Quality Parameters, Caffeine and Theobromine Contents and Antioxidant Activity of Artisan and Commercial Chocolate from Brazil ()
1. Introduction
Chocolate is a solid dispersion of cocoa seeds (Theobroma cacao), sugar particles, additives, cocoa butter, lecithin and flavourings, giving an average of 568 kcal/100g (Leite, 2012 [1]; Godočiková et al., 2019 [2] ). The Mayan and Aztec people, in South America, where the first to document its use. Chocolate and cocoa seeds have several medicinal applications, such as: stimulating the nervous system, the stomach and kidneys activity, treating fever, tuberculosis, treating low virility and anaemia, against mental fatigue, including cardiovascular benefits, among others (Efraim et al., 2011 [3]; Vicentim & Marcellino, 2012 [4]; Kerimi & Williamson, 2015 [5]; Magrone et al., 2017 [6]; Anwar et al., 2018 [7]; Kardum & Glibetic, 2018 [8]; Stanley et al., 2018 [9]; Coutinho et al., 2019 [10]; Godočiková et al., 2020 [11] ). Chocolate is known to contain polyphenols and flavonoids (Counet et al., 2006 [12]; Alvarez-Barreto et al., 2018 [13]; Muhammad et al, 2018 [14] ) and these compounds are in higher concentrations in bitter chocolate, presenting greater health benefits and antioxidant activity (Lee et al., 2003 [15]; Hu et al., 2016 [16]; Mudenuti et al., 2018 [17]; Tafurt et al., 2021 [18] ).
Chocolate is a good source of polyphenols (Torres-Moreno et al., 2012 [19]; Hu et al., 2016 [16] ) such as epicatechin and catechin, along with the proanthocyanin oligomers (Gu et al., 2004 [20]; Gu et al., 2006 [21]; Vinson et al., 2006 [22]; De Oliveira & Genovese, 2013 [23]; Cambrai et al., 2017 [24]; Martini et al., 2018 [25]; Fayeulle et al., 2019 [26] ) and alkaloids, such as theobromine and phenethylamine (Gopalakrishnan et al., 2021) [27]. These molecules can function as primary antioxidants, directly reducing the formation of free radicals (Hu et al., 2016 [16]; Mudenuti et al., 2018 [17] ). Chocolate also contains vitamins A, B, C, D, and potassium, sodium, iron, calcium, magnesium, phosphorous, manganese, copper, zinc and selenium (Leite, 2012 [1]; Torres-Moreno, Torrescasana et al., 2015 [28]; Oliveira et al., 2021 [29] ). In milk chocolate, the cocoa mass is partly replaced by milk powder, giving it a sweeter taste. Milk chocolate or half-bitter chocolate contain between 40% and 53% cocoa butter. A bar of chocolate (30 g) contains an average of 165 calories and between 2 to 3 g of protein. Bitter chocolate, has a high concentration of cocoa mass and low sugar content. It is made with cocoa paste and cocoa butter. It is less refined, having a characteristic dark colour and bitter taste, are due to the reduced sugar content. Half-bitter chocolate is slightly refined with a dark colour. It does not contain milk, but contains sugar, and has a less bitter taste, presenting and about 550 kcal/100g. Some people use chocolate as a form of self-medication to compensate for nutrient deficiency, probably magnesium. Chocolate and cocoa have an exceptionally high concentration of this mineral (100 mg/100g and 250 mg/100g, respectively) and in some Mg cases supplementation reduces chocolate compulsions and decreases the symptoms of premenstrual tension (Del Rei & Medeiros, 2011 [30]; Grassia et al., 2019 [31] ).
CEPLAC (Executive Committee for Cacao Crops), in Ilhéus, Bahia, Brazil, has the mission of promoting the competitiveness and sustainability of the agricultural, agroforestry and agroindustry segments for the development of cocoa producing regions, producing artisan chocolates with high cocoa content. It is an organ of the Brazilian Ministry of Agriculture, created on February 20, 1957, when the cocoa economy was going through a serious crisis, initially focusing on supporting cacao farmers. It succeeded in increasing the Brazilian cocoa production by 310%; from 220 kg/ha in 1962 to 740 kg/ha in 2016, among others. The current priority is the recovery of the regional economy, with emphasis on fighting a disease that has been decimating the cacao plantations known as "witch-broom" and aiming to develop sustainable agroeconomic activities and preserving the remaining forest fragments of two strategic ecosystems in Brazil, the Atlantic Forest and the Amazon Forest (Ceplac, 2021) [32].
