Article pubs.acs.org/JAFC
Easy Extraction Method To Evaluate δ13C Vanillin by Liquid Chromatography−Isotopic Ratio Mass Spectrometry in Chocolate Bars and Chocolate Snack Foods Monica Bononi,*,† Giancarlo Quaglia,‡ and Fernando Tateo† †
Department of Agricultural and Environmental SciencesProduction, Landscape, Agroenergy, University of Milan, Via Celoria 2, 20133 Milan, Italy ‡ Floramo Corporation Laboratories, Via Lime 4, 12047 Rocca de’ Baldi, Cuneo, Italy S Supporting Information *
ABSTRACT: An easy extraction method that permits the use of a liquid chromatography−isotopic ratio mass spectrometry (LC−IRMS) system to evaluate δ13C of vanillin in chocolate products and industrial flavorings is presented. The method applies the determination of stable isotopes of carbon to discriminate between natural vanillin from vanilla beans and vanillin from other sources (mixtures from beans, synthesis, or biotechnology). A series of 13 chocolate bars and chocolate snack foods available on the Italian market and 8 vanilla flavorings derived from industrial quality control processes were analyzed. Only 30% of products considered in this work that declared “vanilla” on the label showed data that permitted the declaration “vanilla” according to European Union (EU) Regulation 1334/2008. All samples not citing “vanilla” or “natural flavoring” on the label gave the correct declaration. The extraction method is presented with data useful for statistical evaluation. KEYWORDS: LC−IRMS, δ13C, vanillin, chocolate, snack foods
■
INTRODUCTION
synthetic or natural origin of vanillin as well as providing information about the starting materials.6,8−12 The determination of stable isotope ratio mass spectrometry of carbon is presently a useful method for discrimination between natural vanillin from vanilla beans and synthetic vanillin or that produced from biotechnological methods. Indeed, vanillin from vanilla beans is naturally enriched in 13C and can be distinguished from vanillin produced from other sources. Various authors have therefore suggested the application of the isotope ratio measurements of 13C/12C, expressed as δ13C,6,9,13−15 to demonstrate the origin to the various vanillins. The European Union (EU) Regulation 1334/200816 permits the citation of the flavoring source (in our case “vanilla”) only if the flavoring component has been obtained exclusively or by at least 95% (w/w) from the referenced source material. The citation in the label “natural flavoring”, following the same EU Regulation,16 may only be used if the flavoring component is derived from a different source material. δ13C can be determined by elemental analysis−isotopic ratio mass spectrometry (EA−IRMS) or gas chromatography− isotopic ratio mass spectrometry (GC−IRMS). Reference values for δ13C ratios concerning vanillin from various origins can be found in several publications.9,14,15,17 According to the range reported in the literature, values of δ13C greater than −21.518 or −21.817 for vanillin are used as a criterion of authenticity (vanillin from vanilla beans).
