Environ Geochem Health

DOI 10.1007/s10653-007-9091-3

ORIGINAL PAPER

An assessment of sampling, preservation, and analytical

procedures for arsenic speciation in potentially
contaminated waters
Youn-Tae Kim Æ Hyeon Yoon Æ Cheolho Yoon Æ
Nam-Chil Woo

Springer Science+Business Media B.V. 2007

Abstract This study was undertaken to ascertain the best resolution and lowest detection limits.
optimal methods of sampling, preserving, separating, However, the procedure using a solid phase extrac-
and analyzing arsenic species in potentially contam- tion (SPE) cartridge can be used economically and
inated waters. Arsenic species are readily trans- conveniently for analyzing samples containing only
formed in nature by slight changes in conditions. inorganic arsenic species, such as groundwater,
Each species has a different toxicity and mobility. especially that related to mine activity.
The conventional field sampling method using filters
of 0.45 mm in size could overestimate the dissolved Keywords Arsenic speciation
Sampling

arsenic concentrations, as passing suspended particles Preservation
Analytical methods
Sample condition
that can act as a sink or source of arsenic depending
on the site condition. For arsenic species in neutral
pH and iron-poor waters, the precipitation can be Introduction
stable for up to 3 days without any treatment, but for
longer periods, a preservative, such as phosphoric Arsenic is a problematic toxic contaminant, which
acid, is required. Also, the analytical procedure must has many chemical species, each with a different
be selected carefully because the levels and hydride toxicity and mobility (Mandal and Suzuki 2002). In
generation efficiencies of arsenic in different species natural water, arsenite [As(III)] and arsenate [As(V)]
can vary, even for the same amount of arsenic. For predominate (Plant et al. 2005). Organic arsenic
arsenic speciation in samples that also include species, such as monomethyl arsonate (MMA) and
organic species, a hybrid high-performance liquid dimethyl arsinate (DMA), are only present at low
chromatography (HPLC) column and inductively levels (Bednar et al. 2004). Inorganic arsenic species
coupled plasma mass spectrometry (ICP-MS) gave include known carcinogens (USEPA, accessed in
June 2006). Of these, As(III) is more toxic than
As(V), while the toxicities of the organic arsenic
Y.-T. Kim
N.-C. Woo
species have not been fully evaluated (Thomas et al.
Department of Earth System Sciences, Yonsei University,
Seoul, Korea 2001; Hughes 2002; Aposhian et al. 2003). As with
their toxicities, the behaviors of arsenic species differ
Y.-T. Kim
H. Yoon (&)
C. Yoon markedly. The ionic charges of inorganic arsenic
Korea Basic Science Institute, Seoul Center, 136-600
species vary from 0 to 3, depending on the pH and
Seongbuk P.O. Box, 12, 88-3, Hawolgok 1-dong,
Seongbuk-gu, Seoul, Korea redox conditions (Raposo et al. 2006). For accurate
e-mail: dunee@kbsi.re.kr risk assessment of arsenic in the environment, the

