Speciation of Inorganic Arsenic in Retorted Oil Shale Slurry by Hydride Generation Atomic Absorption Spectrometry (HG AAS) after Ultrasound-Assisted Acid Extraction

Abstract An analytical method for determination of inorganic arsenic species in retorted oil shale slurry by atomic absorption spectrometry coupled to hydride technique with ultrasound assisted acidic extraction is reported. Trivalent arsenic measurement employed sample stabilization with a pH 4.5 prior to reaction with the reducing agent and analysis. In contrast, total inorganic As depended upon the pre-reduction of As(V) to As(III) with KI and ascorbic acid in 6.0 M HCl and of As(V) was estimated by the difference of the total inorganic As and As(III) concentrations. The limit of quantification was 3.7 mg kg−1 for As(III) and 1.3 mg kg−1 for itAs.


Introduction
The pyrobetum shale, informally known as shale, is rock of sedimentary origin containing kerogen, which can be extracted as oil and gas under conditions of high temperature and high pressure (process of pyrolysis or retort).The kerogen consists of long condensed aromatic chains, alkanes and fatty acids.Its mineral composition varies in concentrations of quartz, feldspar, clays (mainly ilite and chlorite), carbonates (calcite and dolomite), pyrite and other minerals (Na et al. 2012;Snape 1993;Speight 2012).Brazil is the second largest producer of shale, behind only the Green River in the United States (Altun et al. 2006;Han et al. 2014;Pimentel et al. 2006).
The chemical and mineralogical composition of retorted oil shale makes this by-product a promising raw material for the formulation of agricultural inputs (Martinazzo et al. 2020;Ribas 2012) Ground rocks are used to remineralize soils (also known as 'rocks for crops') (Van Straaten 2002, 2007) in order to improve their fertility profile.
Several studies have demonstrated that retorted oil shale may contribute to soil and plant fertility, both as a source of organic matter and of essential elements, especially when the use is combined with fertilizers and soil conditioners without damage to crop growth and development (Chaves and Vasconcelos 2006;Doumer 2011;Leão et al. 2014;Pereira and Vitti 2004).However, these by-products may contain potentially toxic elements, such as arsenic, mercury, cadmium and lead, which may contaminate the soil and, consequently, the plants (Dos Passos et al. 2018;Dos Santos et al. 2017).Therefore, further examination of these by-products for trace elements is needed in order to comply with environmental safety as well as food security requirements, which go hand to hand with agronomic efficiency when aiming at the agricultural input formulations.
The toxic effects of As and its compounds decrease in the sequence As(III) > As(V) > monomethylarsonate > dimethylarsinate > arsenobetaine.The toxic, inorganic forms are also preferentially accumulated in living organisms.Arsenic may be determined by hydride generation atomic absorption spectrometry (HG AAS), providing simplicity and much lower costs of operation and equipment acquisition with similar sensitivity as HPLC-HG-ICP-MS (P etursd ottir et al. 2012).In addition, HG AAS provides adequate sensitivity and detection limits.The atomization signal is nevertheless dependent on the oxidation state and the hydride system used.These differences are also pH dependent, which offers the possibility of selectively determining As(III) from pH 4 to 5 (Welz and Sperling 1999).
The objective of this work was to develop a simple, fast and reliable analytical method for inorganic arsenic species determination (As(III) and As(V)) in retorted oil shale by HG AAS following ultrasound assisted acidic extraction.The As(V) concentration was estimated by the difference of the total inorganic arsenic and trivalent arsenic quantities.

Experimental
Instrumental A flame atomic absorption spectrometer (AA-6300, Shimadzu, Japan) was employed for arsenic quantification equipped with an accessory for chemical vapor generation (HVG-1, Shimadzu).A deuterium lamp performed the background correction, and the radiation source was an arsenic hollow cathode lamp (Hamamatsu Photonic, Japan).Instrumental conditions used in the determination of As species are provided in Table S1 in the supplementary material.Argon with a purity of 99,996% (Linde) was used as the carrier gas for the arsine at a pressure of 400 kPa.A quartz cell with a path-length of 147 mm and a diameter of 24 mm was used.Peak area was used for signal evaluation.
The pH adjustment for the buffer solutions was done by using a PHS-3B meter (PHTek, Brazil).Samples were homogenized using a vortex mixer(VELP Scientifica, Italy) before and after the sample extraction in an ultrasound bath at a frequency of 37 kHz.A centrifuge (Excelsa Baby, FANEM, Brazil) was employed for decantation of the slurry particles resulting from the extraction.All samples were weighed using an analytical Ohaus Adventurer (AR 2140, Pine Brook, NJ, EUA).

