Whisky Analysis by Electrospray Ionization-Fourier Transform Mass Spectrometry

Food Research International 51 (2013) 98–106

Contents lists available at SciVerse ScienceDirect

Food Research International

journal homepage: www.elsevier.com/locate/foodres

Whisky analysis by electrospray ionization-Fourier transform mass spectrometry

Jerusa S. Garcia a,b,⁎, Boniek G. Vaz a, Yuri E. Corilo a, Christina F. Ramires a, Sérgio A. Saraiva a, Gustavo B. Sanvido a, Eduardo M. Schmidt a, Denison R.J. Maia c, Ricardo G. Cosso c, Jorge J. Zacca d, Marcos Nogueira Eberlin a

a University of Campinas, UNICAMP, Institute of Chemistry, ThoMSon Mass Spectrometry Laboratory, 13084-971, Campinas, SP, Brazil b Federal University of Alfenas, UNIFAL, Institute of Chemistry, 37130-000, Alfenas, MG, Brazil c Brazilian Federal Police, Ministério da Justiça, Superintendência Regional de São Paulo, 05038-090, São Paulo, SP, Brazil d Brazilian Federal Police, Ministry of Justice, National Institute of Criminalistics — INC, 70390-145, Brasilia, DF, Brazil

article info abstract

Article history: Electrospray ionization (ESI) coupled to ultra-high resolution and accuracy Fourier transform ion cyclotron Received 15 October 2012 resonance mass spectrometry (FT-ICR MS) was employed for the direct analysis of whisky samples. No Accepted 15 November 2012 pre-separation or extraction steps were employed and, owing to the gentleness of ESI, characteristic profiles of polar constituents of authentic whiskies from five different brands were obtained, which greatly contrast Keywords: to those of counterfeit samples. Owing to the accuracy of FT-ICR MS mass measurements, elemental compo- Whisky sition of the main ions detected were also securely determined, and related to a series of carboxylic acids, MS fingerprinting Authenticity phenols, saccharides, fat acids and sulfur or nitrogen constituents of whisky. The accurate mass measure- High resolution mass spectrometry ments and the number of detected ions, together with the direct, no sample preparation procedure employed, make FT-ICR MS a highly reliable approach to fingerprint whisky and to control its quality, and to screen for aging, counterfeiting and adulteration at the molecular level. © 2012 Elsevier Ltd. Open access under the Elsevier OA license.

1. Introduction Scotch whisky is a prime spirit drink and a leading category of choice between spirit consumers (Gordon, 2003). The Scotch whisky in UK is Fourier transform ion cyclotron resonance mass spectrometry on the range of several billions of dollars (Rhodes, Heaton, Goodall, (FT-ICR MS) is a powerful technique for complex mixture analysis & Brereton, 2009) and whisky is one of the top export products of due to its unsurpassed ability to determine chemical formulas with Scotland. Due to its large commercialization and relatively high prices, ultra high resolution and accuracy (better than 1 ppm) (Marshall et Scotch whisky counterfeiting and/or adulteration is quite common al., 2007). Due to its unique characteristics, FT-ICR MS is a powerful worldwide. Whisky contains a great variety of constituents from differ- View metadata, citation and similar papers at core.ac.uk brought to you by CORE tool for resolving the elemental composition of large molecules and ent chemical classes such as alcohols, ethyl and isoamyl esters, acetates, can be used in distinct applications. ketones, fatty acids,provided monoterpenes, by Elsevier - Publisher and phenols. Connector This variable chemical The use of the gentle ESI technique or ambient ionization tech- composition is mainly influenced by the cereals used in fermentation niques (Corilo et al., 2010) also permit the detection of constituents and by distillation, maturation and blending regimes. Some of these in intact molecular forms, and when FT-ICR MS analysis is employed, compounds originate from raw materials and during the production direct characterization of very complex mixtures such as of crude oils steps (mashing, fermentation and distillation), while others are related can be directly performed without pre-separation methods. For crude to the spirit maturation in oak casks. Whisky constituents can be pres- oils, ESI FT-ICR MS have allowed unambiguous assignments of polar ent in a wide range of concentrations, with contrasting volatilities and heteroatom-containing organic components of crude oils with 20,000 polarities, being common in distinct whisky brands but differing consid- or more distinct elemental compositions (Rodgers, Schaub, & Marshall, erably in relative quantities (Câmara et al., 2007). 2005) and for wine around 30 compounds were identified employing Due to this complex chemical composition, whisky authenticity negative-ion ESI FT-ICR analysis (Cooper & Marshall, 2001). Herein, analysis normally requires sample extraction and pre-separation pro- we describe an investigation aimed at testing the use of ESI FT-ICR MS cedures. The most common strategy for establishing whisky brand in the analysis of whisky samples. authenticity is to determine the profile of volatile organic compounds (VOC congeners) such as acetaldehyde, methanol, ethyl acetate, propanol, isobutanol, diethyl acetal and the amyl alcohols: 2- and 3-methyl butanol (Aylott, Clyne, Fox, & Walker, 1994). Usually, ⁎ Corresponding author at: ThoMSon Mass Spectrometry Laboratory 3084-971, fi Campinas, São Paulo, Brazil. Tel./fax: +55 19 3521 3073. VOC pro les of whisky are obtained via gas chromatography (GC) anal- E-mail address: [email protected] (J.S. Garcia). ysis (MacKenzie & Aylott, 2004; Nascimento, Cardoso, & Franco, 2008;