123
 Environ Geochem Health

electro-chemical state of dissolved arsenic must be distinguishing the different organic species is gener-
retained during sample handling. ally difficult due to their low concentrations.
The stability of aqueous arsenic species is also In this study, our primary objective is to identify
dependent on the presence of other dissolved metal an optimal procedure for arsenic speciation in natural
ions. For example, iron is a common element in water samples without deforming chemical species.
arsenic-contaminated waters, which can precipitate if Sampling and preservation methods commonly used
oxidation occurs during sampling. Arsenic can co- for arsenic-bearing waters were thoroughly tested
precipitate with any Fe-oxyhydroxides or even under various conditions. Then, laboratory separation
change its oxidation state (Daus et al. 2002; Bednar procedures were compared with those employed in
et al. 2002). Pre-treatment of water samples involving the field and the limitations of the ICP-AES and ICP-
the complexation of metal ions or acidification of MS analytical methods for the determination of
samples has been proposed to prevent any precipita- arsenic species in natural samples under various
tion of metal oxyhydroxides. Daus et al. (2002) used conditions were discussed.
nitrilotriacetic acid (NTA), hydrochloric acid (HCl),
phosphoric acid (H3PO4), and acetic acid for this
procedure, and found the best results with Materials and methods
0.01 mol l1 H3PO4. Bednar et al. (2002) used
ethylenedinitrilotetraacetic acid (EDTA), sulfuric Water sampling and the application
acid (H2SO4), nitric acid (HNO3), and HCl to of the preservation method
preserve inorganic arsenic species in groundwater
and acid mine drainage samples. EDTA has been Groundwater samples were collected from the aban-
found to minimize precipitate formation and redox doned mines at Guryong and Ulsan, Korea. The
reactions. However, both of the above studies focused Guryong mine was an Au-Ag-Cu mine. The tailings
on iron-rich waters. Preferred pre-treatment proce- from this mine have been found to contain 7.04–
dures for iron-poor waters remain to be determined. 27.27 wt% Fe2O3, with As2O3 < 0.020 wt% (Kim
Various analytical techniques have been devel- 2004). Groundwater samples were taken from both
oped to distinguish and reduce the detection limits of monitoring wells (NX size, 3.500 inside diameter) and
different arsenic species. Different ion-exchange piezometers (100 inside diameter), and surface waters
resins have also been tried (Yu et al. 2003), and from a pond in the mine tailings. The Ulsan Fe-W
hydride generation procedures have been employed, mine was dug into a skarn that contains a consider-
in conjunction with atomic absorption spectrometry able amount of arsenopyrite (Choi and Youm 2000).
(AAS), atomic fluorescence spectrometry (AFS), Groundwater samples were collected from monitor-
inductively coupled plasma-atomic emission spectro- ing wells in the tailings and from domestic wells
photometry (ICP-AES) or ICP mass spectrometry around the mine. Surface waters were sampled from
(ICP-MS), not only to reduce the detection limits, but creeks and ponds near the Ulsan mine. All water
also to overcome interferences (Hung et al. 2004). samples showed high arsenic concentrations, which
Among these methods, a hybrid high-performance were generally in excess of the World Health
liquid chromatography (HPLC)/ICP-MS method has Organization (WHO) arsenic guideline of 10 mg l1
been most widely applied (B’Hymer and Caruso (WHO 1993) for drinking water, with the pH ranging
2004). In the majority of water analyses, most from neutral to weakly alkaline (pH 6.2–8.6).
attention has been paid to the predominant inorganic Along with groundwater sampling for a total
arsenic species, which are also recognized carcino- analysis of metal ions, various preservation methods
gens (Thirunavukkarasu et al. 2002). Analytical were applied to the separate samples to investigate
methods for the separation and detection of organic the changes in stability of dissolved ionic species,
arsenic species have been tested in the laboratory, but especially ‘‘arsenic,’’ by the duration of time after
these have only recently been applied to natural sampling has been done. Selective filter pore size
waters (Hung et al. 2004; Bednar et al. 2004; Akter and preserving chemicals have been applied to
et al. 2005; Sánchez-Rodas et al. 2005), where different samples for these purposes. Groundwater

