Speciation by capillary ion chromatography
First, standard solutions of As(III), As(V), MMA, and DMA and their mixture were prepared. Each of the standard solutions contained 5 μg/kg of arsenic. In the mixture solution, arsenic concentration from each species was set to 5 μg/kg. The prepared standard solutions were injected into the capillary column. The separated arsenic species were detected by ICP-MS. As shown in Fig. 1a, for the mixture solution, only two peaks were observed at around retention times 222 and 441 s. The two peaks could be identified by comparing the chromatogram with those of the standard solutions of MMA, DMA, As(III), and As(V) which are shown in Fig. 1b–e, respectively. In the chromatograms of the standard solutions, the peaks corresponding to MMA, DMA, As(III), and As(V) were observed at 227, 216, 215, and 440 s, respectively. This indicates that As(V) was well separated from the other species, but MMA, DMA, and As(III) were not resolved by the capillary column. Therefore, in the chromatogram of the mixture solution, the peak at 222 s was assigned to the mixture of As(III), MMA, and DMA, and that at 441 s was identified as the separated As(V). Also, it should be noted that the sensitivity was different among the four arsenic species (compare the intensities of the four peaks in Fig. 1b–e). The four chromatograms in Fig. 1b–e were added and plotted along with the chromatogram of the mixture solution in Fig. 1a. The two chromatograms agreed with each other. This indicates that the sensitivity difference was not due to experimental error but to the intrinsic properties of the different species.
In our previous works, the four arsenic species were completely separated by conventional column (Lee et al. 2019; Son et al. 2019; Nam et al. 2016). The lower resolving power of the capillary column used in this work could be attributed to its smaller inner diameter (0.4 mm) and particle size (4 μm) as compared to the conventional column (inner diameter = 4.6 mm and particle size = 10 μm). However, the aim of this work is to determine the total inorganic arsenics (As(III) and As(V)) using capillary column since the toxicity of the organic ones is ignorable in comparison with that of the inorganics. In this regard, hydrogen peroxide was used to selectively oxidize As(III) to As(V). Then, the solution was injected into the capillary column to separate As(V). The detected intensity for As(V) resulted from the sum of As(V) and As(III) oxidized to As(V) and thus represented the total inorganic arsenics (As(III) + As(V)). The selective oxidation of As(III) to As(V) could be confirmed as explained thereafter. Figure 2 compares the chromatograms of the non-oxidized (blue) and oxidized (red) mixture solutions. As observed in the chromatogram of the oxidized mixture solution, the peak at 222 s decreased and that at 441 s increased in contrast to the non-oxidized solution. This can be attributed to the oxidation of As(III) to As(V). Thus, total inorganic arsenics could be separated and determined by the capillary column of ion chromatography coupled with ICP-MS.
Calibration and detection limit
For the calibration curve of inorganic arsenic, 1.0, 2.0, 10, 20, and 50 µg/kg inorganic arsenic standard solutions were prepared from As(III) and As(V) stock solutions (100 mg/kg). Hydrogen peroxide was added to oxidize As(III) to As(V). A linear calibration curve was obtained using the capillary column of IC coupled with ICP-MS as shown in Fig. 3. The correlation coefficient was 0.9999. The linear dynamic range was more than two orders of magnitude. The detection limit was estimated as the concentration giving a signal equivalent to three times the noise, the standard deviation of three repetitive measurements of the background intensity. The detection limit of the inorganic arsenic was 0.13 μg/kg. In order to compare the value obtained using the capillary column with that of the conventional column, a separate experiment was performed. The volume of sample loop combined with the conventional column was 100 μL, which is 20 times larger than that (5 μL) using the capillary column. The detection limit, 0.033 μg/kg, from the experiment using the conventional column was much lower than that obtained using the capillary column. However, it might be unfair to compare the detection limits of the methods using capillary and conventional columns with different amounts of samples. Thus, instead of using concentrations, masses (= amount of sample × detection limit in concentration) were considered:
$${\text{Capillary:}}\quad 5.0 \times 10^{ - 6} \;{\text{kg}} \times \frac{{0.13\;\upmu {\text{g}}}}{{1\;{\text{kg}}}} = 0.65 \times 10^{ - 6} \;\upmu {\text{g}}$$
(1)
$${\text{Conventional:}}\quad 100 \times 10^{ - 6} \;{\text{kg}} \times \frac{{0.033\;\upmu {\text{g}}}}{{1\;{\text{kg}}}} = 3.33 \times 10^{ - 6} \;\upmu {\text{g}}$$
(2)
As a result, the capillary-column method was found to show even better detection-limit performance than the conventional column method.
Quantitation of total inorganic arsenic in water and human urine
For the validation of the developed analytical method, the determination of inorganic arsenic species in the standard reference material of water was performed. A 2.706 mL of the water SRM was mixed with 0.800 mL of hydrogen peroxide to oxidize As(III) to As(V), and the solution was diluted to 8.119 mL with the eluent. This process led to the dilution factor of 3. The resulting solution was analyzed by IC with the capillary column coupled with ICP-MS. As shown in Fig. 4a, the arsenic species present in the water SRM was identified as inorganic ones as compared to Fig. 1. The measured concentration of inorganic arsenic, 57.61 ± 0.90 μg/kg, was in agreement with the certified concentration of As, 56.85 ± 0.37 μg/kg (Fig. 4) in the SRM. The recovery efficiency was almost 101.3%. The result was obtained based on three analyses.
Due to the toxicity of arsenics (particularly inorganic arsenics), human exposure to arsenic needs to be investigated by an appropriate method capable of determining the quantity and the species of arsenic. In this regard, arsenics in food and drinking water have been analyzed (Nam et al. 2016; Lai et al. 2004). Ingested arsenics undergo various metabolisms in the human body and are finally excreted. Thus, analysis of arsenics in human urine is known to be the most reliable medical diagnosis to check if a person has been exposed to arsenics (Lai et al. 2004; Scheer et al. 2012). The method developed in this work was also applied to analyze total inorganic arsenic in human urine. A 0.498 mL of the human urine SRM was mixed with 0.500 mL of hydrogen peroxide to oxidize As(III) to As(V), and the solution was diluted to 1.448 mL with the eluent. This process led to dilution factor of 3. The resulting solution was analyzed by IC with the capillary column coupled with ICP-MS. As shown in Fig. 5a, the inorganic arsenic species were well separated from the organic ones in the human urine. The measured concentration of total inorganic arsenic was 4.69 ± 0.47 μg/kg, which agreed with the certified value, 3.88 ± 0.40 μg/kg, within the experimental uncertainties (Fig. 5b). The result was obtained based on three analyses. The recovery efficiency was 121% and inferior to that obtained in the analysis of the water SRM. This could be attributed to the more complex matrix of human urine than that of water. The matrix effect on the recovery efficiency needs further detailed investigations. However, the bias (= measured concentration − certified concentration) values are very close to each other. The bias of the inorganic arsenic analysis in the human urine SRM is + 0.81 (= 4.69–3.88) μg/kg, and that in the water SRM is + 0.76 (= 57.61–56.85) μg/kg. Thus, the larger recovery efficiency observed in the analysis of the human urine SRM could be an accuracy issue raised by the instrument and not the sample matrix. The advantages of the method developed compared to other conventional methods including HPLC-ICP-MS were the improved detection capability, low sample volume, and fast analysis time.