Food Chemistry 128 (2011) 371379

Contents lists available at ScienceDirect

Food Chemistry
journal homepage: www.elsevier.com/locate/foodchem

Storage stability of cauliower soup powder: The effect of lipid oxidation and protein degradation reactions
Riikka Raitio, Vibeke Orlien, Leif H. Skibsted
University of Copenhagen, Department of Food Science, Faculty of Life Sciences, Rolighedsvej 30, DK-1930 Frederiksberg C, Denmark

a r t i c l e

i n f o

a b s t r a c t
Soups based on cauliower soup powders, prepared by dry mixing of ingredients and rapeseed oil, showed a decrease in quality, as evaluated by a sensory panel, during the storage of the soup powder in the dark for up to 12 weeks under mildly accelerated conditions of 40 C and 75% relative humidity. Antioxidant, shown to be effective in protecting the rapeseed bulk oil, used for the powder preparation, had no effect on storage stability of the soup powder. The freshly prepared soup powder had a relatively high concentration of free radicals, as measured by electron spin resonance spectroscopy, which decreased during storage, and most remarkably during the rst two weeks of storage, with only marginal increase in lipid hydroperoxides as primary lipid oxidation products, and without any increase in secondary lipid oxidation products. Analyses of volatiles by SPME-GCMS revealed a signicant increase in concentrations of 2-methyl- and 3-methyl butanals, related to Maillard reactions, together with an increase in 2-acetylpyrrole concentration. The soup powders became more brown during storage, as indicated by a decreasing Hunter L-value, in accord with non-enzymatic browning reactions. A signicant increase in the concentrations of dimethyl disulde in soup powder headspace indicated free radical-initiated protein oxidation. Protein degradation, including Maillard reactions and protein oxidation, is concluded to be more important than lipid oxidation in determining the shelf-life of dry cauliower soup powder. 2011 Elsevier Ltd. All rights reserved.

Article history: Received 16 July 2010 Received in revised form 7 February 2011 Accepted 8 March 2011 Available online 15 March 2011 Keywords: Dry food powder Storage stability ESR spectroscopy Lipid oxidation Protein degradation reactions

1. Introduction A common method for preparing soup powders is based on batch operations, in which preproduced dry ingredients are mixed together with addition of oil or fat powder. These dry ingredients are usually powders made of dried vegetables, spices, avours and yeast extracts. The microbial and enzymatic spoilage of dry products is prevented by a low water activity (Gibbs, 1986), thus leaving chemical and physical deterioration reactions as most important for shelf-life. Among the chemical reactions, lipid oxidation and non-enzymatic browning reactions are often linked and further depend on phase transitions in the product (Thomsen, Lauridsen, Skibsted, & Risbo, 2005a). Lipid oxidation, one of the major causes of the degradation of food products during storage, leads to formation of off-avours and furthermore affects the nutritional value and safety. Free radicals are formed in the early stage of lipid autoxidation (Frankel,
Abbreviations: AP, ascorbyl palmitate; aw, water activity; CaA, carnosic acid; Cia, citric acid; DMDS, dimethyl disulde; DUS, degree of unsaturation; ESR, electron spin resonance; IP, induction period; RE, rosemary extract; RH, relative humidity; TP, tocopherols extract. Corresponding author. Address: University of Copenhagen, IFV, Rolighedsvej 30, 1958 Frederiksberg C, Denmark. Tel.: +45 3533 3221; fax: +45 3533 3344. E-mail address: ls@life.ku.dk (L.H. Skibsted). 0308-8146/$ - see front matter 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.foodchem.2011.03.038

2005), and electron spin resonance (ESR) spectroscopy is a specic and sensitive technique for detection of such free radicals (Schaich, 2002). In dry systems, free radicals can be directly detected using ESR spectroscopy, since free radicals are stabilised by low molecular mobility and therefore are sufciently long-lived for detection. Recent studies have indicated a correlation between free radical formation and lipid oxidation during the storage of dry food systems (Nissen, Huynh-Ba, Agerlin Petersen, Bertelsen, & Skibsted, 2002; Nissen, Mnsson, Bertelsen, Huynh-Ba, & Skibsted, 2000; Stapelfeldt, Nielsen, & Skibsted, 1997; Thomsen et al., 2005a), thus allowing ESR spectroscopy to be used to study the initial stages of lipid oxidation in dry foods and for prediction of shelf-life of such products. Non-enzymatic browning reactions, including Maillard reactions, occurring in foods during thermal processing and long storage, are complex and involve both carbohydrates and proteins. Ultimately, the Maillard reactions lead to the formation of high molecular weight coloured compounds known as melanoidins, resulting in a discolouration of pale food products. Besides discoloration, other negative effects related to the Maillard reactions in foods are formation of off-avours, decrease in solubility and nutritional value, and formation of toxic compounds (Nursten, 2005). ESR spectroscopy may also be employed to study the Maillard reactions, as several studies have shown the appearance of free radicals