123
Environ Geochem Health

samples from the Guryong mine were filtered Reagents

through both a 0.2-mm and 0.45-mm pore membrane
syringe filter at the sampling site, and then acidified A standard stock solution, containing 1,000 mg l1
with HNO3 to lower the sample pH to below 2. The As(V) in 2–3% HNO3, was purchased from Merck
sample volume was about 10 ml. This sampling was (Darmstadt, Germany). As(III) was prepared by
designed to see whether the conventional filtering dissolving 0.0660 g As2O3 (>>99.5% [RT] Fluka,
method in field water sampling, using 0.45-mm pore Buchs, Switzerland) in 50 ml of 0.3% (w/w) NaOH.
filters to differentiate elements in dissolved and Monomethyl arsonate (MMA) and dimethyl arsinate
suspended particulate forms, is appropriate for (DMA) were made by dissolving 0.1058 g Na2CH3A-
arsenic assessment. sO3
6H2O (99%, Chem Service, West Chester, PA,
To test the stability of arsenic species followed by USA) and 0.0782 g Na(CH3)2AsO2 (For Synthesis;
sampling, water samples from the Ulsan mine were Merck) respectively in de-ionized water. A standard
filtered though a 0.45-mm pore membrane syringe stock solution of 1031 mg kg1 arsenobetaine (AsB)
filter, divided into four sub-samples, and then (CRM626) was purchased from BCR. Working
preserved in four different ways: the addition of: standard solutions were prepared by appropriate
dilution at the time of each measurement by diluting
1. H3PO4
0.1 ml of each stock solution to 100 ml. These
2. HCl
solutions were diluted to specific concentrations with
3. Ethylenedinitrilotetraacetic acid (EDTA) to a
de-ionized water.
final concentration of 0.01 M
4. No treatment
Speciation analysis of arsenic by on-line
All samples were stored at 48C. The effect of hyphenated system analysis
suspended particles on the arsenic concentration was
tested by changing the order of the H3PO4 addition, For field separation of the inorganic arsenic species,
which is known to be a better preserving reagent for As(III) and As(V), SPE cartridges (Accell Plus QMA;
arsenic, either before or after filtration. Waters, Milford, MA, USA) were used. Various types
of SPE cartridges can be used for routine arsenic
speciation analysis of water. In this study, a strong
Instrumentation anion exchange SPE cartridge (Accell Plus QMA;
Waters) was used. In the field, 10 ml of the filtered
The arsenic concentrations were primarily deter- sample was passed through the cartridge at a rate of
mined by ICP-AES (Ultima 2C; Jobin Yvon, Edison, 1–2 drop s1. The leachate was acidified with HNO3,
NJ, USA), with a radio frequency (RF) power of and the cartridge then sealed and cooled. Upon return
1000 W. All arsenic measurements were made at to the laboratory, each cartridge sample was extracted
wavelength 193.7 nm. Arsenic was also analyzed with 10 ml of 0.16 M HNO3.
with an ICP-MS (Elan 6100; Perkin Elmer, Waltham, Laboratory separations were undertaken using a
MA, USA). The instrumental conditions were opti- SPE cartridge (Accell Plus QMA cartridge; Waters)
mized before each analytical run, with either the and an HPLC column (PRP X-100; Hamilton, Reno,
HPLC or short column (solid phase extraction, SPE) NV, USA). The Accell Plus QMA SPE cartridge
incorporated to permit arsenic speciation analysis. All provided a short column, which allowed primary
ICP-MS measurements of arsenic were made at m/z analysis of the inorganic arsenic species at low
75. pressure. The eluent used consisted of 20 mM of pH
A conventional hydride generation (HG) method 7.0 phosphate buffer, applied at a flow rate of
was selectively applied to lower the detection limit. 1 ml min1 using a peristaltic pump, with a sample
For metal hydride generation of arsenic species, the injection volume of 500 ml. The PRP X-100 HPLC
sample was mixed with strong acid (6N HCl) on-line column allowed for analysis of several arsenic
at a volume ratio of 1:1, then followed by mixing species, including organic species. The eluent used
with 1% NaBH4 (Merck) solution in 0.1% NaOH for consisted of a 10 mM, pH 9.0 phosphate buffer for
arsine gas (AsH3) generation. the Ulsan samples and a 30 mM, pH 6.0 buffer for the

123
 Environ Geochem Health

Guryong samples. The flow rate was adjusted to

1 ml min1 using a P680 HPLC Pump (Dionex,
Sunnyvale, CA, USA), with a sample injection
volume of 20 ml.

Results and discussion

Sampling and preservation methods

When iron-rich groundwater is exposed to air, iron

oxyhydroxide can readily precipitate, which can
significantly alter the total concentration of arsenic
if the pH is not sufficiently low (Bang et al. 2005).
This is the main reason why researchers must prevent
their samples from coming into contact with air
during sampling. To quantify the effect of precipita-
tion on the arsenic concentration, we tested both the
filter pore size and acidification procedures during
groundwater sampling. The former factor can help
clarify the interaction with suspended submicron
particles, and provide a clue to the stability of the
water sample. The latter factor is pertinent for
suspended particles in rich water samples, and gives
information on the role of suspended particles as a
sink for or source of arsenic.
A total of 22 groundwater samples from the
Guryong mine, with a mean Fe concentration of Fig. 1 Total arsenic concentrations of samples from (a) NX-
208 mg l1 (min 29 mg l1, max 737 mg l1), were size monitoring well, and (b) 1@ diameter piezometer filtered
through 0.2- or 0.45-mm membrane pores
filtered through two different membrane filters, with
pore sizes of 0.2 and 0.45 mm. Each sub-sample was
then acidified and analyzed by HG-ICP-MS. As sampling method using filters of 0.45 mm in size
shown in Fig. 1, the total arsenic concentrations could overestimate the dissolved arsenic concentra-
displayed two different trends, reflecting the diverse tions while suspended submicron-sized particles were
characteristics of the two types of well. Groundwater present.
samples from the monitoring wells show no differ- Interactions between suspended particles and dis-
ence in arsenic concentrations with regard to chang- solved arsenic species eventually lead either to a sink
ing the filter pore size (Fig. 1a). However, in the or to regeneration of arsenic within many natural
samples collected from the piezometers (Fig. 1b), environments. Waters containing iron oxides, or
those filtered through the 0.2-mm pore membrane hydroxides can precipitate during sampling or upon
show roughly half the arsenic concentration of those preservation, and need treating to prevent co-precip-
filtered through the 0.45-mm pore filter. This indicates itation of arsenic (Gallagher et al. 2001). If filtered
that the samples from the piezometers contained water samples having high Fe contents and insuffi-
more arsenic in suspended particles, ranging from 0.2 ciently low pH were not acidified, an abrupt decrease
to 0.45 mm in size, than those from the monitoring in arsenic concentration could occur as a result of the
wells. Suspended particles in natural waters could co-precipitation of arsenic. For example, in a ground-
function as a sink for or source of arsenic depending water sample from the Guryong mine (pH 5.8, Eh
on ambient conditions. The conventional field 235 mV, Fe 154 mg l1, and As 11.0 mg l1), arsenic

