Bioimaging of the elemental distribution in cocoa beans by means of LA-ICP-TQMS

Georgina M. Thyssena, Claudia Keilb, Melanie Wolffb, Michael Sperlingac, Daniel Kadowd, Hajo Haaseb and Uwe Karst*a
aUniversity of Münster, Institute of Inorganic and Analytical Chemistry, Corrensstraße 30, 48149 Münster, Germany. E-mail: UK@uni-muenster.de
bTechnische Universität Berlin, Department of Food Chemistry and Toxicology, Gustav-Meyer-Allee 25, 13355 Berlin, Germany
cEuropean Virtual Institute for Speciation Analysis (EVISA), Mendelstraße 11, 48149 Münster, Germany
dAugust Storck KG, R&D Chocolates, Waldstraße 27, 13403 Berlin, Germany

Received 25th October 2017 , Accepted 18th December 2017

First published on 18th December 2017


Cocoa beans (Theobroma cacao L.) are a key raw material for the confectionary industry. Therefore, their ingredients, nutrients as well as contaminants, have to be thoroughly analysed. In this work, a method for spatially resolved analysis of the tissue specific nutrient and contaminant distribution in cocoa beans by laser ablation-inductively coupled plasma-triple quadrupole mass spectrometry (LA-ICP-TQMS) is presented. This method provides high spatial resolution for fine structures and a wide linear range to detect macronutrients and lower concentrated toxic elements – such as Cd – in a single experiment. By using a triple quadrupole analyser, elements that are subject to isobaric or polyatomic interferences in ICP-MS analysis, such as Ca or Fe, can also be determined accurately. The distribution of Na, Mg, P, K, Ca, Fe, Cu, Zn, Cd and Pb in cocoa beans was successfully visualized for the first time. The method represents a new approach for studying contaminants in cocoa beans and ultimately the development of innovative strategies to minimize contamination in the product.


Introduction

Cocoa beans are obtained from the tropical tree Theobroma cacao L. They are an essential export product for many African, South and Central American and Asian countries, and fundamental for the confectionary industry. Currently, a total of 4.2 million tons of cocoa beans is produced worldwide.1 With 37% of the grindings, corresponding to 1.5 million tons of cocoa, Europe is the largest processor.2 In 2012, the European chocolate manufacturers including Switzerland and Norway produced approximately 3.4 million tons of end-consumer chocolate. The consumption per year and person is as high as 5.2 kg in the United States and 5.9 kg in Europe.2 Because of this high consumption, cocoa bean contamination with toxic compounds and elements is a matter of concern. One of the major issues is contamination with cadmium (Cd). Cadmium occurs naturally in soils of volcanic origin. It is taken up by plants including Theobroma cacao, even though it is not a nutrient, and is partially translocated into the seed tissue. In addition to naturally occurring cadmium, a contamination of the soil through flood water and industrial and agricultural activities may take place.3 Cd has severe adverse effects on human health, also when being consumed from food as the main source of exposure for the majority of people.4 Even though the probability of unknowingly consuming amounts of Cd that are acutely toxic is low, the long term effects of lower Cd doses, especially for vegetarians, adolescents and children,5 are alarming: it can cause renal and bone damage, even leading to renal failure after long term exposure.4–6 Additionally, it is classified as group 1 human carcinogen by the International Agency for Research on Cancer (IARC).7 Recent meta-analyses also provide evidence for positive association between dietary cadmium intake and the occurrence of hormone-associated cancers in Western populations.8 Thus, the European Union has enacted new regulations regarding the maximum Cd content in foodstuffs, including cocoa products, in 2014 (applicable in 2019). According to the new regulations, the maximum allowed Cd concentration is related to the total dry cocoa solids amount in the product, with a maximum concentration of 800 μg kg−1 Cd in chocolate with ≥50% total dry cocoa solids and 100 μg kg−1 Cd in milk chocolate with <30% dry cocoa solids.9 This may have a major impact on the way cocoa is produced and controlled.