372

R. Raitio et al. / Food Chemistry 128 (2011) 371379

in the early stages of non-enzymatic browning reactions (Hayashi, Mase, & Namiki, 1985; Hoffman, Bors, & Stettmaier, 1999a; Namiki & Hayashi, 1975; Roberts & Lloyd, 1997), as well as in the Maillard reaction end-products, melanoidins (Hoffman, Bors, & Stettmaier, 1999b). Furthermore, Thomsen, Lauridsen, Skibsted, and Risbo (2005b) found a correlation between the formation of late-stage Maillard reaction products and change in the ESR line shape when studying milk powder. In general, product water activity (aw) is a major issue in relation to chemical stability of dry food products. Very low values of aw are related to high lipid oxidation rates while between aw values of 0.2 and 0.4, lipids have been suggested to have optimal stability and oxidation rates increase again with increasing aw (Labuza, 1971). However, the oxidative stability of dry food products does not always follow this general scheme, but is more productrelated (Jensen & Risbo, 2007; Stapelfeldt et al., 1997). Non-enzymatic browning reactions are also strongly dependent on aw, and the rate of Maillard reactions increases with increasing values of aw to reach a maximum around an aw of 0.7 (Labuza, 1971). Lipid oxidation and Maillard reactions may further be coupled through aw, since water is produced in the Maillard reactions (Nursten, 1986), which increases the aw value of the product and further enhances lipid oxidation. In turn, the volatiles (Elizalde, Dalla Rosa, & Lerici, 1991) and melanoidins (Borrelli, Visconti, Mennella, Anese, & Fogliano, 2002) formed in Maillard reactions may act as antioxidants, protecting against lipid peroxidation. Storage stability of dry food products may be enhanced by using antioxidants to delay the lipid oxidation (Nissen et al., 2000; Sarkardei & Howell, 2008). In the model systems, antioxidants have also been found to prevent the formation of mutagenic/carcinogenic compounds due to the Maillard reaction (Kato, Harashima, Moriya, Kikugawa, & Hiramoto, 1996). The overall chemical stability of dry food products accordingly depends on both composition and storage conditions of each particular dry food system. The purpose of the present study was to study the mechanisms of quality deterioration in cauliower soup powder, adapting methods, which also could be of use for similar sensitive dry food products. In addition, the effects of different antioxidants on storage stability of cauliower soup powder during accelerated storage tests were studied. The antioxidants selected for the investigations differed by action mechanism and/or by solubility. The selected antioxidants were: lipophilic rosemary extract (containing radical-scavenging compounds), hydrophilic carnosic acid (radical-scavenger), hydrophilic citric acid (metal chelator), lipophilic tocopherol extract (radical-scavenger), and ampiphilic ascorbyl palmitate (e.g. regeneration of phenolic antioxidants) (Burton & Ingold, 1981; Frankel, 2005; Hall & Cuppett, 1997).

Fatty acid methyl esters used as standards for GC were obtained from SigmaAldrich (Steinheim, Germany). 2.2. DSC analysis The induction periods for oxidation (IP) were determined for rapeseed oil with the following antioxidant additions: control (no antioxidant addition), rosemary extract (RE) 500 ppm and 1000 ppm, carnosic acid (CaA) 45 ppm, citric acid (CiA) 1000 ppm, tocopherol extract (TP) 1000 ppm, and ascorbyl palmitate (AP) 50 and 300 ppm. Rosemary and tocopherol extracts were lipid-soluble and were added directly to rapeseed oil. Citric acid, ascorbyl palmitate, and carnosic acid were dissolved in absolute ethanol and dispersed into the oil. The DSC experiments were performed with a DSC 1 Stare system (Metler Toledo, Schwerzenbach, Switzerland) isothermally at 100 C. The samples (5.00 mg 0.05 mg) were placed in 40 ll aluminium crucibles which were closed with lids with a hole (diameter % 1 mm) to ensure the oxygen availability inside the crucible. A hermetically sealed empty crucible was used as a reference. The experimental curves were simulated with software Origin 7 (OriginLab Corporation, Northampton, MA, USA), and the induction periods (IP) for lipid oxidation were calculated as previously described (Velasco, Andersen, & Skibsted, 2004), but now from the data simulations. The analyses were made in duplicate. 2.3. Preparing soup powders Cauliower soup powders with different antioxidant additions were prepared by mixing preproduced dry ingredients and by adding 0.5% of rapeseed oil (RO). The ready soup powder contained approximately 20% of fat, calculated from the information given by raw material suppliers. The antioxidant additions to rapeseed oil before powder preparation were the same as used with the DSC analysis (see Section 2.2) and the concentrations of antioxidants were expressed in relation to rapeseed oil. The soup powders (70 g) were stored in closed, 110 ml polypropylene containers in a climate cupboard (Climacell, Planegg, Germany) at 40 C and 75% relative humidity (RH) for up to 12 weeks, and one container was removed from the climate cupboard at each sampling time. The powders were stored frozen at 40 C until analysed. Two separate batches of each soup powder were prepared. The cauliower soup powder with carnosic acid after 4 weeks of storage and cauliower soup powder with 300 ppm of ascorbyl palmitate after 12 weeks of storage from the one storage series were discarded from the analyses because the packaging had been slightly open during the storage. 2.4. Fatty acid composition

2. Materials and methods 2.1. Chemicals BaCl22H2O, hexane, absolute ethanol, 2-methylbutanal, pentanal, dimethyl disulde, a-pinene, and carnosic acid were purchased from SigmaAldrich (Steinheim, Germany); hexanal was from Fluka (Steinheim, Germany), and HPLC grade chloroform was from Lab-Scan (Dublin, Ireland). HPLC grade isooctane, analytical grade methanol, dried methanol, NaCl, and FeSO47H2O were from Merck (Darmstadt, Germany). Analytical grade NH4SCN was from Riedel-De-Han (Seelze, Germany). Food grade rened rapeseed oil, rosemary extract, tocopherol extract, ascorbyl palmitate, and citric acid were used for preparing the soup powder. Rosemary extract contained 4.5% of carnosic acid and tocopherol extract a minimum of 50% of tocopherols according to the data provided by the manufacturer.

The fatty acid compositions of acyl glycerols of the soup powders were determined from fresh powders and after 12 weeks of storage and for rapeseed oil. The analyses were done for cauliower soup powder only from other storage series which did not have any discarded samples. Fatty acid determinations were also done for rapeseed oil with addition of antioxidants (RE 1000 ppm, CaA 45 ppm and AP 300 ppm). The methylations were carried out directly on 40 mg of dry soup powder or on 5 mg of oil with 0.087 M sodium methylate solution at 60 C for 40 min. The formed fatty acid methyl esters were extracted with 1.5 ml of hexane prior to chromatographic analysis (HP 6890 Series, Hewlet-Packard, Palo Alto, CA, USA; FFAP column (25 m 0.20 mm 0.33 lm) from Agilent Technologies, Waldbronn, Germany) with FID detection. The oven temperature programme was: 50 C for 1 min; from 50 C to 180 C at 15 C/min; from 180 C to 195 C at 2 C/min, held for 15 min; from 195 C to 240 C at 4 C/min, held for 10 min. Injected volume was 1 ll and split ratio was 1:25. Helium was used as carrier