Brazilian laws determine that chocolate must have at least 25% cocoa. One in three chocolate-based products in Brazil does not contain this amount of cocoa and therefore cannot be considered chocolate (Lima et al., 2011) [33]. It is estimated that roughly 35% of the common chocolates on the shelves, produced by large food companies, are really chocolate candy. Many products claim to have a high percentage of cocoa, which is not always the case, since there is no supervision to guarantee the information on the label. Furthermore, it is not compulsory by law to declare the amount of cocoa in the product. Instead, these candies contain high amounts of sugar and fat (Leite, 2012 [1]; Seem et al., 2019 [34] ). Countries in Europe and the United States have informative legislation on the composition of chocolates (European Union, 2000 [35]; Rodriguez-Negrette et al., 2019 [36] ).
The present work aimed at analyzing the properties of chocolates, especially that produced by CEPLAC, determining, among other parameters, the fat content, caffeine, theobromine and antioxidant activity, as well as the correlations between the evaluated parameters.
2. Materials and Methods
2.1. Chemicals and Reagents
All reagents employed were at least analytical grade: ethyl ether, ethanol, 16N HCl, sodium carbonate and sodium tartrate (Synth, Brazil), barium oxide (Vetec), universal indicator, Folin-Ciocalteau reagent, Merck (Darmstadt, Germany), 0.45 μm filters (Sartorius, USA). Gallic acid, caffeine and DPPH (radical diphenylpicrylhydrazyl) were from Sigma-Aldrich (St. Louis, USA). Methanol and formic acid were HPLC (High-Performance Liquid Chromatography) grade from (Merck, Darmstadt, Germany), deionized water (18.2 MΩ/cm) was prepared using a Banrsted apparatus (VWR, USA).
2.2. Chocolate Samples
The following nine samples of commercial chocolates were analyzed: bitter 70% cocoa (AD70), white C (BC), white D (BD), Belgian milk (BLC), milk 20% cocoa (CS20), milk 70% of cocoa (CS70), milk H (LH), milk HE (LHE), bitter medium (MAH); as well as artisan chocolate (80% bitter CEPLAC (AC80)) and cocoa powder CEPLAC (SCC).
2.3. Fat Extraction
An aliquot of 15 g of each sample was packed in a pre-weighed cartridge. 200 mL of ethyl ether was used in a soxhlet apparatus for two hours. After extraction, the solvent was evaporated leaving only the fat, which was quantified. The fat-free dry matter, called insoluble solids, was also quantified. The procedure was performed in triplicate.
2.4. Preparation of Solutions
The dried fat-free samples were ground and exactly about 0.1 g of the sample powder was weighed and dissolved in ethanol to give the 0.1% solutions used in tests for antioxidant activity and total phenolics. For the chromatographic analysis, the solutions were prepared as follows: 1.25 g of barium oxide (BaO) were weighed for each chocolate sample plus 2.5 g of the chocolate powders, then 1.25 mL of water was added, and boiled on the heating plate. After 10 minutes the mixture was removed and the pH adjusted to 7 with concentrated HCL. It was then transferred to a 250 mL volumetric flask completed with water and filtered through paper.