Reagents and samples
All chemicals were of analytical grade, and the solutions were prepared using deionized water (DW), obtained from a glass distillatory system (MA078/5, Marconi, Brazil) with a deionizer (CS1800 Evolution, Permution, Brazil).
The arsenic standard solutions were 1002 mg L À1 arsenic(III) Standard for ICP in 2.0% v/v HCl (Sigma-Aldrich) and 1013 mg L À1 H 3 AsO 4 , in 0.5 mol L À1 HNO 3 (Certipur, Merck, Germany).The working solutions were prepared by appropriate dilution of the stock solutions.The reducing agent was 0.4% m/v NaBH 4 (Fluka Analytical) stabilized in 0.5% m/v NaOH (Vetec, Brazil).Citric acid (C 6 H 8 O 7 , Sigma-Aldrich) and trisodium citrate (Na 3 C 6 H 5 O 7 .2H 2 O, minimum purity of 99.0%, Vetec) were used to prepare the buffer solutions.Potassium iodide (KI, Synth, Brazil), at a concentration of 4.0% m/v, preserved in 1.6% m/v ascorbic acid (C 6 H 8 O 6 , ACS, Vetec) was used to achieve the pre-reduction of As(V) to As(III) before the reaction with the reducing agent NaBH 4 for total inorganic arsenic measurement.

Factorial design to underpin slurry preparation
A 2 6-2 fractional factorial design was applied to provide quantitative arsenic extraction with three central points.The program Statistica 7 supported the factorial design evaluation.The investigated factors in this step were HCl concentration; slurry exposure time in the ultrasound bath; HF concentration; concentration of the pre-reducer KI to convert As(V) to As(III) prior to analysis; HCl concentration in the pre-reduction step, and time for the re-reduction of As(V) to As(III).
Table S2 provides additional details of the factorial design.The sample mass was set at 0.125 grams for all experiments.A second factorial design was applied to further simplify the determination of arsenic species.The concentrations of HCl for both sample preparation and as carrier solution were fixed, and the other four parameters were reevaluated, as shown in Table S3.

Samples and sample preparation
Samples of retorted oil shale were provided by Embrapa Clima Temperado (Pelotas, RS, Brazil).The shale rock originates from Irati formation (São Mateus do Sul, PR, Brazil).
The conditions for retorted oil shale slurry preparation included a sample aliquot of 0.125 g ground to 0.3 mm that was weighed directly in 15.0 mL polypropylene (PP) flasks in triplicate.3.0 mL of 5.0 M HCl and 0.5 mL of concentrated HF were added, vortexed for one minute, and transferred to the ultrasonic bath at room temperature for thirty minutes.A sample volume of 14.0 mL was prepared with deionized water (DW) followed by centrifugation for ten minutes at 4000 rpm to separate the solid particles of the slurry and obtain the leachate for analysis.
For total inorganic arsenic determination, 0.15 mL of leachate were transferred to a 50 mL polypropylene flask, treated with 2.0 mL of 4.0% m/v KI stabilized in 1.6% m/v ascorbic acid and 2.0 mL of 6.0 M HCl, and allowed to react for 25 min to reduce As(V) to As(III) dilution to 30.0 mL with deionized water.For As(III) determination, aleachate volume of 0.15 mL was transferred to a 50.0 mL polypropylene flask to which 3.0 mL of citric acid/trisodium citrate buffer were added and the final volume of 30.0 mL was obtained with deionized water.The sample preparation for the itAs and As(III) blank measurements were prepared as above but without analyte.
Spiking and recovery were investigated for the evaluated arsenic species in order to assess the method trueness and precision at three concentration levels due to the absence of a shale certified material.Additionally, a sample of the retorted oil shale from the same batch as in the present work was analyzed in parallel by an accredited laboratory, ACME Analytical Laboratories (Vancouver) Ltd. (Canada), for total arsenic by ICP-MS using hot digestion (Method 1DX).

