Identification of a medicinal off-flavour in mineral water


Раздел науки:

Читать онлайн: 
Текст статьи: 
Andrea Strube a,*, Helmut Guth b, Andrea Buettner a,c a Fraunhofer Institute for Process Engineering and Packaging (IVV), Giggenhauser Str. 35, D-85354 Freising, Germany b Bergische University of Wuppertal, Department of Food Chemistry, Gaußstraße 20, D-42119 Wuppertal, Germany c Institute of Pharmacy and Food Chemistry, Department of Food Chemistry, University Erlangen-Nu ¨rnberg, Schuhstr. 19, D-91052 Erlangen, Germany a r t i c l e i n f o Article history: Received 29 June 2009 Received in revised form 18 August 2009 Accepted 20 August 2009 Available online 26 August 2009 Keywords: Off-odour Sensory analyses Odour threshold 2-iodophenol 2-iodo-4-methylphenol 1D and 2D gas chromatography–mass spectrometry/olfactometry a b s t r a c t The aim of the present study was to identify the compounds responsible for a characteristic medicinal off-odour in mineral water. 2-Dimensional high resolution gas chromatography–mass spectrometry analysis, in combination with olfactometry (HRGC–MS/O), led to the identification of 2-iodophenol and 2-iodo-4-methylphenol as the two compounds exhibiting the medicinal odour notes in mineral water. Determination of odour thresholds of these two compounds in air and water, as well as investigation of their odour characteristics in specific structure–odour activity assays, showed that 2-iodophenol and 2-iodo- 4-methylphenol were both extremely potent odour substances with odour thresholds down to 0.03 and 0.0003 ng/L air and 0.313 and 0.009 mg/L water, respectively. The higher odour potency of 2-iodo-4-methylphenol suggests that this compound is primarily responsible for the medicinal off-odour. Furthermore, the same proportion of the panellists who were not able to detect the off-odour in mineral water exhibited drastically decreased sensibility to 2-iodo-4-methylphenol, but not to 2-iodophenol, which supports this hypothesis. ª 2009 Elsevier Ltd. All rights reserved. 1. Introduction The formation of a medicinal off-flavour in food and drinks has beenextensivelyreportedintheliteratureandisalwaysdescribed as being medicinal, phenolic, iodoform-like and/or chemical. This taint has been postulated, for example, to result from a series of haloorganic compounds such as halogenated phenols (Suffet et al., 1999; Peter and von Gunten, 2007). 2-chloro-6-methylphenol, for example, has been linked to a medicinal off-flavour in biscuits, empty drink cans and chicken meat, and 2-bromo-4-chlorophenol has been reported to induce a chemical, phenolic offflavour in brine-salted Gouda cheese (Mottram, 1998). Additionally, Sanchez Saez et al. (1991) identified 4-bromo-2-chlorophenol as the reason for a medicinal off-odour in melons, and Sevenants and Sanders (1984) described iodocresol (unspecified isomer) as the cause of a medicinal off-flavour in a commercial cake mix. Similarly, the medicinal off-flavour of fish and prawns has been supposedtobecausedbythepresenceofdifferentbromophenols, and a phenolic off-flavour of table wine was said to be due to 2-methoxy-4-vinylphenol, which is formed during alcoholic fermentation via decarboxylation of ferulic acid by yeast (van Wyk and Rogers, 2000). Furthermore, Gocmen et al. (2004) identified guaiacol, 2,6-dichlorophenol and 2,6-dibromophenol as the major contributors to the medicinal taint in orange juice, and Steeg and co-workers reported 2,4,6-trichlorophenol as the responsible compound for a medicinal taint in orange juice (Steeg et al., 1990). The halogenated phenols 2,4-dichlorophenol, 2- bromo-4-chlorophenol, 4-bromo-2-chlorophenol, 2,4,6-trichlorophenol, 2,4-dibromophenol, 4-bromo-2,6-dichlorophenol and 2,4,6-tribromophenol have also been identified as the * Corresponding author. Fax: þ49 (0)8161 491 777. E-mail address: (A. Strube). Available at journal homepage: 0043-1354/$ – see front matter ª 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.watres.2009.08.026 w a t e r r e s e a r c h 4 3 ( 2 0 0 9 ) 5 2 1 6 – 5 2 2 4 substances responsible for a medicinal off-flavour in fish-sauce (Bisling and Schreiber, 1990). In the drinking water industry, the formation of odour problems due to microorganisms (e.g. responsible for musty odours) can be minimised with oxidation processes, such as ozonation and chlorination. A well-known and repeatedly occurring side-effect of these very effective water treatment methods is the formation of by-products such as brominated or chlorinated phenols and/or trihalomethanes, e.g. iodoform or bromoform (Bruchet et al., 1991). Other investigators have reported the presence of iodinated halo forms – at concentrations between 0.3 