Numerous studies have been devoted to the authentication of natural flavoring extracts derived from the tropical orchid of the genus Vanilla, of which Vanilla planifolia and Vanilla tahitensis are the most widely used.1 The increased demand for natural vanilla extracts2 has encouraged research for new sources of vanillin (4-hydroxy-3-methoxybenzaldehyde), the major flavoring component of vanilla.1,3 Commercial vanillin available on the market can be derived from chemically synthesized processes, starting from eugenol, guaiacol, or lignin. Another commercial vanillin is the so-called biovanillin, produced through a biotechnological pathway and classified as a “natural flavor” because of the utilization of natural precursors or raw materials.4 The biovanillin production by microbial fermentation or enzymatic reaction starts from naturally derived precursors, such as ferulic acid, isoeugenol, or curcumin.1,3,5,6 Agro-wastes are potential sources for biovanillin production, including cereal bran, sugar beet pulp, rice bran oil, and palm oil biomass.5 To discriminate between vanillin from authentic vanilla extracts and vanillin derived from extracts adulterated with vanillin from different sources, investigations by chromatographic methods have been proposed and, recently, the Direction Générale de la Concurrence, de la Consommation et de la Répression des Fraudes (DGCCRF) revised the ranges for the vanilla aromatic ratios.7 Site-specific natural isotope fractionation combined with nuclear magnetic resonance (SNIF−NMR) measurements has produced data that define the site-specific abundance of deuterium or carbon-13.8,9 Several reports have shown that 2 H NMR spectroscopy allows for the determination of the © 2015 American Chemical Society
Received: Revised: Accepted: Published: 4777
January 22, 2015 May 2, 2015 May 4, 2015 May 12, 2015 DOI: 10.1021/acs.jafc.5b02136 J. Agric. Food Chem. 2015, 63, 4777−4781
Article
Journal of Agricultural and Food Chemistry Application of a precautionary judgment and consideration of the possible implications of the curing stage,18 the lower limit of −22.6 may be adopted. Data referring to vanillin of different origins, i.e., from isoeugenol, lignin, guaiacol, and ferulic acid, result in ratio values lower than −26.8.6,9,14 Herein, we report the data produced by analysis of δ13C for vanillin in 13 chocolate bars and chocolate snack foods available on the Italian market. Applying the suggested extraction method and the LC−IRMS analytical technique made it possible to confirm the lawful use of the term “vanilla” in these labels. Moreover, we report the results from 8 vanilla flavorings from the European flavoring market used for quality control in three food industries. We also included extracts from two sources of authentic vanilla beans. The label declarations of the considered samples are reported.
■
Table 2. LC−IRMS Data (Averages of Two Measurements) for δ13C (‰) of Vanillin in a Series of Industrial Flavorings and Two Laboratory Extracts from Vanilla Beans for a Useful Comparison sample
MATERIALS AND METHODS
Samples. Samples of snack foods and chocolate were directly purchased on the Italian market and represent well-known consumer products. The industrial vanilla flavorings were obtained from industrial suppliers. Two samples of genuine vanilla beans (source Madagascar) were obtained from a direct trader, and the corresponding extracts were produced by a percolation method (50:50 ethanol/water solution) under vacuum for 3 days. Chemicals. Hexane, ethanol, water, sodium persulfate, and orthophosphoric acid (Sigma-Aldrich, Italy) were analytical-reagentgrade and used without any purification. Helium (5.6 grade) as the carrier gas and carbon dioxide (4.5 grade) as the reference gas were produced by Messer (Turin, Italy); carbon dioxide was calibrated and controlled using calibrated materials. Solutions and dilutions were prepared with ultrapure Milli-Q water (Milan, Italy). Synthetic vanillin (99%) was a Sigma-Aldrich product. Sample Preparation. The chocolate bar or chocolate snack food matrix (Table 1) was finely crushed after freezing (T = −20 °C). A portion (50−100 g) of matrix was extracted for 6 h with a Soxhlet device using hexane as the solvent (150−300 mL). About 10 mL of industrial flavoring samples (samples 14−21 in Table 2) and two vanilla bean extracts produced in our laboratory (samples 22 and 23 in Table 2) were extracted 3 times for 5 min with hexane (30 mL each time) in a 100 mL graduated cylinder. The samples were also subjected to ultrasonic extraction (USE) each time
1 2 3 4 5 6 7 8 9 10 11 12 13
label
δ13C (‰)
judgmenta
chocolate snack chocolate snack chocolate snack chocolate snack chocolate snack chocolate nutmeg chocolate nutmeg chocolate nutmeg dark chocolate milk chocolate milk chocolate milk chocolate white chocolate
natural vanilla extract vanillin vanilla natural flavoring flavoring natural flavoring flavorings vanilla flavoring vanilla extract vanilla beans vanilla extract vanilla extract vanilla Bourbon natural flavoring
−23.7 −30.3 −22.5 −27.5 −32.5 −27.6 −24.0 −32.3 −28.3 −29.7 −21.8 −23.4 −26.9
not correct correct correct correct ? correct not correct not correct not correct not correct correct not correct ?
industrial flavor
15
industrial flavor
16
industrial flavor
17
industrial flavor
18
industrial flavor
19
industrial flavor
20
vanillin (powder)
21
industrial flavor
22
laboratory extract from vanilla beans laboratory extract from vanilla beans
23
vanilla extract natural flavoring vanilla extract natural flavoring vanilla extract vanilla extract natural flavoring vanilla extract
δ13C (‰) −23.1
judgmenta
−36.0
not correct correct
−21.4
correct
−30.3
?