Accuracy and selectivity
To characterize the accuracy and selectivity, the standard solutions were submitted to the same As species extraction followed by HG AAS measurement of As(III) and itAs.The standards were sonicated for 30 min with 3.0 mL of 5.0 M HCl and 0.5 mL of concentrated HF, vortexed for one minute, and centrifuged for 10 min.Next, the standards were diluted with deionized water to final volume of 14.0 mL.Two 0.150 mL aliquots of each diluted standard were obtained.To the first, 3.0 mL of the citric acid/citrate buffer at pH 4.5 were added followed by dilution to 30.0 mL deionized water.The As(III) concentration was measured by HG AAS using 0.4% m/v NaBH 4 as reducing agent and deionized water as the carrier.
To the second 0.150 mL aliquot, a reduction reaction to convert As(V) to As(III) was achieved by adding 2.0 mL of 6.0 M HCl and 2.0 mL of 4.0% KI solution stabilized in 1.6% ascorbic acid.A final volume of 30.0 mL obtained using deionized water following a reaction period of 30 min.The total inorganic arsenic determination was done by HG AAS with 0.4% m/v NaBH 4 as the reducing agent and 4.0 M HCl as the carrier.

Reducing agent concentration
To optimize the reducing agent concentration (NaBH 4 ), the investigated concentration range was from 0.05 to 1.0% m/v.The HCl concentration was 5.0 M, according to the manufacturer's recommendation.As shown in Figure 1, the integrated absorbance increased from 0.05% to 0.4% m/v NaBH 4 and decreased at higher concentrations.This drop in the analytical signal is probably due to excess NaBH 4 which forms volatile compounds that increase the pressure of the system and lead to the dispersion and consequent dilution of the arsine.Thus, to ensure the maximum sensitivity and optimize reagent usage, a NaBH 4 concentration of 0.4% m/v was selected for subsequent measurements.

Concentration of HCl carrier
Figure 2 shows the HCl concentration study for the quantitative reaction of the analyte with the reducing agent.Although 6.0 M HCl provided the highest signal, there was also an increase in the standard deviation.HCl concentrations of 4.0 M and 5.0 M also provided high sensitivity but with lower standard deviations.Thus, the 4.0 M concentration was chosen to perform the subsequent experiments.
In order to evaluate the reaction rate for the formation of the inorganic As hydrides, calibration curves were constructed for As(III), As(V), and a combination of both.The resulting equations were y ¼ 0.0581x þ 0.0053 and R 2 ¼ 0.9935 for As(V) þ As(III); y ¼ 0.0512x þ 0.0034 and R 2 ¼ 0.9942 for As(III); and y ¼ 0.0101x þ 0.0002 and R 2 ¼ 0.9961 for As(V).Citric acid/citrate buffer for selective as(III) measurement Solutions of different trisodium citrate and citric acid ratios as buffer were studied, which generated pH values from 3.5 to 6.0, as shown in Figure 3.Only As(III) presents significant analytical signals at pH above 4.5.Therefore, by using a pH 4.5 buffer, the selective hydride generation of As(III) was achieved, with inhibition of the formation of the volatile As(V) species.
Next, the volume of pH 4.5 buffer optimized.Figure 4 shows that no As(V) signal was present for 2.0 mL of the buffer solution and slightly increased using 4.0 mL.However, the As(III) signal was relatively high in all buffer volumes tested.Hence, 3.0 mL of buffer solution were deemed to be optimum because of the high sensitivity for As(III) and inhibition of the As(V) formation.
Two calibration curves were constructed using the optimized conditions for selective generation of As(III) hydride and determination by HG AAS.This experiment was designed to evaluate the selective As(III) hydride generation calibration range from 0.25 to 3.0 mg L À1 .The equation obtained for the curve from As(III) standard solutions was y ¼ 0.0372x þ 0.0047, with R 2 of 0.9905, and for the mixed standard solutions of As(III) and As(V) was y ¼ 0.0375x þ 0.0046, with R 2 of 0.9924.