Due to their essential role in the confectionary industry, cocoa beans and chocolate products have been subject of many studies. These include monitoring of cocoa bean ingredients, the fate of potentially healthy or toxic compounds during processing, quality assurance and origin aspects. For example, Diomande et al.10 examined the possibility to differentiate between varieties as well as between origins using isotope ratios of C and N in cocoa bean samples. Bertoldi et al.11 compared cocoa beans originating from 23 different countries with regard to their multi-element profiles, whilst Jentzsch et al.12 reported on the differentiation of cocoa varieties based on their Raman spectra. This study includes mapping of lipids and lignin. Oligopeptide patterns have also been shown to provide information regarding origin. In addition, these patterns permit evaluating the degree of fermentation (Caligiani et al.13). Ekpa et al.14 compared the mineral profiles of cocoa beans from different plantations and recently, Da Veiga Moreira et al.15 grouped cocoa in different categories, based on the volatile profile and the sensory acceptance of the final chocolate product. The biochemical processes occurring during cocoa bean processing, namely the fermentation, were closely examined by Kumari et al.16 The authors focused on the storage protein vicilin. Its degradation by the seeds' intrinsic proteases yields important flavour precursors. De Taeye et al.17 studied the fate of anthocyanins through cocoa fermentation. Batista et al.18 used Fourier transform infrared spectroscopy (FTIR) to measure antioxidants in cocoa beans and cocoa-derived products. Other publications deal with harmful ingredients such as ochratoxin A19 and their formation during processing.20 In terms of toxic heavy metals, the Cd and Pb content of cocoa beans and related products, their bioaccessibility and relations to soil Cd content have been closely examined by Mounicou et al.,21 Yanus et al.,22 Chavez et al.23,24 and others.

Several plant seeds have been subject of elemental bioimaging studies. Wu et al.,25 e.g., used a very elegant laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS) imaging approach to visualize the distribution of various elements in cross sections of wheat grain thin sections. Van Malderen et al.26 used LA-ICP-MS as well as micro X-ray fluorescence (μXRF) imaging to unravel the distribution of different nutritional and toxic elements in wheat and rye grains using powder pellets for quantification. In a semi-imaging approach, Cakmak et al.27 examined the Zn distribution in wheat cross sections showing different line cross scans. Sunflower seeds were investigated by Pessôa et al. to determine the distribution of Cd, Cu, Fe and Mn with quantification via matrix matched pellets.28 Promchan et al.29 studied the distribution of macro- and micro nutrients in rice grain longitudinal sections using LA-ICP-MS. The quantitative distribution of toxic and essential elements in rice grains has been examined by Basnet et al.,30 and Choi et al.31 used fs LA-ICP-MS to observe the As distribution in rice. Other studies used synchrotron-based X-ray fluorescence spectroscopy (SXRF),32 proton induced X-ray emission (PIXE)33–35 or secondary ion mass spectrometry (SIMS).36–38 All these studies offer fundamental insight into plant nutrition and defence mechanisms against several toxic elements, but are also certainly of high interest for human nutrition regarding grains and other edible seeds.

For cocoa beans, no elemental imaging approach revealing the element distribution within the cocoa seeds has been published up to now. There are several studies looking at the overall concentration in seed and fruit tissues, e.g., the seed shell, the cotyledons and the pod husk.10,23 In contrast to these approaches, an imaging technique will be much more meaningful when questioning local storages of toxic- or nutrition elements within the seed tissue in detail. Furthermore, biovisualization could also be a valuable tool to depict fluctuations of elemental distribution within cacao seeds along the processing chain, allowing deeper insight into the influence of processing techniques on the essential and toxic elements and their general distribution, similar to the aforementioned imaging studies.

Various tasks need to be addressed when developing an elemental imaging method. First, the sample surface needs to be as smooth as possible to avoid intensity differences caused solely by topographic differences rather than variations in elemental concentration. Second, for a multielement analysis within the same sample set, a wide linear range is required, as some elements (e.g., K or P) can be expected to occur in comparably high concentrations, while others (e.g., Cd or Pb) are present in lower concentrations. Third, because of the sample size, the analysis needs to be fast. Finally, the samples have to be analysed interference-free, which might be a challenge for conventional low-resolution LA-ICP-MS, where ions of isotopes such as 24Mg+, 31P+, 40Ca+ or 39K+ suffer from strong interferences (e.g., by 23Na1H+, 15N16O+ 40Ar+ or 38Ar1H+).39,40

Here, we present a LA-ICP-MS approach with triple quadrupole mass analyser for cocoa bean elemental bioimaging. ICP-TQMS as the detection system provides a good sensitivity and linear range and is supporting reasonable short analysis time. Due to the enhanced chemical resolution by the TQMS (triple quadrupole mass spectrometry) approach combined with the use of dedicated collision/reaction cell gases, interference suppression should be suitable even for the more elusive elements. The fast scanning possibilities of the analyser additionally allow for multi-element detection and visualization in cocoa beans.