and 10 mg/L – as the cause for medicinal odour problems in drinking water (Cancho et al., 2001). Analysis of drinking water before and after chlorination has additionally shown that the untreated water exhibited a musty odour, while the chlorinated water was described as medicinal (Bruchet et al., 1991). Furthermore, it was reported that the concentration of dibromochloromethane, dichloroiodomethane, dibromoiodomethane and iodoform increased with chlorination. The author also detailed the odour characters of dichloroiodomethane and dibromoiodomethane, which were described as sweet, solventlike, and perfumed, and that of iodoform as medicinal, sweet, and solvent-like (Bruchet et al., 1991). The medicinal taint in mineral water was first attributed to the presence of iodoform by Lindner in 1936 (Lindner, 1936). Eristavi et al. (1962) identified 2-chlorophenol as the reason for the medicinal taint in mineral water. The formation of 2-chlorophenol has been described as being due to the reaction of active chlorine (present in the tanks after periodic disinfection) with phenols (contained in the water). Actually, the substance responsible for the periodic appearance of the medicinal off-flavour in mineral water has hitherto not been clarified. The aim of the present study was to identify the compounds responsible for the medicinal taint in mineral water. To achieve this goal, mineral water samples that were returned by customers due to the medicinal taint were investigated. Sensory analyses of the samples were carried out by a trained panel using descriptive techniques and were rated according to their odour intensities. Chemical analyses were additionally performed; odour concentrates of the samples were obtained by solvent extraction using dichloromethane and subsequent mild extract concentration, followed by analyses using 2-dimensional high resolution gas chromatography–mass spectrometry in combination with olfactometry (HRGC–MS/O). In addition, the odour thresholds of the compounds identified to be responsible for the medicinal taint were evaluated. 2. Materials and methods 2.1. Samples Five different mineral water samples (from Germany, Switzerland and Austria), returned by customers due to the specific medicinal off-odour, were analysed. Three of these samples were stored in glass and two samples were stored in plastic bottles. Each sample was analysed in comparison with a reference mineral water sample without the specific offodour. Five replicates of each sample were analysed. 2.2. Chemicals The following compounds were obtained from the suppliers shown: dichloromethane (Merck, Darmstadt, Germany), goctalactone 95% (EGA-Chemie, Steinheim, Germany), vanillin 99%, 1-octen-3-one 97% (ABCR Product List, Karlsruhe, Germany), (E,E )-deca-2,4-dienal 85%, (E )-dec-2-enal 95%, (Fluka, Steinheim, Germany), ethyl-vanillin 98%, (E )- non-2-enal 97%, 2-iodophenol 98%, bromoform 99%, 2-chlorophenol 96%, 2-bromophenol 98%, iodoform 99%, 2,6-dibromophenol 99%, bromodichloromethane 99%, bromoiodomethane, chloroiodomethane 97%, chloroform 99% (Aldrich, Steinheim, Germany), and tetrachloromethane methanol solution 1 mg mL1 (TCI, Eschborn, Germany). The following compounds were provided by AromaLab GmbH, Freising, Germany: (E )-4,5-epoxy-(E )-2-decenal 95% and 2-iodo-4-methylphenol 98%. All compounds were freshly distilled prior to analysis. Chemical and sensory purities were checked by HRGC–O as well as by HRGC–MS. 2.3. Sensory analyses Sensory analyses were performed in a sensory panel room at 21  1 C during three different sessions. The samples (in each case 50 ml in covered glass vessels; capacity 45 ml, i.d. 40 mm) were singly presented to the panel for orthonasal sensory evaluation. The order of the samples was randomised and no information about the purpose of the experiment or the exact composition of the samples was given to the panellists. The panellists were required to open the glass vessel and to list the odour impressions for orthonasal analysis. 2.3.1. Panellists Panellists were 12 trained volunteers (2 male, 10 female, age range from 24 to 41 with a mean age of 30) from the Fraunhofer IVV Institute (Freising, Germany), exhibiting no known illness at the time of examination and with normal olfactory and gustatory function. Prior to participation in the experiments panellists were tested for their olfactory functions during weekly training sessions with selected suprathreshold aroma solutions. Subjective aroma perception was normal throughout the training and at the time of the experiments, as tested with a defined set of aroma substances and an internally developed flavour language. Participation in these sessions was for at least half a year prior to participation in the actual sensory experiments. 