Total inorganic arsenic conditions
The efficiency of KI in 1.6% m/v ascorbic acid as a pre-reducer of As(V) to As(III) was evaluated using 10.0 lg L À1 standards of As(III) and As(V) and 2.0 mL of 6.0 M HCl.6.0 M HCl promotes fast and efficient reduction of As(V) to As(III) (Dedina and Tsalev 1995;Takase et al. 2002).The procedure to reduce As(V) to As(III) prior to the reaction and analysis by HG AAS has been recommended in the literature (Frentiu et al.  Figure 5 shows that KI and ascorbic acid were effective in the conversion of As(V) to As(III), since the signal for the standard solutions of both species, analyzed independently, presented similar results for KI concentrations higher than 2.0% m/v.Thus, in order to guarantee the quantitative reduction of As(V) to As(III), a concentration of 4.0% m/v KI in 1.6% ascorbic acid was used for the reduction of the pentavalent species and the quantification of the total inorganic As.
To evaluate the efficiency of 2.0 mL of 4.0% m/v KI in 1.6% m/v ascorbic acid the as pre-reducer with different concentrations of As(V), a calibration curve was prepared from 0.25 to 3.0 lg L À1 .Table S4 shows the calibration curves for both species did not present significant differences, since their slopes were similar with a mean of 0.0528 and relative standard deviation of 1.05%.Table 1 shows the experimental parameters and the optimum conditions for the determination of inorganic arsenic fractions by HG AAS.

Arsenic extraction from retorted oil shale
From the 2 6-2 fractional factorial design, normal probability plots of the effects for determination of itAs in XR slurries by HG AAS were obtained.Figure S1 shows that the concentration of HCl for the As(V) reduction and volume of HF were significant, since they are well apart from zero of the y-axis (expected normal value) and zero of the x-axis (standardized effect).The effects of the KI concentration and reaction time to reduce As(V) to As(III) are at the threshold of significance for the evaluated levels.The levels evaluated for sonication time (US) and extractor HCl do not appear to significantly impact the response.The influence of the HCl concentration in the reduction step of As(V) to As(III) showed a distinct result compared to the other factors.By increasing the HCl concentration from 3.0 M to 6.0 M, there was a significant and positive effect on the response, indicating a higher value maximizes the conversion of As(V) to As(III) in retorted oil shale leachate.
The use of HF in the extraction of As also influenced the response in a significant and positive way, as expected, since this matrix contains more than 40% SiO 2 (Ribas 2012), making the extraction and the release of the analytes more difficult.The HCl was fixed at the highest tested concentration, due to its significant positive effect, therefore ensuring a strongly acidic medium for reducing the pentavalent species and higher sensitivity.
Considering the significance of several of the evaluated factors in the 2 6-2 factorial design, a refinement design was performed.In this sense, the four reevaluated factors Measurement conditions: addition of 2.0 mL of KI solution stabilized in 1.6% ascorbic acid plus 2.0 mL of 6.0 M HCl to the standard, which reacted for 30 min; 0.4% m/v NaBH 4 stabilized in 0.5% m/v NaOH; 4.0 M HCl as carrier; and flow rates of 15.0 mL min À1 for standard and of 6.0 mL min À1 for carrier and reducer solutions.
Table 1.Summary of the optimal values of the experimental conditions for As(III) and total inorganic As determination in standards by HG AAS. and their levels were HF (0.50 or 0.75 mL); KI (3.0 or 4.0% m/v); reaction time with the KI/ascorbic acid (10 or 25 min), and extraction time in the US (0 or 30 min).For a second time, fractional factorial design was employed to undertake the optimization, being this time a 2 4-1 with three central points.Figure S2 shows the probability plot of the standardized effects for the 2 4-1 design.The reaction time factor was the factor that most positively influenced the response.The standardized effect of the sonication time (US) and the KI and HF concentrations are close to the zero points of the x and y axes, and their variations are insignificant and due to random errors.Table 2 summarizes the optimized conditions.