2.5. Determination of the Total Phenolic Content
An aliquot of 0.1 g of the fat-free sample powder was dissolved in a beaker with a little water, transferred to a 50 mL volumetric flask and the volume was made up with water (stock solution: 2 mg/mL). The procedure was performed from this point in triplicate. An aliquot of 200 µL was transferred to a 10 mL volumetric flask (1:50) containing approximately 5 mL of distilled water. 800 mL of the Folin-Ciocalteau reagent was added and stirred for a few seconds and in an interval of 1 to 8 minutes 1.2 mL of the 20% sodium carbonate-tartrate solution was added. The volume of water was completed until near the meniscus. The solution was kept in a 20˚C bath. After 2 hours, the final volume was completed at 20˚C. It was shaken for a few seconds and read at 760 nm (in a Cary 50, Varian Inc spectrophotometer, USA). The value of total phenolics was determined by comparison to a calibration curve with gallic acid. The same procedure was used for the quantification of phenolics in chocolates extracted with BaO. The procedure was performed in triplicate.
2.6. HPLC-DAD Conditions
The samples were stored in vials for HPLC analysis using LaChrom chromatograph, model D-7100, L-7100 pump, L-7200 automatic sampler and L-7455 diode array detector (Merck, Darmstadt, Germany). To determine theobromine and caffeine in the chocolate samples, a reverse phase C18 column (Gemini 00F-4435-E0, 110A RP 18, 12.5 × 0.4 cm, 5 μm particle diameter Merck, Darmstadt, Germany) with pre-column (Safety Guard Cartridges Kit KJO-4282, Merck, Darmstadt, Germany) and a isocratic system (1 mL/min flow) composed of water acidified with 0.5% formic acid (85%) and methanol (15%), was used. The volume injected was 50 μL. The maximum time of analysis was 10 minutes and detection performed at a wavelength of 272 nm. The program used for data analysis was the Merck-Hitachi model D-7100 (Chromatography Data Station DAD Manager, Merck). The samples (1% chocolate ethanolic solution) were dissolved in methanol, filtered on a 0.45 μm filter and injected into HPLC-DAD (Matsume et al., 2000 [37]; De Maria & Moreira, 2007 [38] ).
Caffeine and theobromine (methylxanthines) were quantified using the HPLC method described by comparison to an external calibration curve of caffeine standard, using the least squares method (triangle equation) and averaged, standard deviation, coefficient of variation (%) and uncertainty. In the linearity, the residuals were also calculated (Ribeiro et al., 2008) [39].
2.7. UHPLC-MS Conditions
UHPLC-MS (Ultra-high performance liquid chromatography mass spectrometry) chromatography was performed on a Waters Acquity UPLC instrument using a Waters C-18 column (50 mm × 2 mm, 1.7 μm particle) with oven at 30˚C. Solvent A = milli-Q water with 0.1% formic acid and solvent B = methanol, 3 μL injection, 0.2 mL/min flow, gradient starting with 90% A, ramping to 0% A in 5 minutes, held until 5.50 minutes, then back to 90% A at 5.51 minutes, and equilibrated until 6.00 minutes, with a total run time of 6.00 minutes. Samples were compared to the caffeine, theobromine and theophylline standards by UHPLC-MS (Waters Acquity TQD mass spectrometer (USA). The mass spectrometry conditions were: electrospray ionization, positive ion mode, capillary 3.0 kV, cone 30 V, source temperature 150˚C, desolvation gas temperature 300˚C, collision energy 25V (in MS/MS mode). The presence of caffeine and theobromine were identified due to peaks in the sample with the same retention time as the standards, and confirmed by MS/MS.
The use of the method described above for the UHPLC was carried out to capture data that require more sensitive devices, than the conventional analysis performed by HPLC. Both methods developed in this study for UHPLC-MS and HPLC-DADUV were separately validated.
2.8. Antioxidant Activity
The method was based on a free radical scavenging reaction using the diphenylpicrilhydrazyl radical (DPPH). Wells were numbered from 0 to 10 in a 24-well plate. Corresponding ethanol volumes were added into each of the various tubes and various volumes of the 0.01% chocolate ethanolic solution. A volume of DPPH (0.15 mM) was then added, with a one-minute interval between tubes, and these were shaken sporadically. The absorbance was read at 517 nm on a Cary 50, Varian Inc spectrophotometer (USA) 30 minutes after addition of DPPH in the first tube, and so on. The absorbance vs. chocolate concentration plot was built and the ED 50 (dose eliminating 50% free radicals) calculated by the least squares method. The procedure was performed in triplicate (Veiga et al., 2017) [40].