2.3.2. Descriptive analysis The panellists were asked to describe the samples (collection of sensory attributes) and to score the respective intensities on a scale from 0 (no perception) to 3 (strong perception). The samples were singly presented to the sensory panel for evaluation in covered glass vessels (capacity 45 mL, i.d. 40 mm). The results obtained in three different sessions for each sample were averaged. The values obtained in different w a t e r r e s e a r c h 4 3 ( 2 0 0 9 ) 5 2 1 6 – 5 2 2 4 5217 sessions and for the different assessors differed by no more than 10%. 2.3.3. Determination of odour threshold concentrations (OTCs) The OTCs of 2-iodophenol and 2-iodo-4-chlorophenol in air were determined as previously described by Ullrich and Grosch (1987). The odour thresholds were evaluated by four trained panellists who exhibited no anosmia to the analysed compounds. The OTCs in water were determined as previously described by Czerny et al. (2008). 2.4. Solvent extraction of mineral water samples The most important prerequisite for successful accomplishment of aroma extract dilution analyses (AEDA) is the greatest possible aroma similarity between the odour impression of the original sample and that of the extract (Grosch, 2001). For preparation of the extract, 200 mL of mineral water were filled into an Erlenmeyer flask and extracted three times, each with 80 mL of dichloromethane. After 30 min of extraction at room temperature the dichloromethane phases were separated in each case. The combined dichloromethane phases (240 mL) were subsequently dried over anhydrous sodium sulfate and finally concentrated to a total volume of 100 mL at 50 C by means of Vigreux distillation and distillation according to Bemelmans (1979). 2.5. High resolution gas chromatography–olfactometry (HRGC–O) HRGC was performed with a helium GC type 5160 (Thermo Finnigan, Dreieich, Germany) using the following capillaries: DB-FFAP (30 m  0.32 mm fused silica capillary, free fatty acid phase FFAP, 0.25 mm; Chrompack, Mu ¨ hlheim, Germany) and DB-5 (30 m  0.32 mm fused silica capillary DB-5, 1.5 mm; J&W Scientific, Fisons Instruments, Waldbronn, Germany). Application of the samples into the GC system was performed at 40 C using the cool-on-column technique. After 2 min, the temperature of the oven was raised at 10 C min1 to 240 C and held for 5 min. The flow rate of the helium carrier gas was 2 mL min1. At the end of the capillary the effluent was split into a sniffing port and an FID using two deactivated but uncoated fused silica capillaries (100 cm  0.32 mm). The FID and the sniffing port were held at 300 C and 250 C, respectively. The linear retention indices (RIs) of the compounds were calculated as previously described by Dool and Kratz (1963). 2.6. 2-Dimensional HRGC–MS/O analysis (2D HRGC–GC–MS/O) The 2D HRGC–GC–MS/O system was a coupled system consisting of two helium CP 3800 GCs (Varian, Darmstadt, Germany) in combination with a Saturn 2200 MS (Varian, Darmstadt, Germany). The 2-dimensional measurements were performed using the following capillaries: DB-FFAP (30 m  0.32 mm fused silica capillary, free fatty acid phase FFAP, 0.25 mm; Chrompack) in the first oven and DB-5 (30 m  0.25 mm fused silica capillary DB-5, 1.5 mm; J & W Scientific, Fisons Instruments) in the second oven. Application of the samples to the GC system was performed at 40 C using the cool-on-column technique. After 2 min the temperature of the oven was raised at 6 C min1 to 230 C and held for 5 min in the first oven, and at 10 C min1 to 240 C and held for 5 min in the second oven. The flow rate of the helium carrier gas was 2.5 mL min1. At the end of the capillary the effluent was split into a sniffing port and an FID or alternatively an MS, using two deactivated but uncoated fused silica capillaries (100 cm  0.20 mm). The FID and the sniffing port were both held at 300 C. MS analysis was performed with a Saturn 2200 MS (Varian, Darmstadt, Germany) with the capillaries described above. Mass spectra in the electron impact (MS/EI) mode were generated at 70 eV ionisation energy for a m/z range of 35–399. Mass spectra in the chemical ionisation mode (MS/CI) were acquired with methanol as the reagent gas at a flow of 2.5 mL min1. The scan rate was 500 and the m/z range was 60–600. 2.7. Identification of odorants In all cases, identification of the odour-active substances was based on the odour character and intensity, the retention indices on three capillary columns of different polarity (DBFFAP, DB-5 and DB-1701), and the mass spectral data (EI and CI) in comparison with the reference compounds. 