Electrochemical Evaluation of Cd, Cu, and Fe in Different Brands of Craft Beers from Quito, Ecuador

The presence of heavy metals in craft beers can endanger human health if the total metal content exceeds the exposure limits recommended by sanitary standards; in addition, they can cause damage to the quality of the beer. In this work, the concentration of Cd(II), Cu(II), and Fe(III) was determined in 13 brands of craft beer with the highest consumption in Quito, Ecuador, by differential pulse anodic stripping voltammetry (DPASV), using as boron-doped diamond (BDD) working electrode. The BDD electrode used has favorable morphological and electrochemical properties for the detection of metals such as Cd(II), Cu(II), and Fe(III). A granular morphology with microcrystals with an average size between 300 and 2000 nm could be verified for the BDD electrode using a scanning electron microscope. Double layer capacitance of the BDD electrode was 0.01412 μF cm−2, a relatively low value; Ipox/Ipred ratios were 0.99 for the potassium ferro-ferricyanide system in BDD, demonstrating that the redox process is quasi-reversible. The figures of merit for Cd(II), Cu(II), and Fe(III) were; DL of 6.31, 1.76, and 1.72 μg L−1; QL of 21.04, 5.87, and 5.72 μg L−1, repeatability of 1.06, 2.43, and 1.34%, reproducibility of 1.61, 2.94, and 1.83% and percentage of recovery of 98.18, 91.68, and 91.68%, respectively. It is concluded that the DPASV method on BDD has acceptable precision and accuracy for the quantification of Cd(II), Cu(II), and Fe(III), and it was verified that some beers did not comply with the permissible limits of food standards.


Introduction
Beer is one of the most consumed alcoholic beverages worldwide, reaching 75% of the market compared to other beverages [1]. In South America, the consumption of craft beer has boomed in recent years due to the constant development and innovation of beers with shades of color in the blonde, red, and black scale, as well as different flavors and alcoholic degrees for a variety of consumers [2,3]. Craft micro-breweries are the main producers of this drink and are the promoters of its great variety in Ecuador. Despite the efforts made by producers to generate beer of higher sensory quality, the hygienic and/or toxicological quality is affected in the production-storage processes, and they do not always follow the same rigorous and automated control of internationally certified industrial macro-breweries. One of the factors that intervene in the quality of both craft and industrial beer is the presence of heavy metals. Heavy metal contamination can come from various sources such as the raw material, barley, the use of additives that contain metallic traces during the fermentation process, the beer maturation (e.g., suspended solids from "green beer" at high temperatures and long times could release more heavy metals into "mature beer", and contamination from brewing equipment corrosion of the beer [4][5][6]. The presence of heavy metals in the drink may incur non-compliance with food regulations since, due to its toxicity, it can cause lethal effects on the health of consumers [7]. One element is iron (Fe) which, in addition to generating health problems such as hemosiderosis 2. Materials and Methods 2.1. Materials, Equipment, Reagents, and Samples 2.1. 1

. Materials and Equipment
For the preparation of solutions, calibrated and amber glassware was used, Nextirrer brand micropipette A ± 5 µL. An electrochemical cell with three heart-type electrodes, an Ag/AgCl reference electrode, a graphite rod counter electrode, and a BDD working electrode (doping level 3000~5000 ppm, 0.3 cm) was used. To weigh the samples, an analytical balance was used, RADWAG model AS 220.R2 A ± 0.0001 g. For electrochemical measurements, a Metrohm Autolab B.V. potentiostat was used. In the treatment of the samples, a heating plate MTOP MS300HS at ± 1 • C was used. In addition, the following was used: SensionTM MM374 potentiometer, BRANSON 1800 brand ultrasound, Thermo Scientific Phenom ProX brand scanning electron microscope (SEM), Perkinelmer AAnalist 400 brand atomic absorption equipment.

Samples
Red beers with alcohol were chosen, and the sampling was carried out in the City of Quito in 13 local breweries in the north-central and northern sectors. A type of random, Foods 2023, 12, 2264 4 of 24 stratified sampling was applied; for this, different areas of the city of Quito were chosen: center-north in Foch, Orellana, Reina Victoria, Avenue 12 de Octubre, Diego de Almagro, and north of Quito in the Mariana de Jesús, Avenue 6 de Diciembre, Oswaldo Guayasamín, Vaca de Castro, Río Coca, Bicentenario, La Pradera and Whymper, see Table 1. The sample was taken directly from the tap of the drum or barrel (lot number is not included). The beers were stored in a refrigerator at 5.5 • C until analysis in 500 mL amber glass bottles.

Characterization of the Boron-Doped Diamond Electrode by Scanning Electron Microscope and
Electrochemically Using the K 3 [Fe(CN) 6 ]/K 4 [Fe(CN) 6

