Food Chemistry 169 (2015) 102–113

Contents lists available at ScienceDirect

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

Can volatile organic compounds be markers of sea salt? Isabel Silva a, Manuel A. Coimbra a, António S. Barros a, Philip J. Marriott b, Sílvia M. Rocha a,⇑ a b

Department of Chemistry, QOPNA, University of Aveiro, Campus Universitário de Santiago, 3810-193 Aveiro, Portugal School of Chemistry, Monash University, Wellington Rd, Clayton, VIC 3800, Australia

a r t i c l e

i n f o

Article history: Received 5 May 2014 Received in revised form 23 July 2014 Accepted 25 July 2014 Available online 4 August 2014 Keywords: Sea salt Food product Volatile markers Saltpan surrounding environment HS-SPME/GC  GC–ToFMS

a b s t r a c t Sea salt is a handmade food product that is obtained by evaporation of seawater in saltpans. During the crystallisation process, organic compounds from surroundings can be incorporated into sea salt crystals. The aim of this study is to search for potential volatile markers of sea salt. Thus, sea salts from seven north-east Atlantic Ocean locations (France, Portugal, Continental Spain, Canary Islands, and Cape Verde) were analysed by headspace solid-phase microextraction combined with comprehensive two-dimensional gas chromatography–time-of-flight mass spectrometry. A total of 165 compounds were detected, ranging from 32 to 71 compounds per salt. The volatile composition revealed the variability and individuality of each salt, and a set of ten compounds were detected in all samples. From these, seven are carotenoid-derived compounds that can be associated with the typical natural surroundings of ocean hypersaline environment. These ten compounds are proposed as potential volatile markers of sea salt. Ó 2014 Published by Elsevier Ltd.

1. Introduction The demand for handmade food products from natural sources has increased considerably in recent years. Emotions evoked by handmade products mainly enhance the pleasure of buying, owning, and using them. More recently, food-elicited emotion is increasingly becoming critical for product differentiation as many food products are produced with similar characteristics, packaging, and price (Jiang, King, & Prinyawiwatkul, 2014). Thus, focus should be taken on the establishment of objective parameters that contribute to product differentiation, contributing to consumer information and helping companies and producers to gain a competitive edge. Sea salt is a handmade food product obtained from seawater in saltpans that are often man-made systems where the salt is produced by crystallisation due to the combined effects of wind and sunlight. Before sea salt crystallisation, seawater circulates along a series of successive ponds with increasing levels of salinity due to continuous water evaporation. Contact with the surrounding environment is a potential source of volatile compounds that may affect the sea salt composition. The presence of volatile compounds in sea salt was demonstrated in 2009 using headspace solid-phase microextraction (HS-SPME) combined with gas chromatography–quadrupole mass spectrometry (GC–qMS) methodology (Silva, Rocha, & Coimbra, 2009). More recently, a high-resolution methodology based on ⇑ Corresponding author. Tel.: +351 234401524; fax: +351 234370084. E-mail address: [email protected] (S.M. Rocha). http://dx.doi.org/10.1016/j.foodchem.2014.07.120 0308-8146/Ó 2014 Published by Elsevier Ltd.

comprehensive two-dimensional gas chromatography was applied (GC  GC–ToFMS), which enabled the tentative identification of a greater number of volatile compounds from sea salt (Silva, Rocha, Coimbra, & Marriott, 2010). In the meantime, other studies reporting the presence of volatile compounds in sea salt have been published (Donadio, Bialecki, Valla, & Dufossé, 2011; Serrano, Nácher-Mestre, Portolés, Amat, & Hernández, 2011; Silva, Rocha, & Coimbra, 2010). Volatiles identified were distributed over the chemical groups of hydrocarbons, aldehydes, esters, furans, haloalkanes, ketones, ethers, alcohols, phenols, terpenoids, norisoprenoids, and lactones. Among these, the norisoprenoid b-ionone (a carotenoid-derived aroma compound), which exhibits a violet odour descriptor, was considered a potential contributor to sea salt aroma (Silva, Rocha, & Coimbra, 2010). Carotenoids are significant potential sources of volatile compounds in nature. Several sources were reported for the volatile compounds identified in sea salt, such as algae (Donadio et al., 2011; Silva, Rocha, & Coimbra, 2009, 2010; Silva, Rocha, Coimbra, & Marriott, 2010), the surrounding bacterial community (Donadio et al., 2011; Silva, Rocha, et al., 2009; Silva, Rocha, & Coimbra, 2010; Silva, Rocha, Coimbra, & Marriott, 2010), and environmental pollution as a consequence of anthropogenic activities (Donadio et al., 2011; Serrano et al., 2011; Silva, Rocha, et al., 2009; Silva, Rocha, & Coimbra, 2010; Silva, Rocha, Coimbra, & Marriott, 2010). The presence of norisoprenoids in fleur de sel (first crystals of sea salt formed at the water surface of saltpans) was related to the concentration of the microalgae Dunaliella salina (Donadio et al., 2011). Different species of plants, animals, algae, bacteria, and other organisms, may be present, according to saltpan location; however,

I. Silva et al. / Food Chemistry 169 (2015) 102–113

a typical biota exists associated with the hypersaline environments. Several studies focused on the flora of hypersaline environments identified in different geographical areas, such as halophytes, namely Salicornia europaea, Halimione portulacoides (Meziane, Bodineau, Retiere, & Thoumelin, 1997), Sarcocornia fruticosa (Válega et al., 2008), Spartina maritime (Silva, Dias, & Caçador, 2009), Limoniastrum monopetalum, and Spartina densiflora (Simões, Calado, Madeira, & Gazarini, 2011), and algae, namely Fucus serratus (Beauchêne, Grua-Priol, Lamer, Demaimay, & Quémeneur, 2000), Polysiphonia denudata, and Laurencia papilosa (Kamenarska et al., 2006). This typical biota may contribute to the volatile composition pattern of sea salt, i.e., the existence of potential sea salt volatile markers. In the last decade, an increased consumption of sea salt has been observed. At the same time, there is a growing interest for protection, and recognising the value of saltpans intrinsically associated with the quality of sea salt. From 2005 to 2007 the project SAL – Salt from Atlantic (Re-valorisation of the Atlantic traditional saltpans identity. Recovery and promotion of the biological, economical, and cultural potential of the humid zones from the coast), supported by the European Commission (INTERREG IIIB Programme), focused on this natural food product (Silva, Rocha, et al., 2009; Silva, Rocha, Coimbra, & Marriott, 2010). Volatile components present in sea salt could play an important role in differentiating this handmade food product from industrial salt, and valuing sea salt as a distinct and desirable product. Therefore, the present work aims to search for potential volatile markers of sea salt. To fulfil this objective, north-east Atlantic Ocean salts from 7 geographical origins were analysed by a methodology previously described (Silva, Rocha, Coimbra, & Marriott, 2010). In addition, the potential impact of the saltpan environment as a source of volatiles is discussed. 2. Materials and methods 2.1. Sea salt samples Sea salt samples from several geographical origins were analysed in this study (Fig. 1). The samples, produced in 2007, were collected (ca. 2–5 kg per sea salt) at different locations of the north-east Atlantic Ocean: Île de Ré (IR), on the west Coast of France; Aveiro (AV) and Figueira da Foz (FF), on the north coast of Portugal; Castro Marim (CM), in the Algarve, south Portugal with Mediterranean influence; Cádiz (CD), in Andalucia, south-western Spain, also with Mediterranean influence; La Palma Island (LP), in the Canary Islands; and Sal Island (S), in Cape Verde. With the exception of sea salt from Cape Verde, obtained from a local store, all the other samples were supplied by the participants of project SAL – Salt from Atlantic, supported by the European Commission (INTERREG IIIB Programme). Four saltpans are located in estuarine areas: (i) Aveiro saltpan (AV) is located in Vouga river margins (Ria de Aveiro), 8 km from the sea; (ii) Figueira da Foz saltpan (FF) is located on Murraceira Island, in Mondego river margins, 3 km from the sea, (iii) Castro Marim saltpan (CM) is located in Guadiana river margins, 5 km from the sea, and (iv) saltpan of Cádiz (CD) is located on the Island of Léon, in Zurraque river margins, 7 km from the sea. Three saltpans are located on north-east Atlantic Ocean islands: (v) Île de Ré saltpan (IR) is located 1 km from the ocean; (vi) La Palma (LP) saltpan is located 2 km from the ocean, and (vii) Sal Island saltpan (S) is an inland saltpan located 2 km from the ocean. These two last saltpans are located in arid zones, i.e., surrounded by much less vegetation than the others. The reported distances were estimated from local maps. For a comparative study two salts from an inland origin, far from the sea (>200 km), were also analysed. These samples were collected in saline aquifers of the Murray Darling Basin, in

103

Fig. 1. Map showing the sea salt sampling sites in the North-east Atlantic Ocean. Île de Ré – IR; Aveiro – AV; Figueira da Foz – FF; Castro Marim – CM; Cádiz – CD; La Palma island – LP; Sal island – S. The Atlantic Ocean currents are also highlighted: North Atlantic Drift, Canary, North Equatorial.

