m/z 59, 69, 71, 87 and 89 are shown in Fig. 5 along with m/z 61. The M + 1 ion of acetic acid contributes largely to m/z 61 (Lasekan & Otto, 2009) and could therefore be interesting as an indicator of spoilage. In Fig. 5 it is seen that m/z 69, 71, 87 and 89 increased significantly during storage for the pork loin samples. The pork loin samples furthermore produced a higher amount of m/z 69 and 87 than the other tested products at the end of the storage period. These changes suggest an increase in the potential spoilage markers 2and 3-methylbutanal, 2- and 3-methylbutanol, diacetyl and acetoin for the pork loin samples. The herbal saveloy samples also contained

relatively high amounts of m/z 71, 87 and 89 at the end of the shelf-life period, but the increase of these ions was not significant. However, a significant increase in m/z 61, corresponding to acetic acid, was observed for the herbal saveloy samples. The amount of m/z 61 found in the herbal saveloy samples at the end of the storage period was also significantly higher than that found in the pork loin samples. In Fig. 5 it is seen that the there was no significant increase in m/z 61, 69, 71, 87 and 89 for the samples of sliced pork flank sausage and sliced sandwich ham. No significant changes were observed for m/z 59 in any of the tested products. As seen in Table 3 both diacetyl, acetone and propanal contribute to m/z 59, and this could mask an increase in this ion caused by diacetyl. The variation in the concentration of m/z 61, 69, 71, 87 and 89 seen in Fig. 5, suggests that there were major differences in the microbial activity and composition of the four types of meat products at the end of the shelf-life period. These differences are ascribed to variation in the composition and growth rate of the specific spoilage flora of the products. The specific spoilage flora consists of the small fraction of bacteria present in the microbial flora of the processing facility, which are able to grow and dominate the product under the given environmental conditions (Mataragas, Skandamis, Nychas, & Drosinos, 2007; Nychas, Skandamis, Tassou, & Koutsoumanis, 2008). Variation in factors such as the pH, water activity, availability of nutrients as well as the type and concentration of preservation agents used, could also contribute to explain the differences in the level of microbially produced VOCs between the four products (Holm et al., 2012; Nychas et al., 2008; Samelis et al., 2000). 3.5. Evaluation of PTR-MS for detection of quality changes in cooked and sliced meat products The objective of this study was to investigate PTR-MS as a tool for detection of quality changes in cooked and sliced meat products. The saveloy samples, investigated in the first part of the experiment, were

0.3

Pork loin End

PC 2 (27%)

0.2

87

43 45 46 89

71

69 83

55

57

0.1 59

73 31

29 47

61 Herbal saveloy End 48 62 Pork flank sausage End 35

0

Sandwich ham End

51

Pork loin Middle

Sandwich ham Middle

-0.1

-0.2

-0.1

95

78

Pork flank sausage Middle

0

63

0.1

33 39 41 81 137 138

75 49

Herbal saveloy Middle

0.2

0.3

PC 1 (34%) Fig. 4. Biplot (PC 1 vs. PC 2) from the PCA model including the PTR-MS ion fragments and samples of 4 types of cooked and sliced meat product measured in the middle and the end of their 4 week shelf-life period. The data was auto-scaled for before calculating the PCA model.

E.S. Holm et al. / Meat Science 95 (2013) 302–310

309

m/z 61

m/z 59 300

abc

a

c

bc

abc

a

abc

ab

200

c

bc

bc

ab

c

bc

c

a

b

a

b

ab

b

a

b

ab

200 100 100 0

0

m/z 71

m/z 69 b

b

b

b

b

a

b

b

300

120

b

b

b

b

b

b

200

80 40

100

0

0

m/z 89

m/z 87 100

b

b

b

b

b

b

a

b

b

300

b

200 50 0

100 0

Fig. 5. The development in m/z 59, 61, 69, 71, 87 and 89 between the middle and the end of the four week shelf-life period for cooked and sliced sandwich ham, pork flank sausage, pork loin and herbal saveloy. The m/z 61 was assigned to acetic acid. m/z 69 was assigned to 2- and 3-methylbutanal, m/z 71 was assigned to 2- and 3-methylbutanol whereas diacetyl, acetoin and 2-and 3-methylbutanol all contribute to m/z 87. The m/z 89 is the M + 1 ion of acetoin. Letters a to c are used to indicate significant differences, based on a Tukey HSD test, in the PTR-MS ion concentrations measured in the different samples.