Rhodes et al., 2009). However, spirit adulteration is becoming more whisky “brands” were also analyzed. The counterfeit street samples sophisticated and sometimes counterfeit samples may display simi- were provided by the Brazilian Federal Police. lar VOC profiles. GC coupled to isotope ratio mass spectrometry (GC/IRMS) has also been used to characterize whisky and to screen 2.2. ESI FT-ICR MS analysis for adulteration (Parker, Kelly, Sharman, & Howie, 1998). Direct mass spectrometry (MS) analysis has also been employed Samples were analyzed using a hybrid 9-T Fourier transform ion to characterize spirits and screen for adulteration and counterfeiting. cyclotron resonance mass spectrometer (LTQ FT; Thermo Scientific, We, for instance, have employed direct “unit-resolution” MS analysis Bremen, Germany) equipped with a chip-based direct infusion nano- with electrospray ionization (ESI-MS) for quality control of spirits, electrospray ionization source (Triversa; Advion Biosciences, Ithaca, including the Brazilian cachaça (de Souza, Augusti, et al., 2007; de NY, USA). Nanoelectrospray conditions comprised a flow rate of Souza, Siebald, et al., 2007; de Souza et al., 2009), wine (Catharino et 200 nL/min, backing pressure of ca. 0.3 psi, and electrospray voltages al., 2006), beer (Araujo et al., 2005)andwhisky(Moller, Catharino, & of 1.5 to 2.0 kV during 120 s, controlled by ChipSoft software (version Eberlin, 2005). 8.1.0, Advion Biosciences, Ithaca, NY, USA). Mass resolution was fixed at 200,000 at m/z 400. Data were obtained as transient files (scans recorded in the time domain). 2. Experimental details All the samples were evaluated in negative ESI(−) and positive ESI(+) ion modes and spectra were acquired in the m/z 100–800 2.1. Whisky samples range. For ESI(+), the samples were analyzed directly (without any sample treatment or dilution); for ESI(−), samples were diluted 1:1 A total of 80 samples, divided into two groups (50 authentic and (v/v) using a solution containing methanol (Mallinckrodt Baker,

30 counterfeit) were used. The authentic whiskies include Scotch Phillipsburg, NJ, USA) and 0.1% (v/v) NH4OH (Mallinckrodt Baker, whiskies from different brands: Johnnie Walker Red Label (n=10), Phillipsburg, NJ, USA). Johnnie Walker Black Label (n=12), White Horse (n=12), Buchanan's The spectra were processed by the Xcalibur Analysis software (n=6) and J&B (n=10). Counterfeit samples (n=30) from different package (version 2.0, Service Release 2, Thermo Electron Corporation).