Australia, near Mildura, in the north-west Victoria. These samples, identified as Coarse Gold (CG) and Pink Flakes (PF), were produced in 2008 and supplied by the Australian trading company SunSalt. All samples were stored in sealed glass bottles in the dark and at room temperature until analysis. 2.2. Sea salt volatile determination by HS-SPME/GC  GC–ToFMS The HS-SPME/GC  GC–ToFMS methodology used in this research study was based on previous work developed by Silva, Rocha, et al. (2009), Silva, Rocha, Coimbra, and Marriott (2010). Briefly, 6 g (1/b = 0.5) of sea salt were added to a 22-mL vial. The vial was capped with a PTFE septum and an aluminium cap (Chromacol, Fisher Scientific, Loughborough, UK), and the sample was thermostated overnight using a dry heat block adjusted to 60.0 ± 0.1 °C. The SPME fibre coated with 1 cm divinylbenzene/ Carboxen/polydimethylsiloxane (DVB/CAR/PDMS, 50/30 lm; Supelco, Bellefonte, PA) was then manually inserted into the sample vial headspace for 60 min. In order to avoid any cross-over contamination, blanks, corresponding to analysis of the coated fibre not submitted to any extraction procedure, were run between sets of three analyses. The SPME fibre with sorbed sea salt volatile compounds was manually introduced into the GC  GC injection port at 240 °C and desorbed for 5 min in splitless mode. The desorbed volatile compounds were separated using a GC  GC–ToFMS system comprising of an HP 6890 (Agilent Technologies, Burwood, Australia) gas chromatograph and a Pegasus III time-of-flight mass spectrometer (LECO, St. Joseph, MI,). To implement the modulation process, a longitudinally-modulated cryogenic system (LMCS; Chromatography Concepts, Doncaster, Australia) was used, operated at a modulation period of 5 s with a cryotrap temperature of 20 °C. The ToFMS was operated at a storage rate of 100 Hz,

104

I. Silva et al. / Food Chemistry 169 (2015) 102–113

using a mass range of m/z 41–415 and a multi-channel plate voltage of 1600 V. The column set used for GC  GC experiments comprised a BPX5 (5% phenyl polysilphenylenesiloxane phase) primary column; 30 m  0.25 mm I.D., 0.25 lm film thickness (df), directly coupled to a BP20 (polyethylene glycol phase) second column of 1.0 m  0.1 mm I.D., 0.1 lm df (both columns from SGE International, Ringwood, Australia). The GC oven temperature program was: initial temperature 50 °C (hold 3 min), raised to 130 °C (10 °C min 1), then raised to 230 °C (5 °C min 1) (hold 5 min). Helium was used at a flow rate of 1.0 mL min 1. The transfer line for the GC  GC–ToFMS system was a 0.50 m deactivated fused silica column of 0.1 mm I.D. (0.21 m inside the interface and 0.29 m inside the oven) from SGE International. Data were processed using LECO Corp. ChromaTOF™ software. Contour plots were used to evaluate the general quality of the separation and for manual peak identification. A signal-to-noise threshold of 100 was used. Tentative identification of compounds was achieved by comparing the experimental mass spectra with database libraries (Wiley 275 and National Institute of Science and Technology Mass Spectra Library (NIST 2.0, 2005) – Mainlib and Replib), and supported by experimentally determined retention index (RI) values that were compared, when available, with values reported in the literature for chromatographic columns, equivalent to that used by the 1D column in the present work (Adams, 1995; Babushok & Zenkevich, 2009; Hoet, Stévigny, Hérent, & Quetin-Leclercq, 2006; Leffingwell & Alford, 2005; McGinitie & Harynuk, 2012; Vasta et al., 2012; Zeng et al., 2007). For determination of RI values a C8–C20 n-alkanes series was used. These were calculated according to the Van den Dool and Kratz equation (Van den Dool & Kratz, 1963). The majority (>80%) of the identified compounds presented similarity matches >800. The DTIC (Deconvoluted Total Ion Current) GC  GC area data were used as an approach to estimate the relative content of each volatile component. Three independent aliquots of each sea salt were analysed. Reproducibility was expressed as relative standard deviation (RSD). 2.3. Data analysis Hierarchical cluster analysis (HCA), using Ward’s method (agglomerative hierarchical clustering procedure), was applied with the aim of helping to characterise the data set. HCA is an exploratory tool designed to reveal natural groupings (or clusters) within a data set, by means of a dendrogram (tree diagram), which would otherwise not be apparent. A full data set comprising 165 volatile components (variables) and a sub-set of ten compounds common to all sea salts under study were considered. Each data set consisted of 21 observations, i.e. 3 replicate analyses from 7 sea salts. HCA was applied to auto-scaled GC peak areas. Autoscaling is a data pre-treatment process that makes variables of different scales comparable. Each variable is autoscaled separately by subtracting its mean value and dividing by its standard deviation. A heat-map visualisation of the full data set (7 sea salts and 165 volatile components), normalised by applying a logarithm function, was also performed. HCA and heat-map were performed using the R (version 2.12.0) statistical software package (R: A Language and Environment for Statistical Computing; R Foundation for Statistical Computing; Vienna, Austria, 2009). 3. Results and discussion 3.1. Volatile profile of sea salts The volatile composition of salts from the north-east Atlantic Ocean, collected at different locations, namely Île de Ré, Aveiro, Figueira da Foz, Castro Marim, Cádiz, La Palma and Sal islands,

comprised a total of 165 compounds, ranging from 32 (LP salt) to 71 (S salt) compounds per salt sample. These were distributed over several chemical groups: 64 hydrocarbons, 8 aldehydes, 7 esters, 2 haloalkanes, 19 ketones, 2 ethers, 8 alcohols, 30 terpenic compounds, 19 carotenoid derivatives, 3 lactones, 1 furan, 1 sulphide, and 1 pyridine (Fig. 2 and Table 1). From these, 37 were identified for the first time in sea salt (Table 1). The most reliable way to confirm the identification of each compound is based on authentic standard co-injection, which in several cases is economically prohibitive, and often unachievable in the time available for analysis, or because standards are not commercially available. Hence many compounds are tentatively identified based on library match and RI. The full data matrix (Table 1) includes a list of all 165 compounds, corresponding retention times in both dimensions, retention index (RI) obtained through the modulated chromatogram, and the RI reported in the literature for one dimensional GC with a 5%-phenyl-methylpolysiloxane GC column or equivalent. These chromatographic data are crucial for identification purposes. Furthermore, GC  GC is an ideal technique for the analysis of complex mixtures where compounds of similar chemical structure are grouped into distinct patterns in the 2D chromatographic plane providing useful information on their boiling point and polarity (as a non-polar/polar column set was used), and relationships of structured retentions have proved especially useful for compound identification (Table 1). This unique characteristic of the GC  GC chromatogram is a powerful tool in the identification step (Petronilho, Coimbra, & Rocha, 2014). GC peak area reproducibility, expressed as relative standard deviation (RSD), varied among sea salts volatile components, ranging from 1% to 163%, which may be explained by the fact that sea salt is a natural heterogeneous product (Silva, Rocha, et al., 2009; Silva, Rocha, & Coimbra, 2010; Silva, Rocha, Coimbra, & Marriott, 2010). In order to facilitate the analysis of the full data set concerning the volatile profile of the sea salts under study, a heat-map representation and a hierarchical cluster analysis (HCA) were performed. The heat-map (Fig. 2) shows a graphical representation of the in-depth data present in Table 1 and allows a rapid visual evaluation of the similarities and differences between samples, whereas the dendrogram (Fig. 3A) built from the HCA, is an exploratory tool designed to reveal natural groupings, and groups sea salts according to their similarities. Fig. 2 heat-map shows that S and IR samples exhibited the greater number of compounds, while a fewer number of volatile compounds was detected for LP. This was mainly related to the major presence of hydrocarbons in IR and S, and also with the presence of the sesquiterpenoids (included in the group of terpenic compounds in Table 1), only identified in S. The HCA shows that the carotenoid derivatives profile was very similar among all the samples under study. The occurrence of ketones, also perceptible in the heat-map, was greater for AV and CD. Among the major chemical groups identified, FF possessed fewer terpenic compounds, and S fewer carotenoid derivatives. For IR and S, the hydrocarbons was the chemical group showing the largest GC peak area, whereas for CM it was the aldehydes, and for the other salts the carotenoid derivatives (Fig. 2 and Table 1). Fig. 3A presents the dendrogram built from the hierarchical cluster analysis of the GC peak area of all 165 volatile compounds. The vertical axis of the dendrograms measures the similarity between samples: lower height corresponds to higher similarity between the samples. Fig. 3A shows that the replicates of each sea salt exhibit less variability, i.e., the intra-variability among replicates was lower than inter-variability among the 7 samples. Although the volatile composition of each sea salt seems to be unique, which may be related to the particular saltpan