2.9. Statistical Analysis
The fat content, insoluble solids, phenol content, theobromine content, caffeine content and antioxidant activity (ED50) data were submitted to the Shapiro-Wilk normality test. Once the normality of the distribution was verified, the data were submitted to analysis of variance (ANOVA) and Tukey test with significance level of 5%, considering the type of chocolate. PAST software Version 3.22 was used (Hammer, Harper & Ryan, 2001) [41].
A matrix containing the data from the various tests, with the lines representing the samples and the columns the variables, was analyzed by Principal Component Analysis. In the present work the software Uscrambler X10.4, 2016 (CAMO AS, Trondheim, Norway) was used.
3. Results and Discussion
The following parameters of chocolate quality were evaluated in the present work: fat content, total phenolics, theobromine and caffeine (methylxanthines) content and antioxidant activity. Table 1 shows the values obtained for chocolate samples and cocoa bean powder.
It can be seen (Table 1) that SCC presented a significantly higher fat concentration (% w/w) than all the chocolate samples. The AC80 presented significantly lower fat concentration than SCC and the others, with the exception of MAH and BLC. Among the chocolate samples, AD70, CS70 and CS20 presented the highest concentrations of fat (% w/w). Cocoa bean powder showed the highest fat content, 33.40%. Leite (2012) [1] evaluated the fat content in toasted cocoa beans, finding values between 32.60% and 36.52%, similar to our result. For the cocoa mass, the same author found values from 52.95% to 58.17% and for chocolates, values between 37.87% and 45.50% were found. De Melo et al. (2020) [42] reported that the percentage of fat varies between 35.1% to 41.8%. The fat content found for chocolates in our study was between 15.51% to 25.51%, values
![]()
Table 1. Amount of fat (% w/w), dried residue (DR) (% w/w) and total phenolic content (% w/w) in chocolate samples and cocoa beans.
*Average of three measurements.CV% is the percentile variation coefficient and Unc. the uncertainty of the measurements, calculated by the standard deviation/square root of the number of measurements. The same letters in each column indicate that there is no statistical difference at a significance level of 5%, according to the Tukey Test results.
slightly lower than those reported by Leite (2012) [1] and De Melo et al. (2020) [42]. As expected, the bitter and half-bitter chocolates presented the lowest fat contents. In the same extraction process with ethyl ether, the solid residue content was measured after extraction. The values are also presented in Table 1. These values ranged from 59.53% to 82.60%. SCC presented insoluble solids concentration (% w/w) which was significantly lower than all chocolates. AC80 showed a significantly higher concentration of insoluble solids than SCC and AD70, CS70, LHE, CS20 and BD chocolate samples. The HAM sample showed a higher concentration of insoluble solids than the AC80 sample. The CS70 sample, on the other hand, had the lowest concentration of insoluble solids.
The total phenolic content was initially measured in the dry extract, after extraction with ethyl ether however, only five samples were evaluated: AD70 (total phenolic content of 2.48% ± 0.11%), BD (0.00% ± 0.00%), BLC (0.31% ± 0.17%), LH (0.36% ± 0.01%) and MAH (0.77% ± 0.36%). After this initial evaluation, the use of BaO for defatting the samples was implemented (Alves & Bragagnolo, 2002) [43] and all the samples were extracted. Total phenolic, caffeine and theobromine contents were quantified in these BaO extracted samples. Table 2 shows the total phenolic content for the samples extracted with BaO. AC80 total phenolic content significantly higher than SCC and all other samples. SCC presented significantly higher total phenolic content than all samples, with the exception of CS70 and AC80. The bitter (AD70 and CS70) and the half-bitter (MAH) samples presented significantly higher total phenolic content than the milk chocolate samples (LH, LHE BLC and CS20) and white chocolate samples (BD and BC). The white chocolate samples BC and BD presented significantly lower total phenolic content than all the other chocolate samples tested. The samples that presented the highest total phenolic content were cocoa seed powder (3.70%) and those with higher cocoa content, which was expected (from 3.17% to 4.21%). The milk chocolates presented intermediate values and the whites almost no phenolic content. White chocolates do not have cocoa mass in their composition, only cocoa butter, which explains a virtually zero content of polyphenols. Ramli et al. (2001) [44] determined the total phenolic content in different samples of cocoa from Malaysia and Ghana, finding for beans, from 3.49% and 6.02%. Vertuani et al. (2014) [45] found values between 2.0% to 3.5% for cocoa. In the present study, the amount found is approximately the same. The values presented in Table 1 are in accordance with the data of these researchers, except as previously mentioned, for white chocolate samples, which do not have the cocoa mass.
