Demethylation of Methylated Arsenic Species during Generation of Arsanes with Tetrahydridoborate(1-) in Acidic Media

Article abstract of DOI:10.1021/acs.analchem.6b00735

Demethylation during generation of volatile hydrides (HG), i.e. formation of noncorresponding arsanes from monomethylarsonic acid (MAs^V), dimethylarsinic acid (DMAs^V), and trimethylarsine oxide (TMAs^VO) by the reaction of

Full text of DOI:10.1021/acs.analchem.6b00735

Article
pubs.acs.org/ac
Demethylation of Methylated Arsenic Species during Generation of
Arsanes with Tetrahydridoborate(1−) in Acidic Media
Karel Marschner,*,†,‡ Stanislav Musil,† and Jirí Dedina†
̌
̌
†Institute of Analytical Chemistry of the CAS, v. v. i., Veverí
̌
97, 602 00 Brno, Czech Republic
‡Department of Analytical Chemistry, Faculty of Science, Charles University in Prague, Albertov 8, 128 43 Prague, Czech Republic
S
* Supporting Information
ABSTRACT: Demethylation during generation of volatile
hydrides (HG), i.e. formation of noncorresponding arsanes
from monomethylarsonic acid (MAsV), dimethylarsinic acid
(DMAsV), and trimethylarsine oxide (TMAsVO) by the reaction
of sodium tetrahydridoborate(1−) (THB) with different acids
under analytical conditions, was investigated and characterized.
Pronounced demethylation of MAsV, DMAsV, and TMAsVO was
found during the reaction of THB with HCl, H2SO4, and HClO4,
while HG from CH3COOH or TRIS buffer after prereduction
with L-cysteine resulted in the formation of only the
corresponding hydrides. In the case of HNO3 formation of
corresponding hydrides was preserved for MAsV and DMAsV but
not for TMAsVO. The extent of demethylation strongly depends
on concentration of the acid and THB. It can be strongly suppressed in HCl medium by partial hydrolysis of THB with optimal
concentration of acid before it reacts with MAsV, DMAsV, or TMAsVO. It appears that the demethylation is due to the action of
specific hydrolytic products of THB (most probably by the first and second one).
eneration of volatile hydrides (HG) is a popular sample
introduction technique for trace element and speciation
(H2O)BH(OH)2 → B(OH)3 + H2
(6)
G
In fact this scheme is simplified. For example dihydridoboron
analysis of “hydride forming” elements As, Se, Sb, Sn, Bi, Pb,
Te, and Ge.1 The reaction between tetrahydridoborate(1−)
(THB) and acid is almost exclusively used for HG when an
analyte in the liquid phase is converted to volatile hydride that
is released from the liquid matrix and subsequently introduced
in the gaseous phase into a detector, typically atomic or mass
spectrometers.
THB in acidic medium is stepwise hydrolyzed to boric acid
and hydrogen. Hydride active As species are converted to
arsanes by reactions with THB and/or with its hydrolyzed
intermediates. This can be described by the following schemes
1 and 22
n
species should be written as L2BH2 , where L could be one or
more groups among H2O, OH−, and Cl− and n is the charge
(−1, 0, or +1). Other hydridoboron species may be treated
similarly.3
Wang and Jolly4 studied THB hydrolysis in water−methanol
at lower temperature, and they found that BH3(H2O) could
+
generate two hydridoboron species (H2O)2BH2 and (H2O)-
BH2OH, with an equilibrium between these species
H3O+ + (H2O)BH2OH ⇄ (H2O)2BH2 + H2O
+
(7)
Only (H2O)BH2OH species undergoes hydrolysis. There-
fore, hydrolysis of THB is much faster when [H3O+] < 0.5 mol
L−1. That is in agreement with results of D’Ulivo et al.5 who
found that increasing acidity from 0.2 to 10 mol L−1 H3O+
decreased the decomposition rate of THB.
Among approaches of employing HG for As speciation
analysis, the generation of substituted arsanes provides several
essential advantages.6 Its general scheme consists of two steps.
The first step is generation of (substituted) arsanes from
hydride active species, i.e. AsH3 is formed from arsenite (iAsIII)
and arsenate (iAsV), CH3AsH2 from monomethylarsonic acid
BH4 + H3O+ + 2H2O → intermediates → B(OH)3 + 4H2
−
(1)
As (hydride active species) + BH4−/intermediates → arsanes + B(OH)3
(2)