I. Silva et al. / Food Chemistry 169 (2015) 102–113

105

Fig. 2. Heat-map representation of GC  GC peak areas from sea salt volatile components. Île de Ré – IR; Aveiro – AV; Figueira da Foz – FF; Castro Marim – CM; Cádiz – CD; La Palma island – LP; Sal island – S. Areas are normalised by applying a logarithm function.

environment where it crystallises and is then harvested, similarities among all sea salts can be observed. Three principal clusters may be observed (Fig. 3A): S salt, IR salt, and a cluster with the remaining salts (CM, LP, FF, CD, and AV). The oceans are dynamic systems, with currents a paramount factor to be considered, and they may promote a genetic connectivity according to the marine organisms among regions provided by the same current and/or marine organisms’ migration (Sequeira, Mellin, Floch, Williams, & Bradshaw, 2014). The spatial dispersion of the marine populations is modulated by their migration promoted by the oceanic currents, and can be assessed by the genetic data related with the connectivity established among the populations. The Atlantic Ocean currents are shown in Fig. 1, and an analogy can be suggested between the three clusters previously defined (Fig. 3A) and the North Equatorial current, the Canary current, and North Atlantic Drift current (Fig. 1). S, situated in a more southerly location, seems to be influenced by the North Equatorial current, coming from east to north-west; IR, located in a more northerly geographical area, seems to be influenced by the North Atlantic drift coming from west to north-east, and the

remaining saltpans, with a central position nearest to the Strait of Gibraltar, appear to be influenced by a descending current starting as a North Atlantic Drift and finishing as the Canary current. These different currents seem to influence the volatile composition of sea salt. 3.2. Potential volatile markers of sea salt From the 165 volatile components of sea salt, a sub-set of ten compounds was identified in all samples under study (Fig. 4): seven carotenoid derivatives, i.e. 6-methyl-5-hepten-2-one (6MHO) (1), 2,2,6-trimethylcyclohexanone (TCH) (2), 3,5,5-trimethyl-2cyclohexenone (isophorone) (3), 2,6,6-trimethyl-2-cyclohexene-1, 4-dione (ketoisophorone) (4), 4-(2,2,6-trimethyl-7-oxabicyclo [4.1.0]hept-1-yl)-3-buten-2-one (b-ionone 5,6-epoxide) (5), dihydroactinidiolide (DHA) (6), and 6,10,14-trimethyl-2-pentadecanone (TPD) (7), two esters, i.e. 3-hydroxy-2,4,4-trimethylpentyl 2-methylpropanoate (3HPP) (8), and 2,4,4-trimethylpentane-1, 3-diyl bis(2-methylpropanoate) (TMPP) (9), and the alcohol 2-ethyl-1-hexanol (2EH) (10). Thus, although the volatile

tR a

1

2

tRa

Compound

RIlitb

RIcalcc

Previously reportedd

Peak areae (10 IR

5

106

Table 1 Volatile composition of sea salts from several origins (IR, AV, FF, CM, CD, LP, and S). For comparison, common volatiles found in inland salts (CG and PF) are also shown. ) and RSD (%) AV

FF

CM

CD

LP

S

– – – – – – – – –

– – – – – – – – –

– – – – – – – – –

– – – – – – – – –

– – – – – – – – –

– 2.21 – – – – – – 0.65

– – –

–

– – – – – – – – – – – – – – – – – – – – – – –

– – – – – – –

–

CG

PF

– – – – – – – – –

– – – – – – – – –

– – – – – – – – – – – – – – – – – – – – – – – –

– – – – – – – – – – – – – – – – – – – – – – – –

– – – – – – – – – – – – – – – – – – – – – –

– – – – – – – – – – – – – – – – – – – – – –

Hydrocarbons

1.73 1.59 1.50 1.45 1.62 1.76

Aromatics (1-Butylheptyl)-benzene (1-Propyloctyl)-benzene (1-Pentylheptyl)-benzene (1-Butyloctyl)-benzene (1-Propylnonyl)-benzene (1-Ethyldecyl)-benzene

0.95 1.14 1.14 1.17 1.22 1.03 1.05 1.05 1.16

1060 1085 1150 1160 1170 1190 1205 1215 1215 1235 1240 1265 1265 1280 1290 1310 1320 1330 1340 1350 1375 1390 1405 1410

1.06 1.08 1.10 1.09 1.10 1.11 1.10 1.14 1.93 1.13 1.30 1.13 1.34 1.16 1.15 1.12 1.32 1.14 1.37 1.13 1.16 1.16 1.14 1.38

1420 1425 1440 1450 1460 1470 1475 1500 1525 1535 1550 1555 1575 1585 1630 1680 1695 1705 1715 1725 1735 1780 1405 1415 1525 1530 1545 1580

1100 1200 – – – – – – –

1101 1201 1268 1280 1293 1312 1324 1335 1366

B A,B – – – – – – B

– 1400 – – – – – – – 1500 – – – – – – – – – 1600 – – – 1629

1374 1397 1441 1449 1456 1471 1482 1490 1490 1505 1509 1528 1528 1539 1547 1562 1570 1578 1586 1593 1613 1625 1637 1641

– A,B – – – – – – – A – – – – – – – – – A – – – B

– – – – – – 1700 – – – – – – – 1800 – – – – – – –

1649 1653 1665 1673 1681 1689 1693 1710 1728 1731 1744 1748 1758 1768 1798 1840 1853 1862 1871 1879 1888 1929

– – – – B – A,B – – – – – – – – – – – – – – –

1631 1643 1711 1716 1725 1790

1637 1645 1728 1731 1741 1764

B B – B B –

1.6

(67)

– 26.4 7.7 3.4 16.7 6.1 5.1

(97) (15) (13) (13) (22) (13)

– 1.5

(64)

3.2 4.1 3.3

(16) (8) (15)

25.7 8.1

(11) (16)

6.8 11.0 50.9 21.2 20.7 13.7 6.4 2.3 1.5 2.0

(12) (17) (14) (15) (4) (26) (10) (10) (20) (13)

–

–

–

– – – – 1.2

(33)

1.4 1.5 2.1 3.6

(8) (16) (12) (24)

– 19.7

(19)

33.6 13.5 16.3 15.4 7.9 4.9 1.3

(20) (26) (22) (11) (40) (21) (16)

2.1

(41)

4.1

(53)

– –

– – – – –

– – – – – –

1.46 1.20

(12)

– – – – – – – – – – – – – – – – – – – –

1.09

(23)

–

– – – – 17.23 – – – – 1.17 – – – – – – – – – – – – 0.47 0.19 0.97 – – –

(9)

– – – – – – – – – – – – – – – – – – – –

(14)

(44)

(46) (49) (67)

0.88

(21)

– – – – – – – – – – – – – – – – – 0.73 0.35

(17) (30)

0.75 0.16

(53) (15)

(18) (27)

(11)

(26)

– – – – – 4.65

(26)

4.57 1.31

(25) (31)

1.95

(18)

–

– – – – – – – – – – – – – 0.48 0.20 0.96

(13) (15) (11)

– – 0.17

(15)

– – – – – –

(13)

3.24

(7)

– – – – – – – – – – – 1.49

2.48

– – – – – – – – – – – – – – – – – – – – – –

–

–

2.26 2.10 – – – – – – – – – – – – – 1.99

– – – –

1.66 – – – – –

(7)

– – – – – – – – – –

(28)

(11) (5) (5) (4) (7)

(43)

(14) (13) (25) (6) (11) (13)

(16) (16) (18)

– 10.25

(18)

4.79

(19)

6.35 14.57 8.83 4.87 3.88

(22) (8) (13) (18) (11)

2.22 9.32 1.84 4.91 2.83 6.69 4.07 6.74 6.28

(19) (17) (20) (17) (81) (13) (29) (22) (40)

–

3.56

(16)

– – – – – – – – – – – – – – –

– – –

–

0.38

(5)

0.59

(11)

– – – –

– – 1.58 3.18 2.13 7.74 2.15 – – 2.07 – – 7.33 8.60 3.70 2.34 1.78 2.12 – – 8.14 3.66 – 12.14