383.12

365.11

527.16

409.16

425.14 545.17 509.15 623.23 443.17 589.23 467.10 689.21 671.20

350 450 550 650 157.08 100 203.05 383.12 527.16 347.09 409.16 (A) 50 231.08 509.15 545.17 301.14 623.23 689.21 737.52 0 383.12 100 157.08 203.05 231.08 365.11 409.16 527.16 347.09 (B) 50 271.08 509.15 545.17 623.23 311.07 689.21 737.50 0 425.14 100 409.16

50 157.08 203.05 (C) 443.17 589.23 271.08 383.12 527.16 639.20 737.47 785.28 0 157.08 100 Relative Abundance 409.16 203.05 365.11 527.16 50 347.09 589.23 (D) 231.08 425.14 301.14 617.26 689.21 737.46 0 383.12 100 157.08 203.05 288.29 365.11 409.16 527.16 50 231.08 623.23 (E) 347.09 509.15 545.17 639.20 737.46 785.28 0 200 300 400 500 600 700 800 m/z

Fig. 1. Representative ESI(+) FT-ICR MS of ﬁve distinct Scotch whisky brands: (A) Johnnie Walker Red Label, (B) Johnnie Walker Black Label, (C) J&B, (D) White Horse and (E) Buchanan's. 100 J.S. Garcia et al. / Food Research International 51 (2013) 98–106

The elemental compositions were generated using the Quali Browser ESI(−) FT-ICR MS of the same samples (ten spectra are shown in software (version 2.0, Thermo Electron Corporation). Elemental com- both figures). Both spectra demonstrate the complexity of the polar position was acceptable only when the average differences between chemical profiles of whisky provided by the technique, with more calculated and experimental masses were less than 1.0 ppm. Molecular than 500 ions observed with more than 1% of relative abundance. formula search was done using 12C, 1H, 16O, 14N, 23Na and 32S All authentic samples share many common ions which is beneficial isotopologue ions. for establishing a common chemical signature of Scotch whisky in terms of polar components. Another typical characteristic common 2.3. Multivariate analysis to all the authentic whisky samples is a “grass ion” with a somewhat Gaussian distribution around m/z 300–700 for ESI(+) and m/z To classify the whisky samples, principal component analysis (PCA) 250–600 for ESI(−), as shown by the insets in Figs. 1 and 2. But whis- was performed using the Pirouette software (version 3.11, Infometrix ky samples from different brands also display characteristic features Inc., Woodinville, WA, USA). The ESI-MS experimental data were com- in terms of relative abundances and unique ions that permit differen- piled to generate a final matrix containing samples and variables (m/z tiation between samples, using both ESI(+) and ESI(−), more partic- values with their respective relative intensities). To construct the ma- ularly via ESI(−), as clearly demonstrated by the PCA plots of Fig. 3. trix, no pre-processing method was performed and only the 50 more Three principal components were necessary to describe more abundant ions were considered. than 58.0% of the data variance in the ESI(+) results (PC1=40.4%, PC2=10.5% and PC3=7.1%). In the positive ion mode the different 3. Results and discussion whisky brands are characterized by certain ions: m/z 269, 323 and 503 for Black Label; m/z 323, 431, 458 and 521 for Red Label; 239, 3.1. Profiles in positive and negative ion modes 351 and 359 for Buchanan's; 172, 343, 399 and 519 for J&B; and, finally, m/z 255, 341, 421, 443 for White Horse. Fig. 1 exhibits the ESI(+) FT-ICR MS of a representative authentic The total variability explained by three components in ESI(−) was sample from the five Scotch whisky brands investigated (one spec- 30.3% (PC1=14.6%, PC2=8.4% and PC3=7.8%). In the negative ion trum from each brand) whereas Fig. 2 shows, for comparison, the mode the characteristic ions are: m/z 359 (Black Label); m/z 503

341.11 359.12 301.00

239.08 503.16

255.23 379.23 517.32 323.10

485.15 401.13 431.14

250 350 450 550

171.14 100 199.17 341.11 (A) 50 143.11 503.16 255.23 323.10 401.13 485.15 521.17 575.18 665.22 737.52 775.21 0 171.14 100