] Redox System
The morphology of the BDD electrode was characterized by SEM, Thermo Scientific Phenom ProX brand; for this, the BDD electrode, clean and dry, was deposited in the sample holder and analyzed at a voltage of 15 kV.
For the electrochemical characterization, a 3-electrode heart cell was used: BDD was used as the working electrode, the Ag/AgCl was used as the reference electrode, and a graphite bar was used as the counter electrode. The electrode was cleaned with 0.2 mol L −1 HNO 3 by cyclic voltammetry (CV) using a Metrohm Autolab B.V. potentiostat. Then a 1 mol L −1 KCl solution was prepared and adjusted to pH = 1 with 37% m/m HCl, in which a working window was found by CV, with a scanning speed of 100 mV s −1 , once the working window was found, 30 cycles were run to condition the electrode before each run, in all cases.
Double layer capacitance (C cl ) was then obtained by CV, in 1 mol L −1 KCl and pH = 1 (adjusted with HCl as indicated above), in a window from 0.5 V to 1.2 V 6 scan rates were applied: 20, 40, 60, 80, 100 and 120 mV s −1 . The capacitance was determined by applying Equation (1). Where ν = sweep speed and J p is the current density J p = i/A (i is the current and A is the area of the electrode).
Subsequently, BDD's electrochemical response was studied against the redox couple K 3 [Fe(CN) 6 ]/K 4 [Fe(CN) 6 ] 4 mmol L −1 , in KCl 1 mol L −1 as electrolyte adjusted to pH = 1 with HCl, and different ν: 20, 40, 60, 80, 100 y 120 mV s −1 . The Randles-Sevcik equation was applied, which describes the effect of the ν at maximum response current or peak current I p (I p vs. ν 1/2 ), Equation (2), where I p : is the maximum current, n: number of electrons transferred in the redox process, A: is the area of the electrode, F: Faraday's constant, D: diffusion coefficient, C: concentration). In addition, ∆E p = (Ep ox − Ep red ), half-wave potential (Ep 1/2 = (Ep ox − Ep red )/2), and the ratio (Ip ox /Ip red ) were evaluated. In addition, the electron exchange rate constant was assessed, k • according to the Nicholson equation, Equation (3) (where k • : rate constant of electron exchange, Ψ: function of the heterogeneous rate constant of the electron, D 0 : diffusion constant of the chemical species, F: constant of Faraday, υ: sweep speed, R: universal gas constant, T: temperature (25 • C) [35].

Preparation of the Craft Beer Sampler
Preparation of the beer sample followed the following steps: (i) degassing 25 mL of craft beer by ultrasound (BRANSON 1800 brand) at 30 • C for 20 min; (ii) acid digestion of the degassed craft beer sample with 5 mL of 65% m/m concentrated HNO 3 and 2 mL of 30% m/m H 2 O 2 . Heating at 100 • C on the MTOP MS300HS heating plate until a yellow coloration is achieved; (iii) the digestion product was measured, with the corresponding support electrolyte solution (the type of electrolyte, the concentration, and the pH depending on the metal to be determined, see Section 2.4.1), up to a final volume of 25 mL. The final pH was controlled by adding 2 mol L −1 NaOH and using the SensionTM MM374 pH meter potentiometer; (iv) prior to the measurements, it was purged for 10 min with 99.99% N 2 to eliminate O 2 molecules that could cause interference in the analysis. To evaluate Cd(II), 0.1 mol L −1 acetic acid/0.055 mol L −1 sodium acetate at pH 4.5 was used as an electrolyte, according to what was reported by Loaiza (2020) [30].
To evaluate Fe(III) were tested: (i) 0.1 mol L −1 sodium citrate, (ii) 0.1 mol L −1 EDTA and (iii) 0.1 mol L −1 KNO 3 . Once the electrolytes with the best responses for each analyte were selected using DPASV, a study of the effect of pH was carried out using different concentrations of acid: in HNO 3 0.01 mol L −1 (pH = 2.10), HNO 3 0.1 mol L −1 (pH = 1.20), and no acid (pH = 7.00).
Appropriate electrolytes for Cu(II) and Fe(III) were chosen by DPASV using a Metrohm Autolab B.V potentiostat.

Determination of Optimal Differential Pulse Adsorption Stripping Voltammetry Parameters
Standard analyte solutions of concentration 100 µg L −1 of Cd(II), 90 µg L −1 of Cu(II), and 80 µg L −1 of Fe (III) were prepared in the medium of selected support electrolyte in the Section 2.4.1. DPASV was applied to each solution in the potential range of −2.4 to 1.0 V. Modulation amplitude (MA), modulation time (MT), and time interval (TI) properties were varied from 0.05 V, 0.05 s, and 0.05 s (respectively) until a well-defined signal is achieved. In addition, the pre-concentration potential and the pre-concentration time of DPASV were evaluated.

Construction of the Calibration Plot and Determination of the Detection and Quantification Limit
Standard solutions of Cd(II), Cu(II), and Fe(III) were prepared; of concentration 40, 80, 120, 160, and 200 µg L −1 for Cd(II), and 30, 60, 90, 120, and 150 µg L −1 for Cu(II) and Fe(III), for separated, in supporting electrolyte selected in Section 2.4.1. The current intensity signal was measured by DPASV of each solution in the potential range of −2.0 to −0.5 V after measuring the blanks. Then the analyte oxidation peaks were detected, and the calibration plots expressed as current intensity versus analyte concentration in µg L −1 were performed. Subsequently, the equation of the plot for each analyte was determined.
Then 10 support electrolyte solutions were prepared, and the current intensity signal was measured by DPASV of each solution at the same potential detected for the current peaks of the calibration plot. Once the readings were obtained, the detection limit and quantification limit (DL and QL) were calculated. To evaluate the repeatability, the samples were run in triplicate by means of DPASV on the same day (total 9 determinations). With the current values obtained and the equations of the linear regressions by standard addition (Section 2.4.5), the concentrations of the sample + added standard (Ym + x) were evaluated. Subsequently, the standard deviation (S d ) and the percentage of the relative standard deviation (RSD%) were calculated.
To evaluate the reproducibility, the samples were run in triplicate on 3 different days (total of 27 determinations) using DPASV. To find the sample concentration, Sd, and RSD %, the above procedure was followed.
To evaluate the recovery percentage (R%), the concentrations obtained in repeatability and reproducibility (Y m+x ), and the respective percentages were calculated according to Samples of craft beer (digested) fortified with analyte standards (Section 2.4.5) and measured with the corresponding electrolyte were evaluated in triplicate. DPASV was applied, and the current intensity peaks of each solution were measured. Current intensity peaks (I) of the voltammograms were measured, and by linearization I vs. C, the equation of the plot was determined, which allowed the calculation of the metal concentration in the beer. The process was repeated for the different craft beers (n = 13).