Leite (2012) [1] analyzed the phenolic content in two varieties of cocoa mass from the region of Ilhéus-BA and chocolates prepared with this same mass. For the cocoa mass, the author phenolic content in the range of 2.39% to 2.58% and for chocolates between 1.55% to 1.91%. Several authors have reported between 2.00% and 8.37% of phenolic content in natural cocoa powder (Waterhouse et al., 1996 [46]; Serra-Bonvehí & Ventura Coll, 1997 [47]; Kealey et al., 2000 [48]; Natsume et al., 2000 [49]; Lee et al., 2003 [15], Vinson et al., 2006 [22]; Efraim et al., 2011 [3]; Vertuani et al., 2014 [45] ). For chocolates, from 0.26 to 4.85% (Hu et al., 2016) [16] and 1.00% to 4.50% (Mudenutti et al., 2018) [17] have been reported; while for bitter chocolate, between 0.246% and 2.39% (Waterhouse et al., 1996 [46]; Natsume et al., 2000 [49]; Vinson et al., 2006 [22]; Counet et al., 2006 [12] ), and for milk chocolate from 0.192 to 0.990% (Waterhouse et al., 1996 [46]; Gu et al., 2006 [21]; Vinson et al., 2006 [22] ). For cocoa liquor, values of 0.367% to 8.52% (Adamson et al., 1999 [50]; Kealey et al., 2000 [48]; Luna et al., 2002 [51] ) and 2.84 ± 0.24 (De Oliveira & Genovese, 2013) [23] were reported. Natsume et al. (2000) [49] found 6.04% of phenolic content in Brazilian cacau liquor, but its geographic origin was not reported.
The range of total polyphenol content for milk chocolate and bitter chocolate found by Efraim et al. (2011) [3] are lower than the values found in the present work for the same types of chocolates. However, the reported value for cocoa is within the range reported by the same authors for this product. Efraim et al. (2011) [3] citing Ramirez-Sanchez et al. (2010) [52] reported that phenolic content in cocoa may vary according to its geographical origin, plant variety, climate, soil type and planting area. The different steps of the transformation of cocoa into chocolate may also influence the phenolic content in the final product. With respect to the quantification of theobromine and caffeine, Figure 1 shows the chromatogram of one of the samples analyzed.
After using barium oxide (BaO) as a fat extracting agent, caffeine and theobromine were quantified using a calibration curve based on caffeine (Figure 1). Theobromine has the same UV-Vis spectrum of caffeine (in the HPLC spectra
![]()
Figure 1. Chromatogram of the 80% cocoa chocolate sample from CEPLAC. The arrows indicate the peaks for mehtylxanthines (theobromine and caffeine) and their retention times in minutes (Rt).
library), but with different retention time, therefore, the proposal was to quantify theobromine by the caffeine area. Figure 2(a) shows the UHPLC-MS that confirm of the presence of both theobromine and caffeine in the samples of chocolates, in comparison with standards. In Figure 2(b) are the mass spectra (MS/MS) of the chocolate sample compared to the caffeine (Figure 2(c)) and theobromine (Figure 2(d)) standards. Theobromine has a lower retention time than caffeine in reverse phase HPLC because the former has a methyl group less than the latter, facilitating its elution from the chromatographic column (Figure 1).
A calibration curve of the caffeine standard was constructed for quantification of caffeine and theobromine by HPLC-DAD. From the confirmation of the theobromine and caffeine peaks and also of the caffeine calibration curve, the contents of these compounds were calculated in the chocolate samples. Table 2 shows the levels of theobromine and caffeine (g/100g) obtained in samples after the BaO extraction.