(B) 50 359.12 503.16 143.11 301.00 239.08 379.23 485.15 521.17 575.18 665.22 737.45 777.23 0 171.14 100 199.17

(C) 50 143.11 255.23 301.00 341.11 503.16 419.17 603.01 665.22 737.47 783.77 0 171.14 100 Relative Abundance 199.17 341.11 (D) 50 503.16 143.11 323.10 359.12 255.23 431.14 521.17 575.18 665.22 737.45 793.22 0 171.14 100

(E) 50 239.08 143.11 341.11 301.00 379.23 503.16 545.17 679.37 737.45 777.23 0 100 200 300 400 500 600 700 800 m/z

Fig. 2. Representative ESI(−)FT-ICRMSofﬁve distinct Scotch whisky brands: (A) Johnnie Walker Red Label, (B) Johnnie Walker Black Label, (C) J&B, (D) White Horse and (E) Buchanan's. J.S. Garcia et al. / Food Research International 51 (2013) 98–106 101

3.2. Elemental composition and formula

(A) Tables 1 and 2 show the ESI FT-ICR MS elemental composition as- 8 Buchanan’s signments of major ions (20 ions with higher relative abundance) detected in the whisky samples. In the negative ion mode (Table 1), 6 White Horse most ions correspond to deprotonated forms of carboxylic acids, phenols, saccharides derivates and other sulfur or nitrogen compounds. 4 Red Label Scheme 1 proposes the chemical structures for such ions; those of 2 m/z 143.1078 and 171.1391 correspond to caprylic and capric acid,

PC3 respectively. These compounds were also detected in cachaça samples 0 J&B Black Label (Moller et al., 2005), and are common in food processing (Beare-Rogers, -2 10 Dieffenbacher, & Holm, 2001). other identiﬁed fatty acids may be related 5 to hydrolysis of fatty acid esters or lactones (Poisson & Schieberle, -4 0 2008). The ion of m/z 197.04574 was attributed to ethyl gallate (3), a -6 -5 food additive also found naturally in a variety of plants. Most of the 10 PC1 8 6 compounds identiﬁed in Table 1 are degradation products of maltose 4 2 -10 0 -2 and glucose or products from their reaction with phenolic compounds, PC2 -4 -15 -6 -8 or carboxylic acids extracted from the wood casks during the aging and maturation process, such as 11 (ellagic acid) and 15 (an isoprenoid derivate). Oxidoreductions and polymerizations occurring during maturation involve compounds present in the raw distillate and (B) wood-derived compounds, such as the lignin-derived p-coumaryl alcohol, coniferyl alcohol and sinapyl alcohol, and could account for forma- 10 Black Label Buchanan’s tions of 12 and 17 (Scheme 1). 5 The ions of m/z 179.05625 and 341.10943 were also detected via Red Label ESI(−)-MS as abundant ions of Scotch whisky samples (Moller et 0 J&B al., 2005). These ions correspond to deprotonated molecules of mono- White Horse f saccharides such as glucose (180 Da) and disaccharides such as sucrose -5

PC3 (342 Da). The spectra in the positive ion mode (Table 2) provided composi- -10 15 tion information complementary to that obtained in the negative -15 10 ion mode. The majority of compounds were detected by ESI(+) as 5 sodium adducts [M+Na]+ and Scheme 2 presents propositions for 0 -20 1 their structures. As discussed before, major compounds are likely 10 -5 5 PC formed from oxidoreduction processes such as 8 and 9; lignin- 0 -10 PC2 -5 -15 derivate products were also found such as 7 and 8,whicharederi- -10 vates of the sinapyl alcohol (Scheme 2). In some cases, secure molecular formula determination was not possible particularly for the Fig. 3. PCA from the ﬁve authentic Scotch whisky samples for the (A) ESI(−) and (B) ESI(+) FT-ICR MS data. heavier ions and the exponential increase of elemental composition variation that is particularly severe above 450 Da (Cooper & Marshall, 2001). (Red Label); m/z 177, 197, 239 and 517 (Buchanan's); m/z 200 and 347 (J&B) and m/z 251 (White Horse). Some of these ions are listed 3.3. Adulteration and counterfeiting in Table 1 (m/z 239 and 503) and Scheme 1 (m/z 197 and 359).

