Article pubs.acs.org/JAFC
Quantitation of Polycyclic Aromatic Hydrocarbons (PAH4) in Cocoa and Chocolate Samples by an HPLC-FD Method Marion Raters* and Reinhard Matissek Food Chemistry Institute (LCI) of the Association of the German Confectionery Industry, Adamsstraße 52-54, D-51063 Köln, Germany ABSTRACT: As a consequence of the PAH4 (sum of four different polycyclic aromatic hydrocarbons, named benzo[a]anthracene, chrysene, benzo[b]fluoranthene, and benzo[a]pyrene) maximum levels permitted in cocoa beans and derived products as of 2013, an high-performance liquid chromatography with fluorescence detection method (HPLC-FD) was developed and adapted to the complex cocoa butter matrix to enable a simultaneous determination of PAH4. The resulting analysis method was subsequently successfully validated. This method meets the requirements of Regulation (EU) No. 836/2011 regarding analysis methods criteria for determining PAH4 and is hence most suitable for monitoring the observance of the maximum levels applicable under Regulation (EU) No. 835/2011. Within the scope of this work, a total of 218 samples of raw cocoa, cocoa masses, and cocoa butter from several sample years (1999−2012), of various origins and treatments, as well as cocoa and chocolate products were analyzed for the occurrence of PAH4. In summary, it is noted that the current PAH contamination level of cocoa products can be deemed very slight overall. KEYWORDS: PAH4, cocoa, chocolate, HPLC-FD, benzo[a]anthracene (BaA), chrysene (Chr), benzo[b]fluoranthene (BbF), and benzo[a]pyrene (BaP)
■
certain foodstuffs in 2011 in Regulation (EU) No. 835/2011.6 Furthermore, certain performance criteria were set for analytically determining these PAH.8 Cocoa butter has long been considered a foodstuff strongly contaminated with PAH. This is explained by the assumption that, in particular, inappropriate processes were used to dry the cocoa beans and the fact that cocoa butter, unlike other vegetable fats and oils, is not refined.9 Generally speaking, it can be assumed that imported cocoa beans, if at all, are already contaminated with PAH prior to being industrially processed. A study conducted by Lowor et al. in Ghana in 2012, in which specific drying and storage procedures and their impact on PAH content levels in the resulting cocoa were tested, showed that it was especially the postharvest process, particularly the direct drying procedures that used smoke or combustion gases, which contributed to the contamination of cocoa and products derived from its subsequent processing.10 Moreover, this study also showed that the cocoa shell, on which the main proportion of detected PAH was concentrated, had a certain barrier effect. The available literature provides only a few, hardly any, systematic analyses of PAH, particularly BaP, in cocoa and cocoa products. In 2007, for example, Stiftung Warentest (Germany’s consumer watchdog) reported a content level of 10 μg/kg in a dark chocolate.11 Ziegenhals and Speer analyzed a total of 40 chocolate samples from the German market with various cocoa content levels and detected BaP levels ranging from 0.07 to 0.63 μg/kg of chocolate.12 In 2011, analysis of a selection of 25 chocolate sweet samples on the Indian market
INTRODUCTION Polycyclic aromatic hydrocarbons (PAH) is the collective term for organic compounds consisting of condensed aromatic rings. Around 250 different compounds of this kind have been identified to date. A differentiation is made between “light” PAH, consisting of three or four rings, and “heavy” PAH, possessing at least five condensed rings. The most prominent member of the PAH family is benzo[a]pyrene (BaP), possessing five aromatic rings, which makes it one of the heavy PAH.1 PAH are formed through pyrolysis during the incomplete combustion of practically all organic substances and, hence, for example, can be generated by large combustion plants, forest fires, and even barbecues or the smoked curing of foods. They usually occur at temperatures above 450 °C through pyrolytic processes (thermal decomposition of organic substances under the exclusion of air) on incomplete combustion of organic material such as coal, fuel, and tobacco or when barbecuing.1,2 As contaminants, PAH are found ubiquitously. Many PAH are carcinogenic and genotoxic. Heavy