R. Raitio et al. / Food Chemistry 128 (2011) 371379

373

gas and the ow was 1 ml/min. The results were analysed with Chemstation software (Agilent Technologies) and fatty acid methyl esters were identied by comparing retention times with known standards. From the obtained fatty acid composition results, the ratio between unsaturated and saturated fatty acids was calculated and marked as degree of unsaturation (DUS). Validated results indicated identical fatty acid compositions when the methylations were done, either directly on the dry soup powder, or on pre-extracted fat from the powder. The fat extractions were carried out using Folch extraction (Folch, Lees, & Sloane Stanley, 1957) with both 10 g of powder and 100 ml of chloroform:methanol (2:1) and 500 mg of powder and 5 ml of chloroform:methanol. 2.5. ESR spectroscopy The relative free radical concentrations of different soup powders during storage were determined with ESR spectroscopy. For ESR analyses, the soup powder was placed in a fused quartz tube with 4 mm inner diameter and 1 mm wall thickness (710-SQ-250 M, Wilmad Glass, Buena, NJ, USA). The tube was placed in the ESR spectrometer (MiniScope MS 200, Magnettech, Berlin, Germany) equipped with a X-band microwave supply. The microwave frequency of each spectrum was recorded with an external frequency meter (Agilent 53150A, Santa Clara, CA, USA) for the determination of g-values. The optimised parameters for ESR determinations were: centre eld 3337.73 G, sweep width 79.36 G, sweep time 120 s, modulation amplitude 6000 mG, and microwave power 1.58 mW. Software analysis (Magnettech) was employed to obtain the area under the ESR spectra by double integration. The measurements were performed with microwave power, with highest sensitivity, without interference from saturation, under which conditions the double integrated area of the ESR spectrum is proportional to the radical concentration in the sample (Weil, Bolton, & Wertz, 1994). The sample densities were calculated from the sample masses and heights in the ESR tubes and these were used to calculate the relative free radical concentrations. Furthermore, the software analysis was employed to determine the g-values of the ESR signals. The free radical concentrations of different soup powders were determined after 0, 2, 4, 8, and 12 weeks of storage at 40 C and 75% RH, the ESR determinations were done in duplicate. The ESR method for soup powder was validated by measuring seven separate samples of fresh soup powder and calculating the relative standard deviation (RSD). To standardise the ESR instrument, a weak pitch standard (Bruker BioSpin GmbH, Rheinstetten, Germany), which had a nominal concentration of 1013 spins per centimetre (Weber & Heiss, 1992), was measured on 8 different days. As with soup powder, the RSD was determined from the double integrated ESR signal. 2.6. Peroxide value determinations The IDF standard method (1991), with modication to sample size, was used to determine peroxide values (POV) of pre-extracted fat from cauliower soup powder and of rapeseed oil (515 mg). The Folch method (Folch et al., 1957) was used for fat extractions of soup powders. The absorbances were recorded with a spectrophotometer (Cary 3, Varian, Mulgrave, Australia). The analyses were done in duplicate after 0, 4, and 12 weeks of storage for soup powder fat extracts. 2.7. SPME-GCMS SPME-GCMS was employed to determine the relative changes in volatiles formed in soup powder during the storage. The analyses were done in duplicate for the samples after 0, 4, and 12 weeks of storage. The 20 ml vials were used with a 1 g (0.1 g) sample size. Before the extraction, the sample was equilibrated for

20 min at 37 C with 500 rpm agitation. The headspace volatiles were extracted using a divinylbenzene/carboxen/polydimethylsiloxane (DVB/CAR/PDMS) 30/50 lm SPME bre assembly (Supelco, Bellafonte, PA, USA) at 37 C for 30 min. All the SPMEextractions were performed with an automated Combi PAL-system (CTC Analytics AG, Zwingen, Switzerland). After the extraction, the bre was desorbed in the injection port of the gas chromatograph for 10 min at 250 C. The GCMS (Scan mode) (G1530A, Agilent 5973 Network Mass Selective Detector, Agilent Technologies) with DB-Wax capillary column (30 m 0.25 mm 0.25 lm) was used. The oven temperature programme for GC was: 45 C for 10 min; from 45 C to 240 C at 6 C/min, held for 10 min. The split ratio was 1:10 and helium was used as carrier gas with the ow of 1 ml/min. The volatile compounds were tentatively identied by matching their mass spectra with those of commercial database (Wiley275.L, HP product G1035A) using software MSD Chemstation (Agilent Technologies) and only the compounds with high quality in identication were selected. The identications of pentanal, hexanal, 2-methylbutanal, dimethyl disulde and a-pinene were further ascertained by analysing the standard mixture of these compounds with the same GCMS equipment and parameters. The liquid injection was applied for a standard mixture with 1 ll sample size and with split ratio 50:1. The quantication of selected compounds was achieved by using the area of specic ion for each compound (Table 1). The optimised SPME condition ensured equilibrium between powder sample and headspace volume and, subsequently, between the volatile samples in headspace and in the SPME layer. However, the optimisation of SPME conditions was only done relative to hexanal and pentanal responses. The SPME-GCMS analysing technique for soup powder was validated by analysing seven separate samples of 12 week-stored control soup powder and calculating RSD from the integration results for each compound. The results were expressed as relative concentrations = area/sample mass. 2.8. Sensory analyses Sensory analyses of the cauliower soup, prepared from soup powders from storage experiments, were done by ranking test with a trained panel (either 6 or 8 judges). Only the soups with food grade antioxidants and control were analysed by sensory panel and sensory tests were done only with the one batch of soup powders after 4 and 12 weeks of storage. The tests were performed by introducing known references of soup prepared from freshly produced powder without antioxidant corresponding to the 0 week control sample. This particular soup was pointed out to the sensory panel in each tasting series as a reference sample in relation to
Table 1 The volatiles selected for the SPME-GCMS analyses, their retention times (RT), target ions for identication and quantication, relative standard deviation (RSD) in quantication obtained in method validation, and grouping of volatiles, where letters refer to Fig. 2. Identication RT (min) 4-Methylheptane 2,4-Dimethylheptane 4-Methyloctanea 2,4-Dimethylheptene 2-Methylbutanal 3-Methylbutanal Pentanal a-Pinene Dimethyl disulde Hexanal b-Pinene 2-Acetylpyrrole
a

Validation RSD (%) 6.1 4.5 3.7 4.7 11.2 13.7 12.7 2.9 6.4 11.8 3 2.3

Group

Target ion (m/z) 71.10 43.10 43.10 43.10 57.05 44.10 44.10 93.10 94.00 56.15 93.10 94.05

2.03 2.27 2.64 2.84 3.01 3.05 4.00 5.29 6.39 7.00 8.56 30.76

B B B B C C A F E A F D

Identication based on average mass spectrum of the peak.