studied in most detail. As seen in Tables 2 and 3 the main change observed in the saveloy samples was a decrease in VOCs contributing to the fresh aroma of the product. Based on these results it is difficult to evaluate the ability of PTR-MS to detect microbial spoilage during storage. However, the results of the saveloy storage experiments did show that there generally was good correspondence between the measurements done with PTR-MS, the HS-GC–MS measurements and the sensory evaluations. In the second part of the experiment PTR-MS measurements were made on samples of pork loin, sliced sandwich ham, herbal saveloy and pork flank sausage. For the pork loin samples, which appeared spoiled during sample preparation, these measurements showed an increase in m/z 69, 71, 87 and 89 from the middle to the end of the shelf-life period. As previously mentioned, these PTR-MS ions correspond to fragments or M + 1 of the spoilage markers 2- and 3-methylbutanal, 2- and 3-methylbutanol, acetoin and diacetyl. The molecular ion of acetic acid (m/z 61) was moreover found to increase significantly in the herbal saveloy samples. These five ions are therefore very interesting as

potential spoilage markers in future PTR-MS measurements of cooked and sliced meat products. However, it should be considered that other VOCs also contribute to the suggested fragment ions. Fragments of several aldehydes beside 2- and 3-methylbutanal also contribute to m/z 69 whereas m/z 89 is M + 1 of both ethyl acetate and acetoin. This could potentially mask the changes in the relevant spoilage markers and give a misleading result. The contribution of other VOCs to m/z 61 69, 71, 87 and 89 therefore needs to be investigated before these ions can be used as spoilage markers in PTR-MS measurements of a given cooked and sliced meat product. However, when relating the results of this study with previous studies made on similar meat products it seems highly probable that the increase in the levels of the five selected ions, particularly in the pork loin samples, is related to microbial spoilage (Holm et al., 2012; Leroy et al., 2009; Stanley, Shaw, & Egan, 1981). Diacetyl, acetoin and 2- and 3-methylbutanal all contribute to m/z 87, and this ion is therefore considered as the most promising in relation to measurement of microbial spoilage in cooked and sliced meat

310

E.S. Holm et al. / Meat Science 95 (2013) 302–310

products. Moreover, no compounds deriving from other sources were found to contribute to m/z 87. This further supports the use of this ion as a quality marker in cooked and sliced meat products. The degree of fragmentation of the potential chemical markers during PTR-MS measurement could pose a problem when using this method for detection of quality changes in complex food matrices. Major fragments of 3-methylbutanal and 3-methylbutanol have been found at m/z 39, 41, 43 and 45 (Buhr et al., 2002; Lasekan & Otto, 2009). At these low m/z-ratios the fragments of several other alcohols and aldehydes also contribute to the total signal and therefore interfere with the compounds of interest (Brown et al., 2010; Buhr et al., 2002). The fragmentation therefore results in a loss of sensitivity of the chemical markers when using PTR-MS. The degree of fragmentation can be lowered by reducing the voltage of the drift tube. However, this will instead result in an increased formation of water clusters and lower the concentration of H3O +-ions. The concentration of H3O+-ions is a critical factor for proton transfer reaction. A decreased concentration of H3O +-ions lowers the possibility of a proton transfer reaction between H3O + and the VOCs. It is desirable to use an excess of H3O +ions to ensure that all VOCs are ionized in the proton transfer reaction (Brown et al., 2010; Lindinger et al., 1998). Therefore manipulation of the drift tube voltage should be used with care and must not compromise the reaction conditions in the drift tube. The volatility of the VOCs found in the headspace of the tested meat products should be considered in relation to the conclusions made in this study. The sample preparation step, in which the packages were opened and the samples cut, could lead to a loss of a fraction of some of the most volatile compounds in the sample headspace. This could moreover result in an underestimation of the importance of these VOCs in relation to shelf-life. With PTR-MS it would be possible to measure directly in the closed packages using a setup with a hypodermic needle. This could be part of a future setup for at-line measurements of the VOC composition of cooked and sliced meat products.