The capability of ESI FT-ICR MS to identify whisky counterfeiting was tested by the addition of a common low cost “distillate whisky” Table 1 commercialized in Brazil, herein referred as “Brazilian distilled whisky”. Elemental compositions assigned to ions in ESI(−) FT-ICR MS data (20 most abundant ions). This “whisky” is known to be frequently used in Brazil to adulterate and a m/z Error (ppm) RDB Elemental composition counterfeit the much more valuable Scotch whiskies and was added at

143.10784 0.64 1.5 C8H16O2 different amounts to an authentic sample of Johnnie Walker Red 161.04568 0.8 2.5 C6H10O5 Label. Fig. 4 displays the resulting ESI(−) spectra. Note that adulteration 171.13918 0.75 1.5 C10H20O2 is easily perceived by the appearance of some characteristic “marker” 179.05625 0.79 1.5 C6H12O6 199.17052 0.85 1.5 C H O ions already at 25% of adulteration, most particularly that of m/z 12 24 2 − 221.06689 0.97 2.5 C8H14O7 269.09. The PCA analysis (Fig. 5) shows that ESI( ) FT-ICR is indeed 227.20189 1.04 1.5 C14H28O2 promising as an effective technique for adulteration and counterfeiting 239.07750 0.35 10 C15H14NS screening of whisky and also to estimate the level of adulteration. The 255.23324 1.12 1.5 C H O 16 32 2 PCA scores plot clearly splitting the authentic and adulterated samples 283.26457 1.12 1.5 C18H36O2 into two distinct groups. In ESI(+), PC1, PC2 and PC3 were found to 323.09883 0.88 12 C19H18O2NS

341.10943 0.94 11 C12H22O11 account for 76.7%, 9.0% and 4.2% of the total variance, respectively. 351.20294 0.92 10 C23H30NS Therefore, the total variance explained by the three ﬁrst PCs is 89.8% 359.12001 0.94 10 C19H22O4NS at a conﬁdence level of 95%. Additionally, in the ESI(−)resultsPC1, 431.14141 −0.22 1.5 C H O S 16 32 9 2 PC2 and PC3 accounts for 81.8%, 7.8% and 5.5% of the total variance, 467.14151 0.01 4.5 C19H32O9S2

485.15215 0.17 3.5 C19H34O10S2 respectively. 503.16264 0.02 2.5 C19H36O11S2 Fig. 6(A) and (B) shows a representative pair of spectra for an au- 521.17323 0.07 1.5 C19H38O12S2 thentic and counterfeit sample of the same brand (Red Label). Note a RDB — ring/double bond equivalents. the immediate recognition of counterfeiting such as via the lack of 102 J.S. Garcia et al. / Food Research International 51 (2013) 98–106

O O

OH OH OH OH HO OH OH HO O O O O OH O 4, 200.17763 OH 1,144.11503 2, 172.14633 3, 198.05282 5, 220.05830 OH - - - [ 4 -H] m/ z 199.17053 [ 1 -H]- m/ z 143.10785 [ 2 -H] m/ z 171.13919 [ 3 -H] m/ z 197.04574 [ 5 -H]- m/ z 219.05125

OH OH O HO HO OH OH O OH OH O O O O HO OH OH HO OH OH 6,228.20893 7,256.24023 8, 270.09508 9,270.07395

- [ 6 -H] m/ z 227.20190 [ 7 -H]- m/ z 255.23326 [ 8 -H]- m/ z 269.08814 [ 9 -H]- m/ z 269.06698

OH OH O OH O OH O OH H C HO OH O 3 HO OH HO O O OH OH O O HO OH HO O O O O O O O HO O HO HO OH HO HO OH HO OH OH HO OH OH OH OH OH 10,298.10525 11 302.00627 12 323.09885 13 342.11621 14 360.12678