2.8. Aroma extract dilution analysis (AEDA) The flavour dilution (FD) factors of the key aroma compounds were determined by AEDA (Guth and Grosch, 1999) from the following dilution series: the original extract of 50 mL (prepared as described in ‘‘solvent extraction of mineral water samples’’) was diluted stepwise (1 þ 1, v/v) with dichloromethane. HRGC–O was then performed on the original extract (FD ¼ 1) and on aliquots of 1 mL of 1 þ 1 dilutions using DB-5. 3. Results and discussion 3.1. Sensory evaluation All of the five tested samples were identified with a distinct offflavour. Orthonasal sensory evaluation (descriptive analysis) showed the attribute medicinal to be the most intense (odour intensity 3) and characteristic in all five samples, with minor ratings of the phenolic and iodo-like impressions (odour intensities 1 and 2, respectively). Also, the odour characters fruity and fatty (odour intensities between 1 and 2, respectively) were detected. Generally, most panel members confirmed the presence of the intense medicinal off-odour note in all effected (returned) water samples, with seven of the twelve panellists describing the off-odour of the mineral water consistently. However, five panellists described the smell of the water samples as fruity and fatty, and did not notice any medicinal smell. This observation was consistent for all identified medicinal smelling water samples for these five panellists. 5218 w a t e r r e s e a r c h 4 3 ( 2 0 0 9 ) 5 2 1 6 – 5 2 2 4 3.2. Identification of the aroma active compounds 3.2.1. 1D HRGC–O/FID analysis Characterisation of odour-active substances by means of onedimensional HRGC–FID/O, both for solvent-extracted samples of references and the off-odour mineral waters led to the olfactory detection of 10 and 12 odour-active substances, respectively. Ten odour-active substances with non-medicinal odour notes, such as fatty, metallic, and sweet, were detected in all water samples. By comparison, two odouractive substances with the specific medicinal odour note were detected only in the respective medicinal smelling water samples. The medicinal notes were present in high intensities in all five off-flavour samples (FD 256, respectively, cf. Table 1), but were not present in the control sample. In Fig. 1, a representative 1D HRGC–O/FID-chromatogram of an effected sample shows that no signal detection via FID was obtained for the two key odorants with the specific medicinal odour character due to the low concentrations of the target compounds. 3.2.2. 2D HRGC–O/MS analysis For further characterisation additional analyses on solvent extracts from 4 L mineral water, after concentration of these extracts to a volume of 50 mL, were carried out using a 2D HRGC–MS/O system. This enabled a higher resolution of the target compounds to be achieved and eliminated background contaminations of the analytes (cf. Fig. 2). In doing so, mass spectral detection both in CI and EI mode was achieved for one medicinal target substance. The mass spectral detection of the second medicinal target compound was achieved after the injection of 40 mL overall (five injections of 8 mL each into the 2-dim GC–MS system). The five cuts from the first oven run were all pooled by collection in the cryotrap, and then transferred as one collective sample into the second oven for analysis. The representative mass spectrum is shown in Fig. 3. Based on these identification parameters, as well as the respective RIs on three capillaries of different polarities and mass spectral data (EI and CI, respectively), the two medicinal odour compounds were unequivocally identified as 2-iodophenol and 2-iodo-4-methylphenol (Table 2). Additional compounds detected were (Z )-non-2-enal, (E )- non-2-enal, (E )-dec-2-enal, (E,E )-deca-2,4-dienal and (tr)- 4,5-epoxy-(E )-dec-2-enal with fatty and plastic-like odour impressions, and oct-1-en-3-one, non-1-en-3-one (mushroomlike), g-octalactone (sweet, coconut-like), vanillin and ethylvanillin. However, none of these substances elicited the specific medicinal smell and the odour intensities of these compounds were lower (Table 1). These carbonyl odorants have just recently been described to be responsible for the socalled ‘‘sunlight’’ off-odour in mineral water (Strube et al., in press). This off-odour compounds can be generated, most specifically, from ultra-violet light exposure and extended storage of mineral water and is induced by decomposition of packaging constituents. Traces of these odorants are generally found in mineral waters, even if no distinct off-odour is detectable, so that detection in water samples, as reported in this study, is well in line with our previous