4. Conclusion The results of this study show that PTR-MS is capable of detecting changes in the VOC composition of cooked and sliced meat products during storage. In the first part of the experiment measurements of the PTR-MS fingerprint of the saveloy samples corresponded well with measurements of the VOC composition with HS-GC–MS and with a sensory evaluation of the product. In the second part of the experiment m/z 61, 69, 71, 87 and 89 were argued to account for the chemical markers 2- and 3-methylbutanal, 2- and 3-methylbutanol, acetoin, diacetyl and acetic acid. Based on these findings PTR-MS is considered as a promising technique for the measurement of microbially induced quality changes in cooked and sliced meat products. The ability of the PTR-MS to provide direct, rapid and continuous measurements makes it suitable for at-line measurements in the production facility. Measurements of the volatile profile packages of cooked and sliced meat products with PTR-MS could therefore have potential as a quality assurance tool in the meat industry.

Acknowledgments The Danish Center for Advanced Food Studies (LMC) is acknowledged for financial support for this study. Laboratory technician Ms. Claudia Nagy is acknowledged for support during the practical execution of this study.

References Acree, T., & Arn, H. (2004), http://www.flavornet.org/ Ref Type: Internet Communication. Andersen, K. B., Hansen, M. J., & Feilberg, A. (2012). Minimisation of artefact formation of dimethyl disulphide during sampling and analysis of methanethiol in air using solid sorbent materials. Journal of Chromatography A, 1245, 24–31. Biasioli, F., Yeretzian, C., Gasperi, F., & Märk, T. D. (2011). PTR-MS monitoring of VOCs and BVOCs in food science and technology. Trends in Analytical Chemistry, 30, 968–977. Bjorkroth, K. J., Vandamme, P., & Korkeala, H. J. (1998). Identification and characterization of Leuconostoc carnosum, associated with production and spoilage of vacuumpackaged, sliced, cooked ham. Applied and Environmental Microbiology, 64, 3313–3319. Borch, E., KantMuermans, M. L., & Blixt, Y. (1996). Bacterial spoilage of meat and cured meat products. International Journal of Food Microbiology, 33, 103–120. Brown, P., Watts, P., Märk, T. D., & Mayhew, C. A. (2010). Proton transfer reaction mass spectrometry investigations of the effects of reduced electric field and reagent ion internal energy on product ion branching ratios for saturated alcohols. International Journal of Mass Spectrometry, 294, 103–111. Buhr, K., van Ruth, S., & Delahunty, C. (2002). Analysis of volatile flavour compounds by proton transfer reaction-mass spectrometry: Fragmentation patterns and discrimination between isobaric and isomeric compounds. International Journal of Mass Spectrometry, 221, 1–7. Dainty, R. H., & Hibbard, C. M. (1983). Precursors of the major end products of aerobic metabolism of Brochothrix-thermosphacta. Journal of Applied Bacteriology, 55, 127–133. Dainty, R. H., & Mackey, B. M. (1992). The relationship between the phenotypic properties of bacteria from chill-stored meat and spoilage processes. Journal of Applied Bacteriology, 73, S103–S114. Feilberg, A., Liu, D., Adamsen, A. P. S., Hansen, M. J., & Jonassen, K. E. N. (2010). Odorant emissions from intensive pig production measured by online proton-transfer-reaction mass spectrometry. Environmental Science, 44, 5894–5900. Gounadaki, A. S., Skandamis, P. N., Drosinos, E. H., & Nychas, G. J. E. (2008). Microbial