Talanta 77 (2008) 189–194

Contents lists available at ScienceDirect

Talanta
journal homepage: www.elsevier.com/locate/talanta

Speciated isotope dilution analysis of Cr(III) and Cr(VI) in water by ICP-DRC-MS

H.-L. Ma, P.A. Tanner ∗
Department of Biology and Chemistry, City University of Hong Kong, Tat Chee Avenue, Kowloon, Hong Kong S.A.R., PR China

a r t i c l e i n f o a b s t r a c t

Article history: An isotope dilution method has been developed for the speciation analysis of chromium in natural waters
Received 30 March 2008 which accounts for species interconversions without the requirement of a separation instrument con-
Received in revised form 5 June 2008 nected to the mass spectrometer. The method involves (i) in-situ spiking of the sample with isotopically
Accepted 5 June 2008
enriched chromium species; (ii) separation of chromium species by precipitation with iron hydroxide; (iii)
Available online 12 June 2008
careful measurement of isotope ratios using an inductively coupled plasma mass spectrometer (ICP-MS)
with a dynamic reaction cell (DRC) to remove isobaric polyatomic interferences. The method detection
Keywords:
limits are 0.4 ␮g L−1 for Cr(III) and 0.04 ␮g L−1 for Cr(VI). The method is demonstrated for the speciation
Chromium
Speciation
of Cr(III) and Cr(VI) in local nullah and synthetically spiked water samples. The percentage of conversion
Natural waters from Cr(III) to Cr(VI) increased from 5.9% to 9.3% with increase of the concentration of Cr(VI) and Cr(III)
Speciated isotope dilution from 1 to 100 ␮g L−1 , while the reverse conversion from Cr(VI) to Cr(III) was observed within a range
Dynamic reaction cell between 0.9% and 1.9%. The equilibrium constant for the conversion was found to be independent of the
Precipitation method initial concentrations of Cr(III) and Cr(VI) and in the range of 1.0 (at pH 3) to 1.8 (at pH 10). The precision of
ICP-MS the method is better than that of the DPC method for Cr(VI) analysis, with the added bonuses of freedom
from interferences and simultaneous Cr(III) determination.
© 2008 Elsevier B.V. All rights reserved.

1. Introduction
Chromium mainly occurs as the hexavalent and trivalent oxi-
dation states in the environment and a brief survey of their
chemistries is now given. Cr(III) is a hard acid which forms octa-
hedral coordination complexes with ligands such as water, sulfate,
urea, ammonia, and organic acids [1]. When Cr(III) exists as
Cr(H2 O)6 3+ , its deprotonated forms (which are formulated shortly
as CrOH2+ , Cr(OH)2 + and Cr(OH)3 ) are dominant within pH 4–10.
Cr(OH)3 is sparingly soluble, however, it forms the readily soluble
Cr(OH)4 − complex at higher pH [2]. Manganese oxides were found
to be the most effective oxidation pathway for Cr(III) rather than
molecular oxygen and H2 O2 in environmental systems [3].
Cr(VI) is present in the aqueous phase as chromate (CrO4 2− ,
HCrO4 − and H2 CrO4 ) or dichromate (Cr2 O7 2− ) ions, the propor-
tions of which depend on both pH and the total concentration of
Cr(VI). HCrO4 − is the predominant form within the pH range 2–6,
since H2 CrO4 is deprotonated at pH > 1 and CrO4 2− only exists in
the solution above pH 6 [4].
Above a certain concentration of Cr(VI), orange-red Cr2 O7 2− ions
are formed but with increase in the basicity, the color changes from
orange-red to yellow due to the formation of CrO4 2− ions. Cr(VI)
∗ Corresponding author. Tel.: +852 2788 7840; fax: +852 27887406.
E-mail address: bhtan@cityu.edu.hk (P.A. Tanner).
ions are easily reduced to Cr(III) by electron donors such as organic
matter, e.g. humic acid at low pH [5], and inorganic species such as
Fe2+ , which is thermodynamically possible over the entire pH field
[6]. Phosphate and sulfide also influence the rate of reduction of
Cr(VI) [7,8].
Because of differences in chemical behavior, toxicity and
bioavailability of (essential) Cr(III) and (carcinogenic) Cr(VI),
speciation is important in their environmental analyses. The
1,5-diphenylcarbazide (DPC) method is a well-known Cr(VI) deter-
mination method [9,10]. It is based on the reaction of DPC with
Cr(VI) at the pH of 1.0 ± 0.3 forming a colored complex which
can be detected by UV–vis spectrophotometry at 540 nm. How-
ever, isolation of Cr(VI) from sample extracts containing Fe(II) is
needed before spectrophotometric measurement is made because
the acidic environment would facilitate reduction of Cr(VI) by Fe(II)
[11] which would result in low recovery of Cr(VI). This method
also suffers from the presence of interfering species such as humic
acid, which is able to compete with DPC for Cr(VI) and which
also absorbs at 540 nm [8]. Other separation techniques for Cr(VI)
and/or Cr(III) have utilized ion chromatography [2,11,12], capillary
electrophoresis [13–16], and in particular, high performance liq-
uid chromatography coupled with an inductively coupled plasma
mass spectrometer (ICP-MS) is a convenient technique for simulta-
neously distinguishing both chromium species with low detection
limit [17–19]. The conventional speciation methods including the
colorimetric method, capillary electrophoresis and chromatogra-