Experimental section

Instrumentation

A 213 nm laser ablation system (LSX-213, Teledyne CETAC Technologies, Omaha, NE, USA) with frequency quintupled Nd:YAG laser head was coupled to an ICP-MS with triple quadrupole mass analyser (iCAP TQ, Thermo Fisher Scientific, Bremen, Germany). For the laser ablation system, a home-built ablation cell with low cell volume was used for wash-out times well below 500 ms, and DigiLaz software G2 version 4.1.2 (Teledyne CETAC Technologies) was used to control the system. The sample was ablated line-wise with an energy of 3.3 J cm−2, a laser shot frequency of 20 Hz, a 100 μm laser spot and 200 μm s−1 ablation speed. A carrier gas flow of 0.8 L min−1 He was guided through the cell, with 1 L min−1 Ar (tuned on a daily basis) added after the cell via a T-piece for plasma stabilization. Coupling between LA and ICP-TQMS was realized with a spherical glass joint for direct introduction of the sample aerosol into the plasma torch injector. The ICP-TQMS was operated with a cool gas flow of 14 L min−1 and an auxiliary gas flow of 0.8 L min−1. A quartz injector tube with an inner diameter of 3.5 mm, a nickel sampler and a nickel skimmer were used. The plasma power was set to 1550 W. Parameters with dependence on the introduction system were tuned on a daily basis using NIST-612 glass standard and an automated laser ablation tune. The ICP-TQMS was operated using Qtegra 2.8 (Thermo Fisher Scientific). For an ideal interference removal, the ICP-TQMS was operated in the so-called S-TQ-O2 mode, which means a high sensitivity interface (with 2.8 mm skimmer insert) and the flatapole between the quadrupole pressurized with oxygen reaction gas (0.3 mL min−1). The dwell time was set to 36 ms per isotope for 14 mass-to-charge ratios: 23Na|23Na, 24Mg|24Mg, 31P|31P16O, 39K|39K, 44Ca|44Ca16O, 56Fe|56Fe16O, 65Cu|65Cu, 66Zn|66Zn, 110Cd|110Cd, 111Cd|111Cd, 112Cd|112Cd, 113Cd|113Cd, 114Cd|114Cd and 208Pb|208Pb. Here, the first isotope (e.g., 56Fe in 56Fe|56Fe16O) represents the set mass of the first quadrupole, while the second isotope combination (e.g. 56Fe16O in 56Fe|56Fe16O) represents the set mass for the third quadrupole. Elements which tend to react with oxygen in the collision/reaction cell were detected as oxides. This mass shift reaction supports the elimination of interferences on elements like Fe, P or Ca. Elemental distribution maps were created by export of the raw data as intensity matrices and image creation using ImageJ 1.47v software (National Institutes of Health, Bethesda, MD, USA). Here, intensities were matched to a sample coordinate according to their acquisition time, the ablation speed and laser spot size. Cut-off and contrast of the images were chosen to represent the structures of the samples in the most detailed way possible.

Materials

Seed treatment. Commercial cocoa (Theobroma cacao L.) beans were fermented for 6 days and dried at their place of origin (Santander, Colombia) before transport to Germany.
Chemicals and Consumables. Hydroxyethyl cellulose was purchased from Sigma-Aldrich Chemie GmbH (Steinheim, Germany). Aluminium oxide abrasive paper with P400 grain size was obtained from C. F. Schröder Schleifmittelwerke GmbH (Hann. Münden, Germany). Water was purified by an Aquatron Water Still purification system model A4000D (Barloworld Scientific, Nemours Cedex, France). All chemicals were used in the highest quality available.

Results and discussion

Sample preparation

For a reliable analysis of the elemental distribution, a planar surface of the sample is an important factor to avoid artefacts based on issues with focusing or sample thickness. The interior of the cocoa beans (schematic representation in Fig. 1) proved to be of varying hardness and consistency, leading to severe tissue damage after cryosectioning. The approach to cut the beans with a scalpel blade proved not to be suitable as well, because the tissues broke. During the embedding process in synthetic resin, necessary for microsectioning, several soluble nutritional elements were washed out (data not shown). Therefore, to obtain a planar surface, a polishing process was developed. For polishing, abrasive paper with P400 grain size was used. The technique was fine-tuned with regard to the longitudinal orientation of the beans, the polishing pressure and the smoothing direction. Pressure was reduced to a minimum and smoothing was only performed in one direction, as this resulted in increased sample stability. The technique leads to a satisfactorily plane surface of the samples with intact embryo and seed shell. Abrasive paper was used only once to avoid cross contamination.
image file: c7ja00354d-f1.tif
Fig. 1 Schematic representation of the longitudinal section through a cocoa bean with indication of the most important structures. Embryo and seed shell together represent the seed.