(14)

– – – – – –

– – – – – – –

I. Silva et al. / Food Chemistry 169 (2015) 102–113

1.20 1.38 1.18 1.17 1.45 1.37 1.19 1.21 1.18 1.42 1.42 1.21 1.38 1.22 1.46 1.22 1.21 1.22 1.20 1.23 1.43 1.44

Aliphatics Undecane Dodecane C12 isomer C12 isomer C12 isomer C13 isomer C13 isomer C13 isomer 2,6,10-Trimethyldodecane C13 isomer Tetradecane C14 isomer C14 isomer C14 isomer C14 isomer C14 isomer C14 isomer C14 isomer Pentadecane C15 isomer C15 isomer C15 isomer C15 isomer C15 isomer C15 isomer C15 isomer C15 isomer C15 isomer Hexadecane C16 isomer C16 isomer C16 isomer 2,6,10-Trimethylpentadecane C16 isomer C16 isomer C16 isomer C16 isomer 1-Heptadecene C16 isomer Heptadecane C17 isomer C17 isomer C17 isomer C17 isomer C17 isomer C17 isomer C17 isomer Octadecane C18 isomer C18 isomer C18 isomer C18 isomer C18 isomer C18 isomer C19 isomer

725 845 925 940 955 980 995 1010 1050

– – – 0.71 – –

(66)

1.77 1.77 1.81

1645 1655 1675

2.42

1.39 2.74

2.50

1.89

1.87

3.12 3.05

1.27 1.25

615

740 1045

1075

1340

1875

1705 1830

700 810

1.48 2.21 1.69 2.33 1.43 2.27 2.12 1.47 1.87

1.50 2.32

1.73 1.02 3.01

3.26

2.49

1.82

2.08

670 690 710 715 730 760 810 830 830

845 890

965 980 1010

1095

1130

1195

1215

Ketones

1.28 1.34 1.44 1.55 1.58 1.66 2.08

525 640 745 860 985 1115 1665

Aldehydes

2

tR a

1

tRa

Aliphatics 1-(2,2-Dimethylcyclopentyl)-ethanone 2,3-Dimethyl-2-cyclopenten-1-one 3,6,6-Trimethylcyclohex-2-enone 3-Methyl-2-cyclohexen-1-one 2-Nonanone 3,5-Dimethyl-2-cyclohexen-1-one 6,6-Dimethylbicyclo[3.1.1]heptan-2-one 3,3,4,4-Tetramethyl-2-pentanone 6,6-Dimethyl-2-methylenebicyclo[2.2.1]heptan-3-one 2-Decanone 1,3,3-Trimethyl-2oxabicyclo[2.2.2]octan-6-one 2-Undecanone 1-Aza-2-cycloheptanone 2-(2-Methylpropylidene)cycloheptanone 3-Acetyl-2,4,4-trimethylcyclohex-2-en1-one 6-Methyl-6-(5-methylfuran-2-yl)heotan2-one 2,6-bis(1,1-Dimethylethyl)-2,5cyclohexadiene-1,4-dione 3-Methyl-4-(2,6,6-trimethyl-2cyclohexen-1-yl)-3-buten-2-one

Subtotal (GC peak area) Subtotal (%)

Esters Aliphatics Isopentyl 3-methylbutanoate 1-Hydroxy-2,4,4-trimethylpentan-3-yl 2methylpropanoate 3-Hydroxy-2,4,4-trimethylpentyl 2methylpropanoateg 2,4,4-Trimethylpentane-1,3-diyl bis(2methylpropanoate) Isopropyl palmitate Aromatics Dibutyl phthalate Bis(2-methoxyethyl) phthalate Subtotal (GC peak area) Subtotal (%) Haloalkanes 1-Chlorooctane 1-Chlorononane

Subtotal (GC peak area) Subtotal (%)

Aliphatics Heptanal Octanal Nonanal Decanal Undecanal Dodecanal 5,9,13-Trimethyl-4,8,12-tetradecatrienal Aromatics Benzaldehyde

Subtotal (GC peak area) Subtotal (%)

(1-Hexylheptyl)-benzene (1-Butylnonyl)-benzene (1-Propyldecyl)-benzene

Compound

Table 1 (continued)

1490

1468

1479

1422

1294 1266 –

1196 –

– – – – 1083 – 1142 – 1114

1044 1159

1897 1990

2011

1605

1381

1101 –

986

916 1012 1127 1216 1316 1411 –

– 1812 1843

RIlitb

1490

1479

1428

1402

1301 1312 1337

1198 1239

1042 1061 1075 1084 1097 1125 1166 1188 1188

1065 1170

1864 1976

2011

1586

1387

1106 1364

993

906 1015 1111 1214 1317 1416 1828

1810 1819 1836

RIcalcc

B

–

B

–

B – B

B B

B B B B B – B B –

B B

B –

B

B

B

– B

B

B B B B B B B

– – –

Previously reportedd

–

–

–

– – –

– –

– – – –

–

– –

–

– –

–

–

–

– – – –

– –

(15)

(31)

0.60

1.15

(11)

(31) (48)

(31)

(47) (33)

(63)

(42)

(144)

(56) (55)

(56)

(20) (5)

5.65

0.70 0.15

0.70

5.07 1.01

1.04

3.89

0.14

0.67 0.14

0.67

378.3 77.13

– – –

IR

Peak areae (10

–

–

–

– –

–

–

– –

–

–

–

– – –

– –

–

1.31

1.02

0.66 0.62

1.26

1.99 1.91 8.41 1.82 1.12 0.24 0.42

0.43 0.09

0.43

15.11 3.18

2.42

11.13

1.56

2.39 0.50

1.21 1.18

22.35 4.66

0.86 0.26

AV

(10)

(26)

(33) (27)

(118)

(25) (5) (16) (4) (19) (3) (19)

(40) (38)

(40)

(13) (20)

(35)

(15)

(11)

(26) (26)

(26) (31)

(10) (6)

(20) (89)

) and RSD (%)

5

FF

0.74 0.41

1.11 4.30 6.64

5.42 3.78

5.83

–

–

–

– –

– –

– – – – – –

– –

–

0.49

0.24

1.72

0.29 0.19

0.29

28.87 1.79 55.68 35.88

–

7.72

11.48

–

13.20 8.34

–

–

–

– – –

(17)

(20)

(10)

(49) (33)

(49)

(143) (107) (84) (61)

(95)

(29)

(10)

(81) (71)

(39) (12)

(105) (94) (79)

(21) (15)

–

–

–

–

– – –

– –

– – – – – –

– –

– –

–

0.98 0.42

0.46 0.52

12.21 5.27

0.52

1.29

9.55

0.86

103.63 44.81

2.81

3.09 13.69f 36.50 41.13 3.55 2.86 –

4.17 1.80

0.24 0.13

CM –

(36) (35)

(20) (51)

(29) (27)

(30)

(68)

(23)

(38)

(7) (5)

(15)

(44) (5) (1) (17) (29) (30)

(1) (0)

(5) (16)

–

– –

–

– –

–

– –

– –

– –

–

–

– –

–

– – – – – – –

0.39

1.74

0.30

0.90 26.70 0.94

0.47

1.02 3.57 1.72 0.79

4.26 0.82

1.76

1.19

1.30

19.32 3.75

CD – – –

(26)

(7)

(29)

(21) (82) (2)

(13)

(9) (5) (13) (10)

(39) (36)

(27)

(54)

(44)

(22) (19)

0.57 0.81

0.57

–

–

–

– – –

– –

0.23

– – 1.26 – 0.19 – – – –

–

– – 4.03 5.64

–

0.71

3.32

– –

5.85 8.19

2.22

– – 2.23 1.40 – – –

11.06 15.51

– –

0.15

LP

(29)

(7)

(6)

(18) (19)

(18)

(19) (18)

(40)

(15)

(20) (19)

(16)

(18) (33)

(10) (10)

(31)

S

0.17

1.81 5.20 5.53

–

–

–

– – –

– –

– – – – – –

– –

– –

– –

– –

0.38

1.25

0.37

2.31 0.92

0.40

1.81

0.10f –

12.71 5.31

–

–

–

–

169.94 72.10

– – –

(33)

(74)

(22)

(116) (110)

(134)

(112)

(123)

(33) (24)

(99)

(19) (27) (42)

(8) (8)

–

–

–

–

– – –

–

– – – –

– – –

– –

–

– – – –

–

–

–

– –

– –

–

–

–

–

– –

3.75

0.53 5.00

3.25

9.58

34.10 65.06 50.14

CG – – –

–

1.75

–

–

– – –

– –

– – – – – – – – –

– –

– –

– – – –

–

4.47

–

– –

– –

–

– 1.45 3.41 2.26 0.51 –

– –

(16)