0039-9140/$ – see front matter © 2008 Elsevier B.V. All rights reserved.
doi:10.1016/j.talanta.2008.06.005
190 H.-L. Ma, P.A. Tanner / Talanta 77 (2008) 189–194
phy, are prone to error due to Cr(III)–Cr(VI) interconversion during
sample processing. Speciated isotope dilution mass spectrometry
(SIDMS) has been developed to address species transformations
that can occur in the steps of sampling, storage, sample prepara-
tion and measurement. Accordingly it provides accurate analysis
of the original sample concentrations of the different chemical
species. No other measurement method can provide this informa-
tion. The procedures of SIDMS include the preparation of enriched
isotopic spikes that have each been chemically converted into a
single species, sample collection and spiking, species separation,
isotope ratio measurement, and deconvolution of species con-
centrations. The isotopically enriched species will undergo the
same reactions as the corresponding naturally occurring species
so that the interconversions that occur after spiking are trace-
able and can be corrected by use of the isotope ratios [20,21].
To illustrate SIDMS, the speciation of Cr(III) and Cr(VI) in water
samples has been carried in this study. We have combined the
SIDMS technique with the species separation method using iron
hydroxide precipitation. The co-precipitation speciation method
is based on the specific scavenging effect of Fe(III) hydroxide on
Cr(III) to form a precipitate which can be subsequently dissolved
in order to obtain the Cr(III) concentration [22–24]. On the other
hand, the total Cr concentration is determined by adding Fe(II)
hydroxide into the sample. The Fe(II) hydroxide reduces the in
situ Cr(VI) to Cr(III), which is subsequently scavenged by Fe(III)
hydroxide formed in the redox reaction, thus resulting in the
removal of all dissolved Cr species. The concentration of Cr(VI)
can be found by the difference of total and Cr(III) concentrations.
Therefore this method is capable of determining both dissolved
Cr(VI) and Cr(III) simultaneously. Besides, it requires minimal sam-
ple preparation and the equipment can easily be employed in
the field. The speciation of Cr(III) and Cr(VI) was further simpli-
fied in our method by only carrying out a single scavenging step
where Fe(III) hydroxide scavenges Cr(III) so that Cr(VI) remains
in the sample. Therefore, by analyzing the dissolved precipitate
and the filtrate, both Cr(III) and Cr(VI) can be determined sep-
arately. Finally, the problems of accurately measuring chromium
isotopic ratios due to polyatomic isobaric interferences such as
34 S16 O+ , 40 Ar12 C+ , 1 H35 Cl16 O+ , 40 Ar13 C+ , 37 Cl16 O+ , etc. have been
surmounted by employing dynamic reaction cell (DRC) technology
[25,26]. The reaction gas ammonia was found to be more efficient
than oxygen in breaking up such molecular ions.
2. Materials and methods
2.1. Instrumentation
An Elan 6100 DRC-ICP-MS system (PerkinElmer SCIEX Instru-
ments, USA) equipped with a peristaltic pump, Meinhard quartz
nebulizer, cyclonic spray chamber, nickel skimmer and sample
cones was employed. The instrument was optimized daily with
consideration of background count, sensitivity, as well as doubly
charged and oxides formation in the standard mode. The pH of
water samples was measured by a Metrohm 744 pH Meter.
2.2. Reagents and materials
Argon gas (99.995%) and ammonia reaction gas (99.999%) were
purchased from the Hong Kong Oxygen Co. and the Hong Kong
Special Gas Co., respectively. Isotopically enriched materials 50 Cr
metal with 96.9% enrichment (Lot# I1-8991B) and 53 Cr oxide with
97.9% enrichment (Lot# I1-8991A) were obtained from Cambridge
Isotope Laboratories Inc., USA. Standard reference material (SRM)
979 (isotopic standard for chromium) was purchased from the
National Institute of Standards and Technology, USA. Purified water
(18.2 M), from a Milli-Q water purification system (Millipore,
Billerica, MA, USA) was used to prepare all the solutions. Other
reagents employed are itemized in the Supplementary information.
2.3. ICP-DRC-MS optimization for isotope ratio measurements
The instrument parameters are listed in Table S1. The DRC con-
ditions were optimized independently from the remainder of the
ICP-MS system. The cell gas flow rate of ammonia and the high-pass
rejection parameter (Rpq) value (the low-pass parameter (Rpa) was
set at 0) were optimized to achieve the highest signal to noise ratio.
Two solutions were used for the DRC optimization. The blank solu-
tion was the solution containing the possible interfering species of
Cr and was composed of 1 mg L−1 S, 400 mg L−1 Cl, 0.5% CH3 OH, 2%
HNO3 . The standard solution contained the interfering matrix (blank
solution) together with a 2 ␮g L−1 Cr spike. Studies of the NH3 flow
rate in the range from 0.1 to 1 ml min−1 using different Rpq val-
ues changing from 0.4 to 0.7 on these solutions were investigated
(Fig. S1).
2.4. Dead-time correction
For accurate isotope ratio determination, both the detector dead
time and mass discrimination effects have to be corrected [27–30].
The dead time concerns the amount of time required for the detec-
tor and its associated electronics to process the signal from an
incident ion and since this effect becomes increasingly stronger as
the ion count rate increases it leads to significant biases in the mea-
sured ion intensity and therefore induces inaccuracy. For dead-time
correction, an optimum dead-time value for which the isotope ratio
is independent of concentration should be determined. The influ-
ence of dead time on the isotope ratios of 50 Cr/52 Cr and 53 Cr/52 Cr
was determined by measuring different concentrations of natural
isotopic abundance Cr standard SRM 979 with respective certified
ratios of 0.052 and 0.113. A sweep is counted each time the instru-
ment scans from the lowest to the highest mass specified. Results
were unsatisfactory using 20 sweeps so that 400 sweeps were
employed and then the instrument dead time of 50 ns was found
to maintain constancy in both interested isotope ratios (Fig. S2).
2.5. Mass discrimination correction
Mass discrimination occurs when ions of different mass to
charge ratios are transmitted or detected through the mass spec-
trometer with different efficiencies within the ICP-MS instrument,
and this results in non-uniform response across the mass range
[29]. The natural isotopic abundance standard SRM 979 was
employed to determine the external correction mass bias factors
from measurements of 50 Cr/52 Cr and 53 Cr/52 Cr isotope ratios at the
determined optimum dead time.
The exact concentrations of the 50 Cr(III) and 53 Cr(VI) enriched
spikes were determined by inverse IDMS. 0.0125 ml of 50 Cr(III)
and 53 Cr(VI) enriched spike solutions were separately spiked into
two 100 ml of 10 ␮g L−1 SRM 979 solutions, i.e. each containing
19.2 nmol of Cr. The dead-time corrected isotope ratios 50 Cr/52 Cr
and 53 Cr/52 Cr were measured in both enriched 50 Cr(III) and 53 Cr(VI)
spiked standard solutions by ICP-DRC-MS, and corrected by the
mass bias factors (Table S2).
2.6. Sample collection and spiking
Local nullah water samples were collected in Shek Kip Mei Nul-
lah on 8 and 15 March 2006, using a plastic bracket attached to a
 H.-L. Ma, P.A. Tanner / Talanta 77 (2008) 189–194 191