To fit the samples into the ablation cell, polishing was performed very carefully from the upper and lower side of the bean to reduce sample thickness. The polished samples were glued onto glass slides with hydroxyethyl cellulose (150 mg mL−1 90 kDa, 5 mg mL−1 1300 kDa) the preferred sticking material to avoid Na or Ca contamination of the area surrounding the sample.

LA-ICP-TQMS analysis

The elemental distribution maps of fermented, dried cocoa bean samples are shown in Fig. 2 (Cd isotopes) and Fig. 3 (other elements). The bean samples were imaged by two subsequent ablation sequences. During the first sequence, the whole area of polished beans was imaged. Before starting the second ablation run, cotyledon tissue was partially removed at the top (marked with * in Fig. 2 and 3), leaving the inner part of the shell partially uncovered for a more detailed detection of the elements in this area. For comparison, this distinct part of the sample is displayed within the figures twice, once before removal of the inner piece (small pictures above), once afterwards (whole bean pictures below) (see Fig. 2 and 3).
image file: c7ja00354d-f2.tif
Fig. 2 Polished cocoa bean: photomicroscopic image (upper left) and element distribution maps of 110Cd, 111Cd, 112Cd, 113Cd, and 114Cd acquired via LA-ICP-MS with oxygen as cell gas. Smaller images above each distribution map show the upper part of the sample before removing a part (marked with an *) of the inner bean. The hypocotyl and radicula part of the sample have been marked with a #. The red square within the photomicroscopic image indicates the area used for the Cd isotope distribution comparison shown in Fig. 5.

image file: c7ja00354d-f3.tif
Fig. 3 Polished cocoa bean: photomicroscopic image (upper left) and element distribution maps of 23Na, 24Mg, 31P16O, 39K, 44Ca16O, 56Fe16O, 65Cu, 66Zn, and 208Pb, acquired via LA-ICP-MS with oxygen as cell gas. Smaller images above each distribution map show the upper part of the sample before removing a part (marked with an *) of the inner bean. The hypocotyl and radicula part of the sample have been marked with a #.

From the element distribution maps, it can be concluded that most of the elements, including Na, Mg, P, K, Cu, Zn, Pb and, to a lesser extent, also Cd are accumulated in the seed shell. This is especially the case for Mg and K, where, in general, higher intensities can be observed in the seed shell all around the embryo. For Cd and Pb, the observed elemental distribution is highly compatible with earlier studies, which examined fractionated, acid digested cocoa tissue samples by means of flame atomic absorption spectroscopy and ICP-MS to capture metal distribution in cocoa seeds.21,41,42 For Na, P, Cu, Zn, Pb and Cd, signal intensities are obviously higher within the right periphery of the sample. Partial removal of the cotyledon tissue (marked with * in Fig. 2 and 3) leads to a clear increase of Mg, P, K, Cu, Zn and Cd signal intensity inside the inner shell parts. Within the same area of the shell, however, Na and Pb intensities are not elevated, even though these two elements are clearly detectable within the shell in other parts of the cocoa bean dissection. This leads to the conclusion that Na and Pb most likely reside in more peripheral parts of the multilayer organised cocoa shell.27 At the down end side of the bean (marked with # in Fig. 2 and 3), containing preferentially meristematic tissues, mainly Mg, P, K and Cu can be detected, but to a certain amount also Ca in the radicula and the upper part of the hypocotyl. Na, Fe and Pb hardly appear in the meristematic region, while traces of Zn and Cd can be observed in this area as well. Within the cotyledons tissue, small vessel-like Fe and Ca-enriched structures are noticeable, most likely due to local element concentration at the sides of the cotyledon interfaces. In addition, the elemental maps also hint to certain substructures inside the cotyledon interior, notably higher in Cu and partly also in K content. A rather global comparison of the elemental distribution within the two cotyledons depicts very similar patterns for Mg, P, K, Zn and Cd, whereas Na, Ca, Fe, Cu and Pb differ remarkably from the other elements. Nevertheless, Cd, at least partially, overlaps with Cu within cotyledon subregions, confirming earlier bioimaging data from sun flower seeds.28