Assessment of the inorganic arsenic species accuracy and selectivity
In addition to the standard solutions of each As species, a series of solutions with different concentrations of each were also evaluated in order to obtain the final concentration of total inorganic arsenic of 2.50 lg L À1 in all evaluations.
The analyte recoveries were between 98 and 106%, as shown in Table 3, demonstrating good accuracy for inorganic arsenic species determination without interconversion and volatilization losses.Additionally, for all tested conditions, As(III) was selectively measured by HG AAS, no matter the As(V) concentration, by fixing the pH at 4.5 by a buffer solution.Likewise, successful total inorganic arsenic determination was achieved with a reduction step for As(V) to As(III) prior to analysis.This approach permitted the determination of the As(V) concentration by subtracting the total inorganic arsenic content (itAs) minus As(III).
The figures of merit for total inorganic arsenic and trivalent arsenic are presented in Table 4.The developed methodology presented adequate linearity determination coefficients greater than 0.998.The limits of detection and quantification were satisfactory and the methodology was demonstrated to be reproducible, with relative standard deviations lower than 6%.

Inorganic arsenic species determination in retorted oil shale
The concentrations of the inorganic arsenic (As(III) or itAs) in retorted oil shale are presented in Table 5. Comparing the total inorganic arsenic content of 29.29 mg kg À1 obtained in this work with the measured value of 31.9 mg kg À1 by ACME Analytical Laboratories shows the developed method for arsenic is accurate.This conclusion is based upon the 92% recovery for total As compared to the standard value.

Addition and recovery tests
In order to evaluate the accuracy, recovery tests were performed at three trivalent arsenic concentrations in triplicate.The levels of fortifications were based upon the natural arsenic in the matrix which were level 1 of 20%, level 2 of 50% and level 3 of 100% of 29.29 mg kg À1 (itAs), meaning 0.259, 0.647 and 1.29 mg L À1 of arsenic in solution following the dilution steps in the method.
Table 6 shows satisfactory recoveries were achieved for all evaluated fortified levels, varying between 101 and 118%, for both As(III) and total inorganic As measurements.These results demonstrate the trueness of the method for species determination in retorted oil shale by HG AAS.Additionally, the obtained recovery values also support   the absence of matrix effects, showing suitable selectivity.The relative standard deviations for the recovery experiments for all fortification levels were between 0.6 and 5.7%.

Conclusions
The ultrasound assisted acidic extraction was a simple and fast approach for the preparation of retorted oil shale as slurry, allowing the extraction of inorganic arsenic species at room temperature without interconversion or losses due to volatilization.The use of hydrochloric and hydrofluoric acids, along with ultrasonic energy, was essential for the quantitative extraction and allowed the analysis of the sample leachate.The accuracy was demonstrated to be reliable, with recovery values close to 100% (trueness), and relative standard deviations below 5.7%.
The concentration of As(III) in the retorted oil shale leachate was selectively measured in the presence of As(V) in buffered medium with citric acid/citrate at pH 4.5 by HG AAS.In contrast, the determination of total inorganic As concentration in the slurry by HG AAS required the reduction of the As(V) to As(III) using 4.0% KI in 1.6% m/v ascorbic acid and 6.0 M HCl.The measured concentration of 29.29 mg kg À1 validates the hypothesis that all arsenic present in the retorted oil shale is either As(III) or As(V).The measured As(III) concentration in the sample was 7.67 mg kg À1 , which implies in an estimated As(V) concentration of 21.62 mg kg À1 (calculated as itAs minus As(III)).This is actually a positive finding of the inorganic arsenic species determination of retorted oil shale, since most arsenic is As(V) (73.8% of the itAs), which is ten times less toxic then trivalent As.