nylon wire, and stored in acid-cleaned 2 L polythene bottles. Sam- The calculated ˛ and ˇ are related to the equilibrium constant
ples were filtered by a cellulose mixed-ester filter in a syringe before (Kc ) between Cr(III) and Cr(VI) by Eq. (5):
spiking. Each sample was then spiked with a known quantity of
NxVI (1 − ˇ) + ˛NxIII
50 CrIII and 53 CrVI . The optimal ratio would be 50 CrIII and 53 CrVI in a Kc = (5)
NxIII (1 − ˛) + ˇNxVI
concentration that approaches approximately a 1:1 ratio with the
natural 50 CrIII and 53 CrVI respectively, and therefore the concentra- The method detection limits (MDL) were determined as three
tions of spikes depend upon the amounts of Cr(III) and Cr(VI) in the times the standard deviation of the concentrations of seven repli-
original sample. cates of blank reagent that were processed through the entire
analytical method with a sample and were 0.4 and 0.04 ␮g L−1
2.7. Species separation by iron hydroxide co-precipitation for the analysis of Cr(III) and Cr(VI), respectively, following the
described procedures.
1 ml of 0.01 M of iron(III) hydroxide solution prepared from
adding 1 ml of 10% (w/w) ammonium hydroxide to 50 ml of 0.01 M 3. Results and discussion
iron(II) ammonium sulfate solution and shaking for 24 h, was added
to the 140 ml sample. The sample was shaken for 1 h by a mechan- 3.1. ICP-DRC-MS optimization for isotope ratio measurements
ical shaker (Julabo SW-21C, Germany) and then suction filtered
through a 0.4 ␮m Nucleopore membrane filter pre-cleaned with The DRC optimization now discussed was used for isotope ratio
10% (v/v) HNO3 and milli-Q water. The sample filtrate was col- measurements on solutions. The first consideration of an accu-
lected for Cr(VI) analysis. The filter paper containing the green rate determination of the isotopic ratios of chromium concerns the
precipitate was sonicated with 5 ml of 50% (v/v) HNO3 for 30 min interferences upon the interested isotopes 50,52,53 Cr. The concen-
to extract Cr(III). The extract was filtered through a cellulose filter trations given above of the predominant interfering species arising
and then diluted to 50 ml with milli-Q water prior to ICP-DRC-MS from Cl, C, S and N were estimated from the average concentra-
measurement for Cr(III). Additional scavenging experiments were tions of Cl− , total organic carbon, S2− , inorganic nitrogen (NO3 −
carried out with Fe(II) hydroxide to check the total chromium con- and NO2 − ) and organic nitrogen from measurements in Hong Kong
centrations and the results were in agreement and not reported inland waters by the Hong Kong Environmental Protection Depart-
herein. ment [31]. In this study, DRC technology was found to successfully
overcome the problem of polyatomic interferences on Cr by using
2.8. Isotope ratio measurement and calculation ammonia gas under optimized operating conditions. Investigations
on varying the NH3 flow rates as well as the Rpq values were car-
The isotope ratios of 50 CrIII /52 CrIII , 53 CrIII /52 CrIII ,
four ried again to obtain reliable DRC operation parameters. Among the
50 CrVI /52 CrVI
and 53 CrVI /52 CrVI were measured and corrected. ranges of Rpq and cell gas flow rate studied, the operation at the
The number of moles of Cr(III) and Cr(VI) in the sample and the high- and low-pass rejection parameters Rpq = 0.45, Rpa = 0, and
degree of conversion between Cr(III) and Cr(VI) that occurred after flow rate of 0.45 ml min−1 provided a good estimated detection
spiking in sample were calculated by Eq. (1)–(4): limit, as well as strong signal intensity for all Cr isotopes, and were
 50  (NxIII 50 Ax + NsIII 50 AIII VI 50 A + N VI 50 AVI )ˇ thus selected to be the subsequent operating conditions (Fig. S1).
s )(1 − ˛) + (Nx x s s
RIII = By consideration of the ionization energies of Cr (6.77 eV) and
52 (NxIII 52 Ax + NsIII 52 AIII VI 52 A + N VI 52 AVI )ˇ
s )(1 − ˛) + (Nx x s s NH3 (10.16 eV), the electron transfer from NH3 to Cr+ is there-
(1)
fore endothermic, so that the interference removal mechanism is
expected to be preferentially by conversion into new species, rather
 53  (NxIII 53 Ax + NsIII 53 AIII VI 53 A + N VI 53 AVI )ˇ than by forming new cluster ions with Cr at different masses. Our
s )(1 − ˛) + (Nx x s s
RIII = results (not shown) demonstrated that Cr does not readily react to
52 (NxIII 52 Ax + NsIII 52 AIII VI 52 A + N VI 52 AVI )ˇ
s )(1 − ˛) + (Nx x s s form Cr(NH3 )n + (n = 2, 3, 4, . . ., 6) species and that only the signal
(2)
intensities of z Cr(NH3 )+ (n = 1; z = 50, 52, 53, 54) increased linearly
with different concentrations of Cr standard solutions. However,
 50  (NxVI 50 Ax + NsVI 50 AVI III 50 A + N III 50 AIII )˛ the signal intensity of 52 Cr(NH3 )+ was too weak to provide an accu-
s )(1 − ˇ) + (Nx x s s
RVI = rate determination.
52 (NxVI 52 Ax + NsVI 52 AVI III 52 A + N III 52 AIII )˛
s )(1 − ˇ) + (Nx x s s Using 400 sweeps of the mass spectrometer and the optimized
(3)
dead time of 50 ns, the isotope ratios obtained were 0.048 and
0.119 for 50 Cr/52 Cr and 53 Cr/52 Cr respectively, which are only cor-
 53  (NxVI 53 Ax + NsVI 53 AVI III 53 A + N III 53 AIII )˛ respondingly 8% and 4% in error with the certified values. The mass
s )(1 − ˇ) + (Nx x s s
RVI = discrimination factors per unit mass of 50 Cr/52 Cr and 53 Cr/52 Cr were
52 (NxVI 52 Ax + NsVI 52 AVI III 52 A + N III 52 AIII )˛
s )(1 − ˇ) + (Nx x s s
both calculated as 4.0%, which were then used for the mass dis-
(4)
crimination corrections of 50 Cr/52 Cr and 53 Cr/52 Cr in the sample by
and since the subscripts x and s refer to the sample and spike, applying Eqs. (S1, S2) (Table S2).
respectively: Rz (y/52) is the isotope ratio of y Cr to 52 Cr of Cr(z) (z = III
for y = 50 and z = VI for y = 53) in the spiked sample; Ax is the relative 3.2. Preparation of 50 Cr(III) and 53 Cr(VI) enriched spike solutions
abundance of the respective isotope of Cr in the original sample; AIII s
and AVI s are the relative abundances of the respective isotopes of Cr The isotopic abundances of the 50 CrIII and 53 CrVI enriched spike
in the enriched 50 CrIII and 53 CrVI isotopic standard spikes, respec- solutions prepared were determined and are shown in Table 1. The
tively; Nsz is the number of moles of Cr(z) in the enriched 50 Crz measured abundances were determined by calculating the ratio of
(z = III) and 53 Crz (z = VI) isotopic standard spikes, respectively; NxIII the net intensity of a specific isotope to the total intensity of the ele-
and NxVI are the numbers of moles of Cr(III) and Cr(VI) initially in ment (sum of intensity for all isotopes). The results showed good
the sample, respectively; ˛ is the percentage of Cr(III) oxidized to agreement between the certified and the measured abundances in
Cr(VI) after spiking; and ˇ is the percentage of Cr(VI) reduced to both enriched spike solutions, with good precision. The concentra-
Cr(III) after spiking. tions of the 50 CrIII and 53 CrVI enriched spikes were determined by
192 H.-L. Ma, P.A. Tanner / Talanta 77 (2008) 189–194