−26.1 −18.5
not correct correct
−32.6
?
−25.5
not correct
−20.3 −20.7
a
Judgment concerning the outcomes with respect to EU Regulation 1334/2008 is also reported.
for 15 min (T = 25 °C) at a fixed frequency of 35 kHz. The three hexane fractions are collected and combined. The hexane extracts were concentrated by rotary evaporation under vacuum to a final volume of 90−100 mL and treated in a separatory funnel 3 times with 5 mL of water/ethanol (30:70), shaking for 5 min. The three water/ethanol phases were collected and pooled. The aqueous−alcoholic phase was concentrated by rotary evaporation to a final volume of 0.1−0.2 mL. For a better control of the concentration and to recover the residue, the round-bottom flask can be fitted in the bottom with a graduated conical tube. Dependent upon the vanillin concentration, the aqueous residue may be appropriately diluted with water and filtered (0.45 μm) before the IRMS analysis. Instrumentation. LC−IRMS. A Delta V Advantage (Thermo Fisher, Bremen, Germany) isotope ratio mass spectrometer was used coupled to a high-performance liquid chromatography (HPLC) system and the LC IsoLink interface (Thermo Fisher, Bremen, Germany). A P 1000 Thermo Fisher HPLC (Bremen, Germany) was used in conjunction with a Surveyor MS pump, a HTC PAL autosampler (CTC, Zwingen, Switzerland), and a column oven (Mistral, Spark Holland B.V., Emmen, Netherlands). The oxidation reagent consisted of sodium persulfate in water (1.5 M, 30 μL min−1), and the acid reagent was phosphoric acid (1.5 M, 50 μL min−1) pumped by two pump heads. The mixture of these reagents, added by a T piece to the mobile phase, flowed through a capillary oxidation reactor at 100 °C, as described by Krummen et al.19 for the LC IsoLink. After the oxidation reactor, the mobile phase was cooled and the individual CO2 peaks were transferred to a counter flow of helium; thereby, the liquid phase was completely degassed. The fluxing helium introduced CO2 to the mass spectrometer. For chromatographic separation, a Phenomenex HPX-87C Bio-Rad (Italy) column (300 × 7.8 mm) was used, with ultrapure water as the mobile phase; the flow rate was 0.3 mL min−1, and the injection volume was 15 μL. The column temperature was 50 °C. Water was used to prepare samples and standard solutions for LC−IRMS analysis; all samples were filtered through a polypropylene membrane filter with 0.45 μm porosity and 25 mm diameter. The Isodat 3.0 software (Thermo Scientific Bremen, Germany) was the controlling system. The analysis was repeated twice for each sample, and the average values of these two measurements are listed in Tables 1 and 2.
Table 1. LC−IRMS Data (Averages of Two Measurements) for δ13C (‰) of Vanillin in a Series of Chocolate Bars and Chocolate Snack Food Samples Available on the Italian Market sample
14
label
a
Judgment concerning the outcomes with respect to EU Regulation 1334/2008 is also reported. 4778
DOI: 10.1021/acs.jafc.5b02136 J. Agric. Food Chem. 2015, 63, 4777−4781
Article
Journal of Agricultural and Food Chemistry The δ13C values (‰) were calibrated to Vienna Pee Dee Belemnite (VPDB) by three pulses of CO2 reference gas, calibrated against the international standard, and performed at the beginning of the elution run. EA−IRMS. δ13C values of synthetic vanillin were measured with an Flash 2000 elemental analyzer coupled to a Delta V Advantage via a ConFlo IV interface. The EA worked under 100 mL min−1 helium flux and temperatures of 950 °C in the oxidation tube and 850 °C in the reduction tube. The outlet was equipped with a column that physically retained CO2 at 70 °C; carbon dioxide was released by increasing the temperature to 210 °C. The overall experiment duration was 600 s.