123
Environ Geochem Health

was not detected in water if it had not been acidified. of arsenic species are the same. However, different
In this case, precipitation can act as a sink for arsenic. behavior was observed when with the use of ICP-
H3PO4 can prevent the As co-precipitation with iron AES as the analytical method, the organic arsenic
oxides or hydroxides, and can also extract arsenic species in de-ionized water showed only half the
bound to these species (Ko et al. 2005). Therefore, intensities of the inorganic species. This problem can
the test changing the order of the H3PO4 addition, be resolved by acidifying the samples with nitric acid
before or after filtration, can inform us with regard to (Table 2).
the role of suspended particles. Groundwater samples Similarly, different intensities were obtained with
from the Ulsan mine contained some suspended different species during the HG procedure, even
particles; only one sample showed a difference in though the same concentrations of each arsenic
arsenic concentrations by changing the order of species were measured (Maity et al. 2004). The on-
H3PO4 addition: 4.15 mg l1 in the sub-sample of line HG method was tested for both inorganic and
H3PO4 addition before filtration, and 0.09 mg l1 in organic arsenic species by applying ICP-AES in order
the sub-sample of the H3PO4 addition after filtration. to achieve the lower detection limit. Figure 2 shows
This means that these suspended particles containing the hydride generation efficiencies of various arsenic
absorbed arsenic prior to sample collection can act as species in the same arsenic concentrations analyzed
a source of arsenic by condition change. by the on-line HG-ICP-AES. The efficiency of the
on-line HG procedure was best with the reduced
Analytical methods form, As(III), confirming the result of Hung et al.
(2004). According to the results in Fig. 2, the relative
Various techniques have been used for the determi- intensity of As(III), As(V), and MMA showed about
nation of arsenic speciation. Widely used procedures 50% intensities, and DMA about 10% intensities
have employed exchange resins to separate arsenic compared with 100%, or even. AsB failed to generate
species, coupled with spectroscopic and ICP analyses hydride gas.
to measure the arsenic concentrations (Hung et al. Until now, a conventional HG method, without
2004). In our study, both the SPE cartridge and HPLC any pre-reduction, has been most frequently used to
column were used for both field and laboratory measure the total arsenic concentrations in most
separation of arsenic species to examine any prob- applications, largely because organic species and
lems that could arise during analysis. As(III) are normally rare in natural waters (Maity
et al. 2004). However, where these species are
Analytical implications for various arsenic species present at high levels, the use of conventional
methods for the determination of the total arsenic
The measured intensities of different inorganic concentrations of waters would lead to under- or
arsenic species yield almost the same values during over-estimations. Consequently, whenever the con-
ICP-MS analysis (Table 1) if the absolute concentrations ventional HG method is used to measure the total

Table 1 Calculated intensities of arsenic species containing the same amount of arsenic, as measured by inductively coupled plasma
mass spectrometry (ICP-MS). All intensities were corrected using an internal standard
Species Intensity for 10 mg l1 Calibration
of arsenic