Table 1
Determination of isotopic abundance of the 50 Cr(III) and 53 Cr(VI) spike solutions prepared (identical values were obtained for two replicates)
50
Isotope CrIII spike solution 53
CrVI spike solution

Certified abundance (%) Measured abundance (%) Certified abundance (%) Measured abundance (%)
50
Cr 96.9 96.6 – 0.1
52
Cr – 3.2 – 1.9
53
Cr – 0.2 97.9 97.9
54
Cr – 0 – 0.1

inverse IDMS using SRM 979 and were found to be 90 and 67 mg L−1 , sample analyzed by ICP-DRC-MS. Thus about 90% of Cr in the nul-
respectively. lah water existed as Cr(VI). This may be due to the immobility or
absence of Cr(III) in this water system since it: (i) has a strong ten-
3.3. Assessment of interconversion between Cr(III) and Cr(VI) dency to complex with many inorganic or organic ligands; (ii) is
adsorbed by naturally occurring solids; and (iii) precipitates out
To investigate the influence of initial concentrations upon under neutral to basic condition. Thus Cr(VI) is generally the most
the distribution between Cr(III) and Cr(VI) in water, six sets of mobile Cr form in natural waters. The calculated concentrations of
iron hydroxide co-precipitation experiment were performed using both Cr(III), Cr(VI) and total Cr (i.e. labeled as Cx : corresponding
140 ml nullah water samples (pH 7.52). The water sample was fil- to the amount of natural abundance spikes plus the original small
tered by a syringe cellulose mixed-ester filter before spiking. Since amount present in the nullah water) are in good agreement with
the total chromium concentration in the neat nullah water samples their true concentrations. Subtraction of the natural isotopic abun-
(labeled A, B) was found to be only 0.9 ␮g L−1 , additional solutions dance spike concentrations (nat Cstotal ) from the total determined Cr
in duplicate (labeled C. . .L) were made up with natural abundance concentration (Cxtotal ) gives the concentration of Cr present in the
Cr(III) and natural abundance Cr(VI) spikes of up to 100 ␮g L−1 . original nullah water in the last column of Table 2. Good agreement
Appropriate concentrations (in a roughly 1:1 species ratio) of iso- is obtained for this small concentration (samples A–F inclusive)
topically enriched 50 CrIII and enriched 53 CrVI were also spiked except when two large numbers are being subtracted (samples
into the samples. An aliquot of 140 ml milli-Q water without any G–L).
spikes acted as a blank sample. Following iron(III) hydroxide treat- Satisfactory recoveries (of between 94% and 107%) were
ment and subsequent manipulations described above, the isotope obtained by using the proposed speciation method as shown in
ratio measurements of 50 CrIII /52 CrIII , 53 CrIII /52 CrIII , 50 CrVI /52 CrVI and Table 3. The percentage of conversion from Cr(III) to Cr(VI) (˛)
53 CrVI /52 CrVI were made separately for the speciated Cr(III) and increases from ∼6% to ∼10% with increase of the initial concen-
Cr(VI) samples. These isotope ratios were employed in the set of trations of Cr(III) and Cr(VI) each from <1 to >50 ␮g L−1 , while the
Eqs. (1)–(4) to calculate the initial amounts of Cr(III) (NxIII ) and Cr(VI) percentage of conversion from Cr(VI) to Cr(III) (ˇ) was between 1.6%
(NxVI ) in the sample, as well as the respective concentrations which and 0.9% in these ranges (Table 3, columns 7 and 8). This relatively
are listed in Table 2. small degree of interconversion from Cr(VI) to Cr(III) is expected
A and B are the samples without any natural Cr spikes (i.e. only under the basic medium since Cr(VI) then has a negative redox
nullah water with subsequent enriched isotopic spikes). The initial potential (E0 = −0.13 V) and exhibits thermodynamic stability under
Cr(VI) concentrations were found to be 0.7 ␮g L−1 , which is simi- aerobic conditions [32] whereas dissolved oxygen rapidly reacts
lar to the total Cr concentration of 0.9 ␮g L−1 in the nullah water with reducing agents such as S2− , Fe2+ and organic matter usually

Table 2
Concentrations of Cr(III), Cr(VI) and total Cr in the samples and the determined mean ± standard deviation values from replicate samples