ecology of food contact surfaces and products of small-scale facilities producing traditional sausages. Food Microbiology, 25, 313–323. Ho, C. T., Oh, Y. C., & Bae-Lee, M. (1994). The flavour of pork. In F. Shahidi (Ed.), Flavour of meat and meat products (pp. 38–51). Blackie Academic & Professional. Holm, E. S., Schäfer, A., Skov, T., Koch, A. G., & Petersen, M. A. (2012). Identification of chemical markers for the sensory shelf-life of saveloy. Meat Science, 90, 314–322. Järvenpää, E. P., Zhang, Z., Houpalahti, R., & King, F. W. (1998). Determination of fresh onion (Allium cepa L.) volatiles by solid phase microextraction combined with gas chromatography–mass spectrometry. Zeitschrift fur Lebensmittel-Untersuchung UndForschung, 207, 39–43. Kováts, E. (1958). Gas-chromatographische Charakterisierung Organischer Verbindungen. Teil 1: Retentionsindices aliphatischer Halogenide, Alkohole, Aldehyde und Ketone. Helvetica Chimica Acta, 41, 1915–1932. Lasekan, O., & Otto, S. (2009). In vivo analysis of palm wine (Elaeis guineensis) volatile organic compounds (VOCs) by proton transfer reaction-mass spectrometry. International Journal of Mass Spectrometry, 282, 45–49. Laursen, B. G., Byrne, D. V., Kirkegaard, J. B., & Leisner, J. J. (2009). Lactic acid bacteria associated with a heat-processed pork product and sources of variation affecting chemical indices of spoilage and sensory characteristics. Journal of Applied Microbiology, 106, 543–553. Leroy, F., Vasilopoulos, C., Van Hemelryck, S., Falony, G., & De Vuyst, L. (2009). Volatile analysis of spoiled, artisan-type, modified-atmosphere-packaged cooked ham stored under different temperatures. Food Microbiology, 26, 94–102. Lindinger, W., Hansel, A., & Jordan, A. (1998). On-line monitoring of volatile organic compounds at pptv levels by means of proton-transfer-reaction mass spectrometry (PTR-MS) — Medical applications, food control and environmental research. International Journal of Mass Spectrometry, 173, 191–241. Løkke, M. M., Edelenbos, M., Larsen, E., & Feilberg, A. (2012). Investigation of volatiles emitted from freshly cut onions (Allium cepa L.) by real time proton-transfer reaction-mass spectrometry (PTR-MS). Sensors, 12, 16060–16076. Maleknia, S. D., Bell, T. L., & Adams, M. A. (2007). PTR-MS analysis of reference and plantemitted volatile organic compounds. International Journal of Mass Spectrometry, 262, 203–210. Mataragas, M., Skandamis, P., Nychas, G. J. E., & Drosinos, E. H. (2007). Modeling and predicting spoilage of cooked, cured meat products by multivariate analysis. Meat Science, 77, 348–356. Meynier, A., Novelli, E., Chizzolini, R., Zanardi, E., & Gandemer, G. (1999). Volatile compounds of commercial Milano salami. Meat Science, 51, 175–183. Mottram, D. S. (1998). Flavour formation in meat and meat products: A review. Food Chemistry, 62, 415–424. Nychas, G. J. E., Skandamis, P. N., Tassou, C. C., & Koutsoumanis, K. P. (2008). Meat spoilage during distribution. Meat Science, 78, 77–89. Samelis, J., Kakouri, A., & Rementzis, J. (2000). Selective effect of the product type and the packaging conditions on the species of lactic acid bacteria dominating the spoilage microbial association of cooked meats at 4 degrees C. Food Microbiology, 17, 329–340. Stanley, G., Shaw, K. J., & Egan, A. F. (1981). Volatile compounds associated with spoilage of vacuum-packaged sliced luncheon meat by Brochothrix-thermosphacta. Applied and Environmental Microbiology, 41, 816–818.