Determination of the Concentration of Heavy Metals Cd, Cu, and Fe in Craft Beers by Flame Atomic Absorption Spectrometry
FAAS was applied to the craft beer samples because it is the most widely used technique for the quantification of heavy metals due to its ability to determine more than 70 elements in solution and in different matrices [36]. The beer samples were prepared in the same way as for the electrochemical method, as described in Section 2.3. Standard addition calibration plots were prepared in the same way as for DPASV. However, other linear ranges were applied due to the difference in sensitivity of the FAAS equipment (Perkinelmer brand AAnalist 400) compared to the DPASV method. Regarding the calibration of Cd(II), a linear range of 0.01 to 1 mg L −1 was used; for Cu(II) from 0.05 to 1.5 mg L −1 and for Fe(III) from 0.3 to 5 mg L −1 . The DL values for Cd, Cu, and Fe are 0.0045, 0.0164, and 0.1337 mg L −1 , respectively; the QL values for Cd, Cu, and Fe are 0.0151, 0.0546, and 0.4457 mg L −1 , respectively. Cd was measured at a wavelength (λ) of 228.8 nm, Cu at 324.75 nm, and Fe at 302.06 nm, and three measurements of each type of sample were made. Once the readings were obtained, the arithmetic mean, the standard deviation, and the percentage of the real standard deviation were calculated. In addition, Student's "t" was calculated using SPSS Statistics to compare whether there are significant differences between the two applied methods (DPASV vs. FAAS) for the quantification of the 3 heavy metals (Cd, Cu, and Fe). Figure 1 shows the granular morphology of the BDD electrode by SEM; it presents microcrystals with an average size between 300 and 2000 nm. A surface free of impurities is observed. The crystals present a relatively uniform distribution; they do not present holes or cracks on their surface; that is, the BDD presents a characteristic micrograph of a clean electrode and is suitable for use in electroanalysis [37]. The study by energy dispersive spectrometer (EDS) of the BDD yields a composition: carbon 94.11% atomic, 94.26% w/w; boron 4.91% atomic, 4.43% w/w; 0.98% atomic oxygen, 1.31% w/w, 0.052 B/C. The B/C value recorded in this study is within the range determined by Xu and Einaga (2020), from 0.03 to 2.20, which is characteristic of boron-doped diamond electrodes [38,39].

Microscopic Characterization of the Boron-Doped Diamond Electrode
is observed. The crystals present a relatively uniform distribution; they do not present holes or cracks on their surface; that is, the BDD presents a characteristic micrograph of a clean electrode and is suitable for use in electroanalysis [37]. The study by energy dispersive spectrometer (EDS) of the BDD yields a composition: carbon 94.11% atomic, 94.26% w/w; boron 4.91% atomic, 4.43% w/w; 0.98% atomic oxygen, 1.31% w/w, 0.052 B/C. The B/C value recorded in this study is within the range determined by Xu and Einaga (2020), from 0.03 to 2.20, which is characteristic of boron-doped diamond electrodes [38,39].  Figure 2a shows the characteristic electrochemical behavior of the BDD when its surface is clean. A working window potential is observed between −1.15 V and 1.60 V in 1 mol L −1 KCl as electrolyte at pH = 1 (adjusted with HCl 1 M); the work window is similar to the report done by Bogdanowicz et al. (2020), −1.5 to 1.5 V [40]. Therefore, in this potential interval, any electroanalytical study can be carried out without interference from the medium [41,42]. In addition, the electrode response was evaluated against the K4[Fe(CN)6] 4 mmol L −1 -K3[Fe(CN)6] 4 mmol L −1 system; the oxidation and reduction signals appear at 0.42 and 0.32 V, respectively. These responses are characteristic of BDD when it is clean and ready to be used in electroanalysis [43,44]. Figure 2b shows the electrochemical response of the BDD at different scan rates (v) in 1.0 mol L −1 KCl at pH = 1, in a potential interval of 0.5 V to 1.2. V. It is observed that as v increases, the capacitive current increases. To determine the double layer capacitance, a potential of 1.2 V was set, and a plot was constructed, maximum currents (Ip) vs. v (mV s −1 ), insert in Figure 2b. From the slope of the linear fit, the double layer capacitance (Cdl) was calculated according to Equation (1). A value of Cdl = 0.01412 μF cm −2 was obtained; this value corresponds to that reported in the literature for the case of BDD with a low concentration of C-sp 2 [45].   Figure 2a shows the characteristic electrochemical behavior of the BDD when its surface is clean. A working window potential is observed between −1.15 V and 1.60 V in 1 mol L −1 KCl as electrolyte at pH = 1 (adjusted with HCl 1 M); the work window is similar to the report done by Bogdanowicz et al. (2020), −1.5 to 1.5 V [40]. Therefore, in this potential interval, any electroanalytical study can be carried out without interference from the medium [41,42]. In addition, the electrode response was evaluated against the K 4 [Fe(CN) 6 ] 4 mmol L −1 -K 3 [Fe(CN) 6 ] 4 mmol L −1 system; the oxidation and reduction signals appear at 0.42 and 0.32 V, respectively. These responses are characteristic of BDD when it is clean and ready to be used in electroanalysis [43,44]. Figure 2b shows the electrochemical response of the BDD at different scan rates (v) in 1.0 mol L −1 KCl at pH = 1, in a potential interval of 0.5 V to 1.2. V. It is observed that as v increases, the capacitive current increases. To determine the double layer capacitance, a potential of 1.2 V was set, and a plot was constructed, maximum currents (I p ) vs. v (mV s −1 ), insert in Figure 2b. From the slope of the linear fit, the double layer capacitance (C dl ) was calculated according to Equation (1). A value of C dl = 0.01412 µF cm −2 was obtained; this value corresponds to that reported in the literature for the case of BDD with a low concentration of C-sp 2 [45]. Figure 2c shows the electrochemical response of BDD, at different v, against the redox couple K 4 [Fe(CN) 6 6 ] in KCl 1 mol L −1 pH = 1. In the insert of Figure 2c, the I p increases linearly with υ 1/2 , characteristic behavior of BDD when its surface is clean and there are no diffusional complications. The standard rate constant (k • ) was determined with the Nicholson equation [35], Equation (3), using an average value of ∆E of all the scan rates studied. The average value obtained from k • = 2.44 × 10 −2 ± 4.67 × 10 −3 cm s −1 , and it is similar to the values obtained by Rehascek et al. (2020), 1.01 × 10 −2 to 3.60 × 10 −3 cm s −1 [46]. This indicates that the redox process is fast and quasi-reversible [47][48][49]; see Table S1.