As(V) 0.0829 y = 0.0082x + 0.0009, R2 = 1.0000

As(III) 0.0843 y = 0.0086x  0.0017, R2 = 0.9999
MMA 0.0878 y = 0.0087x + 0.0008, R2 = 1.0000
DMA 0.0836 y = 0.0083x + 0.0006, R2 = 1.0000
AsB 0.1195 y = 0.0119x + 0.0005, R2 = 0.9998
As(V) arsenate, AS(III) arsenite, MMA monomethyl arsonate, DMA dimethyl arsinate, AsB arsenobetaine
Inorganic species were calibrated using 0, 0.5, 1.0, 10, and 50 mg l1 standards, and organic species using 0, 1.0, 10, 50, and
100 mg l1 standards

123
 Environ Geochem Health

Table 2 Relative spectral intensities produced by the same Moreover, in natural waters, organic arsenic species,
amount of arsenic for different species within different matri- such as MMA and DMA, are present as anions.
ces, as measured by inductively coupled plasma-atomic emis-
Consequently, As(III) will not be trapped in an SPE
sion spectrophotometry (ICP-AES)
cartridge, while As(V), MMA and some of the
Species Intensity for 50 mg l1 of arsenic organic species present as anions will (Yu et al.
De-ionized water 2% HNO3 2003), and can be extracted in the laboratory prior to
analysis. In the results of Yu et al. (2003), the
As(V) 1,817.64 1,881.95 percentages of As(III), As(V), DMA, MMA, and AsB
As(III) 1,597.83 1,772.39 retained on 2 g of strong anion exchange sorbent are
MMA 700.67 1,835.86 0, 100 ± 3, 41 ± 5, 100 ± 2, and 6.5 ± 3%. By
DMA 550.98 1,689.41 application of this method, Bednar et al. (2004) found
AsB 777.77 2,173.96 a recovery of 99% for 25 mg l1 As(V) after 47 days.
However, this technique can only be applied to
separate inorganic species if there are, as two or more
arsenic species, present as anions, cannot be sepa-
rated within the SPE cartridge. Further, it is useful to
know the rough arsenic concentrations involved due
to the limited cartridge capacity. For the analysis of
extractants with elevated anion contents, it is better to
avoid ICP-MS analysis, as arsenic when measured at
m/z 75 suffers from 40Ar35Cl+ interference.

Laboratory separation

Laboratory separation procedures were conducted

using a SPE cartridge (Accell Plus QMA cartridge;
Waters), as a short column and an HPLC column
(PRP X-100; Hamilton). The results obtained are
shown in Fig. 3. A 100 mg l1 standard solution,
Fig. 2 Hydride generation (HG) efficiencies for different including As(III), As(V), MMA, and DMA, was
arsenic species, as measured by on-line hydride generation analyzed by SPE-HG-ICP-AES (Fig. 3a). The DMA
inductively coupled plasma-atomic emission spectrophotome- peak is not shown, and the MMA peak overlaps with
try (HG-ICP-AES). As(V) arsenate, AS(III) arsenite, MMA
monomethyl arsonate, DMA dimethyl arsinate, AsB arsenobe-
that of As(III). This result indicates that short
taine columns are only capable of separating inorganic
species. If ICP-MS is used, the detection limit might
arsenic concentration in water, the arsenic species be reduced. Figure 3b shows the peaks for a 5 mg l1
present must be considered, and pre-reduction of standard solution, containing both As(III) and As(V),
arsenic species to As(III) should be carried out prior analyzed by SPE-ICP-MS, while Fig. 3c shows the
to applying the HG procedure. results for a 5 mg l1 standard solution, which
included AsB, As(III), As(V), MMA, and DMA,
Field separation analyzed by HPLC-ICP-MS.
Consequently, when arsenic speciation has a
Employing field separation procedures using SPE significant impact on researchers’ projects such as
cartridges can avoid the changes in the chemical state toxicity or risk assessment of potentially As-contam-
of arsenic that might result from sample storage. The inated waters, the HPLC-ICP-MS method can be
ionic charges of inorganic arsenic species in solution applied because of its capacity to separate more
can readily occur with changes in pH (Raposo et al. arsenic species present in the solution and to detect to
2006). As(III) has a zero charge under neutral pH low concentrations. However, since the HPLC col-
conditions, whereas the change for As(V) is negative. umn will take more time and more effort to optimize

123
Environ Geochem Health

periods of within 18, 72, 120, 192, and 360 hours.