(130)

(87)

(102) (7) (28)

(continued on next page)

(29)

(26) (33)

(27)

(19)

(32) (14) (27)

PF – – –

I. Silva et al. / Food Chemistry 169 (2015) 102–113 107

tR a

2.59

1.49 1.48

725

Ethers 565 1435

Subtotal (GC peak area) Subtotal (%)

Aliphatics 2-Ethyl-1-hexanol 1-Undecyn-4-ol 2,6-Dimethylcyclohexanol 1-Undecanol 2-Dodecanol 1-Tridecanol 3,7,11-Trimethyl-1-dodecanol 1-Tetradecanol

1-Methoxy-4-methylbenzene Dioctyl ether Subtotal (GC peak area) Subtotal (%)

Carotenoid derivatives 580 1.43 6-Methyl-2-heptanone 600 1.64 2-Methyl-1-hepten-6-one 620 1.49 6-Methyl-5-hepten-2-one 645 1.61 2,4,4-Trimethylcyclopentanone 680 1.51 2,2,6-Trimethylcyclohexanone 730 1.80 3,4,4-trimethyl-2-cyclohexen-1-one 740 2.01 3,4,4-Trimethyl-2-cyclopenten-1-one 785 2.24 Isophorone (isomer) 810 2.46 Ketoisophorone 840 3.02 2,2,6-Trimethyl-1,4-cyclohexanedione

Subtotal (GC peak area) Subtotal (%)

953 966 1003 – 1036 1097 – 1121 1169 1190

994 1038 1032 1072 1087 1087 1098 1152 1143 1154 1237 1171 – 1178 1206 1204 1248 1242 1365 1506 1525 1524 1546 1609 – – 1651 1661 1674 –

1029 – 1112 1370 1387 1583 – 1673

– –

1065

Aromatic Acetophenone

Subtotal (GC peak area) Subtotal (%)

RIlitb

Compound

Terpenic compounds 625 1.21 2,3-Dehydro-1,8-cineole 665 1.27 Cymene 675 1.23 1,8-Cineole 710 1.83 Dihydromyrcenol 715 1.76 Linalool oxide (isomer) 730 1.88 Linalool oxide (isomer) 740 1.89 Linalool 810 1.97 Dihydro terpineol 815 1.69 Camphor 820 1.48 Menthone (isomer) 830 3.10 Pulegone (isomer) 835 2.27 Umbellulone 845 2.01 Menthan-1-ol 845 2.22 Isomenthol a-Terpineol 865 2.23 890 2.49 Verbenone 905 2.50 Eucarvone 930 2.45 Carvone 1040 1.88 Neryl acetate 1245 1.74 Muurolene 1270 1.78 Cadinene 1280 1.76 Calamenene 1315 2.12 Calacorene 1405 2.27 Ledol 1410 2.34 C15 (m/z 41, 43, 55, 119, 105) 1430 2.33 C15 (m/z 41, 119, 43, 105, 55) 1445 2.62 Cadinol 1450 2.67 Muurolol 1490 2.53 Cadalene 1605 1.18 2,6,10,14-Tetramethylhexadecane

665 735 765 1080 1110 1340 1415 1465

Alcohols

1.97 2.35 2.22 1.36 2.09 1.41 1.29 2.37

2

1

tRa

Table 1 (continued)

958 983 997 1020 1051 1097 1106 1147 1171 1197

1001 1042 1042 1079 1083 1097 1106 1170 1174 1179 1189 1193 1202 1197 1219 1235 1252 1281 1359 1513 1532 1540 1567 1637 1642 1658 1670 1674 1708 1781

1038 1102 1129 1386 1413 1586 1642 1686

940 1657

1093

RIcalcc

A, B – A,B B A,B,C B – B,C B B,C

– A,B A,B B – – B – B – – – – B – B B B – – – – – – – – – – – B

A,B B – B B B B B

B B

B

Previously reportedd

(12) (8) (21) (4) (4) (12) (9) (150)

– 15.83 3.35 3.35 0.27 2.44 0.09

(14) (30)

1.67 5.73

–

37.53 7.94

(14)

– – – – – – – – – – – – – – – 36.36

(28) (42)

(13)

(88) (76)

(88)

(83) (83) (82)

(153)

0.34 0.32

0.33

3.09 0.61

3.09

0.51 0.51 0.11

(7) (21)

0.18

– – –

– – – –

– – –

– – – – – – –

–

– 7.41 1.55

Peak areae (10 IR

5

–

–

– – – – – – – – – – – –

– – –

–

– – –

–

– –

– –

– – – –

–

5.11 50.70 9.83

7.72 8.57 12.10 10.01

15.93

28.71 5.77

12.27

0.79 1.01 0.46

0.29

0.71

0.81

12.37

8.93 1.88

0.33

8.60

1.11 0.54 1.66 0.36

20.77 4.41

AV

(12) (1) (9)

(7) (9) (9) (9)

(7)

(60) (49)

(14)

(9) (14) (14)

(19)

(22)

(27)

(128)

(9) (19)

(19)

(10)

(51) (25) (43) (53)

(7) (23)

) and RSD (%)

2.45 1.80

–

–

–

– – – – – – – – – – – – – – – – – – – – – – – – – –

– –

– – – – – – –

1.13 11.05 6.00

6.79 1.47

3.95

17.71

6.55 4.66

6.34

0.21

1.84 1.26

1.84

– 0.43f 0.43 0.26

–

FF

(21) (27) (33)

(19) (29)

(5)

(111)

(15) (24)

(12)

(101)

(47) (31)

(47)

(117) (117) (121)

(8) (39)

0.47

0.39

9.13 0.18 1.03 0.42

1.39

0.36

0.45

39.38 – 4.11 – 0.24 – – 0.22 2.89 1.25

35.28 15.26

2.40 – – – – – – – – – – – 29.77

– –

–

– – – – – – – –

–

0.54 0.37

11.63 5.05

–

–

– – – –

– –

–

CM

(4) (6) (18)

(6)

(15)

(15)

(14) (14)

(19)

(13)

(12)

(41)

(46)

(22) (8)

(29) (30)

(22)

(10)

(37) (3) (3) (5)

– –

–

– – – – – – – – – –

– – –

– –

– –

–

–

– – –

–

–

–

– – – –

1.17 20.04 6.02

0.87 4.02 5.42

36.75

12.55 2.45

5.37

0.45

0.25

0.41

0.59

1.26 0.42

3.80

67.75 13.23

1.44 1.45

0.73

62.08

2.04

39.89 7.64

1.36

CD

(8) (10) (13)

(37) (21) (35)

(21)

(13) (13)

(15)

(12)

(52)

(14)

(10)

(2) (58)

(22)

(6) (8)

(23) (29)

(15)

(7)

(7)

(55) (49)

(26)

4.13 5.79

3.25

0.33 0.33

0.23

2.16 3.03

2.16

1.68 2.35

27.39 – 5.57 – 1.46 0.36 – 0.33 1.74 –

– – – – – – – – – – – – – – – – – – –

– – – –

– – –

– – – – – – –

– – – –

–

LP

(1) (2)

(4) (13)

(10)

(1)

(10) (11)

(13)

(8) (17)

(9)

(12) (11)

(12)

(1) (1)

–

– –

–

– –

–

– – –

– – – –

–

– – – –

– – – –

– –

– – – –

–

S

0.54 0.11

0.23

4.66

32.14 13.52

1.43 7.39 5.08 6.44 1.73 1.40 1.98 0.61 1.17 1.46

0.73 0.70

0.09 0.31 0.36

0.68 0.57

9.61 3.96

0.36

9.24

2.00 0.82

(29) (79)

(12)

(36)

(19) (10)

(30) (20) (10) (5) (60) (58) (85) (67) (81) (38)

(16) (17)

(5) (3) (3)

(31) (54)

(66) (58)

(23)

(68)

(49) (40)

3.79

–

–

– – –

– –

– –

– – – – – –

–

– – – –

–

– –

–

– – – –

–

– –

0.48

0.84

4.97

1.70 22.40

1.25

1.02

1.29

1.89

10.00 14.72

1.08

1.72

1206.55 0.70f – – – – – –

– – – –

– –

CG

(20)

(16)

(64)

(12) (14)

(17)

(43)

(16)

(36)

(19) (27)

(26)

(64)

(17) (112)

(27)

– – – – – – – – – –

– –

– – – – – – – – – – – – – – – – – – – – – – – – – – – – –

– –

2.55

94.52 – – – – – – –

– – – –

– –

–

PF

(30)

(31)