Figure 1 .
Figure 1.Optimization of NaBH 4 concentration for the determination of As by HG AAS.Measurement conditions: 10 mg L À1 As(III) standard solution; 5.0 M HCl as carrier solution; flow rates of 15.0 mL min À1for the sample and of 6.0 mL min À1 for both carrier and reducer solutions.

Figure 2 .
Figure 2. Optimization of HCl concentration as the carrier solution in CVG for As determination by HG AAS.Measurement conditions: 10 mg L À1 As(III) standard solution; 0.4% m/v NaBH 4 stabilized in 0.5% m/v NaOH; flow rates of 15.0 mL min À1 for the standard and of 6.0 mL min À1 for both carrier and reducer solutions.

Figure 3 .
Figure3.Optimization of the citric acid/citrate buffer pH for the selective generation of As(III) hydride by HG AAS from standard solutions of As(III) and As(V).Measurement conditions: 10 mg L À1 As(III) or As(V) standard; 3.0 mL of buffer added to the standard; 0.4% m/v NaBH 4 stabilized in 0.5% m/v NaOH; deionized water as the carrier; flow rates of 15.0 mL min À1 for standard and of 6.0 mL min À1 for carrier and reducer solutions.

Figure 4 .
Figure 4. Evaluation of the volume of the pH 4.5 citric acid buffer for selective generation of As(III)hydride from standard solutions of As(III) and As(V) and analysis by HG AAS.Measurement conditions: 10 mg L À1 As(III) or As(V) standard; 0.5 M buffer of citric acid and trisodium citrate added to the standards; 0.4% m/v NaBH 4 stabilized in 0.5% m/v NaOH; deionized water as the carrier; flow rates of 15.0 mL min À1 for the standard and of 6.0 mL min À1 for both carrier and reducer solutions.

Figure 5 .
Figure5.Evaluation of the KI concentration in ascorbic acid for the reduction of As(V) to As(III) from 10.0 mg L À1 standards of As(III) and As(V) and determination of total inorganic As by HG AAS.Measurement conditions: addition of 2.0 mL of KI solution stabilized in 1.6% ascorbic acid plus 2.0 mL of 6.0 M HCl to the standard, which reacted for 30 min; 0.4% m/v NaBH 4 stabilized in 0.5% m/v NaOH; 4.0 M HCl as carrier; and flow rates of 15.0 mL min À1 for standard and of 6.0 mL min À1 for carrier and reducer solutions.
.: not applicable; a Citric acid/trisodium citrate buffer solution of pH 4.5; b NaBH 4 in 0.5% m/v NaOH; c KI in 1.6% m/v ascorbic acid; d Line of the chemical vapor generation system.

Table 2 .
Optimized conditions for ultrasound assisted acidic extraction of arsenic from retorted oil shale and total inorganic arsenic determination by HG AAS.

Table 3 .
Assessment of the accuracy, selectivity, and arsenic species integrity under the optimized conditions for sample preparation and inorganic arsenic species determination by HG AAS.Nominal concentration of the standard solutions, prepared by addition of the two As species, in the reported concentrations on the table.b As(V) concentration estimated as the difference between the itAs and As(III) concentrations.c Concentration of total inorganic arsenic (sum of As(III) and As(V)).n.a.: not applicable.

Table 4 .
Analytical figures of merit for the optimized conditions for inorganic arsenic species applied to retorted oil shale samples by HG AAS following ultrasound assisted acidic extraction.

Table 5 .
Concentration of total inorganic As determined with pre-reduction of As(V) to As(III) by KI, ascorbic acid and HCl, trivalent arsenic (obtained with citrate buffer) and pentavalent arsenic (obtained by difference) in retorted oil shale leachate by HG AAS in mg kg À1 (n ¼ 3).Calculated as the difference between total inorganic As and As(III) concentrations.b Sum of the trivalent and pentavalent arsenic species in the sample.

Table 6 .
Determination for As(III) and total inorganic As for the recovery in retorted oil shale leachate by HG AAS following ultrasound assisted acidic extraction.The values are in mg L À1 (n ¼ 3).Fortification level related to the natural As concentration in the retorted oil shale of the present work: of 1-20%;2-50% and 3-100%.b Fortification with As(III) standard solution.