observations. Accordingly, contribution of these substances to the medicinal off-odour described herein can be excluded. To check the presence of iodoform and 2-iodophenol, described previously by other groups to be responsible for a medicinal taint in mineral water, the respective standard solutions were analysed in parallels by means of HRGC–FID/ MS/O measurements. However, it was shown that none of these substances was present in any of the investigated offflavour samples, most specifically not at concentrations representing their odour thresholds in water. Also, the analytical parameters such as RIs and mass spectral data did not agree with the detected medicinal odorants of this study. In this context it is interesting to note that Thomas et al. (1980) reported limited stability of trihalomethanes in solvents or Table 1 – Most odour-active volatiles (FD ‡ 4) in solvent extracts from mineral water (comparison of off-flavour sample with control). No Odour charactera Retention indexb FD-factor Odorantc DB-5 DB-FFAP Reference sample Effected sample 1 Mushroom-like 976 1296 8 8 Oct-1-en-3-one 2 Mushroom-like 1079 1393 8 8 Non-1-en-3-one 3 Fatty 1145 1497 8 16 (Z )-Non-2-enal 4 Fatty 1162 1527 8 8 (E )-Non-2-enal 5 Medicinal 1203 2219 n. d. 256 2-iodophenol 7 Sweet, coconut-like 1253 1911 16 16 g-octalactone 6 Fatty 1256 1629 16 8 (E )-Dec-2-enal 8 Medicinal 1308 2322 n. d. 256 2-iodo-4-methylphenol 9 Plastic-like, fatty 1322 1802 16 16 (E,E )-Deca-2,4-dienalc 10 Metallic 1375 1998 128 64 (tr)-4,5-Epoxy-(E )-dec-2-enal 11 Vanilla-like 1400 2580 32 64 Vanillin 12 Vanilla-like 1477 2502 8 8 Ethyl-vanillin n.d. – Odour impression was not detected. a Odour quality as perceived at the sniffing port. b Linear retention indices (RIs) were calculated as described by DooI and Kratz (1963). c Odorants were identified by MS, comparison of retention indices of reference standards and odour perceived by sniffing analyses. w a t e r r e s e a r c h 4 3 ( 2 0 0 9 ) 5 2 1 6 – 5 2 2 4 5219 water. Accordingly, the stability of trihalomethanes in solvents and drinking water has been reported to be two months for a concentration of 5 ppb and one month for a solution containing a concentration of 0.5 ppb (Thomas et al., 1980). In contrast to this, our studies show that even in samples that were stored for more than three years, the medicinal odour was still clearly detectable. Accordingly, the responsible odour substances must be extremely stable in the water samples under standard storage conditions so that contributions from instable trihalomethanes are unlikely. 3.3. Sensory properties of 2-iodophenol and 2-iodo-4-methylphenol The odour characters of the two identified compounds were evaluated in different concentration steps (concentration range: 0.02–51.4 mg/L water for 2-iodophenol and 0.0012– 2.77 mg/L for 2-iodo-4-methylphenol) via triangular testing by the panel members. The odour character of an aqueous solution of 2-iodophenol was always described as medicinal and band-aid-like in all concentrations tested by all twelve panellists, i.e. no odour quality change was observed as function of concentration. The OTC of 2-iodophenol was determined for all panellists to be 0.02-ng/L air and 0.3-mg/L water, respectively (Table 2). The odour impression of the solution containing 2-iodo-4- methylphenol changed with increasing concentration. The panellists who detected the medicinal odour in the mineral water described the odour impression as medicinal at concentrations between 0.01 and 0.10 mg/L. The impression at higher concentration was described as phenolic. The five panellists who did not elicit the medicinal odour in the mineral water did not respond at concentrations lower than 0.3 mg/L. These panellists described the odour character between 0.3 and 2.7 mg/L as medicinal and band-aid-like. The odour threshold for 2-iodo-4-methylphenol in water was evaluated with 0.01 mg/L (sensitive panellists) and 0.3 mg/L (insensitive panellists), respectively. The OTC in air was exclusive evaluated by sensitive persons and was determined with 0.0003 ng/L air. Differential detection for OTCs was well in line with successful or unsuccessful detection of the odour in the offodour mineral water samples. In addition, GC–O analysis of the reference odorant on three columns of different polarities clearly showed that only 2-iodo-4-methylphenol was present as an odour-active compound in the reference solution and that no other trace impurities elicited the differential response. In contrast to this, the OTC of 2-iodophenol did not show such a markedly differential sensory behaviour, resulting in comparable OTC values for all panellists for this substance. This might indicate that 2-iodo-4-methylphenol is the main compound responsible for the medicinal off-odour in mineral water. However, proof of this assumption needs to be carried out by additional quantification and reconstitution experiments. Also, possible synergistic effects of both identified compounds need to be checked in further studies (Strube, in preparation). 