108 I. Silva et al. / Food Chemistry 169 (2015) 102–113

Table 1 (continued) tR a

1

2

tRa

895 1100 1145 1160 1220 1230 1325 1330 1665

2.05 1.61 2.31 2.21 2.46 2.64 4.34 4.11 1.76

1010

2.25

1215

4.23

1285

3.73

Compound

b-Cyclocitral 6,10-Dimethyl-2-undecanone a-Ionone Dihydro-b-ionone b-Ionone b-Ionone 5,6-epoxide Dihydroactinidiolide (isomer) Dihydroactinidiolide (isomer) 6,10,14-Trimethyl-2-pentadecanone Subtotal (GC peak area) Subtotal (%) Lactones 2,2,4-Trimethyl-5-(2,2-dimethylpropyl)3(2H)-furanone 7a-Methyl-3methylenehexahydrobenzofuran-2-one Tetrahydroactinidiolide

RIlitb

RIcalcc

Previously reportedd

Peak areae (10 IR

1223 1410 1426 1424 1485 1463 1539 1539 1843

1243 1405 1439 1450 1494 1498 1580 1588 1832

A,B,C A,B A,B,C B A,B B A,B,C B A,B

–

1336

–

– –

–

1492

B

–

1545

–

1.17 1.79 1.66

2-Pentylfuran Dimethyl trisulphide 2,4,6-Trimethylpyridine Subtotal (GC peak area) Subtotal (%) Total Number of identified compounds

996 972 1014

997 997 1006

B B –

(6)

3.08

(12)

– 1.09 6.77 3.41

(33) (26) (35)

0.86 51.16 10.60

(42) (8) (12)

–

3.90 – 21.69 4.89 3.00 24.17 198.96 – 2.23 378.83 78.57

FF (10)

3.00 0.63

(23) (31)

0.63

(7)

–

2.37

(5)

– – –

(31) (15) (4)

8.84

– 0.62 5.07 1.92 0.56f 1.88 61.70 43.72

(14)

–

(18)

(9)

3.46

(9)

– –

0.26 1.02

(18) (52)

3.75 61.97 26.83

(55) (10) (11)

–

0.56

(7)

0.48

(9)

1.66 1.81

(18) (12)

– 0.50 41.85 58.68

(38) (2) (3)

–

2.43 76.04 187.78 – 5.22 351.67 68.66

(7) (7) (5) (36) (1) (6)

–

– –

S

CG

PF

– – – – –

– – – – – – – – – – – –

0.52 0.02

(102) (74)

0.32 6.40 2.75

(45) (12) (22)

– – – – – – – – – – –

0.72

(45)

–

–

–

–

–

2.18

(20)

–

–

–

–

–

–

5.95

(81)

–

–

–

–

– –

8.13 1.62

(57) (59)

– –

(163) (30) (83)

– – – –

– – 0.71

(42)

(78) (5)

– 71.34

0.71 236.54

(42) (10)

(14) (2)

– –

0.33

(48)

(108)

1.29

(14)

5.45 4.16 9.62

(112) (31)

0.56 231.15

(12) (2)

1.82 513.19

0.63

(7)

0.33

(48)

0.13 488.05

(12) (16)

0.07 481.96

(44) (15)

0.12 147.75

55

LP

2.46

2.44 0.51

– –

69

CD

–

0.20f – – 0.20

– –

(6) (23) (49) (90) (16) (43) (40)

– 2.44

(23)

(10)

– (7) (10) (12) (19) (27)

–

3.00

CM 1.19

40

(108)

1.29

(14)

– –

44

–

52

32

(1)

0.72 0.30

(45) (37)

– –

– –

0.71

(42)

– – – –

– – –

–

–

0.58

(59)

71

Sea salts code: Île de Ré – IR; Aveiro – AV; Figueira da Foz – FF; Castro Marim – CM; Cádiz – CD; La Palma island – LP; Sal island – S. Inland Salts Code: Coarse Gold – CG, Pink Flakes – PF. For inland salts was only included data about compounds common with sea salts. a Retention times in seconds (s) for first (1tR) and second (2tR) dimensions. b RI: retention index reported in the literature for 5% phenyl-dimethyl polysilphenylene-siloxane GC column or equivalents (Adams, 1995; Hoet et al., 2006; Leffingwell & Alford, 2005; McGinitie & Harynuk, 2012; Vasta et al., 2012; Zeng et al., 2007). c RI: retention index obtained through the modulated chromatogram. d A – Silva, Dias, and Caçador (2009), Silva, Rocha, and Coimbra (2009); B – Silva, Rocha, and Coimbra (2010), Silva, Rocha, Coimbra, and Marriott (2010); C – Donadio et al. (2011); D – Serrano et al. (2011). e Mean of three replicates. f The compound was detected in two replicates. g Compounds highlighted in bold are common to all sea salts.

I. Silva et al. / Food Chemistry 169 (2015) 102–113

Others 620 620 630

) and RSD (%) AV

3.23 –

Subtotal (GC peak area) Subtotal (%)

5

109

110

I. Silva et al. / Food Chemistry 169 (2015) 102–113

composition of sea salts is related with the surrounding environment where these are collected, there is like a base composition characteristic of all sea salts. To explore possible natural groupings (or clusters) within sea salts using this sub-data set, HCA was performed (Fig. 3B). This dendrogram showed a higher similarity between samples compared to the dendrogram comprising all the volatile compounds (Fig. 3A). It was also verified that sea salt replicates of FF, S, and CM were not closely grouped. Based on the HCA of the sub-set of ten common compounds, it was not possible to group salts by saltpan, suggesting that they characterise the sea salt itself rather than their origin/environment. Thus, this sub-set of compounds was defined as sea salt potential volatile markers. As they are always present, no discrimination of origin could be made based on them. In that sense, the discrimination of origin should be made using the global sea salt volatile composition. To confirm the hypothesis that the 10 volatile compounds identified are always present in sea salt, their presence was checked in several sea salts and non-sea salts (inland salts) analysed by HS-SPME/GC  GC–ToFMS. Concerning the sea salts, this sub-set was detected in five sea salts from a different year (2010), namely from Aveiro, Figueira da Foz, Castro Marim, Île de Ré, and Guérande (also situated off the Western Coast of France) (data not shown). The occurrence of this sub-set was also checked in sea salts previously analysed (Silva, Rocha, Coimbra, & Marriott, 2010), comprising two saltpans of Aveiro, harvested in 2007, and the ten compounds were detected in both sea salts. Moreover, this study also comprised the analysis of sea salts stored for two and

three years, and only 3HPP, 2EH, 6MHO, ketoisophorone, DHA, and TPD were detected. These results can be explained by the longer storage time, which contributes for the loss of volatile compounds, including these potential markers (Silva, Rocha, Coimbra, & Marriott, 2010). To verify if potential markers of sea salt might also occur in inland salt, i.e., salt produced in saltpans away from the coast (where groundwater sourced salt that is released from natural brine aquifers), two inland salts from Australia (GC and PF) were analysed by the same HS-SPME/GC  GC–ToFMS methodology. Among the sub-set of ten compounds, only 4 potential volatile markers were identified, and only 2-ethyl-1-hexanol was detected in both inland salts. 6-Methyl-5-hepten-2-one and ketoisophorone were detected only in CG sample, and 2,4,4-trimethyl-1-(2methylpropanoyloxy)pentan-3-yl] 2-methylpropanoate in PF (Table 1). These results corroborate the proposed hypothesis about the existence of potential markers of sea salt. In addition, it may be pointed out that only 28 compounds were detected in common with the sea salts under study, 24 for CG, and 10 for PF (Table 1), although other volatiles have been detected in the inland salts. Although unlikely, an ancient sea source for the inland salt might still contains some of its volatile compounds possibly preserved in its crystalline structure. An in-depth understanding of the origin of volatiles from sea salt is crucial to interpret the surrounding environment of saltpans, taking into consideration the microbial communities, plants, and algae. Although useful, this information is very scarce and may be unavailable in the literature for the locations under study.

Fig. 3. Dendrograms from GC peak areas of sea salt volatile compounds. (A) All volatile compounds (165) detected in north-east Atlantic Ocean sea salts under study. (B) Subset of ten potential volatile markers common to all north-east Atlantic Ocean salts (Île de Ré – IR; Aveiro – AV; Figueira da Foz – FF; Castro Marim – CM; Cádiz – CD; La Palma island – LP; Sal island – S).

I. Silva et al. / Food Chemistry 169 (2015) 102–113

Fig. 4. Chemical structures of the ten tentatively identified compounds common to all north-east Atlantic Ocean salts. (1) 6-methyl-5-hepten-2-one; (2) 2,2, 6-trimethylcyclohexanone; (3) isophorone; (4) ketoisophorone; (5) b-ionone 5, 6-epoxide; (6) dihydroactinidiolide; (7) 6,10,14-trimethyl-2-pentadecanone; (8) 3-hydroxy-2,4,4-trimethylpentyl 2-methylpropanoate; (9) 2,4,4-trimethylpentane-1, 3-diyl bis(2-methylpropanoate); (10) 2-ethyl-1-hexanol.