Sample labels Concentration (␮g L−1 )

nat
CsIII nat
CsVI nat
Cstotal CxIII CxVI Cxtotal Cxtotal − nat Cstotal

A, B 0 0 0 <DLa 0.7 ± 0 0.8 ± 0.1 0.8 ± 0.1

C, D 5.2 5.4 10.6 5.2 ± 0.1 6.2 ± 0 11.4 ± 0.1 0.9 ± 0.1
E, F 10.7 10.5 21.2 10.5 ± 0.1 11.6 ± 0.1 22.1 ± 0.1 0.9 ± 0.1
G, H 20.7 20.9 41.6 18.9 ± 0.4 23.2 ± 0.5 42.1 ± 0.1 0.5 ± 0.1
I, J 49.1 52.3 103.9 46.9 ± 1.7 53.7 ± 0.7 100.7 ± 2.5 −3.4 ± 2.5
K, L 103.3 103.2 206.5 97.8 ± 0.5 106.6 ± 1.4 204.4 ± 1.9 −2.2 ± 1.9
nat z
Cs : initial concentration of natural abundance Cr(z) spiked into the sample; nat
Cstotal : total natural abundance Cr concentration in the sample; CxIII , CxVI and Cxtotal are the
determined concentrations of Cr(III), Cr(VI) and total Cr in the sample.
a
The detection limit for Cr(III) is 0.4 ␮g L−1 (3 value; N = 7).

Table 3
Chromium recoveries, interconversions and equilibrium constants for samples A–L (mean ± S.D.; N = 2)

Sample Recovery of Cr (%) Cr existing in the sample (%) Percentage of conversion (%) Equilibrium constant, Kc

Cr(III) Cr(VI) Cr(total) Cr(III) Cr(VI) ˛ ˇ

A, B – – 94.2 ± 4 (9.8 ± 0.4) 90.3 ± 0.4 5.9 ± 1.8 1.6 ± 0.1 nda
C, D 97.9 ± 1 102.4 ± 0 99.7 ± 0.5 45.1 ± 0.3 54.9 ± 0.3 8.5 ± 0.2 1.1 ± 0.1 1.4 ± 0
E, F 97.5 ± 1.2 103 ± 1.4 99.9 ± 0.1 47.7 ± 0.7 52.4 ± 0.7 8.0 ± 0.5 1.9 ± 0.6 1.2 ± 0
G, H 91 ± 2.2 107.3 ± 2.2 99.1 ± 0.1 44.9 ± 1.1 55.1 ± 1.1 8.4 ± 1.2 1.0 ± 0.1 1.4 ± 0.1
I, J 90.7 ± 3.3 101.4 ± 1.4 96.1 ± 2.3 46.6 ± 0.6 53.4 ± 0.6 12.1 ± 0.4 0.9 ± 0.1 1.4 ± 0
K, L 94.6 ± 0.5 102.6 ± 1.3 98.6 ± 0.9 47.9 ± 0.2 52.2 ± 0.2 9.3 ± 2 0.9 ± 0.1 1.3 ± 0
a
This value was not determined since the initial concentrations of Cr(III) (CxIII ) in samples A and B were below detection limit.
 H.-L. Ma, P.A. Tanner / Talanta 77 (2008) 189–194 193

Table 4
Summary of the determined Cr(III) and Cr(VI) in samples M–R, the percentage of interconversion, and the equilibrium constant between Cr(VI) and Cr(III) at different days
of analysis under different pH conditions

Sample label Day of storage pH Concentration of Cr in the sample (␮g L−1 ) Percentage of Conversion (%) Kc

CxIII CxVI ˛ ˇ

M, N 1 3.00 201.6 ± 1.3 204.5 ± 0.3 1.4 ± 0.1 2.7 ± 0.1 1.0 ± 0.0
2 2.89 201.1 ± 0.4 205.0 ± 0.3 1.6 ± 0.1 2.8 ± 0.1 1.0 ± 0.0
5 2.88 201.7 ± 0.1 203.5 ± 0.1 1.7 ± 0.1 3.3 ± 0.1 1.0 ± 0.0

O, P 1 7.00 201.0 ± 0.6 205.2 ± 0.1 1.1 ± 0.0 1.7 ± 0.1 1.0 ± 0.0
2 6.36 200.4 ± 0.0 206.7 ± 0.0 6.6 ± 0.0 2.4 ± 0.1 1.1 ± 0.0
5 6.26 200.5 ± 0.2 204.8 ± 0.5 3.7 ± 0.5 3.4 ± 0.0 1.0 ± 0.0

Q, R 1 10.02 202.6 ± 3.2 204.7 ± 1.7 30.8 ± 0.1 1.2 ± 0.0 1.8 ± 0.0
2 9.45 201.3 ± 0.2 207.2 ± 1.1 25.4 ± 2.5 1.1 ± 0.2 1.7 ± 0.1
5 9.35 201.8 ± 0.3 205.1 ± 0.1 30.0 ± 2.5 1.3 ± 0.0 1.8 ± 0.1

Values represent the mean ± S.D. (N = 2). The initial concentrations of Cr(III) and Cr(VI) were 201.3 ␮g L−1 and 204.0 ␮g L−1 , respectively.