When overlaying the elemental distribution maps of the two toxic metals 114Cd and 208Pb, (Fig. 4), it turns out that the distributions of Pb and Cd are predominantly counter correlating. Only some areas of the outer cell layers of the seed shell (marked “1” in Fig. 4) share higher intensities for both elements, whereas close to the hypocotyl and radicula, the highest Cd intensities are clearly located on the inner side of the shell. However, the highest Pb intensities were observed in the outer shell. This fits to the observations from Fig. 2 and 3, where the inner shell reveals elevated Cd signal intensity upon partial cotyledon removal, whereas Pb intensities are decreased. Once more, these results emphasize cocoa shell multilayer anatomy, but also hint to certain differences in elemental translocation into the cocoa bean tissues either during intrapod development or over the period of fermentation or drying.


image file: c7ja00354d-f4.tif
Fig. 4 Overlay of the element distribution maps of 208Pb (yellow) and 114Cd (blue) in the cocoa bean. Total signal intensities obtained by LA-ICP-MS were adapted to allow visualisation of both elements.

Due to ongoing global discussions on the regulation of the cadmium threshold in cocoa beans and products, the elemental analyses were further extended on Cd detection. All of the different Cd isotopes measured share a very similar distribution pattern (Fig. 2), so that – also considering the measurement in triple quad mode with oxygen cell gas – any form of spectral interferences (e.g. fragments of organic molecules) can be excluded, as this would have altered certain isotope distribution maps compared to others. A comparison of the Cd isotopic pattern over 80 pixels within the middle inner part of the cocoa bean with the theoretical pattern of natural Cd isotope occurrence shows a good agreement (Fig. 5, normalized to 114Cd as it represents the isotope with the highest intensity). This provides a suitable representation for the whole sample and supports the finding that no spectral interferences were detected. Minor deviations can be ascribed to the fact that neither a detector dead time determination nor a mass bias correction have been performed and that the dwell time of 26 ms per isotope is perfect for the analysis of numerous elements during a transient measurement with 2 Hz, but not for a quasi-simultaneous measurement of isotopic ratios. The m/z 110 isotope shows a noticeably reduced result of about 5% below the expected value, as the mass bias compared to m/z 114 used for normalization is more pronounced here.


image file: c7ja00354d-f5.tif
Fig. 5 Isotopic pattern of Cd normalised to 114Cd: isotopic pattern in an area of 80 pixels within the middle inner part of the cocoa bean (see Fig. 2) compared to the theoretical pattern of naturally occurring Cd isotopes.

Reflecting the function of the different elements for plants and especially seeds, summarized by Wu et al.,25 the distribution within the fermented and dried cocoa beans is to a large extent of functional origin. Within the cocoa bean dissections, Mg, P, K, Ca, Cu and Zn were clearly detectable in the meristematic part of cocoa seeds, which is in good agreement with their biochemical functions. Zn is essential for thousands of proteins, also for root growth and stress tolerance during growth.27 Mg plays an important role for energy metabolism (e.g., ATP synthesis), while P is a key constituent of nucleic acids and phospholipids.25 K is an important growth- and reproduction factor, Ca is vital for the structural maintenance of membranes and cell walls, and Cu serves in numerous metabolic processes from photosynthesis to oxidative stress protection.25 Considering Fe as a nutrient with indispensable functions for enzymes, hormone biosynthesis or pathogen defence, location in these parts of the seed would have been expected as well, while toxic Cd, Na and Pb should be absent.25,43,44 Nevertheless, small amounts of Cd seemed to enter the hypocotyl and radicula, while Fe was not detectable in this part of the sample. Na has been proven to be replaceable by K in plants,25 though the distinctly varied distribution of both elements in our cocoa bean dissections does not support this possibility here. The related distribution of Zn and Cd, though, is somewhat expectable considering their chemical resemblance.

The toxic elements location in peripheral tissues, like the cocoa shell already during seed development, would be a clear benefit to minimize any entry into subdividing meristematic cells. However, keeping in mind that the cocoa beans were already processed before LA-ICP-MS application, the elemental distribution mapped might be attributed, at least to a certain extent, to cocoa fermentation and drying. So far, biochemistry of cocoa bean ingredients throughout the processing has been a matter of concern in numerous studies.45,46 However, there is a clear gap of knowledge to what extent the cocoa process management, and thus process-driven fluctuations in possible metal ligands, de facto affect elemental distribution within cocoa beans. This is of great relevance for restricting bioaccessibility of toxic elements, while keeping the essential nutrients available for consumers.