Electrochemical Characterization of the Boron-Doped Diamond Electrode
there are no diffusional complications. The standard rate constant (k°) was determined with the Nicholson equation [35], Equation (3), using an average value of ΔE of all the scan rates studied. The average value obtained from k° = 2.44 × 10 −2 ± 4.67 × 10 −3 cm s −1 , and it is similar to the values obtained by Rehascek et al. (2020), 1.01 × 10 −2 to 3.60 × 10 −3 cm s −1 [46]. This indicates that the redox process is fast and quasi-reversible [47][48][49]; see Table S1. Once the KNO3 electrolyte was selected, the pH was changed by adding HNO3 in order to improve the response current signal in the detection of Cu(II) by DPASV. Figure  3b   In order to improve the Fe(III) signal, different concentrations of HNO3 were assessed for the selected electrolyte, KNO3. Figure 4b shows the DPASV voltammogramas of 80 μg L −1 of Fe(III) in: KNO3 0.1 mol L −1 , KNO3 0.1 mol L −1 /HNO3 0.01 mol L −1 (pH = 2.10) and KNO3 0.1 mol L −1 /HNO3 0.1 mol L −1 (pH = 1.20) [50,51]. The DPASV parameters were the same used previously. A more defined and more intense current signal was achieved for the case of KNO3 0.1 mol L −1 /HNO3 0.01 mol L −1 , between −1.78 V to −1.47 V, so this support electrolyte was selected for the determination of Fe(III).   The parameters of DPASV at the BDD electrode for the quantification of Cd(II) 100 μg L −1 of Cd(II) was used in electrolyte support of 0.1 mol L −1 acetic acid/sodium acetate pH 4.5. In Figure 5a, a cadmium stripping signal between −1.4 V to −0.7 V is observed.
In order to improve the Cd(II) response signal, the MA, MT, and TI parameters were varied from 0.05 V, 0.05 s, and 0.05 s to 0.5 V, 0.5 s, and 0.5 s, respectively. When the values of MA, MT, and TI were larger, a stronger and more defined stripping signal was Once the KNO 3 electrolyte was selected, the pH was changed by adding HNO 3 in order to improve the response current signal in the detection of Cu(II) by DPASV. Figure 3b reports   Figure 4a shows the voltammograms of 80 μg L −1 of Fe(III) in different electrolytes, using DPASV, with the following parameters: MA 0.1 V, MT 0.1 s, TI 0.1 s, pre-concentration of −1.5 V for 1 min. An oxidation signal was achieved in KNO3 between 1.20 V and −0.90 V. No signal was achieved in the other electrolytes [50].
In order to improve the Fe(III) signal, different concentrations of HNO3 were assessed for the selected electrolyte, KNO3. Figure 4b shows the DPASV voltammogramas of 80 μg L −1 of Fe(III) in: KNO3 0.1 mol L −1 , KNO3 0.1 mol L −1 /HNO3 0.01 mol L −1 (pH = 2.10) and KNO3 0.1 mol L −1 /HNO3 0.1 mol L −1 (pH = 1.20) [50,51]. The DPASV parameters were the same used previously. A more defined and more intense current signal was achieved for the case of KNO3 0.1 mol L −1 /HNO3 0.01 mol L −1 , between −1.78 V to −1.47 V, so this support electrolyte was selected for the determination of Fe(III).   The parameters of DPASV at the BDD electrode for the quantification of Cd(II) 100 μg L −1 of Cd(II) was used in electrolyte support of 0.1 mol L −1 acetic acid/sodium acetate pH 4.5. In Figure 5a, a cadmium stripping signal between −1.4 V to −0.7 V is observed.
In order to improve the Cd(II) response signal, the MA, MT, and TI parameters were varied from 0.05 V, 0.05 s, and 0.05 s to 0.5 V, 0.5 s, and 0.5 s, respectively. When the values of MA, MT, and TI were larger, a stronger and more defined stripping signal was In order to improve the Fe(III) signal, different concentrations of HNO 3 were assessed for the selected electrolyte, KNO 3 . Figure 4b shows the DPASV voltammogramas of 80 µg L −1 of Fe(III) in: KNO 3 0.1 mol L −1 , KNO 3 0.1 mol L −1 /HNO 3 0.01 mol L −1 (pH = 2.10) and KNO 3 0.1 mol L −1 /HNO 3 0.1 mol L −1 (pH = 1.20) [50,51]. The DPASV parameters were the same used previously. A more defined and more intense current signal was achieved for the case of KNO 3 0.1 mol L −1 /HNO 3 0.01 mol L −1 , between −1.78 V to −1.47 V, so this support electrolyte was selected for the determination of Fe(III).