present. On the other hand, the basic medium destabilizes Cr(III) Table 5
Concentrations of Cr(VI) (␮g L−1 ) determined in samples analyzed by SIDMS and
and converts it to Cr(VI) in the presence of oxidizing agents. How-
DPC methods. The initial concentrations were 405 ␮g L−1
ever, the manganese oxidation state existing in natural waters is
generally Mn(II) and not the potential oxidant Mn(IV). Besides, the Sample label Day of storage pH SIDMS method DPC method
oxidation of Cr(III) by dissolved oxygen was reported to be slow nat+iso
CxVI nat+iso
CxVI
[32] and the reaction was found to be insignificant when compared
S, T 1 3.00 404.5 ± 0.3 386.2 ± 0.1
with MnO2 [33]. 2 2.89 405.0 ± 0.3 411.2 ± 0.6
As expected, the equilibrium constant (Kc : Table 3, last column) 5 2.88 403.5 ± 0.1 407.3 ± 0.2
calculated from ˛ and ˇ was independent of the initial concen- U, V 1 7.00 405.2 ± 0.1 378.7 ± 0.5
trations of Cr(III) and Cr(VI) and maintained at the constant value 2 6.36 406.7 ± 0.0 403.6 ± 0.4
of 1.3 at room temperature and pH 7.52. Thus, the percentage of 5 6.26 404.8 ± 0.5 396.8 ± 0.4
conversion from Cr(III) to Cr(VI) increases more significantly than W, X 1 10.02 404.7 ± 1.7 381.3 ± 0.2
the conversion from Cr(VI) to Cr(III) with the increase of initial 2 9.45 407.2 ± 1.1 403.3 ± 0.5
concentrations of Cr(III) and Cr(VI). 5 9.35 405.1 ± 0.1 398.7 ± 0.1