An important point during sample preparation was to obtain a smooth and plain surface to avoid topographical effects. The distribution of the elements within the samples indicates that the sample preparation has been successful. Although some of the elements were predominantly concentrated on the right bean shell side, the distribution of all elements in the inner embryo does not show any considerable gradient that would imply an oblique sample surface. The signal validity within the area where a part of the inner cotyledon tissue had been removed is rather complicated. The laser beam had not been fully focused in this area for the second measurement, as the focus was orientated on the flat inner bean surface. Therefore, a lower element intensity than in other parts of the bean would have been expected for this area, as it is the case for Ca, Fe and Pb. Contrarily, all other elements (except for Na, where only the shell is significantly visible) show an enhanced intensity in this area after removal of the inner cotyledon part. This could be a very first indication that in this inner shell, the concentration is high enough to counterbalance the hypothetical lower intensity of the not fully focused sample part. Nevertheless, for a confident statement, first, the laser beam would have to be focused on the inner seed shell, which is almost impossible due to its curved nature. Second, an approach for elemental quantification needs to be found. This, however, is extremely challenging due to the complexity of the cocoa bean consistence and diverse matrix composition. Though internal standardization could help to address this issue, the most commonly used approach of 13C standardization is still under discussion,47 as the inhomogeneous distribution of carbon within the samples can seriously falsify the results, especially for samples with varying density as used in this study.

For the analysis, it was of high importance to ensure the absence of interferences for the analytes of interest. As most of the interferences, such as 40Ar16O at 56Fe or organic fragments at several Cd isotopes, could be expected to occur within the whole analysis and not only in certain areas of the sample, the low background intensities of 0 cps for all elements can be seen as an evidence that the interference suppression worked out sufficiently. Another indication for this is the similar appearance of all Cd isotope distribution maps and the natural isotopic pattern of the observed Cd distributions.

Conclusions

Within this study, a sample preparation protocol and LA-ICP-TQMS method to analyse the distribution of several nutrients as well as toxic elements in fermented cocoa beans was established. The application of a triple quadrupole analyser and an extended analysis on distribution of various Cd isotopes within the cocoa beans led to the conclusion that any spectral interference disturbing the analysis was eliminated. Topographic influences could be also minimized without destruction of the sample (e.g., loss of the husk) by a special polishing process during sample preparation.

It could be found that several elements, including the essential elements Na, Mg, P, K, Cu and Zn as well as toxic elements like Pb and Cd, were enriched in the seed shell of the processed cocoa bean. Elevated intensities of Mg, P, K, Cu and Ca could be detected in the hypocotyl and radicula area. The structure of the cotyledon interface revealed increased intensities of Ca and Fe as well.

Thus, in the future, the LA ICP-MS approach can be a useful tool to screen for genetic or geographic parameters (e.g., geographic origin or cultivation conditions of Theobroma cacao L. tree) that might affect the distribution of either toxic or essential elements within the cocoa seeds. Moreover, LA-ICP-MS could be applied to improve the cocoa bean manufacturing process in a manner that is beneficial for the customer, with an advantageous reduction of contaminants in cocoa-derived products, while nutrients promoting human health, such as polyphenols or essential trace elements, are preserved.

Nevertheless, for a further improvement of the current LA-ICP-MS method, a way to ensure a planar sample after sample preparation, e.g., by profile depth measurements, would be beneficial to grant a higher degree of automation. Additionally, a reliable method for metal quantification would be beneficial as well. However, the application of matrix matched standards, which are often used in LA-ICP-MS bioimaging,25 tends to be very difficult especially in the case of cocoa beans, where the matrix is inhomogeneous and even variable in molecular composition. Certified reference materials with matrices appropriate for cocoa seed are hard to find, especially in cases where nutrients like Na shall be quantified, which naturally occur in high concentrations in natural samples. Thus, for a relative comparison between different processing methods or cultivation conditions as has been sought in this study, a qualitative method can provide sufficient information.

Overall, it can be concluded that imaging by LA-ICP-TQMS allows determining the local distribution of nutrients and toxic elements in cocoa bean subsets.

Conflicts of interest

The authors do not report any conflicts of interest.

Acknowledgements

The authors would like to thank the staff of the precision mechanics department of the Institute for Inorganic and Analytical Chemistry (University of Münster) for their vital help in the development of the sample preparation method.