PAH are generally seen as toxic, but even light PAH have proven to include carcinogenic substances.3,4 In the past it was generally assumed that the ratios of toxicologically relevant PAH in relation to each other were relatively constant both in feedstuffs and in vegetable foodstuffs and hence that BaP was a suitable guide substance for indicating the entire group of PAH.5 In 2008, however, after evaluating extensive scientific data from the member states on the occurrence of carcinogenic PAH in foodstuffs, the EFSA CONTAM panel found that a summary system of the four PAH (PAH4) benzo[a]anthracene (BaA), chrysene (Chr), benzo[b]fluoranthene (BbF), andBaP was best suited as a PAH marker in foodstuffs.6,7 On the basis of these findings, maximum limits for BaP and PAH4 were fixed for © XXXX American Chemical Society
Received: June 24, 2014 Revised: October 6, 2014 Accepted: October 11, 2014
A
dx.doi.org/10.1021/jf5028729 | J. Agric. Food Chem. XXXX, XXX, XXX−XXX
Journal of Agricultural and Food Chemistry
Article
showed an average BaP content level of 1.62 μg/kg of chocolate sweet.13 Representative analyses on the occurrence of PAH4 in cocoa, derived products, and chocolate products on the German market were not available at the start of this study. The pertinent literature contains a plethora of analysis methods for detecting the various relevant PAH individual substances. These methods are based on gas chromatography measurement procedures21−26 and detection procedures using UV, fluorescence,14−19 and MS.20 GC-MS is the method most frequently used to detect PAH.27 HPLC coupled with mass spectrometry (LC-MS) is only seldom used due to the nonionic and unpolar property of PAH.20,28 Currently, the European cocoa and chocolate industry is aware of the recent discussion at the European level about the possibility of a review into how to express PAH levels detected in cocoa beans and derived products. The ongoing discussion was triggered by the results of a recent proficiency test carried out by the European Reference Laboratory (EU-RL).29 This showed that the results obtained in laboratories are less favorable if the analyses are conducted on a fat basis rather than on a product basis. It seems that converting the analytical results obtained on a product basis into results on a fat basis leads to variations in the final results obtained. For this reason it is considered that the best approach to addressing this issue is to use a standardized “fat extraction method” for cocoa beans and derived products. The current study consistently sidestepped this issue by using only cocoa butter obtained from products by way of cold extraction for PAH analysis. This did away with the need to use an additional fat determination method. The objective of this research work was to develop a suitable internal standard (ISTD) method of routine analysis for quantitating the relevant PAH in cocoa and cocoa products and to give a situation assessment of the occurrence of these substances in the tested products. Due to the entry into force on April 1, 2013, of Regulation (EU) No. 835/2011, amending the EU Contaminants Regulation with regard to “cocoa beans and derived products” and stipulating a maximum level of 5.0 μg/kg fat for BaP and 35 μg/kg fat for PAH4 (30 μg/kg fat for PAH4 as of 01.04.2015), the focus of analytical methods was especially placed on BaP and PAH4 levels in cocoa, derived products, and chocolate products. In addition, a combined pool of 218 samples was analyzed, composed of samples of cocoa butter, cocoa masses, and raw cocoa plus monitoring samples of finished cocoa and chocolate products from the German market. Specific examination of cocoa butter samples from various growing and harvest years was also aimed at tracing the development and pattern of PAH4 content levels from 1999 to 2012.
■