374

R. Raitio et al. / Food Chemistry 128 (2011) 371379

which the other soup samples were ranked. In addition, in each tasting series, a blind control of this particular fresh soup was added. In each ranking series, 3 or 4 samples were introduced, besides the blind control sample. The signicance of the results was calculated by using R-index, which gives the statistical difference of each sample from the reference sample (Moskowitz, 1988). However, as the sensory panel sizes were small, the results from sensory analysis should be considered as tentative. 2.9. Colour determination The colour development of the soup powders during storage was measured using the Hunter L, a and b colour system with BYK-Gardner colour guide (BYK-Gardner GmbH, Gerestried, Germany). Each sample was measured, in triplicate, after 0, 2, 4, 8, and 12 weeks of storage. In this study the colour changes were expressed by the L-value, which measures sample lightness as white (+) or black (). The complete colour difference (DE) between 0 and 12 week samples, was further calculated by the following equation (Francis & Clydesdale, 1975):

2.11. Water content The water contents of powders were determined by placing approximately 3 g of powder on a dry aluminium tray and subsequently drying the sample in an oven at 103 C overnight. The results were calculated as g (water)/g (dry weight). The analyses were done in duplicate for samples after 0, 2, and 12 weeks of storage. 2.12. Statistical analysis Signicant differences between the samples were determined using the two-sided, heteroscedastic t-test at signicance level p < 0.05. 3. Results and discussion 3.1. General Cauliower soup powder is a sensitive product and little is known about its protection against oxidative deterioration. In order to optimise protection during storage, the lipid added, i.e. rapeseed oil, was investigated, separately, with the purpose of evaluating the effects of different types of added antioxidants and to identify the mechanism for initiation of lipid oxidation. 3.2. Antioxidants in rapeseed oil

DE DL Da Db

2 1=2

DL is lightness difference, Da is redness difference, and Db is yellowness difference.

2.10. Water activity aw The water activities of soup powders were determined at room temperature after 0, 2, 4, 8 and 12 weeks of storage with Aqualab CX2 equipment (Aqualab, Washington, USA).

To study the effect of antioxidants on oxidative stability of rapeseed oil, the induction periods (IP) for oxidation were obtained

Fig. 1. Cauliower soup powder with different antioxidant additions during 12 weeks of storage at 40 C and 75% RH. Control; j Rosemary extract (500 ppm); h Rosemary extract (1000 ppm); d Carnosic acid (45 ppm); s Citric acid (1000 ppm); N Tocopherols extract (1000 ppm); Ascorbyl palmitate (50 ppm); x Ascorbyl palmitate (300 ppm). The dashed lines are used for the lower concentration if two different concentrations of the same antioxidant were applied. (A) Relative free radical concentrations determined with ESR spectroscopy, (B) water activity aw, (C) peroxide values (POV), (D) Hunter L-value, indicating lightness of the sample.

R. Raitio et al. / Food Chemistry 128 (2011) 371379

375

using DSC (Table 3). Isothermal DSC has previously been used to study the oxidative stability of different oils and the results have been found to correlate well with the oxidative stability index (OSI) method (Tan, Che Man, Selamat, & Yusoff, 2002; Velasco et al., 2004). No lipid oxidation was observed for up to 900 min when citric acid (CiA), corresponding to 1000 ppm, was added to rapeseed oil and oxidation was not observed even when the CiA concentration was further reduced to 100 ppm, indicating that the initiation of oxidation in bulk rapeseed oil is transition metal-ion catalysed, as CiA is an efcient metal chelator (Frankel, 2005). Soup powder may accordingly be stabilised against oxidation by addition of CiA, as metal-ion catalysed decomposition of peroxides initiates oxidation in complex food systems, e.g. emulsions, in which several components may be the source of metal ions (Jacobsen & Nielsen, 2007). Ascorbyl palmitate (AP), added corresponding to 300 ppm, increased IP of rapeseed oil oxidation signicantly. AP has been suggested to be located at the oilwater or oilair interface, where the ascorbic moiety regenerates phenolic antioxidants from their one-electron oxidised radicals, chelates metal ions, or reduces lipid hydroperoxides to more stable alcohols (Frankel, 2005). However, AP has also been suggested to have a radical-scavenging activity of its own (Beddows, Jagait, & Kelly, 2001). In the present study, unstripped rapeseed oil was used, and tocopherols are not removed completely in the rening process (Ferrari, Schulte, Esteves, Brhl, & Mukherjee, 1996). It is therefore not possible to conclude whether the antioxidant activity of AP in rapeseed oil was due to radical-scavenging, regeneration of tocopherols or metal chelation. The effectiveness of CiA could indicate that the metal chelation is also important for AP. For the rapeseed oil, addition of antioxidants other than AP and CiA increased IP only marginally. Both rosemary extracts (RE) and carnosic acid (CaA) have in previous studies been found to act as antioxidants in bulk oil systems (Frankel, Huang, Aeschbach, &