References

  1. ICCO Quarterly Bulletin of Cocoa Statistics, Production of cocoa beans, vol. XLIII(2), Cocoa, 2016/17 Search PubMed.
  2. ICCO Quarterly Bulletin of Cocoa Statistics, Grinding of cocoa beans, vol. XLIII(2), Cocoa, 2016/17 Search PubMed.
  3. G. Ramtahal, I. Chang Yen, N. Ahmad, I. Bekele, F. Bekele, K. Maharaj, L. Wilson and L. Harrynanan, Commun. Soil Sci. Plant Anal., 2015, 46, 2585–2594 CrossRef CAS.
  4. L. Järup and A. Åkesson, Toxicol. Appl. Pharmacol., 2009, 238, 201–208 CrossRef PubMed.
  5. EFSA, EFSA J., 2009, 980, 1–139 Search PubMed.
  6. F. Thévenod and W. K. Lee, Toxicology of Cadmium and Its Damage to Mammalian Organs, in Cadmium: From Toxicity to Essentiality, ed. A. Sigel, H. Sigel and R. Sigel, Springer, Dordrecht, 2013, vol. 11, pp. 415–490 Search PubMed.
  7. IARC, IARC Monographs on the Evaluation of Carcinogenic Risks to Humans, 2012, vol. 100C, pp. 121–146 Search PubMed.
  8. Y. A. Cho, J. Kim, H. D. Woo and M. Kang, PLoS One, 2013, 8(9), e75087 CAS.
  9. The European Commission, Off. J. Eur. Communities: Legis., 2014, 138, 75–79 Search PubMed.
  10. D. Diomande, I. Antheaume, M. Leroux, J. Lalande, S. Balayssac, G. S. Remaud and I. Tea, Food Chem., 2015, 188, 576–582 CrossRef CAS PubMed.
  11. D. Bertoldi, A. Barbero, F. Camin, A. Caligiani and R. Larcher, Food Control, 2016, 65, 46–53 CrossRef CAS.
  12. P. Vargas Jentzsch, V. Ciabotă, W. Salinas, B. Kampe, P. M. Aponte, P. Rösch, J. Popp and L. A. Ramos, Food Chem., 2016, 211, 274–280 CrossRef CAS PubMed.
  13. A. Caligiani, A. Marseglia, B. Prandi, G. Palla and S. Sforza, Food Chem., 2016, 211, 431–439 CrossRef CAS PubMed.
  14. O. Ekpa, A. A. Akpan and A. E. Udo, J. Sci. Food Agric., 1993, 61, 47–50 CrossRef CAS.
  15. I. M. Da Veiga Moreira, M. G. Da Cruz Pedrozo Miguel, C. L. Ramos, W. F. Duarte, P. Efraim and R. F. Schwan, J. Food Qual., 2016, 39, 839–849 CrossRef.
  16. N. Kumari, K. Jean Kofi, S. Grimbs, R. N. D'Souza, N. Kuhnert, G. Vrancken and M. S. Ullrich, Food Res. Int., 2016, 90, 53–65 CrossRef CAS PubMed.
  17. C. De Taeye, V. J. E. Evina, G. Caullet, N. Niemenak and S. Collin, J. Agric. Food Chem., 2016, 64, 8876–8885 CrossRef CAS PubMed.
  18. N. N. Batista, D. P. de Andrade, C. L. Ramos, D. R. Dias and R. F. Schwan, Food Res. Int., 2016, 90, 313–319 CrossRef CAS PubMed.
  19. F. Malvano, D. Albanese, R. Pilloton and M. Di Matteo, Food Chem., 2016, 212, 688–694 CrossRef CAS PubMed.
  20. C. Bortolini, V. Patrone, E. Puglisi and L. Morelli, Int. J. Food Microbiol., 2016, 236, 98–106 CrossRef CAS PubMed.
  21. S. Mounicou, J. Szpunar, D. Andrey, C. Blake and R. Lobinski, Food Addit. Contam., 2003, 20(4), 343–352 CrossRef CAS PubMed.
  22. R. L. Yanus, H. Sela, E. J. Borojovich, Y. Zakon, M. Saphier, A. Nikolski, E. Gutflais, A. Lorber and Z. Karpas, Talanta, 2014, 119, 1–4 CrossRef CAS PubMed.
  23. E. Chavez, Z. L. He, P. J. Stoffella, R. S. Mylavarapu, Y. C. Li, B. Moyano and V. C. Baligar, Sci. Total Environ., 2015, 533, 205–214 CrossRef CAS PubMed.
  24. E. Chavez, Z. L. He, P. J. Stoffella, R. S. Mylavarapu, Y. C. Li and V. C. Baligar, Chemosphere, 2016, 150, 57–62 CrossRef CAS PubMed.