The concentrations of other metals, such as Al, Mn,
Cd, Cr, Cu, Co, Ni, and Pb, were all below
0.1 mg l1, but the total arsenic concentrations were
73.3 mg l1 and 23.7 mg l1 respectively.
The results from the arsenic speciation analysis are
shown in Fig. 4. Concentrations of arsenic species are
converted into the compositional ratio in percentage
of total arsenic concentrations measured with ICP-
MS. Both samples have only inorganic arsenic
species, and As(V) revealed as the dominant arsenic
species, with a minor As(III) content. The dashed
lines show the results of the field separations, and the
four point symbols represent the As(III) (bottom) and
As(V) (top) concentrations in the samples treated
with each of the preservatives.
In the groundwater sample (Fig. 4a), As(III) is
present at a very low concentration (1.17 mg l1), and
thus it is difficult to detect via laboratory separation
Fig. 3 Spectra of standard solutions analyzed using different
procedures and various analytical conditions. (a) 100 mg l1 using HPLC-ICP-MS analysis. In the surface water
standard of As(III), As(V), MMA, and DMA analyzed by solid sample, As(III) at a concentration of (4.77 mg l1)
phase extraction HG-ICP-AES. (b) 5 mg l1 standard of As(III) appears to be relatively stable for up to 360 h
and As(V) measured by SPE inductively coupled plasma mass (15 days) of storage with no significant effects of the
spectrometry (SPE-ICP-MS). (c) 5 mg l1 standard of AsB,
As(III), As(V), MMA, and DMA determined by high- preservatives on the concentration. On the contrary,
performance liquid chromatography ICP-MS As(V) continuously decreases over storage time, in
all sub-samples treated with different methods and in
both groundwater and surface water. The decrease
analytical conditions than the short column, the rate of As(V) concentration appears to become stable
procedure using the short column can be used after 192 h (8 days). Regarding the efficiencies of the
economically in samples containing only inorganic individual preservatives, H3PO4 was proved to be the
arsenic species. best.
This result was very different with the previous
Comparison of preservatives in iron-poor water tests on the As(III)/As(V) equilibria. In the artificial
samples spiked with Fe2+ 2 mg l1 and As(III)
Preservation methods are introduced to sampling 200 mg l1 without any preservatives, As(III) con-
procedures to maintain both the amounts and species centration was dramatically decreased: 190 mg l1 at
of the elements of interest (i.e., arsenic in this study) fresh, 69 mg l1 after 24 h, then < 1 mg l1 after
from undergoing chemical changes due to metal 9 days (Daus et al. 2002). In reagent water initially
oxyhydroxide precipitation, photochemical oxidation containing an equal distribution of As(III) and As(V)
or redox reactions. Several reagents have been tested at 10 mg l1, the distribution of the arsenic species in
in the laboratory for the preservation of iron-rich samples without preservatives changes within 2 days
water samples containing arsenic (Daus et al. 2002; with or without light exposure: As(III) was reduced to
Bednar et al. 2002). As(V) (Bednar et al. 2002).
Herein, EDTA, H3PO4 and HCl were tested as Because As(III) concentration was maintained,
preservatives for groundwater from the Ulsan mine As(V) was inferred not to be transformed, but to be
(Fig. 4a) and surface water (Fig. 4b), both of which removed from water. A possible assumption is an
had a neutral pH and low iron contents. These adsorption of negative charged As(V) onto the
samples were analyzed using laboratory separation surface of the bottle or invisible settled particles. In
procedures employing HPLC-ICP-MS after storage this case, As(III), having a zero charge, is not

123
 Environ Geochem Health

Fig. 4 The stability of arsenic species with different preser- 23.7 mg l1 and Fe 0.03 mg l1). Dashed lines were
vatives over time compared using laboratory separation concentrations of As(III) (bottom) and As(V) (top) using field
methods for (a) groundwater (pH 7.8, AsT 73.3 mg l1 and separation methods
Fe < 0.01 mg l1), and (b) surface water (pH 7.1, AsT

adsorbed. Therefore, neutral and iron-poor water

samples for the arsenic analysis, even if for total
arsenic concentrations, should be measured within a
short time. To reduce bias, it would be better if
neutral and iron-poor water were not treated if
samples could be analyzed within 3 days (±10%
error compared with field separation); however, if
samples could not be analyzed within 3 days,
preservatives become necessary. The efficiencies of
the individual preservatives were similar, but H3PO4
proved to be the best.