3.4. Discussion of possible pathways of the identified substances Regarding possible sources of 2-iodophenol, several previous studies might give indications of possible pathways: It has been reported, that formation of bromo- and chlorophenols can arise in mixtures of the respective halogen constituents and phenol, and that this process depends predominantly on pH and on the bromine-to-phenol and chlorine-to-phenol ratio, respectively. With regard to the chloro compounds, a high odour intensity was obtained due to the formation of 2,6-dichlorophenol obtained at a chlorine-tophenol ratio of 2:1, which increased even further at a ratio of 4:1. The optimum pH value for the formation of chlorophenols has been reported to be between 8 and 9 (Suffet et al., 1999). Mirlohi (1997) reported that the reaction of iodine in aqueous 10,0 20,0 30,0 40,0 50,0 60,0 1 800 900 1000 1100 1200 1300 1400 1500 1600 1700 1800 RI-DB 5 FID response 1 2 3 4 5 6 7 9 10 11 8 12 Fig. 1 – FID-chromatogram of the GC sniffing analysis: the most odour-active volatiles (FD ‡ 4) in all analysed samples are shown (numbers correspond to Table 1). 5220 w a t e r r e s e a r c h 4 3 ( 2 0 0 9 ) 5 2 1 6 – 5 2 2 4 Minutes -17 0 50 100 150 mVolts c:\varianws\data\strube\eigenforschung\medizinisch\kas\med_curt2220_3_ File:c:\varianws\data\strube\eigenforschung\medizinisch\kas\ Channel:Front = FID Last recalc:NA X: 7.6540 Minutes Y:4.69 mVolts 0.00 0.25 0.50 0.75 1.00 kCounts 35:399 0.00 0.25 0.50 0.75 1.00 1.25 1.50 MCounts 35:399 med_cut3_2220_ _11-19-2008 .sms TIC Filtered Sample Notes: inj 4 µl 1st oven 2nd oven TIC m/z 35-399 SIM m/z 220 this part was transfered to the second oven 2.5 5.0 7.5 10.0 17.5 5 10 15 20 25 minutes med_cut3_2220_ _11-19-2008.sms Ions: 220.0 Filtered Sample Notes: inj 4 µl 5 10 15 20 25 minutes medicinal 1 2 3 Fig. 2 – Example of 2D GC–MS/O analysis for an increased resolution in order to identify the off-flavour compounds; displayed here for 2-iodophenol. (1 [ one-dimensional GC–FID; 2 [ two-dimensional GC–MS in full scan; 3 [ twodimensional GC–MS filtered for m/z 220). w a t e r r e s e a r c h 4 3 ( 2 0 0 9 ) 5 2 1 6 – 5 2 2 4 5221 solutions containing organics is similar to chlorine reactions. However, iodine is the least reactive reagent in comparison with the other halogens. It has been demonstrated that iodine forms iodoform in the presence of alkane series and could react with phenol to phenolic disinfection by-products (e.g. 2- iodophenol, 4-iodophenol), which have also been shown to be odorous. The OTCs for 2-iodopenol and 4-iodophenol were indicated with 1 mg/L (factor five higher than the determined minutes 0 100 200 300 400 500 Counts med_cut_ref_ic__4-16-2009.sms Ions: 234.0 Filtered Sample Notes: Referenz 2-iodo-4-methylphenol 35:249 2A 5 10 15 20 25 50 100 150 200 250 m/z 0% 25% 50% 75% 100% 51 53 77 78 107 127 233 234 Spectrum 2A, Ion: 23463 us, RIC: 578, Scan: 663, 10.669 min BP: 234 (197=100%), med_cut_ref_ic__4-16-2009.sms 0 50 100 150 200 250 300 Counts med_cut_ic_1_3__4-17-2009.sms Ions: 234.0 Filtered Sample 35:249 1A 5 10 15 20 25 minutes 50 100 150 200 250 m/z 0% 25% 50% 75% 51 55 77 78 107 127 235 Spectrum 1A, Ion: 13471 us, RIC: 2369, Scan: 658, 10.616 min BP: 234 (760=100%), med_cut_ic_1_3__4-17-2009.sms 1 1 2-iodo-4-methyl phenol 100% 234 2 2 2-iodo-4-methyl phenol Notes: 8µl Fig. 3 – Identification of 2-iodo-4-methylphenol in the off-flavour mineral water. (EI mass spectra data of 2-iodo-4- methylphenol are presented. Comparison of reference solution (1) and the sample (2). OTCs in the present study) and 0.5 mg/L water, respectively. The odour characters were described as chemical, phenolic and medicinal. It has also been reported that iodination of phenol proceeds by initial substitution of iodine on the 2nd and 4th positions of the aromatic ring. In the second order the iodination has been found to proceed with further substitutions on the aromatic ring to diiodophenol, triiodophenol leading to the formation of non-phenolic smelling oxidation compounds. The formation of iodophenol in the presence of phenol and iodine in drinking water was also reported by Dietrich et al. (1999). Furthermore, it has been shown that the concentration of phenol disappeared in presence of iodine. 