The concentration of both theobromine and caffeine were higher in the samples when the barium oxide extraction method was used, indicating that the barium oxide method is much more efficient in extracting these methylxanthines than when ethyl ether is used. Regarding the theobromine content, SCC presented a significantly higher percentage than all the chocolate samples. AC80 had significantly lower theobromine content than MAH, similar to CS70 and superior to AD70, and to all milk chocolate samples (LH, LHE, BLC and CS20). Theobromine was not detected in white chocolate samples (BD and BC). Regarding caffeine content, SCC presented significantly higher caffeine content than all the other chocolate samples. The AC80 showed significantly higher theobromine and caffeine content than all the other chocolate samples. Milk chocolates (LH, LHE, BLC and CS20) had significantly lower caffeine content
![]()
Figure 2. Positive ion mode UHPLC-MS chromatograms of the CEPLAC chocolate sample, showing retention times of theobromine [M + H]+ ion (m/z 181) at 1.58 minutes and caffeine [M + H]+ ion (m/z 195) at 2.33 minutes (circles). (a) MS/MS spectra of theobromine (m/z 181) and caffeine (m/z 195) in the chocolate sample (b) and of caffeine standard (c) and theobromine standard (d). Full circles indicate the precursor ion of caffeine and dashed lines the precursor ion of theobromine.
than all bitter chocolate samples (AC80, AD70, CS70 and MAH). Caffeine was not detected in white chocolate samples (BD and BC). Cacao seeds (SCC) presented higher caffeine and theobromine contents, higher than all evaluated chocolates.
The results of theobromine and caffeine content presented a strong positive correlation (R = 0.9297). Ramli, Yatim, Said & Hok (2001) [44] evaluated the theobromine and caffeine content in different cocoa products: for cocoa liquor from Malaysia and Ghana, they found 1.720% to 2.530% (w/w) of theobromine and from 0.310% to 0.380% caffeine. For the cocoa beans from the same sites, between 1.620% and 2.660% of theobromine and from 0.250% to 0.500% of caffeine were found. The same authors evaluated these levels in cocoa powder in Malaysia, reporting between 1.960% to 2.940% of theobromine and 0.330% to
![]()
Table 2. Theobromine and caffeine contents (%) in chocolate samples and cocoa beans extracted with BaO.
*Average of three measurements. CV% is the percentile variation coefficient and Unc. the uncertainty of the measurements, calculated by the standard deviation/square root of the number of measurements. The same letters in each column indicate that there is no statistical difference at a significance level of 5%, according to the Tukey Test results.
0.55% of caffeine; for Malaysian chocolate with cocoa butter (brand I) from 0.080% to 0.100% theobromine and from 0.006% to 0.009% caffeine. Chocolates from the same place with cocoa butter (brand II) presented theobromine and caffeine contents from 0.036% to 0.065% and 0.003% to 0.006%, respectively. The results presented in the present study are in line with those reported by Ramli, Yatim, Said & Hok (2001) [44] regarding the theobromine content in cocoa beans. However, the caffeine concentration for the same product was much lower in Malaysian chocolate than in the present study. De Oliveira & Genovese (2013) [23] reported a caffeine content of 0.27% ± 0.01% for cocoa liquor. For the chocolates prepared with cocoa butter, the values mentioned by Ramli et al. (2001) [44] were much lower than those found in the present study. If, as reported by Efraim et al. (2011) [3], the levels of polyphenols in cocoa may vary according to their origin, the same may occur for the methylxanthines: theobromine and caffeine.
The antioxidant activity of chocolate samples was evaluated in the DPPH free radical discoloration method. Table 3 shows the results for all evaluated chocolates; ED50 values ranged from 4.54 to 285.83 μg/mL. Vertuani et al. (2014) [45] reported values of ED50 between 140 to 470 μg/mL for cocoa. It was observed that the SCC and AC80 presented the highest antioxidant activity among the chocolate samples tested, being significantly superior to the other chocolates, with the exception of the MAH, which presented similar values. CS70 also had high antioxidant activity, higher than the other chocolate samples, but lower than SCC and AC80. Bitter chocolate samples (MAH, CS70 and AD70) showed
![]()
Table 3. Antioxidant activity of chocolate samples (ED50 µg/mL).