Arsenic speciation in water samples

Figure 5 shows the analytical results for the real

water samples collected from various environments,
using different sampling methods and with specific
preservation procedures, that were then analyzed by
various arsenic species separation methods. A
groundwater sample from the tailings dam of the
Guryong mine (Fig. 5a) was initially treated with
H3PO4 and then analyzed by SPE-ICP-MS. Arsenic
was mainly present in the As(III) form, with a
concentration of 1.96 mg l1. A similar groundwater
sample from the tailings dam at the Ulsan mine
(Fig. 5b) was also treated with H3PO4, but this
sample was analyzed by HPLC-ICP-MS. This sample
contained both As(III) and As(V) at concentrations of
11.1 and 10.3 mg l1 respectively. A sample from a Fig. 5 Spectra of (a) iron-rich groundwater from the Guryong
mine analyzed by SPE-ICP-MS, (b) neutral groundwater from
pond in the Ulsan area was not treated, but was the Ulsan mine, (c) surface water from the Ulsan pond, (d)
analyzed within 24 h of sampling, which showed surface water from the Ulsan creek. Samples (b), (c), and (d)
As(V) to be the dominant species (26.0 mg l1), with were analyzed by HPLC-ICP-MS

123
Environ Geochem Health

small amounts of AsB, As(III) (1.9 mg l1) and DMA amounts of organic arsenic species. Hence, the
(3.4 mg l1; Fig. 5c). AsB is the predominant species SPE cartridge, used as a short column, is
in marine organisms, but is rare in freshwater appropriate for arsenic speciation in groundwater
(Dembitsky and Rezanka 2003). A creek water samples, but an HPLC column is required to
sample from the Ulsan area (Fig. 5d) showed high speciate arsenic in surface water samples.
arsenic concentrations. Although the dominant ar-
In conclusion, researchers should select their field
senic species in most surface waters is As(V), the
and laboratory procedures with caution and pay
As(III) concentration in this sample was 25.6 mg l1,
attention to the procedures of sample collection,
notably higher than the 11.0 mg l1 for As(V).
preservation, and analysis as well as consideration of
Construction involving the disturbance of surface
site characteristics.
sediments was being undertaken at this creek during
sampling, and the suspended surface sediments could Acknowledgement This research is a part of the Ph.D.
have influenced the arsenic species. dissertation work of Youn-Tae Kim, Yonsei University.
Financial support was provided by the research fund from
the Korea Basic Science Institute (N26051) and the Korea
Research Foundation (KRF-2005-C00081).
Conclusion
References
In order to evaluate the toxicity, mobility, and risk of
arsenic in waters, accurate speciation is required. The Akter, K. F., Chen, Z., Smith, L., Davey, D., & Naidu, R.
following conclusions can be drawn with regard to (2005). Speciation of arsenic in ground water samples: A
comparative study of CE-UV, HG-AAS and LC-ICP-MS.
the sampling and preservation required to enable Talanta, 68, 406–415.
accurate analyses of arsenic species: Aposhian, H. V., Zakharyan, R. A., Avram, M. D., Kopplin, M.
J., & Wollenberg, M.L. (2003). Oxidation and detoxifi-
1. The conventional field sampling method using cation of trivalent arsenic species. Toxicology and Applied
0.45-mmfilters could over-estimate the dissolved Pharmacology, 193, 1–8.
arsenic concentrations. Bang, S., Johnson, M. D., Korfiatis, G. P., & Meng, X. (2005).
Chemical reactions between arsenic and zero-valent iron
2. Various arsenic species show different responses
in water. Water Research, 39, 763–770.
in de-ionized water analyzed with ICP-AES, and Bednar, A. J., Garbarino, J. R., Ranville, J. F., & Wildeman, T.
different hydride generation efficiencies. There- R. (2002). Preserving the distribution of inorganic arsenic
fore, samples containing various arsenic species species in groundwater and acid mine drainage samples.
Environmental Science and Technology, 36, 2213–2218.
should be acidified or pre-reduced to As(III) prior
Bednar, A. J., Garbarino, J. R., Burkhardt, M. R., Ranville, J.
to analyzing total arsenic concentration. F., & Wildeman, T.R. (2004). Field and laboratory arsenic
3. For detailed assessment of arsenic speciation in speciation methods and their application to natural-water
water samples, the HPLC-ICP-MS procedure analysis. Water Research, 38, 355–364.
B’Hymer, C., & Caruso, J. A. (2004). Arsenic and its specia-
should be applied due to its capacity to separate
tion analysis using high-performance liquid chromatog-
more arsenic species, which are present in the raphy and inductively coupled plasma mass spectrometry.
solution, and to detect low concentrations. How- Journal of Chromatography A, 1045, 1–13.
ever, the procedure using SPE cartridges can be Choi, S.-G., & Youm, S.-J. (2000). Compositional variation of
arsenopyrite and fluid evolution at the Ulsan deposit,
used economically and conveniently for samples
Southeastern Korea: A low-sulfidation porphyry system.
containing only inorganic arsenic species. The Canadian Mineralogist, 38, 567–583.
4. Water samples with a neutral pH and low iron Daus, B., Mattusch, J., Wennrich, R., & Weiss, H. (2002).
concentration are stable for up to 3 days, but Investigation on stability and preservation of arsenic
species in iron-rich water samples. Talanta, 58, 57–65.
require preservative treatment if they are to be
Dembitsky, V. M., & Rezanka, T. (2003). Natural occurrence
stored for longer periods. Phosphoric acid of arseno compounds in plants, lichens, fungi, algal spe-
(H3PO4) proved to have the greatest preservative cies, and microorganisms. Plant Science, 165, 1177–1192.
efficiency in this study. Gallagher, P. A., Schwegel, C. A., Wei, X., & Creed, J. T.
(2001). Speciation and preservation of inorganic arsenic
5. Groundwater, especially that related to mine
in drinking water sources using EDTA with IC separation
activity, contained only inorganic arsenic spe- and ICP-MS detection. Journal of Environmental Moni-
cies. However, surface waters also showed small toring, 3, 371–376.