■
RESULTS AND DISCUSSION Method Validation. To evaluate the chemical oxidation of eluted compounds and CO2 extraction from the IRMS system, the linearity was verified from 0.50 to 20.00 mg L−1 synthetic vanillin (99%), analyzing each solution in triplicate. The corresponding area averages by IRMS were reported versus the vanillin concentration in Figure 1, and the correlation indicated the complete oxidation of the eluted molecule to CO2; also, we demonstrated the complete recovery of the gas.
Figure 2. δ13C values of a sample of synthetic vanillin at concentrations from 0.05 to 20.0 mg L−1 produced by LC−IRMS compared to the EA−IRMS mean target value measured on the solid vanillin.
the extraction conditions. Therefore, any influence on the δ13C value by natural pHB (δ13C from −15.2 to −17.0) would be irrelevant on the decision concerning the vanillin origin, even if traces of pHB were eluted with vanillin peak. The same observation is even more valid for all other compounds present in the considered matrices. Moreover, under the chromatographic conditions adopted for LC−IRMS, the retention time of pHB was sufficiently differentiated from that of vanillin. In fact, in the LC−IRMS chromatogram, pHB can be coeluted only with caffeine, when this last compound is present. Moreover, the caffeine peak is well-resolved in LC−IRMS from the vanillin peak. Thus, the extraction method adopted in this work allowed us to selectively avoid the influence of other volatile compounds present in the considered matrices. Figure 3 shows two examples of LC−IRMS traces of the vanillin peak that illustrate the good resolution of vanillin from caffeine when the latter substance is present.
Figure 1. Values of CO2 peak area LC−IRMS produced by vanillin synthetic solutions at concentrations ranging from 0.5 to 20.0 mg L−1.
The agreement between δ13C values produced by a sample of synthetic vanillin, determined from the solid by EA−IRMS, and expressed as a mean of five measurements (target value) and the δ13C values determined by the LC−IRMS method on the same vanillin specimen at different concentration values were also evaluated. The data were distributed in the range ±0.4‰ for concentration values from 0.05 to 20 mg L−1 (Figure 2). Therefore, undetectable isotopic fractionation was observable. Reproducibility was checked adopting the Nordic Committee on Food Analysis (NMKL) method20 and with consideration of the relative repeatability standard deviation (RSDr = 0.011) derived from all data in duplicate obtained for all samples listed in Tables 1 and 2 (see Table S1 of the Supporting Information). The samples considered in the present study, following the suggested extraction method, had all been previously analyzed by HPLC−diode array detection (DAD) to evaluate the possible influence of other compounds; we were able to demonstrate that the vanillin peak and, consequently, δ13C values of vanillin in the analyzed samples were not influenced by the nature of the sample matrix. Furthermore, in all samples analyzed, the p-hydroxybenzaldehyde (pHB) peak was at least 25−30 times lower than the vanillin peak because of the partition coefficient of pHB under
Figure 3. Examples of resolution between vanillin and caffeine peaks in LC−IRMS traces for two samples, where their ratios were different, adopting the suggested operative conditions (samples 2 and 6 in Table 1).