*Average of three measurements. CV% is the percentile variation coefficient and Inc. the uncertainty of the measurements, calculated by the standard deviation/square root of the number of measurements. The same letters in each column indicate that there is no statistical difference at a significance level of 5%, according to the Tukey Test results.
significantly higher antioxidant activity than milk chocolates (LH, LHE, BLC and CS20). The white chocolate samples (BC and BD) had very low antioxidant activity, significantly lower than the other samples.
The data in Table 3 confirms that the antioxidant activity is related to the phenolic content, as reported at the beginning of the present study. The bean powder and the chocolate samples with higher cocoa content are those that presented higher total phenolic content, consequently better antioxidant activity. Through the statistical analysis performed, the following correlations (R) were observed: very high between caffeine x theobromine (0.9274), moderated between ED50 x theobromine (0.6441) and a weak ED50 x caffeine (0.4854).
After collecting all the physical-chemical data and the antioxidant activity, they were submitted to a multivariate statistical analysis (Principal Component Analysis or PCA). According to this, factor 1 explained 99% of the variance and factor 2 plus 1% of the variance, in a total of 100%, which explains the distribution in two PCs (1 and 2). These results can be seen in Figure 3.
The samples were separated into three different groups. In Group 1, the white chocolate samples (BD and BC), which do not present antioxidant activity, total phenolic content is very low and lack theobromine and caffeine (they are in the negative quadrant of Figure 3(b)). Group 2 is the intermediate, with samples AD70, LH and LHE, with intermediate amounts of theobromine and caffeine and phenolic content and moderate antioxidant activity. In Group 3 are AC80, CS20, CS70, BLC and MAH, with better antioxidant activity, higher total phenolic and methylxanthine content. Group 3 stands out from all others for having the highest content of methylxanthines (caffeine and theobromine), consequently presenting the best antioxidant activity. The bean powder was separated
![]()
Figure 3. Principal component analysis of the (PCA) of the chocolate samples. PC1 xPC2 scores (samples) (a) and the loadings (variables) (b). For abbreviations see Materials and Methods.
from the rest because they had a high fat content. The AC80 and MAH chocolate samples have the highest content of ether insoluble solids (Figure 3(a)).
4. Conclusion
The highest fat contents were found in the milk chocolate samples followed by the white chocolate samples. Bitter chocolate has a higher percentage of cocoa in its composition than other types of chocolates and therefore presented higher total phenolic contents than other types (such as white chocolate that has a very low or almost no polyphenols because there is no cocoa paste). To date, cocoa has been studied as a source of polyphenols, with its main components extracted, isolated and identified in bench processes, using organic solvents not allowed for use in food, such as methanol, for example. These solid-liquid extractions leave a residual content in the product, rendering it unsuitable for human consumption. Extractions using permitted organic solvents often promote degradation of the polyphenols during the process due to the temperature and oxygen exposure conditions to which the product is subjected. Polyphenols are compounds that degrade easily in the presence of heat and oxygen. A method of extraction of the alkaloids using barium oxide was used, successfully separating of the methylxanthines, as confirmed by mass spectrometry. The literature describes the use of magnesium oxide for the same analyses. It was possible to evaluate and to prove that the antioxidant activity was closely linked to the phenolic content. Theobromine and caffeine content were also high in these samples, possibly due to the higher percentage pf cacao paste used, confirmed by the multivariate statistical analysis. This work came to contribute to the confirmation of the properties of chocolates for human health, and these can be considered functional foods, such as CEPLAC chocolate with high phenol content, theobromine and caffeine, as well as a good antioxidant activity.
Acknowledgements
CEPLAC (Comissão Executiva de Planejamento da Lavoura Cacaueira) Ilheus, Bahia, Brazil, which provided the samples of cocoa and chocolate seeds to carry out this work.