description: seed leaves [cotyledons] without seed shell) are the roasted, unshelled, mostly already broken up cocoa kernels. Cocoa mass is made up of finely ground cocoa nibs; cocoa powder is the finely ground cocoa press cake, which, in addition to cocoa butter, is what remains after pressing out the cocoa mass.30 Cocoa nibs, cocoa masses, cocoa powder, and cocoa butter are collectively referred to as semifinished cocoa products. The term cocoa products and chocolate products covers products defined under Annex 1 of the German Cocoa Regulation, such as whole milk chocolate, dark chocolate, and chocolate coating.31 The term cocoa and cocoa products used in this research work is a collective term and implies all products defined in this section. Chemicals. The solvents used were petroleum benzene, n-hexane, and ethanol for analysis, as well as dichloromethane and acetonitrile, both of HPLC grade (>99.9%). Other reagents and materials used were potassium hydroxide and Strata EPH (200 μ, 70 Å) 5 g/20 mL (part 8B-SO31_LEG), Giga tube. All of the solvents, reagents, and materials mentioned are commercially available from Phenomenex (Aschaffenburg, Germany), Sigma-Aldrich (Steinheim, Germany), and Merck (Darmstadt, Germany). The following PAH standards were used for the identification and quantitation of the PAH4 present in the samples: benzo[a]anthracene, chrysene, benzo[b]fluoranthene, benzo[a]pyrene, and benzo[b]chrysene, which was used as an internal standard. All standards were obtained from Dr. Ehrenstorfer (Augsburg, Germany) and LCG (Wesel, Germany). Methodology. To extract the cocoa butter, the samples are first defatted; depending on the respective matrix, this means weighing 10− 15 g (raw cocoa, cocoa nibs, cocoa mass, and chocolate) or 50 g (cocoa powder) into a 250 mL centrifuge tube. Seventy milliliters of petroleum benzene and a magnetic stir bar are added to the samples. This mixture is then stirred for 1 h and centrifuged. The supernatant is transferred via pipet into a 500 mL round flask. This process is repeated three times, and the combined petroleum benzene residues are removed via rotary evaporator. The isolated fat, which is dried overnight in a desiccator cabinet (at 50 °C), can subsequently be used for further PAH analysis. Cocoa butter samples are melted at approximately 50 °C and can be immediately used for PAH analysis without additional preparation. For saponification, 2 g of cocoa butter was weighed, spiked with 5 ng of benzo[b]chrysene as internal standard, and incubated for 30 min. After the addition of 8 mL of ethanol and 2 mL of potassium hydroxide solution (10%), the solution was heated at 60 °C for 30 min in an ultrasonic bath. After saponification, 5 mL of demineralized water was added and, after cooling, 10 mL of hexane. After centrifugation (20 °C, 2000 rpm), the upper hexane phase was put on a Strata EPH Giga Tube (conditioned with 30 mL of hexane). This extraction procedure was repeated twice with 8 and 7 mL of hexane. PAH were eluted with 20 mL of dichloromethane, and the solvent was removed using a rotary evaporator and further evaporated to just dryness under a nitrogen stream. The residue was redissolved in 2 mL of acetonitrile and finally transferred into an autosampler vial for HPLC-FD-analysis. For the identification and quantitation of PAH4, extracts were analyzed using HPLC coupled to a fluorescence detector (FD). The HPLC system consists of a gradient pump, a column oven, an autosampler, and a fluorescence detector (Dionex Ultimate 3000, Idstein, Germany). A Multosphere PAH-III column (250 × 4.6 mm i.d., CS-Chromatographie Service, Langerwehe, Germany) was used along with a precolumn filled with the same material (20 × 4 mm). Separation was performed using the following gradient: acetonitrile/ water (85:15, v/v) for 13.5 min, followed by a 2 min ramp to 100% acetonitrile, this solvent being further maintained for 15 min. The total flow rate was 1 mL/min. The analytical column was placed in an oven, and the following temperature program was performed: 25 °C for 4 min, followed by direct heating to 45 °C. Detection was performed at selected excitation and emission wavelengths, programmed as follows: 0−8.3 min, 290/470 nm; 8.3−11 min, 270/390 nm; 11−11.5 min, 290/470 nm.
MATERIALS AND METHODS
Samples. In the period from 2010 to 2012, a total of 218 samples were analyzed for PAH4: of these, 12 were raw cocoa beans, 22 were cocoa masses, 67 were cocoa butter samples, and 117 were cocoa and chocolate products. The cocoa beans and semifinished cocoa products (cocoa masses and butter) came from the main growing regions in Africa, Asia, and America, in the harvest period from 1999 to 2012. The cocoa products and chocolate products were purchased quarterly in retail shops in the period from 2009 to 2012. Definitions of the cocoa and chocolate products we analyzed in our research work are as follows: Cocoa beans (including raw cocoa) are the unshelled, fermented, and dried seeds of Theobroma cacao L. They have a thin shell (cocoa shell, botanical description: seed shell consisting of testa and rest of pulp). Cocoa nibs (botanical B