Prior, 1996; Trojkov, Rblov, Nguyen, & Pokorny, 2001). However, the effectiveness of antioxidants in bulk oil cannot be directly related to their effectiveness in multiphase foods, such as emulsions or dry soup powder, due to interfacial phenomena (Frankel, Huang, Kanner, & German, 1994), and the antioxidants were accordingly further tested in cauliower soup powder. 3.3. Storage tests with cauliower soup powder Radical formation, in cauliower soup powders with different antioxidants added, was followed, in order to identify the mechanism of lipid oxidation in this complex food product. The ESR method validation gave a RSD of 7% for weak pitch standard, which includes the variation caused by the instrument, including the integration of spectra. For cauliower soup powder, the RSD from the validation trial was 5%; thus, the powder sample itself did not cause any larger variation in the results than the instrument and data handling. The optimised ESR method was considered rather reliable with such low values of RSD, as earlier values of up to 15% were considered to be acceptable (Thomsen, Vedstesen, & Skibsted, 1999). In conclusion, no signicant differences in free radical concentration between soup powders with different antioxidant additions were observed (Fig. 1A). A signicant decrease in free radical concentrations in all soup powders was observed after 2 weeks of storage (Fig. 1A). After this storage period, the radical concentrations became stable between 2 and 4 weeks of storage after which they decreased again signicantly in all soup powders except for the powder with AP (300 ppm) after storage for 8 weeks. The higher free radical concentration in soup powder with CaA added, following 4 weeks of storage, may be atypical, since only samples from one batch were analysed. There were no signicant changes in the free radical concentrations between 8 and 12 weeks of storage. A decrease in free radical concentrations in cappuccino powder stored at relative humidity of 75% has previously been related to an increased water content leading to increased mobility of molecules in powder and allowing radicals to combine, in effect reaching a steady state-concentration (Becker, Madsen, & Skibsted, 2009). Similar observations of a decrease in free radical concentration for several types of food products were explained by increased relative humidity during moisture absorption (Jensen & Risbo, 2007). However, cappuccino powder seems to be more hygroscopic than is cauliower soup powder during storage since the nal aw observed for cappuccino powder (aw = 0.64) was higher than was that for cauliower soup powders (Fig. 1B). The cauliower soup powder was accordingly found not to absorb water, since the water content of the soup powder was constant during the storage period (results not shown), and the decrease in the free radical concentration in cauliower soup powders could not be caused by an increased molecular mobility, but may rather indicate the combination reactions of free radicals within powder ingredients or between dry material particles.

Table 2 Fatty acid composition (% of total FA) of cauliower soup powder at 0 and 12 weeks of storage. The relative standard deviation for each value presented in the Table was below 5%. Cauliower soup powder (no antioxidant) 0 weeks C8:0 C10:0 C12:0 C14:0 C16:0 C18:0 C18:1 C18:1 trans C18:2 C18:3 C18:4 2 1 9 4 36 11 27 1 7 1 1 12 weeks 2 1 9 4 36 10 28 1 7 1 1

Table 3 The induction periods (IP) for rapeseed oil oxidation with different antioxidant additions, observed total colour change, DE, in soup powders between 0 and 12 weeks of storage and degree of unsaturation (DUS) of soup powder acylglycerols obtained from fatty acid compositions. The standard deviations are presented after each particular value. Control (no antioxidant) IP (min) DE (012 weeks) DUS (0 weeks) DUS (12 weeks)
a b c

Rosemary extract 500 ppm 233 6 4.34 0.25 0.63 0.01 0.64 0.01b

Rosemary extract 1000 ppm 264 1 4.48 0.42 0.66 0.04 0.64 0.00b

Carnosic acid 45 ppm 272 16 4.80 0.40 0.63 0.01 0.63 0.01b

Citric acid 1000 ppm a 4.33 0.39 0.62 0.02 0.64 0.00b

Tocopherols extract 1000 ppm 287 26 4.51 0.26 0.63 0.01b 0.65 0.00b

Ascorbyl palmitate 50 ppm 368 1 4.72 0.26 0.63 0.01b 0.64 0.01b

Ascorbyl palmitate 300 ppm 516 17b 3.93c 0.63 0.00b 0.63 0.00b

223 20 4.37 0.09 0.60 0.00 0.60 0.00

No oxidation caused induction period observed. Signicantly different (P < 0.05) from control sample. Value based only on determination from one storage series.

376

R. Raitio et al. / Food Chemistry 128 (2011) 371379

The ESR spectra of cauliower soup powders consisted of a single line with g-values close to 2.0045 which may indicate radicals of amino acid residues of proteins or peptides. Any identication of radicals in the powder spectrum is, however, still very hypothetical as no hyperne structures are observed (Schaich, 2002). The observed spectra may actually be a superposition of two or more underlying spectra related to more than one radical species (Thomsen, Lauridsen, Skibsted, & Risbo, 2005b). Lipid oxidation is in general known to be more signicant at interfaces, and lipid oxidation in cauliower soup powder was expected to play a signicant role in the shelf-life as no preventive actions were taken to limit the oxygen exposure. The fatty acid composition of cauliower soup powder, at 0 and 12 weeks of storage, are presented in Table 2. The most abundant fatty acids in cauliower soup powder were palmitic and oleic acids. The fatty acid methyl ester (FAME) analysis did not reveal any changes in the ratio between unsaturated and saturated fatty acids (DUS values, Table 3) due to the storage in any of the soup powders as the decrease of DUS value would have indicated the oxidation of unsaturated fatty acids. However, slightly higher DUS values were observed in powders with antioxidant additions. As this phenomenon was present already at the beginning of the storage experiment and no effect of the actual 12 weeks storage was observed, it is suggested that the antioxidants may protect the unsaturated fatty acids against the oxidation during the methylation process. Similarly Nissen et al. (2000) observed a lower content of linoleic acid in dehydrated chicken meat, already at the beginning of storage, compared to meat with antioxidant additions. To rule out any protective effect of antioxidants during the methylation process, rapeseed oil without antioxidants and with addition of RE (1000 ppm), CaA (45 ppm), and AP (300 ppm) was methylated to obtain fatty acid compositions for calculation of DUS-values. However, no effect of antioxidants added to the rapeseed oil on DUS-values was observed (results not shown). The higher DUS values in antioxidant soup powders accordingly remained unexplained. The concentration of lipid hydroperoxides as primary lipid oxidation products in the soup powders increased between weeks 0 and 12 as a clear indication of lipid oxidation, and the increase was statistically signicant for all soup powders except for powders with CiA and TP (Fig. 1C). The RSD values were, however, very high throughout the storage (on average ca. 15%), due to large variation between the two different batches of soup powders prepared. Furthermore, the SPME-GCMS analysis on volatiles in the soup powders did not show any increase in secondary lipid oxidation products, such as pentanal and hexanal (Fig. 2A) as the relative concentrations of these compounds were rather stable or even decreased during the storage period. Hence, the analysis of secondary lipid oxidation products did not indicate lipid oxidation in the soup powders, and it was not possible to correlate the data obtained from ESR analysis with lipid oxidation, as in the previous studies with other dry food systems (Nissen, Huynh-Ba, Agerlin Petersen, Bertelsen, & Skibsted, 2002; Nissen et al., 2000; Stapelfeldt et al., 1997). The grouping of volatiles is presented in Table 1. The formation of branched alkanes and alkenes (Fig. 2B) dominated the volatile composition of cauliower soup powders, as their concentration in headspace increased as much as 3500% during the 12 weeks of storage compared to lipid oxidation markers (Fig. 2A), the concentrations of which decreased, on average by 40%, during the same storage period. The origin of the branched alkanes and alkenes is not clear as the alkanes and alkenes from lipid oxidation typically have unbranched carbon chains. The sum of the relative concentrations of 2-methylbutanal and 3-methylbutanal (Fig. 2C) in headspace of the cauliower soup powders increased markedly during the rst 4 weeks, while no clear changes were observed between 4 and 12 week samples. At 12 weeks of storage, the concentrations