2- iodophenol has also been shown to be the major iodinated phenol that was produced. de Rossi and Veglia (1986) studied the formation of iodophenols in aqueous solutions that contained iodide and phenol, where it was shown that the formation of the ortho isomer was more prevalent than the formation of the meta and para isomers, and that the formation of diiodinated phenols mainly resulted from the iodination of 2-iodophenol. 2-iodo-4-methylphenol as an odour impact contributor to medicinal taint in mineral water has, to our knowledge, never been reported before. Sevenants and Sanders (1984) identified an unspecified isomer of iodomethylphenol as the cause of a medicinal taint in commercial cake mix. It has been assumed that this formation resulted from the reaction of iodide (ingredient of iodised salt) with 4-methylphenol (flavour constituent of lemon) (Sevenants and Sanders, 1984). Ettinger and Ruchhoft (1951) previously reported that 4- methylphenol reacts in the presence of chlorine to compounds with high odour activities, described as chlorophenol-like. In many studies it has been reported that the reaction of iodine is similar to the reaction of other halogens (Mirlohi, 1997). Therefore, it could be assumed that the presence of 2- iodo-4-methylphenol results from the reaction of iodine with 4-methylphenol. However, 4-methylphenol was not detected in the mineral water samples of the present study. It could not be excluded that the 4-methylphenol was degraded completely during the reaction with iodide. Accordingly, further studies are needed to clarify the pathway for elucidation of 2-iodo-4-methylphenol (Strube et al., in press). The findings of the present study also have broader implication with regard to the rating of consumer complaints and sensory evaluations. The off-odour example demonstrates that sensory taints might occur which are very potent for only a fraction of all consumers while others are not able to detect the problem et all. With regard to statistical interpretation of data in sensory evaluations this effect might led to biased rendering results off-odour detection insignificant. Therefore, differential sensitivity of consumers might mislead the off-odour assessment as discussed previously (Bu ¨ ttner et al., 2007). Accordingly detailed knowledge on structure– activity relationships are hence demonstrated as being crucial for off-odour characterisation and assessment. 4. Conclusions The present study shows that a characteristic medicinal taint that is repeatedly found in mineral water was related to the presence of 2-iodophenol and 2-iodo-4-methylphenol. It was interesting to show, that the medicinal off-flavour was still present after a storage time of more than three years. Sensory experiments and different sensitivities of panel members indicated that specifically the 2-iodo-4-methylphenol is the key contributor to this medicinal off-odour. This observation is supported by the fact that five panel members were not able to detect the intense off-odour samples and that they responded, likewise, with significantly reduced aroma sensitivity to 2-iodo-4-methylphenol. On the other hand, no differences in response were observed between panellists for Table 2 – Sensory and analytical properties of 2-iodophenol and 2-iodo-4-methylphenol, responsible for the medicinal offflavour in mineral water. Odorant CAS MW Odour character a Retention index onb Odour threshold Mass spectra FFAP DB-5 DB-1701 Airc ng L1 Water d mg L1 EI CI 2-iodophenol 533–58–4 220 medicinal, band-aid-like 2219 1203 1379 0.02 0.2e 0.3f 220 (100), 65 (31), 39 (15), 93 (13), 63 (10), 127 (9), 221 (7), 64 (5), 62 (4), 92 (4) 221 (100), 94 (9), 65 (9), 222 (5), 2-iodo-4- methylphenol 16188–57–1 234 medicinal 2322 1308 1485 0.0003 0.003e 0.01f 234 (100), 107 (39), 77 (33), 127 (12), 51 (11), 78 (8), 79 (8), 50 (8), 235 (7), 233 (7), 53 (5), 52 (4) 108 (100), 235 (16), 109 (8) a Odour quality as perceived at the sniffing port. b Linear retention indices (RIs) were calculated as described by DooI and Kratz (1963). c Odour thresholds in air, determined as previously described by Ullrich and Grosch (1987). d Odour thresholds in water, determined as previously described by Czerny et al. (2008). e Detection limit. f Recognition limit. w a t e r r e s e a r c h 4 3 ( 2 0 0 9 ) 5 2 1 6 – 5 2 2 4 5223 2-iodophenol. Furthermore, a medicinal off-flavour formation due to the presence of trihalomethanes such as bromoform and iodoform, as reported previously in literature, could be excluded in the present study. r e f e r e n c e s Bemelmans, J.H.M., 1979. Review of isolation and concentration techniques. In: Land, D.G., Nursten, H.E. (Eds.). Appl. Sci. Publ., London, pp. 79–98. Bisling, M., Schreiber, W., 1990. Auftreten von