3.4. Effect of pH and storage time

The proposed speciation method only allows co-precipitation of
Cr(III) with Fe(OH)3 to form Crx Fe1 − x (OH)3 when the pH is above
4 [32,34]. Thus, to investigate the influence of pH and storage time
on the distribution between Cr(III) and Cr(VI) in water, the pH val-
ues of three bottles each of 1.2 L of synthesized water samples
were adjusted to pH 3, 7 and 10, respectively, by using NaOH or
HNO3 . Then, natural isotopic abundance Cr(III) and Cr(VI), and iso-
topically enriched Cr(III) and Cr(VI) spikes were added, with the
final concentrations of each being 200 ␮g L−1 in each sample. Two
replicates of 140 ml sample at each adjusted pH (together with
an aliquot of 140 ml milli-Q water without any spikes acting as a
blank sample), were analyzed after days 1, 2 and 5 by the SIDMS
method. Immediately before the Fe(III) co-precipitation, the pH of
each sample was adjusted to 7. Then the isotope ratio measure-
ments of 50 CrIII /52 CrIII , 53 CrIII /52 CrIII , 50 CrVI /52 CrVI and 53 CrVI /52 CrVI
were made as described before. The calculated concentrations of
Cr(III) and Cr(VI) in the different samples labeled M–R and their
corresponding ˛ and ˇ are shown in Table 4. The calculated con-
centrations of both Cr(III), Cr(VI) and total Cr in the sample over the
whole pH range studied agreed with their true concentrations (i.e.
corresponding to the amount of natural abundance spikes added).
According to the results, a greater extent of the conversion from
Cr(VI) to Cr(III) was found under an acidic medium, whereas the oxi-
dation from Cr(III) to Cr(VI) dominated at higher pH. These results
are expected since Cr(VI) exhibits a high positive redox potential
under acidic media. By contrast, a negative potential occurs under
basic media which destabilizes Cr(III) and converts it to the Cr(VI)
oxidation state in the presence of oxidizing agents. The equilib-
rium constant, Kc (final column, Table 4) changed with pH from
1 (pH 3) to 1.8 (pH 10) but was temporally constant for a given
pH value.
3.5. Comparison of SIDMS with DPC method for Cr(VI) analysis
The DPC method for Cr(VI) analysis, with experimental details
as described in the Supplementary information, was also employed
for temporal and accuracy comparisons with the proposed SIDMS
analysis method. Table 5 shows the concentration of the natural
abundance Cr spikes in the samples analyzed by the SIDMS and
DPC methods at different pH and after storage for different times.
The Cr(VI) concentrations obtained from the DPC method fluctu-
ated more over the period of study (maximum error 6.5%) under
different pH conditions than those from SIDMS (maximum error
0.5%). Besides the problems of chemical and spectral interferences
mentioned in Section 1, the much simpler spectrophotometric DPC
method has the usual problems of colorimetric methods in ensuring
uniform color development and reproducibility.
4. Conclusions
Species-specific isotope dilution methods for Cr(III)/Cr(VI) anal-
ysis using mass spectrometry have been previously employed
together with the separation techniques HPLC [19] and ion chro-
matography [12]. The present method withdraws the requirement
for a separation technique. Furthermore, the mass spectrometry
component becomes more accurate with the incorporation of a
dynamic reaction cell into the instrument. The advantages of the
present method over the conventional spectrophotometric Cr(VI)
speciation methods are that it is more accurate and precise, and