However, as far as possible, a general overview about these data was undertaken (Fig. 5), with particular emphasis on the 10 potential volatile markers. Considering the potential sources of volatile compounds previously reported for sea salt (Donadio

111

et al., 2011; Serrano et al., 2011; Silva, Rocha, et al., 2009; Silva, Rocha, & Coimbra, 2010; Silva, Rocha, Coimbra, & Marriott, 2010), namely, algae, halophyte, bacteria, and anthropogenic activity, the potential contribution of these sources to the several chemical groups detected (Fig. 5), using references that may support this information was reported (Beauchêne et al., 2000; El Hattab, Al Easa, Tabaries, Piovetti, & Kornprobst, 2007; Elenkov, Georgieva, Hadjieva, Dimitrovakonaklieva, & Popov, 1995; Erakin & Güven, 2008; Evans, 1994; Gao, Zhao, Kong, Chen, & Hu, 2004; Gressler, Colepicolo, & Pinto, 2009; Grossi & Raphel, 2003; Kamenarska et al., 2006; Krock & Wilkins, 1996; Nalli, Horn, Grochowalski, Cooper, & Nicell, 2006; Shibata, Hama, Miyasaki, Ito, & Nakamura, 2006; Sun, Chung, & Shin, 2012). Algae proved to be a potential source for the majority of the chemical groups identified. The occurrence of these potential sources depends on the surrounding environment of saltpans and contributes to the existence of different sea salt volatile profiles (Fig. 2). Seven of the 10 sea salt potential volatile markers are carotenoid derivatives. These may arise from carotenoid degradation, such as from the natural pigment b-carotene (Winterhalter & Rouseff, 2002) present in plants and algae. For instance, halophilic microalga D. salina accumulates large amounts of b-carotene (Lamers, Janssen, De Vos, Bino, & Wijffels, 2008). The compound 6MHO was already reported as a component of the halophytes S. europaea rubra and Puccinellia nuttalliana (Evans, 1994), and marine brown algae F. serratus and Hormophysa cuneiformis (Beauchêne et al., 2000; El Hattab et al., 2007). TDP was also reported to be related to brown algae (El Hattab et al., 2007). Carotenoid derivatives 6MHO, TCH, and TPD were previously identified as volatile components of the marine green alga Capsosiphon fulvescens (Sun et al., 2012). 6MHO, along with isophorone and ketoisophorone, were previously identified in a shore-dwelling cyanobacterial mat community from the hypersaline Wells Lake, in Canada (Evans, 1994). b-Ionone 5, 6-epoxide and TPD were found in the marine red algae Bostrychia radicans and Bostrychia tenella growing on the rocky shore (Oliveira, Silva, Turatti, Yokoya, & Debonsi, 2009). DHA has already

Fig. 5. Potential sources of sea salt volatile compounds ((1) Evans, 1994; (2) Elenkov et al., 1995; (3) Krock & Wilkins, 1996; (4) Beauchêne et al., 2000; (5) Grossi & Raphel, 2003; (6) Gao et al., 2004; (7) Kamenarska et al., 2006; (8) Nalli et al., 2006; (9) Shibata et al., 2006; (10) El Hattab et al., 2007; (11) Erakin & Güven, 2008; (12) Gressler et al., 2009; (13) Sun et al., 2012).

112

I. Silva et al. / Food Chemistry 169 (2015) 102–113

been identified in several kinds of marine algae (Gressler et al., 2009), such as the green alga Cladophora vagabunda (Elenkov et al., 1995), and numerous red algae (Kamenarska et al., 2006). TPD has been found as a component of Zostera marina shoots (Kawasaki et al., 1998), an aquatic plant of marine environments also known as eelgrass, and in the marine green alga Ulva pertusa (Gressler et al., 2011). The majority of the above algae exist in the Atlantic Ocean (http://www.algaebase.org/). Considering the non-carotenoid derivatives, it may be pointed out that 2EH was previously identified in marine brown algae F. serratus (Beauchêne et al., 2000), and 3HPP was also detected as a minor compound in the secretions of marine brown algae Eisenia bicyclis and Ecklonia kurome (Shibata et al., 2006). 3HPP, and also 6MHO, were routinely identified in the coastal atmosphere (oceanic and continental air masses) of Mace Head in Ireland, with 6MHO and 2EH the major oxygenated compounds of these environments (Sartin, Halsall, Davison, Owen, & Hewitt, 2001). In this study 6MHO was considered to be ubiquitous and of biogenic origin while 2EH and 3HPP were of unclear origin. 2EH has been reported as having a non-natural source since it is used in the production of plasticisers for polyvinyl chloride (PVC) resins and as an intermediate in the manufacture of inks, paper, rubber, resins, surfactants, and lubricants (Staples, 2001). It can also be produced by bacteria and fungi from degradation of plasticisers (Nalli et al., 2006). No marine or hypersaline origins have been found for the non-carotenoid derivative TMPP. However, this compound was already identified in natural sources, such as Scapania verrucosa Heeg, and its endophytic fungus Chaetomium fusiforme (Guo et al., 2008). Consequently, the non-carotenoid derivatives present in all the salts analysed may result from the integration of compounds from anthropogenic activity on the metabolism of marine organisms.

4. Conclusions Analysis of the volatile composition of sea salts from Île de Ré, Aveiro, Figueira da Foz, Castro Marim, Cádiz, and from La Palma and Sal islands, produced in 2007, enabled the detection of 165 compounds distributed over the chemical groups of hydrocarbons, aldehydes, esters, haloalkanes, ketones, ethers, alcohols, terpenic compounds, carotenoid derivatives, and lactones. In spite of the particular volatile profile of each salt under study modulated by several factors, namely by ocean currents, and saltpan surrounding environment, it was possible to identify a sub-set of ten compounds common to all, namely 6-methyl-5-hepten-2-one, 2,2,6-trimethylcyclohexanone, isophorone, ketoisophorone, b-ionone 5,6-epoxide, dihydroactinidiolide, 6,10,14-trimethyl-2pentadecanone, 3-hydroxy-2,4,4-trimethylpentyl 2-methylpropanoate, 2,4,4-trimethylpentane-1,3-diyl bis(2-methylpropanoate), and 2-ethyl-1-hexanol. These compounds have their origin in the environmental elements such as plants, algae, bacteria of marine and/or other hypersaline elements. These potential volatile markers seem to characterise sea salt, independently from saltpan, geographic origin and harvest. Thus, the volatile organic compounds can be markers of sea salt. Volatile markers of sea salt could play an important role in the definition of this handmade food product, and so add to the value of sea salt as a distinct and desirable product. A more extensive study should be conducted including samples collected across the world’s oceans. On the other hand, although representing minor components, the volatile compounds of sea salt may be seen as flavour compounds with a potential contribution to the organoleptic and chemical properties of the foodstuffs where sea salt is added, leading to potential sea salt taste modulators.