Optimal Differential Pulse Adsorption Stripping Voltammetry Parameters for the Determination of Cd(II), Cu(II), and Fe(III)
The parameters of DPASV at the BDD electrode for the quantification of Cd(II) 100 µg L −1 of Cd(II) was used in electrolyte support of 0.1 mol L −1 acetic acid/sodium acetate pH 4.5. In Figure 5a, a cadmium stripping signal between −1.4 V to −0.7 V is observed.
achieved. However, the parameters MA = 0.4 V, MT = 0.4 s, and TI = 0.4 s were chosen for the quantification of Cd(II), lower background current is generated with these parameters. Figure 5b shows the effect of the pre-concentration potential of Cd(II) in the previously selected support electrolyte, with 100 μg L −1 of Cd(II); at higher potential values, current peaks are more intense. However, when −1.6 V and −1.7 V pre-concentration are applied, the responses are similar, so −1.6 V was chosen for the sample reading. Figure 5c shows the effect of the pre-concentration time of Cd(II) 100 μg L −1 , from 15 s to 180 s in the previously selected electrolyte. Little variation in the intensity of the peak was observed, so 15 s was selected as the pre-concentration time. This time will allow us to have fast measurements, which is what is sought in analytical methods. Similarly to Cd(II), we proceeded with Cu(II) and Fe(III) in order to find the optimal parameters: In the case of Cu(II), 90 μg L −1 , in the previously selected support electrolyte, KNO3 0,1 mol L −1 /HNO3 0,1 mol L −1 , a potential window from −1.7 to −1.0 V was obtained.
In order to improve the Cu(II) response signal, the MA, MT, and TI parameters were varied from 0.05 V, 0.05 s, and 0.05 s to 0.3 V, 0.3 s, and 0.3 s, respectively. When the values of MA, MT, and TI were larger, the sharpest current signal was achieved. However, at values of MA = 0.3 V, MT = 0.3 s, and TI = 0.3 s, the oxidation signal is distorted, a behavior that does not occur with parameters of MA = 0.2 V, MT = 0, 2 s, and TI = 0.2 s, these being the parameters chosen. Furthermore, these parameters yield a lower background current (see Figure S1a). On the other hand, different pre-concentration voltages were applied from −0.8 V to −1.1 V; it was found that, at higher potential values, the Cu(II) current intensity peak is more intense and defined down to −1.0 V, at −1.1 V a similar signal was obtained. Therefore, −1.0 V was chosen for the pre-concentration (see Figure S1b). Finally, the effect of the Cu pre-concentration time in a range from 20 s to 240 s was analyzed; little In order to improve the Cd(II) response signal, the MA, MT, and TI parameters were varied from 0.05 V, 0.05 s, and 0.05 s to 0.5 V, 0.5 s, and 0.5 s, respectively. When the values of MA, MT, and TI were larger, a stronger and more defined stripping signal was achieved. However, the parameters MA = 0.4 V, MT = 0.4 s, and TI = 0.4 s were chosen for the quantification of Cd(II), lower background current is generated with these parameters. Figure 5b shows the effect of the pre-concentration potential of Cd(II) in the previously selected support electrolyte, with 100 µg L −1 of Cd(II); at higher potential values, current peaks are more intense. However, when −1.6 V and −1.7 V pre-concentration are applied, the responses are similar, so −1.6 V was chosen for the sample reading. Figure 5c shows the effect of the pre-concentration time of Cd(II) 100 µg L −1 , from 15 s to 180 s in the previously selected electrolyte. Little variation in the intensity of the peak was observed, so 15 s was selected as the pre-concentration time. This time will allow us to have fast measurements, which is what is sought in analytical methods.
Similarly to Cd(II), we proceeded with Cu(II) and Fe(III) in order to find the optimal parameters: In the case of Cu(II), 90 µg L −1 , in the previously selected support electrolyte, KNO 3 0.1 mol L −1 /HNO 3 0.1 mol L −1 , a potential window from −1.7 to −1.0 V was obtained.