halogenphenolen in garba ¨rden fu ¨ r marinaden. Z Lebensmittel Untersuchung Forsch A 190, 31–33. Bruchet, A., N’Guyen, K., Mallevialle, J., Anselme, C., 1991. Identification and behavior of iodinated haloform medicinal odor. In: Org. Micropollut. Aquat. Environ., Proc. Eur. Symp., 6th, pp. 371–383. Cancho, B., Fabrellas, C., Diaz, A., Ventura, F., 2001. Determination of the odor threshold concentrations of iodinated trihalomethanes in drinking water. Journal of Agricultural and Food Chemistry 49, 1881–1884. Czerny, M., Christlbauer, M., Christlbauer, M., Fischer, A., Granvogl, M., Hammer, M., Hartl, C., Moran Hernandez, N., Schieberle, P., 2008. Re-investigation on odour thresholds of key food aroma compounds and development of an aroma language based on odour qualities defined aqueous odorant solutions. European Food Research and Technology 228, 265–273. Dietrich, A.M., Mirlohi, S., DaCosta, W.F., Dodd, J.P., Sauer, R., Homan, M., Schultz, J., 1999. Flavor profile analysis and GC/MS detection of phenolic iodinated disinfection byproducts in drinking water for the USA space program. Water Science and Technology 40 (6), 45–51. Dool, v.d.H., Kratz, P., 1963. A generalization of the retention index system including linear temperature programmed gas– liquid partition chromatography. Journal of Chromatography 11, 463–471. Eristavi, D.I., Salukvadze, E.K., Darchashvili, T.V., Kachukhashvili, S.I., Brouchek, F.I., Goguadze, M.I., Dzhincharadze, G.G., 1962. The causes of the periodic appearance of an iodoform odor in Borzhom mineral water. Trudy Gruzinskogo Politekhnicheskogo Instituta 2, 33–43. Ettinger, M.B., Ruchhoft, C.C., 1951. Effect of stepwise chlorination on taste -and odor -producing intensity of some phenolic compounds. Journal - American Water Works Association 43, 561–567. Gocmen, D., Elston, A., Williams, T., Smoot, J., Parish, M., Rouseff, R., 2004. Identification of antiseptic off-flavours formed from thermophilic spore-forming bacteria using GC–O and GC–MS. state-of-the-art in flavour chemistry and biology. In: Proceedings of the Wartburg Symposium on Flavour Chemistry and Biology, 7th, Eisenach, Germany, Apr. 21–23 2004, pp. 240–247. Grosch, W., 2001. Evaluation of the key odorants of food by dilution experiments, aroma models and omission. Chemical Senses 26, 533–545. Guth, H., Grosch, W., 1999. Evaluation of important odorants in foods by dilution techniques. Flavor Chemistry 30, 377–386. Lindner, E., 1936. Iodoform odor of mineral water containing iodine. Chemiker-Zeitung 60, 426. Mirlohi, S., 1997. Chemical Identification and Flavor Profile Analysis of Iodinated Phenols Produced from Disinfection of Spacecraft Drinking Water. Virginia Polytechnic Institute And State University, Blacksburg, Virginia. Mottram, D.S., 1998. Chemical tainting of foods. Journal of Food Science and Technology 33, 19–29. Peter, A., von Gunten, U., 2007. Oxidation kinetics of selected taste and odor compounds during ozonation of drinking water. Environmental Science & Technology 41 (2), 626–631. de Rossi, R.H., Veglia, A.V., 1986. On the product distribution in the iodination of phenol. Tetrahedron Letters 27 (49), 5963–5966. Sanchez Saez, J.J., Herce Garraleta, M.D., Calvo Anton, P., Folgueiras Alonso, M.L., 1991. Identification of 4-bromo-2- chlorophenol as a contaminant responsible for organoleptic taint in melons. Food Additives and Contaminants 8 (5), 627–631. Sevenants, M.R., Sanders, R.A., 1984. Anatomy of an off-flavor investigation: the ‘‘medicinal’’ cake mix. Analytical Chemistry 56 (2), 293A–298A. Steeg, E., Swaczyna, H., Speer, K., 1990. Mass spectrometric investigation of off-aromas in an orange nectar and in a fish marinade. Deutsche Lebensmittel-Rundschau 86 (1), 6–9. Strube, A., Buettner, A., Gro ¨tzinger, C. Characterization and identification of a plastic-like off odor in mineral water. Water Science and Technology: Water Supply, in press. Suffet, I.H., Khiari, D., Bruchet, A., 1999. The drinking water taste and odor wheel for the millennium. Water Science and Technology 40 (6), 1–13. Thomas, R.F., Weisner, M.J., Brass, H.J., 1980. The fifth trihalomethane: dichloroiodomethane, its stability and occurrences in chlorinated drinking water. Water Chlorination: Environmental Impact and Health Effects 3, 161–168. Ullrich, F., Grosch, W., 1987. Identification of the most intense volatile flavour compounds during autoxidation of linoleic acid. Z Lebensm Unters Forsch 184, 277–282. van Wyk, C.J., Rogers, I.M., 2000. A phenolic off-odour in white table wines: causes and methods to diminish its occurrence. South African Journal for Enology and Viticulture 21 (1), 52–57. 5224 w a t e r r e s e a r c h 4 3 ( 2 0 0 9 ) 5 2 1 6 – 5 2 2 4