- - - - [ 10 -H] m/z 297.09836 [ 11 -H]- m/z 300.99938 [ 12 -H] m/z 324.10565 [ 13 -H] m/z 341.10946 [ 14 -H] m/z 359.12005

OH OH HO HO OH OH OH O O O O HO O O O HO OH OH HO OH O O O O O OH HO OH O O OH HO OH O O HO OH OH OH OH O OH HO OH OH OH 17,400.12169 18, 414.13734 15,380.13186 16,380.11073 - - [ 17 -H]- m/z 399.11515 [ 18 -H] m/z 413.13080 [ 15 -H]- m/z 379.12518 [ 16 -H] m/z 379.10404

Scheme 1. Structures proposed for the major compounds detected in the whisky samples by ESI(−) FT-ICR MS.

Table 2 the Gaussian “grass” of polar compounds around m/z 300–700 and Elemental compositions assigned to ions in ESI(+) FT-ICR data (20 most abundant ions). the presence of an abundant sugar ion of m/z 311.07 corresponding m/z Error (ppm) RDBa Elemental composition to sodiated saccharose. Counterfeit samples (n=30) from different

157.08350 0.09 0.5 C6H14O3Na whisky brands were also analyzed, and the resulting spectra, as 203.05261 0.01 0.5 C6H12O6Na expected, displayed quite contrasting and variable proﬁles of polar 213.07336 0.06 1.5 C8H14O5Na constituents (Fig. 6(C)). 219.02655 0.71 5.5 C9H8O5Na

231.08393 0.08 0.5 C8H16O6Na 271.07884 0.05 2.5 C10H16O7Na 3.4. Aging and blending 301.14106 0.1 5.5 C16H22O4Na

347.09490 0.09 2.5 C12H20O10Na The ability of ESI FT-ICR MS to evaluate whisky aging and blending 365.10548 0.12 1.5 C12H22O11Na

383.11602 0.07 0.5 C12H24O12Na was also tested (Fig. 7) using samples of whisky from Johnnie Walker 409.16215 0.03 9.5 C22H26O6Na Black, Gold and Blue Labels. Note that aging and blending is easily rec- 411.14730 0.01 0.5 C14H28O12Na ognized by changes in the polar proﬁles particularly in the relative 425.13613 0.44 14.5 C25H22O5Na “ ” 443.16766 0.05 8.5 C H O Na abundances of the heavier ions and that of the ion grass around 22 28 8 – 509.14777 0 19 C31H26O2NS2 m/z 300 500.

527.15842 0.03 11 C25H30O8NSNa

545.16894 0.07 17 C31H31O4NS2

589.22561 0.11 9.5 C28H38O12Na 4. Conclusion

623.23115 0.06 24 C41H36ONS2 689.21181 0.21 2.5 C25H46O16S2Na ESI FT-ICR MS has been shown to be a powerful technique for di- a RDB — ring/double bond equivalent. rect analysis at the molecular level of Scotch whiskies, being able to J.S. Garcia et al. / Food Research International 51 (2013) 98–106 103

OH HO

O HOH O O O OCH3 H O HO O HO H3C HO H HO OH H OH O H CO OH H OH 3 OH 1,134.094294 2, 150.06808 3, 180.06339 4,200.06847 5, 204.02974

+ + [1 +Na+] m/ z 157.08350 [2 +Na+ ] m/z 173.05774 [3 +Na ] m/z 203.05261 [4 +Na+ ] m/z 223.05768 [5 +Na] m/ z 227.0895

O O OH HO O O O HO HO HO OH O OH OH OH H OH H O O O HO O O OH HO HO O O HO HO H H OH HO H3CO OCH HO HO OH OH H O 3 OH OH OH OH OH

6,208.09468 7,242.07904 8,324.10564 9,324.10564 10,342.11621

+ + [6 +Na+ ] m/z 231.08393 [7 +Na+ ] m/z 265.06827 [8 +Na ] m/z 347.09490 [9 +Na ] m/z 347.09490 [10 +Na+ ] m/ z 365.10548

OH OH OCH3 H CO OCH OH 3 3 O HO H3C O O OH H3C O CH3 O O O O O O O HO H C O 3 O HO OH H3CO OH OCH3 13 14,402.14673 12, 386.17294 ,386.17294 11, 360.126776 + + [14 +Na ] m/ z 425.13614 [12 +Na+ ] m/ z 409.16215 [12 +Na ] m/ z 409.16215 [11 +Na+ ] m/ z 383.11602

Scheme 2. Structures proposed for the major compounds detected in the whisky samples by ESI(+)FT-ICR MS.