of these two volatiles were signicantly larger in powders with RE and AP additions than in control powder without antioxidant addition. The methyl butanals are so called Strecker aldehydes which form in the Maillard reaction, and may contribute to the avour of food. These compounds can further react to form melanoidins, as indicated by more intense browning (Nursten, 2005). Formation of 3-methylbutanal has also been linked to lipid oxidation in dry cream powder as it has been suggested that aldehydes from lipid oxidation may further react with amino compounds, forming dicarbonyls, which in subsequent reactions lead to the formation of Strecker degradation products (Andersson & Lingnert, 1998). On the other hand, Andersson and Lingnert (1998) measured simultaneous increase in the formation of hexanal and 3-methylbutanal, in contrast to what was observed in the present study with cauliower soup powder. The formation of Strecker aldehydes in cauliower soup powders was accordingly concluded not to be coupled directly to lipid oxidation. The formation of 2-acetylpyrrole during the storage of cauliower soup powder (Fig. 2D) further indicated the importance of Maillard reactions in the soup powders, as pyrrole compounds are formed from Amadori compounds (Nursten, 2005). The observed decrease in Hunter L-value (Fig. 1D) and overall change in colour (DE) (Table 3), for all soup powders, provides further evidence for non-enzymatic browning reactions. The overall colour changes, DE (Table 3), did not depend on addition of antioxidant, in agreement with absence of direct coupling between browning reactions and lipid oxidation. One interesting result from analysis of volatiles was the formation of dimethyl disulde (DMDS) in all of the soup powders during storage (Fig. 2E). Volatile organic sulphur compounds are related to the typical avour of cruciferous vegetables, among which dimethyl sulde and methanethiol most likely contribute to some customer rejection for cooked cauliower (Engel, Baty, Le Corre, Souchon, & Martin, 2002), while for hot-water dipped broccoli DMDS and dimethyl trisulde have been found to be responsible for characteristic off-odours (Forney & Jordan, 1998). The formation of DMDS in vegetables can be linked to enzymatic activity (Chin & Lindsay, 1994; Tulio, Yamanaka, Ueda, & Imahori, 2002). The low water activities, characteristic of cauliower soup powder, however, hamper any enzymatic activities and other formation pathways for formation of DMDS in cauliower soup powder must exist. In milk, DMDS is known to be responsible for the burntfeather off-avour following light exposure and photooxidation of methionine side chains of proteins (Jung, Yoon, Lee, & Min, 1998). The cauliower soup powder was stored in the dark but a high level of free radicals initially present in soup powders may initiate free radical oxidations of methionine residues, with a sulphur radical as an intermediate reacting with molecular oxygen (Jung et al., 1998). The sensory analysis indicated a decrease in sensory quality of the cauliower soup following the storage of the soup powders (Fig. 3). However, the soups prepared from powders stored for 4 weeks could not in all cases be differentiated from the blind control (Fig. 3A and E). Moreover, among the antioxidants, RE at 1000 ppm (Fig. 3B), CiA (Fig. 3E), and TP (Fig. 3F) had detrimental effects on the avour of soup following 12 weeks of storage of soup powder, and only the addition of AP (300 ppm) (Fig. 3D) was found to improve the soup compared to the control soup after a 12 week storage of the powder. It is interesting to compare the sensory analysis with the analysis of volatiles. Formations of branched alkanes and alkenes, and methyl butanals were already signicant during the rst 4 weeks of storage, while sensory analyses did not reveal signicant changes in the avour of soups during the same period. These volatiles may accordingly not be among the most signicant contributors to the avour of cauliower soup, probably because they evaporate during the preparation of soups. Instead, DMDS, not present in the fresh cauliower soup powder

R. Raitio et al. / Food Chemistry 128 (2011) 371379

377

Fig. 2. Relative concentrations of volatiles in cauliower soup powder during 12 weeks of storage analysed by SPME-GCMS. Concentration units are arbitrary and not directly comparable to each other. Control; j RE (500 ppm); h RE (1000 ppm); d CaA (45 ppm); s CiA (1000 ppm); N TP (1000 ppm); AP (50 ppm); x AP (300 ppm). The dashed lines are used for the lower concentration when two different concentrations of the same antioxidant were applied. (A) Lipid oxidation markers, (B) branched alkanes and alkenes, (C) methylbutanals, (D) 2-acetylpyrrole, (E) dimethyl disulde, (F) pinenes. The individual volatiles in each group are listed in Table 1.

and with no contribution to the avour of fresh soup, has a moderate increasing concentration in the headspace of the soup powders during the rst 4 weeks, followed by a more signicant increase during storage from week 4 to week 12. As the sensory analysis also indicated that soup made from powders stored for 12 weeks had a signicantly lower avour quality than had the soup made from powder stored for 4 weeks, it may be suggested that DMDS is a good indicator of the quality deterioration of cauliower soup powder. A similar conclusion appears with 2-acetylpyrrole. Decreases in a- and b-pinene concentrations during the accelerated storage (Fig. 2F) were identied as important factors affecting the sensory quality of cauliower soup powders, as a- and b-

pinene are terpines which typically have pleasant avours (Belitz & Grosch, 1999). Although the concentration of pinenes decreased signicantly during the rst 4 weeks of storage, the increases in the concentrations of DMDS and 2-acetylpyrrole were only moderate during this period. This clearly indicates that the sensory quality of cauliower soup is not a result of a single avour substance but rather depends on several factors. Furthermore, it does not seem possible to explain the observed negative effect of some added antioxidants (RE, CiA, and TP) on avour of soups prepared from stored powders by chemical analyses. It should still be emphasised that the sensory analyses were done with the ready cauliower soup and the SPME-GCMS analyses for volatile com-