In order to improve the Cu(II) response signal, the MA, MT, and TI parameters were varied from 0.05 V, 0.05 s, and 0.05 s to 0.3 V, 0.3 s, and 0.3 s, respectively. When the values of MA, MT, and TI were larger, the sharpest current signal was achieved. However, at values of MA = 0.3 V, MT = 0.3 s, and TI = 0.3 s, the oxidation signal is distorted, a behavior that does not occur with parameters of MA = 0.2 V, MT = 0.2 s, and TI = 0.2 s, these being the parameters chosen. Furthermore, these parameters yield a lower background current (see Figure S1a). On the other hand, different pre-concentration voltages were applied from −0.8 V to −1.1 V; it was found that, at higher potential values, the Cu(II) current intensity peak is more intense and defined down to −1.0 V, at −1.1 V a similar signal was obtained. Therefore, −1.0 V was chosen for the pre-concentration (see Figure S1b). Finally, the effect of the Cu pre-concentration time in a range from 20 s to 240 s was analyzed; little variation in the intensity of the peak was observed, so 60 s was selected because, at this time, it generated a better-defined signal (see Figure S1c).
In In addition, at these values, a lower background current is obtained (see Figure S2a). On the other hand, the pre-concentration potential was evaluated from −0.1 V to −1.8 V; at −1.6 V, the sharpest current signal was achieved (see Figure S2b). Finally, the study of the pre-concentration time of Fe(II) was carried out from 20 s to 180 s; little variation in the intensity of the peak was achieved, so 60 s was selected; in this time, the sharpest current signal was achieved (see Figure S2c).  Figure 6c). Figure S3a-c shows the calibration plots for Cd(II), Cu(II), and Fe(III) (respectively); they show a linear behavior for the three metals. Table 2 Table 2.
mol L −1 acetic acid/sodium acetate at pH 4.5, in the potential range from −1.4 V to −1.0 V, see Figure 6a. For Cu(II), we worked in concentrations of 30 to 150 μg L −1 in electrolyte support KNO3 0.1 mol L −1 /HNO3 0.1 mol L −1 , in the potential range of −1.7 V at −1.0 V, Figure 6b. In the case of Fe(III), we worked at concentrations of 30 to 150 μg L −1 of Fe(III) in a supporting electrolyte KNO3 0.1 mol L −1 /HNO3 0.01 mol L −1 , in the potential range of −1.4 to −0.8 V (see Figure 6c).   In the case of Cd, DPASV was applied to previously digested, fortified, and graduated craft beer samples with 200 µg L −1 of Cd(II) in the electrolyte acetic acid 0.1 mol L −1 /sodium acetate 0.055 mol L −1 at pH 4.5, the signal was generated between −1.8 V to −1.0 V, with broader peaks than those generated without matrix effect (−1.4 V to −1.0 V) (see Figure 6a and Figure S5a). To improve the signal, the parameters MA, MT, and TI were modified with respect to the pre-selected ones without matrix effect. MA, MT, and TI were varied from 0.5 V, 0.5 s, and 0.5 s to 0.7 V, 0.7 s, and 0.7 s, respectively; in these last values, the best signal with effect was obtained. Matrix for Cd(II) quantification (see Figure S5a).
In the case of Cu, 90 µg L −1 of Cu(II) was used in 0.1 mol L −1 KNO 3 /0.1 mol L −1 HNO 3 in previously digested craft beer samples. When applying DPASV, Cu(II) signals between −1.8 V to −0.7 V were obtained, in the same way, with broader peaks than those generated without matrix effect (−1.7 V to −1.0 V) (see Figure 6b and Figure S5b). To improve the signal, MA, MT, and TI were varied from 0.5 V, 0.5 s, and 0.5 s to 0.8 V, 0.8 s, and 0.8 s, respectively. When the values of MA, MT, and TI were larger, a stronger and more defined current signal was achieved. However, the parameters MA = 0.7 V, MT = 0.7 s, and TI = 0.7 s were chosen for the quantization of Cu since, with these parameters, there is a lower current of background (see Figure S5b).
Finally, in the case of Fe, 80 µg L −1 of Fe (III) was used in 0.1 mol L −1 KNO 3 / 0.01 mol L −1 HNO 3 in previously digested craft beer samples in which when applying DPASV. Fe(III) signal appeared between −1.95 to −0.80 V, with broader peaks than those generated without the matrix effect (−1.4 V to −0.8 V) (see Figure 6c and Figure S5c). To improve the signal, MA, MT, and TI were varied from 0.4 V, 0.4 s, and 0.4 s to 0.7 V, 0.7 s, and 0.7 s, respectively. When the values of MA, MT, and TI were larger, a stronger and more defined current signal was achieved, see Figure S2a. Therefore, these last parameters were selected as optimal for the quantification of Fe in the solutions prepared from craft beers (see Figure S5c).