Sample Measurements. The results of the stable isotope ratio analysis of the samples (Tables 1 and 2) identified some discrepancies with the product declaration on the label of the food or flavoring agent. Of the chocolate bars and chocolate snack foods in Table 1, the samples 1, 3, 7, 8, 9, 10, 11, and 12 were declared as “vanilla” on the label. Only samples 3 and 11 gave results that agreed with this declaration.16 Of the other samples 2, 4, 5, 6, and 13, only 2, 4, and 6 reported a lawful declaration16 because no reference was made to the terms “vanilla” or “natural flavoring”. The samples 5 and 13 reported the declaration “natural flavoring”, but the data 4779
DOI: 10.1021/acs.jafc.5b02136 J. Agric. Food Chem. 2015, 63, 4777−4781
Article
Journal of Agricultural and Food Chemistry (−32.5 and −26.9, respectively) did not permit the confirmation that a natural source of vanillin was used. These values may be justified if the flavoring agent was derived from a mixture of biovanillin from ferulic acid and vanillin from vanilla beans. Such a designation would be valid if biovanillin is considered as “natural”. Data reported in Table 2 showed that two samples (16 and 19) of five industrial flavoring agents declared as “vanilla” (14, 16, 18, 19, and 21) were really derived from vanilla beans. Samples 14, 18, and 21 produced data lower than −22.6,20 which is the lower limit compatible with vanillin from vanilla beans; therefore, for these three samples, the label declaration was not lawful. The other three samples, 15, 17, and 20, declared “natural flavoring” and showed very low values. Only sample 15 was clearly compatible with biovanillin from ferulic acid,9 while the other two data values permitted the declaration “natural flavoring” only if derived from a mixture of biovanillin from ferulic acid and vanillin from vanilla beans. The same analytical method was applied for the two samples produced by laboratory extraction from two samples of genuine vanilla beans (23 and 24) and the values produced results that were strictly included in the range cited by the literature for vanillin from vanilla beans. From the results of Tables 1 and 2, we deduced that of the 13 samples citing the term “vanilla” on the label, only four of them (3, 11, 16, and 19) were in compliance with the EU Regulation 1334/2008.16 Three samples not citing the terms “vanilla” or “natural” (2, 4, and 6) showed the correct label. Accepting that the biovanillin from ferulic acid, which produces a δ13C value of absolute −36.0, may actually be considered as a “natural flavoring”, we were unable to assign a definite judgment for the four samples 5, 13, 17, and 20, which cited the term “natural flavoring”, on their labels. Indeed, a wide range of data (from −22.7 to −36.1) corresponded to the various putative mixtures of vanillin from vanilla beans, synthetic vanillin, and biovanillin. In conclusion, for the food matrices examined in the present work, the suggested method was able to distinguish between flavorings derived from authentic vanilla beans and those flavorings from other sources. Data reported in Tables 1 and 2 permitted us to conclude that only ca. 30% of the samples citing the term “vanilla” were in accordance with the EU Regulation 1334/200816 and all of the samples not citing the terms “vanilla” or “natural” agreed with the same regulation. Only for sample 15, which showed a δ13C value of −36.0 characteristic of biovanillin from ferulic acid, was it possible to confirm the classification as “natural flavoring”.
■
Notes
The authors declare no competing financial interest.
■