378

R. Raitio et al. / Food Chemistry 128 (2011) 371379

Fig. 3. Sensory analyses of cauliower soup made from soup powder. The blind control in each series was a soup prepared from fresh cauliower soup powder corresponding to control sample at the beginning of the storage experiment. The ranking was based on the closest to reference principle. (A) Control, Rosemary extract (500 ppm and 1000 ppm after 4 weeks), (B) control, rosemary extract (500 ppm and 1000 ppm after 12 weeks), (C) control, ascorbyl palmitate (50 ppm and 300 ppm after 4 weeks), (D) control, ascorbyl palmitate (50 ppm and 300 ppm after 12 weeks), (E) control and citric acid (1000 ppm after 4 and 12 weeks of storage), (F) control and tocopherol extract (1000 ppm after 4 and 12 weeks of storage). Statistically non-signicant samples are marked with asterix. The maximum of each y-axis tells the number of judges in sensory panel for each series.

pounds were done with dry powder. This may affect the signicance of the particular aroma compounds as they might have different partitioning behaviours in the two different systems. 4. Conclusions The storage stability of cauliower soup powder under mildly accelerated conditions was determined by following the chemical and physical changes in soup powders during storage and by sensory analyses done with the ready soup prepared from the stored powder samples. Furthermore, the mechanisms of deterioration

reactions in the soup powders were examined by studying the effects of addition of different types of antioxidants. The concentrations of free radicals were initially high in the freshly prepared powders and a signicant decrease was already observed after two weeks of storage for all the soup powders, with and without added antioxidants. The decrease in the free radical concentration could not be linked to an increase in primary and secondary oxidation products, as in the previous studies with dry food powders, including milk powders. This difference may be understandable since the cauliower soup powder is made by mixing dry ingredients in contrast to most other dry food powders which are made by

R. Raitio et al. / Food Chemistry 128 (2011) 371379

379

drying food emulsions. Lipid oxidation is concluded to be of little importance for quality deterioration during storage of cauliower soup powders, while formation of volatiles characteristic of Maillard reactions during storage indicated the importance of protein degradation. A browning of the powders is further indicative of protein degradation. Antioxidants added, which had protective effects on the rapeseed oil (used as an ingredient for the soup powder) when studied for the bulk oil, did not enhance the storage stability of soup powder. An observed, decrease in the free radical concentration of soup powder during storage is concluded to be linked to free radical reactions with the protein fraction rather than with lipid oxidation in the soup powder. Acknowledgements The nancial support from OSCAR A/S and FOOD Denmark Research School is gratefully acknowledged. We gratefully thank Heidi Graversen from OSCAR A/S for helpful discussions and Mette Tinglev Pindstrup from OSCAR A/S for performing the sensory analyses. References
Andersson, K., & Lingnert, H. (1998). Inuence of oxygen concentration and light on the oxidative stability of cream powder. Lebensmittel Wissenschaft und Technologie, 31, 169176. Becker, E. M., Madsen, L. S., & Skibsted, L. H. (2009). Storage stability of cappucino powder. Milchwissenschaft, 64(4), 413417. Beddows, C. G., Jagait, C., & Kelly, M. J. (2001). Effect of ascorbyl palmitate on the preservation of a-tocopherol in sunower oil, alone and with herbs and spices. Food Chemistry, 73, 255261. Belitz, H.-D., & Grosch, W. (1999). Food chemistry (2nd ed.). Berlin Heidelberg: Springer-Verlag. Borrelli, R. C., Visconti, A., Mennella, C., Anese, M., & Fogliano, V. (2002). Chemical characterization and antioxidant properties of coffee melanoidins. Journal of Agricultural and Food Chemistry, 50, 65276533. Burton, G. W., & Ingold, K. U. (1981). Autoxidation of biological molecules. 1. The antioxidant activity of vitamin E and related chain-breaking antioxidants in vitro. Journal of American Chemical Society, 103, 64726477. Chin, H.-W., & Lindsay, R. C. (1994). Mechanisms of formation of volatile sulfur compounds following the action of cysteine sulfoxide lyases. Journal of Agricultural and Food Chemistry, 42, 15291536. Elizalde, B. E., Dalla Rosa, M., & Lerici, C. R. (1991). Effect of Maillard reaction volatile products on lipid oxidation. Journal of the American Oil Chemists Society, 68(10), 758762. Engel, E., Baty, C., Le Corre, D., Souchon, I., & Martin, N. (2002). Flavour active compounds potentially implicated in cooked cauliower acceptance. Journal of Agricultural and Food Chemistry, 50, 64596467. Ferrari, R. Ap., Schulte, E., Esteves, W., Brhl, L., & Mukherjee, K. D. (1996). Minor constituents of vegetable oils during industrial processing. Journal of American Oil Chemists Society, 73(5), 587592. Folch, J., Lees, M., & Sloane Stanley, G. H. (1957). A simple method for the isolation and purication of total lipides from animal tissues. Journal of Biological Chemistry, 226, 497509. Forney, C. F., & Jordan, M. A. (1998). Induction of volatile compounds in broccoli by postharvest hot-water dips. Journal of Agricultural and Food Chemistry, 46, 52955301. Francis, F. J., & Clydesdale, F. M. (1975). Food colourimetry: Theory and applications. Westport, Connecticut: The AVI Publishing Company (pp. 143151). Frankel, E. N. (2005). Lipid oxidation (2nd ed.). Bridgwater: The Oily Press, PJ Barnes and Associates. Frankel, E. N., Huang, S.-W., Aeschbach, R., & Prior, E. (1996). Antioxidant activity of a rosemary extract and its constituents, carnosic acid, carnosol, and rosmarinic acid, in bulk oil and oil-in water emulsions. Journal of Agricultural and Food Chemistry, 44, 131135. Frankel, E. N., Huang, S.-W., Kanner, J., & German, B. (1994). Interfacial phenomena in evaluation of antioxidants: Bulk oils vs emulsions. Journal of Agricultural and Food Chemistry, 42, 10541059. Gibbs, P. A. (1986). Microbiological quality of dried foods. In D. MacCarthy (Ed.), Concentration and drying of foods (pp. 89111). London: Elsevier applied science publisher. Hall, C. A., & Cuppett, S. L. (1997). Structure-activities of natural antioxidants. In O. I. Aruoma & S. L. Cuppett (Eds.), Antioxidant methodology. Champaign: Springer (pp. 141172).