dx.doi.org/10.1021/jf5028729 | J. Agric. Food Chem. XXXX, XXX, XXX−XXX
Journal of Agricultural and Food Chemistry
Article
Table 1. Validation Parameters Obtained for the Four Individual PAH parameter
BaA
Chr
BbF
BaP
calibration range (μg/kg) recovery rate (%) limit of detection (LOD, μg/kg) limit of quantitation (LOQ, μg/kg) intraday accuracy (N = 12, %) interday precision (N = 12, %) HORRATr
0.02−15.57 (r2 > 0.999) 81−101 0.01 0.02 10 11 0.66
0.02−9.45 (r2 > 0.999) 66−91 0.01 0.02 13 13 0.88
0.13−36.32 (r2 > 0.999) 94−103 0.04 0.13 7 9 0.46
0.17−67.33 (r2 > 0.999) 90−100 0.07 0.17 6 7 0.42
Table 2. Overview of Raw Cocoa and Semifinished Cocoa Products (Cocoa Butter and Cocoa Masses) Examined within the Scope of This Studya treatment
(μg/kg)
all (N = 67)
deodorized (N = 29)
not deodorized (N = 8)
unknown (N = 30)
all (N = 22)
alkalized (N = 3)
unknown (N = 19)
(N = 12)
BaA
Chr
mean median range mean median range mean median range mean median range
4.46 3.40 0.72−32.77 3.07 2.27 0.76−13.42 5.70 5.93 1.74−9.99 5.47 4.16 0.72−32.77
mean median range mean median range mean median range
1.32 1.04 0.71−3.37 0.88 0.80 0.75−1.08 1.39 1.13 0.71−3.37
mean median range
2.18 0.86 0.27−9.55
Cocoa Butter 6.05 4.18 1.12−45.03 3.81 3.27 1.32−11.23 6.57 5.65 2.26−14.16 8.07 5.51 1.12−45.03 Cocoa Mass 2.59 2.48 1.18−5.56 1.50 1.51 1.39−1.60 2.76 2.50 1.18−5.56 Raw Cocoa 3.49 2.03 0.47−15.89
BbF
BaP
PAH4
0.98 0.70 ≤LODb−6.86 0.84 0.55 ≤LODb−2.32 1.08 0.64 0.32−2.28 1.24 1.12 ≤LODb−6.86
0.96 0.66 ≤LODb−7.86 0.86 0.48 0.14−3.82 0.94 0.63 0.33−2.03 1.09 0.92 ≤LODb−7.86
12.44 9.94 2.09−92.53 8.55 6.32 2.67−24.21 14.29 13.56 4.82−25.73 15.72 12.20 2.07−92.53
0.33 0.29 ≤LODb−1.06 0.26 0.26 ≤LODb−0.29 0.36 0.31 0.17−1.06
0.32 0.26 ≤LODb−0.76 0.28 0.34 0.12−0.38 0.35 0.26 ≤LODb−0.84
4.57 3.80 2.59−9.29 2.83 2.61 2.56−3.32 4.84 3.83 2.59−9.29
1.27 0.41 ≤LODb−8.86
1.28 0.24 ≤LODb−9.98
8.21 2.91 0.88−44.28
a All content levels are on a fat basis. bCalculation of mean and median was performed using 50% LOD (LOD: BaA = 0.01 μg/kg fat, chrysene = 0.01 μg/kg fat, BbF = 0.04 μg/kg fat, BaP = 0.07 μg/kg fat).
■
RESULTS AND DISCUSSION
Method Validation. The presented method was validated
on the basis of the guidelines provided by Kromidas.32 For this Table 3. PAH4 Content Levels in Cocoa Butter Samples from Various Sample Yearsa
a
sample year:
1999
2010
N:
28
26
2011/2012 13
PAH4 (μg/kg fat)
maximum mean minimum
92.53 17.12 2.09
24.42 8.81 2.67
21.94 9.63 3.28
BaP (μg/kg fat)
maximum mean minimum
7.86 1.00 ≤LODb
1.61 0.78 0.14
3.82 1.27 0.30
Figure 1. Time line of PAH4 content levels in cocoa butter samples for 1999−2012 (for data see Table 3).
purpose a cocoa butter sample without detectable PAH4 content was spiked with a PAH4 solution, prepared in keeping
All content levels are on a fat basis. bLOD: BaP = 0.07 μg/kg fat. C
dx.doi.org/10.1021/jf5028729 | J. Agric. Food Chem. XXXX, XXX, XXX−XXX
Journal of Agricultural and Food Chemistry
Article
Table 4. PAH4 Content Levels in Cocoa and Chocolate Productsa
a
milk chocolate/chocolate coatings
dark chocolate/chocolate coatings
cocoa powder
cocoa drink powder
N:
27
69
12
9
total 117
PAH4 (μg/kg fat)
maximum mean median 90th percentile
17.08 10.11 8.76 11.10
17.42 5.88 5.25 10.34
21.56 8.86 8.49 13.96
19.98 9.48 8.90 15.08
21.56 6.66 6.22 11.20
BaP (μg/kg fat)
maximum mean median 90th percentile
1.96 0.70 0.65 1.41
1.56 0.57 0.44 1.02
2.44 1.07 0.92 2.07
2.92 1.00 0.81 1.81
2.92 0.69 0.61 1.35
All content levels are on a fat basis.
this observation is the calcination of volatile compounds within the thermal treatment of the deodoration process.10 The maximum PAH4 content of 92.53 μg/kg fat was determined in a cocoa butter sample of unknown origin and treatment. The cocoa mass samples examined within the scope of this study (N = 22) contained a mean of 4.57 μg PAH4/kg fat and 0.32 μg BaP/kg fat (median = 3.80 μg/kg PAH4/kg fat and 0.26 μg BaP/kg fat) within a range of 2.59−9.29 μg PAH4/ kg fat and