Standard Addition Plot of Cd(II), Cu(II), and Fe(III)
Having defined the optimal parameters for the electrochemical detection of Cd, Cu, and Fe in craft beer, the calibration plot was constructed in the concentration interval from 0 to 200 µg L −1 for Cd(II) (see Figure S5a) and in the interval from 0 to 150 µg L −1 for Cu(II) and Fe(III) (see Figure S5b,c), Table 3 summarizes the parameters of the calibration plots, with and without matrix effect. When comparing the calibration plots for Cd(II), Cu(II), and Fe(III), a higher slope is identified when there is a matrix effect, Table 3. The coefficient of determination varies slightly without and with the food matrix, Table 3. Once the calibration plots were established by standard addition, the repeatability, reproducibility, and percentage recovery of the analytes by DPASV were evaluated. Table 4 reports the percentage values of relative standard deviation (RSD%) in three days, repeatability, and reproducibility of the DPASV method expressed RSD% for each element. In Figure 7a-c, box and whisker diagrams of the recovery percentage are shown, where it is observed that by increasing the concentration by standard addition of Cd(II), Cu(II), and Fe (III), there is a tendency to recover more of the analyte. In the interday precision of the measurements (3 days of evaluation), a higher percentage of recovery (R%) was obtained for the addition of 200 µg L −1 Cd(II), 150 µg L −1 Cu(II), and 150 µg L −1 Fe(III). While a lower recovery in the standard addition of 40 µg L −1 of Cd(II), 30 µg L −1 of Cu(II), and 30 µg L −1 of Fe(III). It can also be seen that the recovery tends to vary, increasing slightly as the days of repetition of the analysis progress. Globally, during the days of evaluation, the average R% of Cd(II), Cu(II), and Fe(III) was 94.94%, 91.68%, and 96.79%, respectively. The results indicate that the method for the determination of these three metals is accurate in time (3 days) since the values are within the acceptable range (80-120%) [52,53]. The recovery percentages obtained are similar to those reported for heavy metals by Tefela and Ayele (2020), 90.9 to 104.3% [54].   Table 5 shows the average concentrations of Cd, Cu, and Fe in craft beers obtained by DPASV. The results show some samples with the presence of metals that are not suitable are reported for human consumption because their metal concentrations exceed the limits allowed by the Ecuadorian norm NTE INEN 2263 (concentrations of iron and copper < 0.2 mg L −1 , <1.0 mg L −1 , respectively) [33]. Beer regulations and Brazilian legislation on heavy metals for fermented alcoholic beverages [34] establish cadmium concentrations < 0.5 mg L −1 ; the beers that do not comply with the regulations are CAR-A, CAR-E, CAR-H, CAR-H, CAR-K, and CAR-M. On the contrary, the craft beers CAR-B, CAR-C, CAR-D, CAR-F, CAR-G, CAR-I, CAR-J, and CAR-L, although they contain heavy metals, do not exceed the limits allowed by the reference regulations.         Student's "t" statistical test between the DPASV and FAAS methods in determining Cd resulted in a statistical "t" of 2.080, a value that does not exceed the critical "t" of 2.18 for 12 degrees of freedom and for two tails. In the Student's t-test between the DPASV and FAAS methods in the determination of Cu, the t statistic was 1.649, a value that also did not exceed the critical t of 2.18 for 12 degrees of freedom and for two tails; that is, in both cases (Cd and Cu), there is no statistically significant difference between methods.