(1) Ramachandra Rao, S.; Ravishankar, G. A. Vanilla flavour: Production by conventional and biotechnological routes. J. Sci. Food Agric. 2000, 80, 289−304. (2) Brownell, R. New approaches to a sustainable vanilla supply chain. Perfum. Flavor. 2014, 39, 28−30. (3) Kaur, B.; Chakraborty, D. Biotechnological and molecular approaches for vanillin production: A review. Appl. Biochem. Biotechnol. 2013, 169, 1353−1372. (4) Schreier, P. Bioflavours: An overview. In Biotransformation of Flavours; Patterson, R. L. S., Charlwoods, B. V., MacLeod, G., Williams, A. A., Eds.; Royal Society of Chemistry (RSC): Cambridge, U.K., 1992; pp 1−20. (5) Zamzuri, N. A.; Abd-Aziz, S. Biovanillin from agro wastes as an alternative food flavour. J. Sci. Food Agric. 2013, 93, 429−438. (6) Benz, I.; Muheim, A. Biotechnological production of vanillin. In Flavour ScienceRecent Developments; Taylor, A. G., Mottram, D. S., Eds.; Royal Society of Chemistry (RSC): Cambridge, U.K., 1996; pp 111−117. (7) Ministry of Economy, Finance, and Industry. Directorate General for Competition, Consumption, and Fraud Repression; Ministry of Economy, Finance, and Industry: Paris, France, June 16, 2003; Information Note 2003-61. (8) Remaud, G.; Martin, Y. L.; Martin, G. G.; Martin, G. J. Detection of sophisticated adulterations of natural vanilla flavors and extracts: Application of the SNIF−NMR method to vanillin and phydroxybenzaldehyde. J. Agric. Food Chem. 1997, 45, 859−866. (9) Bensaid, F. F.; Wietzerbin, K.; Martin, G. J. Authentication of natural vanilla flavorings: Isotopic characterization using degradation of vanillin into guaiacol. J. Agric. Food Chem. 2002, 50, 6271−6275. (10) Martin, G.; Remaud, G.; Martin, G. J. Isotopic methods for control of natural flavours authenticity. Flavour Fragrance J. 1993, 8, 97−107. (11) John, T. V.; Jamin, E. Chemical investigation and authenticity of Indian vanilla beans. J. Agric. Food Chem. 2004, 52, 7644−7650. (12) Tenailleau, E. J.; Lancelin, P.; Robins, R. J.; Akoka, S. Authentication of the origin of vanillin using quantitative natural abundance 13C NMR. J. Agric. Food Chem. 2004, 52, 7782−7787. (13) Culp, R. A.; Legat, J. M.; Otero, E. Carbon dioxide composition of selected flavoring compounds for the determination of natural origin by gas chromatography/isotope ratio mass spectrometer. ACS Sym. Ser. 1998, 705, 260−287. (14) Godin, J. P.; MacCullagh, J. S. O. Review: Current applications and challenges for liquid chromatography coupled to isotope ratio mass spectrometry (LC/IRMS). Rapid Commun. Mass Spectrom. 2011, 25, 3019−3028. (15) Lamprecht, G.; Pichlmayer, F.; Schmidt, E. R. Determination of the authenticity of vanilla extracts by stable isotope ratio analysis and component analysis by HPLC. J. Agric. Food Chem. 1994, 42, 1722− 1727. (16) European Union (EU).. Regulation (EC) No 1334/2008 of the European Parliament and of the Council of 16 December 2008 on flavourings and certain food ingredients with flavouring properties for use in and on foods and Amending Council Regulation (EEC) No 1601/91, Regulations (EC) No 2232/96 and (EC) No 110/2008. Off. J. Eur. Union, L: Legis. 2008, 51, 34−50. (17) Hansen, A. M. S.; Fromberg, A.; Frandsen, H. L. Authenticity and traceability of vanilla flavors by analysis of stable isotopes of carbon and hydrogen. J. Agric. Food Chem. 2014, 62, 10326−10331. (18) Gassenmeier, K.; Binggeli, E.; Kirsch, T.; Otiv, S. Modulation of the 13C/12C ratio of vanillin from vanilla beans during curing. Flavour Fragrance J. 2013, 28, 25−29. (19) Krummen, M.; Hilkert, A. W.; Juchelka, D.; Duhr, A.; Schluter, H. J.; Pesch, R. A new concept for isotope ratio monitoring liquid chromatography/mass spectrometry. Rapid Commun. Mass Spectrom. 2004, 18, 2260−2266.
ASSOCIATED CONTENT
S Supporting Information *
Vanillin isotope ratio determined on all of the samples listed in Tables 1 and 2 and RSDr derived from data in duplicate (Table S1) (PDF). The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/ acs.jafc.5b02136.
■
REFERENCES
AUTHOR INFORMATION
Corresponding Author
*Telephone: +39-0250316538. Fax: +39-0250316539. E-mail:
[email protected]. 4780
DOI: 10.1021/acs.jafc.5b02136 J. Agric. Food Chem. 2015, 63, 4777−4781
Article
Journal of Agricultural and Food Chemistry (20) Nordic Committee on Food Analysis (NMKL). Estimation and Expression of Measurement Uncertainty in Chemical Analysis; NMKL: Oslo, Norway, 1997; NMKL Procedure 5.
4781
DOI: 10.1021/acs.jafc.5b02136 J. Agric. Food Chem. 2015, 63, 4777−4781