Hayashi, T., Mase, S., & Namiki, M. (1985). Formation of the N,N0 -dialkylpyrazine cation radical from glyoxal dialkylamine produced on reaction of a sugar with amine or amino acid. Agricultural and Biological Chemistry, 49(11), 31313137. Hoffman, T., Bors, W., & Stettmaier, K. (1999a). Studies on radical intermediates in the early stage of the nonenzymatic browning reaction of carbohydrates and amino acids. Journal of Agricultural and Food Chemistry, 47, 379390. Hoffman, T., Bors, W., & Stettmaier, K. (1999b). Radical-assisted Melanoidin formation during thermal processing of foods as well as under physiological conditions. Journal of Agricultural and Food Chemistry, 47, 391396. International IDF Standard: Determination of peroxide value, 74A:1991, International Dairy Federation. Jacobsen, C., & Nielsen, N. S. (2007). Optimization of oxidative stability of omega-3 enriched foods. In H. Breivik (Ed.), Long-chain omega-3 specialty oils (pp. 197214). Bridgewater: The Oily Press. Jensen, P., & Risbo, J. (2007). Oxidative stability of snack and cereal products in relation to moisture sorption. Food Chemistry, 103, 717724. Jung, M. Y., Yoon, S. H., Lee, H. O., & Min, D. B. (1998). Singlet oxygen and ascorbic acid effects on dimethyl disulde and off-avour in skim milk exposed to light. Journal of Food Science, 63(3), 408412. Kato, T., Harashima, T., Moriya, N., Kikugawa, K., & Hiramoto, K. (1996). Formation of the mutagenic/carcinogenic imidazoquinoxaline-type heterocyclic amines through the unstable free radical Maillard intermediates and its inhibition by phenolic antioxidants. Carcinogenesis, 17(11), 24692476. Labuza, T. P. (1971). Kinetics of lipid oxidation in foods. CRC Critical Reviews in Food Technology, 2, 355405. Moskowitz, H. R. (1988). Applied sensory analysis of foods (Vol. 1). Boca Raton: CRC Press (pp. 157167). Namiki, M., & Hayashi, T. (1975). Development of novel free radicals during the amino-carbonyl reaction of sugars with amino acids. Journal of Agricultural and Food Chemistry, 23(3), 487491. Nissen, L. R., Huynh-Ba, T., Agerlin Petersen, M., Bertelsen, G., & Skibsted, L. H. (2002). Potential use of electron spin resonance spectroscopy for evaluating the oxidative status of potato akes. Food Chemistry, 79, 387394. Nissen, L. R., Mnsson, L., Bertelsen, G., Huynh-Ba, T., & Skibsted, L. H. (2000). Protection of dehydrated chicken meat by natural antioxidants as evaluated by electron spin resonance spectrometry. Journal of Agricultural and Food Chemistry, 48, 55485556. Nursten, H. (2005). The Maillard reaction; chemistry, biochemistry and implications. Cambridge: The Royal Society of Chemistry. Nursten, H. E. (1986). Maillard browning reactions in dried foods. In D. MacCarthy (Ed.), Concentration and drying of foods (pp. 5368). London: Elsevier Applied Science Publisher. Roberts, R. L., & Lloyd, R. V. (1997). Free radical formation from secondary amines in the Maillard reaction. Journal of Agricultural and Food Chemistry, 45, 24132418. Sarkardei, S., & Howell, N. K. (2008). Effect of natural antioxidants on stored freezedried food product using horse mackerel. International Journal of Food Science and Technology, 43, 309315. Schaich, K. M. (2002). EPR methods for studying free radicals in foods. In M. J. Morello, F. Shahidi, & C.-T. Ho (Eds.), Free radicals in food, chemistry, nutrition, and health-effects (pp. 1234). Washington DC: American Chemical Society. Stapelfeldt, H., Nielsen, B. R., & Skibsted, L. H. (1997). Effect of heat treatment, water activity and storage temperature on the oxidative stability of whole milk powder. International Dairy Journal, 7, 331339. Tan, C. P., Che Man, Y. B., Selamat, J., & Yusoff, M. S. A. (2002). Comparative studies of oxidative stability of edible oils by differential scanning calorimetry and oxidative stability index methods. Food Chemistry, 76, 385389. Thomsen, M. K., Lauridsen, L., Skibsted, L. H., & Risbo, J. (2005a). Temperature effect on lactose crystallization, Maillard reactions, and lipid oxidation in whole milk powder. Journal of Agricultural and Food Chemistry, 53, 70827090. Thomsen, M. K., Lauridsen, L., Skibsted, L. H., & Risbo, J. (2005b). Two types of radicals in whole milk powder. Effect of lactose crystallization, lipid oxidation, and browning reactions. Journal of Agricultural and Food Chemistry, 53, 18051811. Thomsen, M. K., Vedstesen, H., & Skibsted, L. H. (1999). Quantication of radical formation in oil-in-water food emulsions by electron spin resonance spectroscopy. Journal of Food Lipids, 6, 149158. Trojkov, L., Rblov, Z., Nguyen, H. T. T., & Pokorny, J. (2001). Antioxidant activity of rosemary and sage extracts in rapeseed oil. Journal of Food Lipids, 8, 113. Tulio, A. Z., Yamanaka, H., Ueda, Y., & Imahori, Y. (2002). Formation of methanethiol and dimethyl disulde in crushed tissues of broccoli orets and their inhibition by freezethawing. Journal of Agricultural and Food Chemistry, 50, 15021507. Velasco, J., Andersen, M. L., & Skibsted, L. H. (2004). Evaluation of oxidative stability of vegetable oils by monitoring the tendency to radical formation. A comparison of electron spin resonance spectroscopy with the rancimat method and differential scanning calorimetry. Food Chemistry, 85, 623632. Weber, R. T. & Heiss, A. A. (1992), ECS 106 EPR Spectrometer Users Manual, Bruker Instruments Inc., (pp. 7.3). Billerica Massachusetts USA. Weil, J. A., Bolton, J. R., & Wertz, J. E. (1994). Electron paramagnetic resonance. Elementary theory and practical applications. New York: John Wiley & Sons (pp. 492519).