Discussion
According to Figure 1, the BDD crystals are on the nanometric scale (300 to 2000 nm), which allows for a larger contact surface and it increases the number of active sites where electrochemical reactions can take place [55,56]. Proper cleaning allows residual carbon impurities to be removed from the BDD surface, which reduces possible interferences and increases the reproducibility of analytical signals. The low percentage of oxygen, close to 1%, determined by EDX, is related to the low amount of C-sp 2 bonds, suitable for electroanalytical studies. In addition, in this study, 4.91 atomic % boron, 4.43 % by weight, was obtained by EDX, which confers conductivity to the diamond electrode.
The working potential window using cyclic voltammetry is the potential interval in which neither the electrolyte nor the solvent reacts. For the BDD electrode in KCl 1 mol L −1 , the window was from −1.15 V to 1.60 V; this interval agrees with what has been reported in the literature [29,45,57].
Double layer capacitance (C dl ) is the amount of charge that an electrode stores when it is polarized. When the C dl has a high value, the capacitive current overlaps the faradaic current during the measurement, which is very detrimental when working with analytes at trace levels. The value obtained for the BBD electrode in this study was 0.01412 µF cm −2 , relatively low according to what was reported by Kim et al. (2013) for the BDD electrode, 15.2 µF cm −2 [45,58]. The C dl is directly linked to the amount of C-sp 2 , and with the boron doping level of the BDD, a higher concentration of C-sp 2 , a higher C dl is expected [59]. It is furthermore considered that the low double layer electrical capacitance of BDD electrodes derives from the low density of electronic states (DOS) at the Fermi level [60].
One of the most widely used redox couples to characterize carbonaceous materials electrode is K 3 [Fe(CN) 6 ]/K 4 [Fe(CN) 6 ]. In the BDD electrode, this redox pair presents characteristic responses, depending on its surface termination; when the surface of BDD contains abundant oxygen (large amount of C-sp 2 ), the redox process is slow and irreversible, ∆E takes large values, the opposite occurs when the surface is less oxygenated [61][62][63]. In the insert of Figure 4, the linear behavior of plot I vs. ν 1/2 can be verified that the process is purely diffusional, and there is no adsorption or other adjacent phenomena, according to the Randles-Sevcik equation [64,65].
On the other hand, k • = 2.44 × 10 −2 ± 4.67 × 10 −3 cm s −1 was determined with the Nicholson equation [35], Equation (3) [48,67], the recommended v to study kinetic parameters are from 20 mV s −1 to 100 mV s −1 , since at v > 100 mV s −1 , the capacitances substantially influence the shape of the voltammogram. On the contrary, at v < 20 mV s −1 , the species formed diffuse into the bulk, and reversibility in the electrochemical response is lost when the reverse process is run [46,68].
When Ip ox /Ip red is equal to 1, it is considered a fast and reversible redox process; when it is close to 1, it is considered a quasi-reversible redox process; and when it is very different from 1, the process is considered slightly reversible or irreversible [69]. In this study, Table S1, a value of Ip ox /Ip red = 0.99 was obtained; that is, the redox process for the potassium ferro-ferricyanide system in BDD is quasi-reversible.
One of the key variables in electroanalysis to achieve a well-defined and sharpest current signal is the electrolyte used and the pH. In this study, the best signals were achieved using 0.1 mol L −1 acetic acid/0.055 mol L −1 sodium acetate at pH 4.5; KCl 0.1 mol L −1 /HCl 0.01 mol L −1 and KNO 3 0.1 mol L −1 /HNO 3 0.01 mol L −1 to quantify Cd(II), Cu(II) and, Fe(III) by DPASV, respectively. The reaction medium directly influences the solvation sphere of the analyte and its activation energy when electrochemically inducing a transformation [30,50,51].
It was evidenced that the response signals do not grow symmetrically when there is an effect of the food matrix; this is due to the presence of analytical interference typical of the matrix [70].
The DL and QL for Cd(II), Cu(II), and Fe(III) by means of DPASV in the BDD electrode are low values and adequate for chemical elements in low concentrations, such as in craft beers. On the other hand, the DL and QL are lower for Fe(III) (DL = 1.72, QL = 5.72, µg L −1 ) compared to Cu(II) (DL = 1.76, QL = 5.87, µg L −1 ) and much lower with respect to Cd(II) (DL = 6.31, QL = 21.04, µg L −1 ). The DL of Cu and Fe are relatively lower than those reported by Passaghe (2015) [71] and Bonilla (2022) [72] (4.12 µg L −1 and 36.0 µg L −1 , respectively); while the DL of Cd is relatively higher than that reported by Marcano (2010) [73] (1.8 µg L −1 ), however, this reflects that the method is acceptable.
Despite not obtaining Fe values in craft beers by FAAS, the results obtained for Fe by DPASV show that there is a presence of the metal, indicating greater sensitivity of the DPASV method, and confirm, in addition, that the electrochemical method is appropriate to detect elements in very low concentrations.
The DPASV method applied on the BDD electrode, having RSD% below 10% and recoveries above 90%, therefore, has acceptable precision and accuracy, respectively, for the quantification of these heavy metals in the ionic state [74].
The concentrations of Cd reported in this study (cervezas de Quito, Ecuador) (0.0083-0.0910 mg L −1 of Cd(II)) are lower than the results of the study by Becerra (2014) [26] in craft beers from Cuenca-Ecuador, where values higher than 0.1 mg L −1 were found for Cd in blonde type beer. The concentrations of Cu (0.1339-0.4660 mg L −1 of Cu) were also in a lower range compared to those of Becerra, where concentrations greater than 17 mg L −1 Cu were found in black type. Regarding the Fe content (0.1250-0.3159 mg L −1 ) from this study, it was lower than that reported by Zambrzycka-Szelewa et al. (2020) [75] in craft beers from Bialystok-Poland, range of 0.0624-1.199 mg L −1 . As can be seen, the concentration of heavy metals is variable depending on the origin, type of beer, or the way in which they were made during production.
From this study, five beers have a concentration above the permissible limits of the food regulation NTE INEN 2262 [33] (0.2 mg L −1 ), so the consumption of these drinks is not accepted. Since iron is an essential element of the human diet, the intake between 10-20 mg day −1 [76,77] is recommended; however, it should not exceed 20 mg day −1 since it can cause damage to health, as stated by Ho Wang (2022) [78].
In the evaluated beers, higher concentrations of heavy metals were found in those with a high degree of alcohol; in addition, it was possible to demonstrate a high Fe(III) content in the CAR-A, CAR-E, CAR-H, CAR-K, and CAR M beers and that within their brewing ingredients, they contained caramelized barley malt, and sometimes roasted barley (CAR-A, CAR-E, CAR-H). The extract from roasted barley involves applying hot water favoring the solubility of the metal. In this sense, it can be corroborated that the ingredients used for the elaboration of the craft beer give the presence and increase of heavy metals. In addition, in an aqueous medium, the high alcoholic degree leads to certain biomolecules of the food matrix dissolving, releasing together with them the metal in an ionic state (Sancho, 2010). Sancho (2010) describes that the concentrations of free iron were lower for those that did not contain alcohol, and the composition of ingredients such as roasted barley influenced the concentration of the metal in the dark beers in their study [28].

Conclusions
The DPASV electrochemical method using BDD as the working electrode allowed the evaluation of Cd(II), Cu(II), and Fe(III) in craft beers. The selection of supporting electrolytes, pH range, and voltage modulation parameters play an important role in obtaining a defined signal for metal quantification. The DPASV method applied on the BDD electrode has acceptable precision and accuracy for the quantification of these heavy metals in the ionic state, which corroborates its good repeatability and reproducibility of the method used compared to other electrochemical studies [27,28,79]. The thirteen Quito craft beers have amounts of Cd (II), Cu (II), and Fe (III) that in most brands tend to be low concentration values; the concentration ranges were 0.0083-0.0910 mg L −1 of Cd(II), 0.1339-0.4660 mg L −1 of Cu(II) and 0.1250-0.3159 mg L −1 of Fe(III). On the other hand, the content of Cd and Cu quantified in the 13 craft beers does not represent a risk to the health of consumers. However, it was verified that some of the beers did not comply with the permissible limits of the NTE INEN 2262 standards. Five of the thirteen brands were outside the acceptable concentration limit in Fe(III); these beers were those coded as