123
 Environ Geochem Health

Hughes, D. F. (2002). Arsenic toxicity and potential mecha- Raposo, J. C., Zuloaga, O., Sanz, J., Villanueva, U., Crea, P.,
nisms of action. Toxicology Letters, 133, 1–16. Etxebarria, N., Olazabal, M. A., & Madariaga, J. M.
Hung, D. Q., Nekrassova, O., & Compton, R. G. (2004). (2006). Analytical and thermodynamical approach to
Analytical methods for inorganic arsenic in water: A re- understand the mobility/retention of arsenic species from
view. Talanta, 64, 269–277. the river to the estuary. The Bilbao case study. Marine
Kim, J. Y. (2004). Mineralogical and physico-chemical prop- Chemistry, 99, 42–51.
erties of the tailings in the Guryoung Mine Area (Master Sánchez-Rodas, D., Gómez-Ariza, J. L., Giráldez, I., Velasco,
Thesis, Yonsie University). A., & Morales, E. (2005). Arsenic speciation in river and
Ko, I., Chang, Y.-Y., Lee, C.-H., & Kim, K.-W. (2005). estuarine waters from southwest Spain. Science of the
Assessment of pilot-scale acid washing of soil contami- Total Environment, 345, 207–217.
nated with As, Zn and Ni using the BCR three-step Thirunavukkarasu, O. S., Viraraghavan, T., Subramanian, K.
sequential extraction. Journal of Hazardous Materials A, S., & Tanjore, S. (2002). Organic arsenic removal from
127, 1–13. drinking water. Urban Water, 4, 425–421.
Maity, S., Chakravarty, S., Thakur, P., Gupta, K. K., Bhatta- Thomas, D. J., Styblo, M., & Lin, S. (2001). The cellular
charjee, S., & Roy, B.C. (2004). Evaluation and stan- metabolism and systemic toxicity of arsenic. Toxicology
dardization of a simple HG-AAS method for rapid and Applied Pharmacology, 176, 127–144.
speciation of As(III) and As(V) in some contaminated USEPA, IRIS, CASRN 7440-38-2, accessed in June 2006.
groundwater samples of West Bengal, India. Chemo- WHO (1993). Guidelines for drinking water quality, Vol. 1.
sphere, 54, 1199–1206. Geneva: World Health Organization.
Mandal, B. K., & Suzuki, K. T. (2002). Arsenic round the Yu, C., Cai, Q., Guo, Z.-X., Yang, Z., & Khoo, S. B. (2003).
world: a review. Talanta, 58, 201–235. Inductively coupled plasma mass spectrometry study of
Plant, J. A., Kinniburgh, D. G., Smedley, P. L., Fordyce, F. M., the retention behavior of arsenic species on various solid
& Klinck, B. A. (2005). Arsenic and selenium. In B. S. phase extraction cartridges and its application in arsenic
Lollar (Ed.), Environmental geochemistry (33 pp). Ox- speciation. Spectrochimica Acta Part B, 58, 1335–1349.
ford: Elsevier-Pergamon.

123