  25. B. Wu, F. Andersch, W. Weschke, H. Weber and J. S. Becker, Metallomics, 2013, 5, 1276–1284 RSC.
  26. S. J. M. Van Malderen, B. Laforce, T. Van Acker, L. Vincze and F. Vanhaecke, J. Anal. At. Spectrom., 2017, 32, 289–298 RSC.
  27. I. Cakmak, M. Kalayci, Y. Kaya, A. A. Torun, A. Aydin, Y. Wang, Z. Arisoy, H. Erdem, A. Yazici, O. Gokmen, L. Ozturk and W. J. Horst, J. Agric. Food Chem., 2010, 58, 9092–9102 CrossRef CAS PubMed.
  28. G. d. S. Pessôa, C. A. Lopes Júnior, K. C. Madrid and M. A. Z. Arruda, Talanta, 2017, 167, 317–324 CrossRef PubMed.
  29. J. Promchan, D. Günther, A. Siripinyanond and J. Shiowatana, J. Cereal Sci., 2016, 71, 198–203 CrossRef CAS.
  30. P. Basnet, D. Amarasiriwardena, F. Wu, Z. Fu and T. Zhang, Environ. Pollut., 2014, 195, 148–156 CrossRef CAS PubMed.
  31. S. H. Choi, J. S. Kim, J. Y. Lee, J. S. Jeon, J. W. Kim, R. E. Russo, J. Gonzalez, J. H. Yoo, K. S. Kim, J. S. Yang and K. S. Park, J. Anal. At. Spectrom., 2014, 29, 1233–1237 RSC.
  32. A. A. Meharg, E. Lombi, P. N. Williams, K. G. Scheckel, J. Feldmann, A. Raab, Y. Zhu and R. Islam, Environ. Sci. Technol., 2008, 42, 1051–1057 CrossRef CAS PubMed.
  33. E. T. F. Witkowski, I. M. Weiersbye-Witkowski, W. J. Przybyłowicz and J. Mesjasz-Przybyłowicz, Nucl. Instrum. Methods Phys. Res., Sect. B, 1997, 130, 381–387 CrossRef CAS.
  34. A. G. Kachenko, N. P. Bhatia, R. Siegele, K. B. Walsh and B. Singh, Nucl. Instrum. Methods Phys. Res., Sect. B, 2009, 267, 2176–2180 CrossRef CAS.
  35. H. L. Malan, J. Mesjasz-Przybylowicz, W. J. Przybylowicz, J. M. Farrant and P. W. Linder, Nucl. Instrum. Methods Phys. Res., Sect. B, 2012, 273, 157–160 CrossRef CAS.
  36. P. J. Heard, K. A. Feeney, G. C. Allen and P. R. Shewry, Plant J., 2001, 30(2), 237–245 CrossRef.
  37. K. L. Moore, M. Schröder, E. Lombi, F.-J. Zhao, S. P. McGrath, M. J. Hawkesford, P. R. Shewry and C. R. M. Grovenor, New Phytol., 2010, 185, 434–445 CrossRef CAS PubMed.
  38. S. Vaidyanathan, M. Salim, C. Hurley and J. Pugh, Surf. Interface Anal., 2013, 45, 290–293 CrossRef CAS.
  39. G. M. Thyssen, M. Holtkamp, H. Kaulfürst-Soboll, C. A. Wehe, M. Sperling, A. von Schaewen and U. Karst, Metallomics, 2017, 9, 676–684 RSC.
  40. C. A. Wehe, A.-C. Niehoff, G. M. Thyssen, M. Sperling and U. Karst, Rapid Commun. Mass Spectrom., 2014, 28, 2627–2635 CrossRef CAS PubMed.
  41. G. Ramtahal, I. Chang Yen, I. Bekele, L. Wilson, K. Maharaj and B. Sukha, Qual. Assur. Saf. Crops Foods, 2015, 7, 731–736 CrossRef CAS.
  42. C. K. Lee and K. S. Low, Pertanika, 1985, 8, 243–248 CAS.
  43. J.-K. Zhu, Trends Plant Sci., 2001, 6, 66–71 CrossRef CAS PubMed.
  44. F. J. M. Maathuis and A. Amtmann, Ann. Bot., 1999, 84, 123–133 CrossRef CAS.
  45. J. Voigt and R. Lieberei, in Biochemistry of Cocoa fermentation, ed. R. F. Schwan and G. H. Fleet, CRC Press, 2014, pp. 193–226 Search PubMed.
  46. J. E. Kongor, M. Hinneh, D. Van de Walle, E. O. Afoakwa, P. Boeckx and K. Dewettinck, Food Res. Int., 2016, 82, 44–52 CrossRef CAS.
  47. D. A. Frick and D. Günther, J. Anal. At. Spectrom., 2012, 27, 1294–1303 RSC.

This journal is © The Royal Society of Chemistry 2018