Acknowledgements Funding is acknowledged from the European Regional Development Fund (FEDER) through the Competitive Factors Thematic Operational Programme (COMPETE) and from the Fundação para a Ciência e Tecnologia (FCT), Portugal, under projects PEst-C/QUI/UI0062/2013 and FCOMP-01-0124-FEDER-037296 (Research Unit 62/94 QOPNA). The authors thank Mr. Paul Morrison for his technical assistance, SGE International for provision of capillary columns, and acknowledge support from Agilent Technologies for gas chromatography facilities, and LECO Corp. for maintenance of the ToFMS system. The authors thank the project SAL – ‘‘Sal do Atlântico’’ (INTERREG IIIB) and both Professor Filomena Martins and Ana Margarida Silva for providing the European sea salt samples. The authors also thank the trading company SunSalt for providing sea salt samples from the Murray Darling Basin, in Australia. I. Silva thanks Fundação para a Ciência e Tecnologia (FCT) for her Ph.D. Grant SFRH/BD/31076/2006. References Adams, R. P. (1995). Identification of essential oil components by gas chromatography/ mass spectrometry (1st ed.). Illinois, USA: Allured Publishing Corp. Babushok, V. I., & Zenkevich, I. G. (2009). Retention indices for most frequently reported essential oil compounds in GC. Chromatographia, 69, 257–269. Beauchêne, D., Grua-Priol, J., Lamer, T., Demaimay, M., & Quémeneur, F. (2000). Concentration by pervaporation of aroma compounds from Fucus serratus. Journal of Chemical Technology and Biotechnology, 75, 451–458. Donadio, C., Bialecki, A., Valla, A., & Dufossé, L. (2011). Carotenoid-derived aroma compounds detected and identified in brines and speciality sea salts (fleur de sel) produced in solar salterns from Saint-Armel (France). Journal of Food Composition and Analysis, 24, 801–810. El Hattab, M., Al Easa, H. S. S., Tabaries, A., Piovetti, L., & Kornprobst, J. M. (2007). Volatile components of the Phaeophyceae Hormophysa cuneiformis growing along Qatar coasts. Journal of Essential Oil Research, 19, 37–39. Elenkov, I., Georgieva, T., Hadjieva, P., Dimitrovakonaklieva, S., & Popov, S. (1995). Terpenoids and sterols in cladophora-vagabunda. Phytochemistry, 38, 457–459. Erakin, S., & Güven, K. C. (2008). The volatile petroleum hydrocarbons in marine algae around Turkish coasts. Acta Pharmaceutica Sciencia, 50, 167–182. Evans, W. G. (1994). Volatile organic-chemicals of a shore-dwelling cyanobacterial mat community. Journal of Chemical Ecology, 20, 219–230. Gao, H., Zhao, T., Kong, Q., Chen, X. G., & Hu, Z. D. (2004). Analysis of unknown organic pollutants in sewage by solid-phase extraction combined with gas chromatography–mass spectrometry. Journal of Chromatographic Science, 42, 91–99. Gressler, V., Colepicolo, P., & Pinto, E. (2009). Useful strategies for algal volatile analysis. Current Analytical Chemistry, 5, 271–292. Gressler, V., Fujii, M. T., Martins, A. P., Colepicolo, P., Mancini, J., & Pinto, E. (2011). Biochemical composition of two red seaweed species grown on the Brazilian coast. Journal of the Science of Food and Agriculture, 91(9), 1687–1692. Grossi, V., & Raphel, D. (2003). Long-chain (C19–C29) 1-chloro-n-alkanes in leaf waxes of halophytes of the Chenopodiaceae. Phytochemistry, 63, 693–698. Guo, L., Wu, J. Z., Han, T., Cao, T., Rahman, K., & Qin, L.-p. (2008). Chemical composition, antifungal and antitumor properties of ether extracts of Scapania verrucosa Heeg. and its endophytic fungus Chaetomium fusiforme. Molecules, 13, 2114–2125. Hoet, S., Stévigny, C., Hérent, M. F., & Quetin-Leclercq, J. (2006). Antitrypanosomal compounds from the leaf essential oil of Strychnos spinosa. Planta Medica, 72, 480–482. Jiang, Y., King, J. M., & Prinyawiwatkul, W. (2014). A review of measurement and relationships between food, eating behavior and emotion. Trends in Food Science & Technology, 36, 15–28. Kamenarska, Z., Ivanova, A., Stancheva, R., Stoyneva, M., Stefanov, K., et al. (2006). Volatile compounds from some Black Sea red algae and their chemotaxonomic application. Botanica Marina, 49, 47–56. Kawasaki, W., Matsui, K., Akakabe, Y., Itai, N., Kajiwara, T., & Dimitrova-Konaklieva, S. (1998). Volatiles from Zostera marina. Phytochemistry, 47, 27–29. Krock, K. A., & Wilkins, C. L. (1996). Qualitative analysis of contaminated environmental extracts by multidimensional gas chromatography with infrared and mass spectral detection (MDGC-IR-MS). Journal of Chromatography A, 726, 167–178. Lamers, P. P., Janssen, M., De Vos, R. C. H., Bino, R. J., & Wijffels, R. H. (2008). Exploring and exploiting carotenoid accumulation in Dunaliella salina for cellfactory applications. Trends in Biotechnology, 26, 631–638. Leffingwell, J. C., & Alford, E. D. (2005). Volatile constituents of Perique tobacco. Electronic Journal of Environmental, Agricultural and Food Chemistry, 4, 899–915. McGinitie, T. M., & Harynuk, J. J. (2012). Prediction of retention times in comprehensive two-dimensional gas chromatography using thermodynamic models. Journal of Chromatography A, 1255, 184–189.

I. Silva et al. / Food Chemistry 169 (2015) 102–113 Meziane, T., Bodineau, L., Retiere, C., & Thoumelin, G. (1997). The use of lipid markers to define sources of organic matter in sediment and food web of the intertidal salt-marsh-flat ecosystem of Mont-Saint-Michel Bay, France. Journal of Sea Research, 38, 47–58. Nalli, S., Horn, O. J., Grochowalski, A. R., Cooper, D. G., & Nicell, J. A. (2006). Origin of 2-ethylhexanol as a VOC. Environmental Pollution, 140, 181–185. Oliveira, A. L. L., Silva, D. B., Turatti, I. C. C., Yokoya, N. S., & Debonsi, H. M. (2009). Volatile constituents of Brazilian Bostrychia species (Rhodomelaceae) from mangrove and rocky shore. Biochemical Systematics and Ecology, 37, 761–765. Petronilho, S., Coimbra, M. A., & Rocha, S. M. (2014). A critical review on extraction techniques and gas chromatography based determination of grapevine derived sesquiterpenes. Analytica Chimica Acta, . Sartin, J. H., Halsall, C., Davison, B., Owen, S., & Hewitt, C. N. (2001). Determination of biogenic volatile organic compounds (C8–C16) in the coastal atmosphere at Mace Head, Ireland. Analytica Chimica Acta, 428, 61–72. Sequeira, A. M. M., Mellin, C., Floch, L., Williams, P. G., & Bradshaw, C. J. A. (2014). Inter-ocean asynchrony in whale shark occurrence patterns. Journal of Experimental Marine Biology and Ecology, 450, 21–29. Serrano, R., Nácher-Mestre, J., Portolés, T., Amat, F., & Hernández, F. (2011). Nontarget screening of organic contaminants in marine salts by gas chromatography coupled to high-resolution time-of-flight mass spectrometry. Talanta, 85, 877–884. Shibata, T., Hama, Y., Miyasaki, T., Ito, M., & Nakamura, T. (2006). Extracellular secretion of phenolic substances from living brown algae. Journal of Applied Phycology, 18, 787–794. Silva, H., Dias, J., & Caçador, I. (2009). The role of salt marsh vegetation in sedimentation processes. Hydrobiologia, 621, 33–47. Silva, I., Rocha, S. M., & Coimbra, M. A. (2009). Headspace solid phase microextraction and gas chromatography-quadrupole mass spectrometry

113

methodology for analysis of volatile compounds of marine salt as potential origin biomarkers. Analytica Chimica Acta, 635, 167–174. Silva, I., Rocha, S. M., & Coimbra, M. A. (2010). Quantification and potential aroma contribution of b-ionone in marine salt. Flavour and Fragrance Journal, 25, 93–97. Silva, I., Rocha, S. M., Coimbra, M. A., & Marriott, P. J. (2010). Headspace solid-phase microextraction combined with comprehensive two-dimensional gas chromatography time-of-flight mass spectrometry for the determination of volatile compounds from marine salt. Journal of Chromatography A, 1217, 5511–5521. Simões, M. P., Calado, M. L., Madeira, M., & Gazarini, L. C. (2011). Decomposition and nutrient release in halophytes of a Mediterranean salt marsh. Aquatic Botany, 94, 119–126. Staples, C. A. (2001). A review of the environmental fate and aquatic effects of a series of C4 and C8 oxo-process chemicals. Chemosphere, 45, 339–346. Sun, S. M., Chung, G. H., & Shin, T. S. (2012). Volatile compounds of the green alga, Capsosiphon fulvescens. Journal of Applied Phycology, 24, 1003–1013. Válega, M., Lillebø, A. I., Caçador, I., Pereira, M. E., Duarte, A. C., & Pardal, M. A. (2008). Mercury mobility in a salt marsh colonised by Halimione portulacoides. Chemosphere, 72, 1607–1613. van den Dool, H., & Kratz, P. D. (1963). A generalization of retention index system including linear temperature programmed gas–liquid partition chromatography. Journal of Chromatography, 11, 463–471. Vasta, V., Ventura, V., Luciano, G., Andronico, V., Pagano, R. I., Scerra, M., et al. (2012). The volatile compounds in lamb fat are affected by the time of grazing. Meat Science, 90, 451–456. Winterhalter, P., & Rouseff, R. L. (2002). Carotenoid-derived aroma compounds: An introduction. Washington, DC: Oxford University Press. Zeng, Y. X., Zhao, C. X., Liang, Y. Z., Yang, H., Fang, H. Z., Yi, L. Z., et al. (2007). Comparative analysis of volatile components from Clematis species growing in China. Analytica Chimica Acta, 595, 328–339.