that species interconversions prior to, and during the analysis, are
eliminated. The present method has the added bonus that there is
a built-in check upon the total chromium concentration in solution
by scavenging with Fe(II) hydroxide.
194 H.-L. Ma, P.A. Tanner / Talanta 77 (2008) 189–194
Appendix A. Supplementary data
Supplementary data associated with this article can be found,
in the online version, at doi:10.1016/j.talanta.2008.06.005.
References
[1] California Air Resources Board, Chromium and compounds, hexavalent
chromium Report ARB/SSD/SES, 1997, Available at www.arb.ca.gov/toxics/
tac/factshts/chromium.
[2] J. Kotas, Z. Stasicka, Environ. Pollut. 107 (2000) 263.
[3] C. Seigneur, E. Constantinou, Environ. Sci. Technol. 29 (1995) 222.
[4] F.A. Cotton, G. Wilkinson, Advanced Inorganic Chemistry, Wiley-Intersci., New
York, 1988 (Chapter 18).
[5] P.R. Wittbrodt, C.D. Palmer, Environ. Sci. Technol. 29 (1995) 255.
[6] D.L. Sedlak, P.G. Chan, Geochim. Cosmochim. Acta 61 (1997) 2167.
[7] M. Pettine, F.J. Millero, R. Passino, Marine Chem. 46 (1994) 335.
[8] M. Pettine, S. Capri, Anal. Chim. Acta 540 (2005) 231.
[9] APHA, AWWA, WEF, Standard Methods for the Examination of Water and
Wastewater, 20th ed., APHA, Washington, DC, 1998, pp. 3–65.
[10] United States Environmental Protection Agency, Method 7196A, 1992.
[11] K. Ashley, A.M. Howe, M. Demange, O. Nygren, J. Environ. Monit. 5 (2003) 707.
[12] K. Tirez, W. Brusten, A. Cluyts, J. Patyn, N. De Brucker, J. Anal. At. Spectrom. 18
(2003) 922.
[13] Y.M. Lui, J.K. Cheng, Electrophoresis 24 (2003) 1993.
[14] B. Baraji, M. Martinez, A. Sastre, M. Aguilar, J. Chromatogr. A 695 (1995) 103.
[15] G.Y. Jung, Y.S. Kim, H.B. Lim, Anal. Sci. 13 (1997) 463.
[16] S. Himeno, T. Nakashima, K.I. Sano, Anal. Sci. 14 (1998) 369.
[17] Y.L. Chang, S.J. Jiang, J. Anal. At. Spectrom. 16 (2001) 858.
[18] Z. Grosser, W. Reuter, P. Perrone, K. Neubauer, Water Environ. Lab. Solut. 11
(2004) 1.
[19] J. Meija, L. Yang, J.A. Caruso, Z. Mester, J. Anal. At. Spectrom. 21 (2006) 1294.
[20] H.M. Kingston, D. Huo, Y. Lu, S. Chalk, Spectrochim. Acta Part B: Atom. Spectrosc.
53 (1998) 299.
[21] United States Environmental Protection Agency, Elemental and speciated iso-
tope dilution mass spectrometry, U.S. Environmental Protection Agency, USEPA
6800, 1998.
[22] H.E. Cranston, J.W. Murray, Anal. Chim. Acta 99 (1978) 275.
[23] R.J. Kieber, G.R. Heiz, Environ. Sci. Technol. 26 (1992) 307.
[24] S.E. Kaczynskl, R.J. Kieber, Environ. Sci. Technol. 27 (1993) 1572.
[25] S.F. Kan, P.A. Tanner, Environ. Chem. 3 (2006) 149.
[26] S.F. Kan, P.A. Tanner, J. Anal. At. Spectrom. 19 (2004) 639.
[27] S.M. Nelms, C.R. Quetel, T. Prohaska, J. Vogl, P.D.P. Taylor, J. Anal. At. Spectrom.
16 (2001) 333.
[28] F. Vanhaecke, L. Balcaen, I. Deconinck, I.D. Schrijver, C.M. Almeida, L. Moens, J.
Anal. At. Spectrom. 18 (2003) 1060.
[29] L. Yang, R.E. Sturgeon, J. Anal. At. Spectrom. 18 (2003) 1452.
[30] K.G. Heumann, S.M. Gallus, G. Rädlinger, J. Vogl, J. Anal. At. Spectrom. 13 (1998)
1001.
[31] Hong Kong Environmental Protection Department, River water quality in
Hong Kong 2003, Report EPD/TR1/04, 2004, Available at http://www.epd.gov.
hk/epd/english/environmentinhk/water/river quality/rwq report03.html.
[32] F.Y. Seleh, T.F. Parkerton, R.V. Lewis, J.H. Huang, K.L. Dickson, Sci. Total Environ.
86 (1989) 25.
[33] D. Rai, L.E. Eary, J.M. Zachara, Sci. Total Environ. 86 (1989) 15.
[34] H. Zhang, R. Bartlett, Environ. Sci. Technol. 33 (1999) 588.