100 171.14 199.17

50 341.11 (A) 219.05 503.16 323.10 375.12 255.23 485.15 521.17 401.13 563.18 0 171.14 100 199.17 269.09 239.08 359.12 50 285.08 (B) 401.13 521.17 323.17 385.14 431.14 563.18 485.22 593.19 0 269.09 100

179.06 (D) 50 341.11 431.14 195.00 251.08 375.12 593.20 277.02 449.15 521.17 0 200 300 400 500 600 m/z

Fig. 4. ESI(−) FT-ICR MS of (A) authentic Johnnie Walker Red Label sample and adulterated samples with different proportions of the “Brazilian distilled whisky”, (B) 25%, (C) 80% and (D) 100% (v/v). 104 J.S. Garcia et al. / Food Research International 51 (2013) 98–106

de Nível Superior (CAPES) and Petrobras for financial support. We — (A) also thank ABRAPE Associação Brasileira de Bebidas for providing authentic whisky samples. 4 Brazilian Distilled Whisky References 3 90% Araujo, A. S., da Rocha, L. L., Tomazela, D. M., Sawaya, A. C. H. F., Almeida, R. R., 2 Catharino, R. R., et al. (2005). Electrospray ionization mass spectrometry finger- Red Label printing of beer. Analyst, 130, 884–889. 1 Aylott, R. I., Clyne, A. H., Fox, A. P., & Walker, D. A. (1994). Analytical strategies to confirm scotch whiskey authenticity. Analyst, 119, 1741–1746. 0 Adulterated Beare-Rogers, J., Dieffenbacher, A., & Holm, J. V. (2001). Lexicon of lipid nutrition. Pure

PC3 and Applied Chemistry, 73, 685–744. -1 30 Câmara, J. S., Marques, J. C., Prestrelo, R. M., Rodrigues, F., Oliveira, L., Andrade, P., et al. 25% 25 (2007). Comparative study of the whisky aroma profile based on headspace solid -2 20 phase microextraction using different fiber coatings. Journal of Chromatography A, 15 1150, 198–207. -3 10 Catharino, R. R., Cunha, I. B. S., Fogaca, A. O., Facco, E. M. P., Godoy, H. T., Daudt, C. E., 8 5 et al. (2006). Characterization of must and wine of six varieties of grapes by direct 6 0 PC1 4 infusion electrospray ionization mass spectrometry. Journal of Mass Spectrometry, 2 -5 – PC2 0 41, 185 190. -2 -10 Cooper, H. J., & Marshall, A. G. (2001). Electrospray ionization Fourier transform mass -4 spectrometric analysis of wine. Journal of Agricultural and Food Chemistry, 49, 5710–5718. Corilo, Y. E., Vaz, B. G., Simas, R. C., Nascimento, H. D. L., Klitzke, C. F., Pereira, R. C. L., et al. (2010). Petroleomics by EASI(+/−) FT-ICR MS. Analytical Chemistry, 82, Adulterated 3990–3996. (B) Brazilian Distilled de Souza, P. P., Augusti, D. V., Catharino, R. R., Siebald, H. G. L., Eberlin, M. N., & Augusti, Whisky 10 R. (2007). Differentiation of rum and Brazilian artisan cachaça via electrospray ionization mass spectrometry fingerprinting. Journal of Mass Spectrometry, 42, 1294–1299. 8 de Souza, P. P., de Oliveira, L. C. A., Catharino, R. R., Eberlin, M. N., Augusti, D. V., Siebald, H. G. L., et al. (2009). Brazilian cachaça: “Single shot” typification of fresh alembic fi 6 and industrial samples via electrospray ionization mass spectrometry ngerprint- ing. Food Chemistry, 115, 1064–1068.

de Souza, P. P., Siebald, H. G. L., Augusti, D. V., Neto, W. B., Amorim, V. M., Catharino, R. PC3 4 R., et al. (2007). Electrospray ionization mass spectrometry fingerprinting of Brazilian artisan cachaça aged in different wood casks. Journal of Agricultural and 2 Red Label Food Chemistry, 55, 2094–2102. 2 Gordon, G. E. (2003). Marketing Scotch whisky. In L. Russel, G. Stewart, & C. Bamforth – 0 0 (Eds.), Whisky: Technology, production and marketing (pp. 309 349). London: Elsevier. -2 MacKenzie, W. M., & Aylott, R. I. (2004). Analytical strategies to confirm Scotch whisky -2 – 8 -4 authenticity. Part II: Mobile brand authentication. Analyst, 129, 607 612. 6 Marshall, A. G., Hendrickson, C. L., Ernmetta, M. R., Rodgers, R. P., Blakney, G. T., & 4 2 -6 PC1 0 Nilsson, C. L. (2007). Fourier transform ion cyclotron resonance: State of the art. -2 – PC2 -4 -8 European Journal of Mass Spectrometry, 13,57 59. -6 Moller, J. K. S., Catharino, R. R., & Eberlin, M. N. (2005). Electrospray ionization mass spectrometry fingerprinting of whisky: Immediate proof of origin and authenticity. Fig. 5. PCA for the (A) ESI(−) and (B) ESI(+) FT-ICR MS data from authentic Johnnie Analyst, 130, 890–897. Walker Red Label whisky samples (n=12) and samples adulterated with different Nascimento, E. S. P., Cardoso, D. R., & Franco, D. W. (2008). Quantitative ester analysis proportions of the “Brazilian distilled whisky” (25, 50, 70, 80 and 90%). in cachaca and distilled spirits by gas chromatography–mass spectrometry (GC–MS). Journal of Agricultural and Food Chemistry, 56,5488–5493. Parker, I. G., Kelly, S. D., Sharman, M., & Howie, D. (1998). Investigation into the use of provide a characteristic profile of polar constituents both in the neg- carbon isotope ratios (C-13/C-12) of Scotch whisky congeners to establish brand ative and positive ion modes. Samples from different brands and authenticity using gas chromatography combustion–isotope ratio mass spectrom- – with different ages and blending were also shown to display contrast- etry. Food Chemistry, 63, 423 428. Poisson, L., & Schieberle, P. (2008). Characterization of the most odor-active com- ing profiles. Adulteration and counterfeiting were also promptly pounds in American bourbon whisky by application of the aroma extract dilution identified. This strategy is also probably applicable to other valuable analysis. Journal of Agricultural and Food Chemistry, 56, 5813–5819. spirits and beverages. Rhodes, C. N., Heaton, K., Goodall, I., & Brereton, P. A. (2009). Gas chromatography carbon isotope ratio mass spectrometry applied to the detection of neutral alcohol in Scotch whisky: An internal reference approach. Food Chemistry, 114, Acknowledgments 697–701. Rodgers, R. P., Schaub, T. M., & Marshall, A. G. (2005). Petroleomics: MS returns to its roots. Analytical Chemistry, 77, 20A–27A. The authors thank the State of São Paulo Research Foundation (FAPESP), Brazilian National Council for Scientific and Technological Development (CNPq), Coordenação de Aperfeiçoamento de Pessoal J.S. Garcia et al. / Food Research International 51 (2013) 98–106 105 (A)

(B)

(C)

Fig. 6. (A) Authentic spectrum, (B) counterfeit spectrum and (C) PCA plots for the ESI(−) FT-ICR data of Johnnie Walker Red Label whisky samples (authentic and counterfeit). 106 J.S. Garcia et al. / Food Research International 51 (2013) 98–106

Fig. 7. ESI(−) FT-ICR for samples of Johnnie Walker (A) Black, (B) Gold and (C) Blue Label whisky samples.