Hydrolysis of THB under strong acidic conditions in
aqueous medium is described by the following reactions 3−6
postulated by D’Ulivo et al.2
−
BH4 + H3O+ → BH3(H2O) + H2
(3)
BH3(H2O) + H2O → (H2O)BH2OH + H2
(4)
Received: February 24, 2016
Accepted: May 31, 2016
(H2O)BH2OH + H2O → (H2O)BH(OH)2 + H2
(5)
© XXXX American Chemical Society
A
DOI: 10.1021/acs.analchem.6b00735
Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry
Article
(MAsV), (CH3)2AsH from dimethylarsinic acid (DMAsV), and
(CH3)3As from trimethylarsine oxide (TMAsVO).7 The second
step is the separation of generated arsanes. There are practically
only two ways employed to separate arsanes: (i) cryotrapping
(CT) when arsanes are trapped in a liquid nitrogen cooled U-
trap and subsequently volatilized gradually according to their
different boiling points,8−15 or (ii) gas chromatography (GC)
when arsanes are injected into a GC column.16−18 The
advantage of HG−CT is possibility to achieve low detection
limits, due to preconcentration of arsanes in CT. The
disadvantage is the laboriousness of this method.
The explicit relation between a hydride active species and its
corresponding arsane, for example between MAsV and
CH3AsH2, is a necessary condition for reaching accurate results
of the speciation analysis. It is generally assumed that an As−C
bond (in methylated species) is not cleaved during reaction 2
under analytical conditions and therefore that As species are
converted only to their corresponding arsanes.7,19 However,
some kind of molecular rearrangement of methylated As
species was investigated under nonanalytical conditions
recently.19 There were also scarce reports from the past
indicating As species rearrangement: Talmi and Bostick20 and
Van Elteren et al.21 observed that DMAsV species was
converted not only to its corresponding hydride, (CH3)2AsH,
but also to AsH3 and CH3AsH2. MAsV was converted not only
to CH3AsH2 but also partly to AsH3 and even (CH3)2AsH.
These experiments were performed in the medium of HCl, and
Talmi and Bostick20 referred to the fact that the conversion to
the noncorresponding hydride was pH dependent. A different
example of cleavage of the As−C bond by HG in HCl medium
is known for arsenosugars that contain, however, a much longer
side chain than methyl.22 The volatile product was identified as
(CH3)2AsH with a very low yield of 6−9%.23 The authors
explained this cleavage by formation of the analyte-borane
complex which withdrew electron density from methylene
making it more susceptible to hydride attack. This indicates that
As speciation analysis based on the generation of substituted
arsanes in the HCl media (which is the most used acid for HG)
could provide inaccurate speciation information.
inorganic arsenic (approximately half of iAsIII and half of iAsV).
The solution of NaBH4 (THB, Fluka, Germany) in 0.1% KOH
(Lach-Ner, Czech Republic) was prepared fresh daily. HCl
(p.a., Merck, Germany), H2SO4 (suprapure, Merck, Germany),
HClO4 (p.a., Merck, Germany), HNO3 (semiconductor grade,
Sigma-Aldrich, Germany), CH3COOH (Lachema, Czech
Republic), and tris(hydroxymethyl)aminomethane (TRIS)
buffer were used for HG. TRIS buffer was prepared from
Trizma hydrochloride (Sigma, Germany) - pH was adjusted to
6.00 by KOH (Lach-Ner, Czech Republic). L-Cysteine
hydrochloride monohydrate (Merck, Germany) for prereduc-
tion of As species was added to the As standard 1 h prior to
analysis. Solid NaOH (Lach-Ner, Czech Republic) in the form
of 3 mm pearls was used as a filling of a NaOH dryer.25
When explicitly stated, solutions of THB, HCl, DIW, and As
standards used for HG were purged by Ar for 30 min to remove
dissolved oxygen.
Hydride Generator. The generator (Figure 1) consisted of
peristaltic pumps (PP1 and PP2, Reglo Digital, Ismatec,
Figure 1. Batch hydride generator (a) without and (b) with hydrolysis
coil (HC); PP1/PP2 − peristaltic pumps, DIW − deionized water,
GLS − gas−liquid separator, HC − hydrolysis coil.
The aim of the present work was to (i) characterize the
noncorresponding volatile products of HG of methylated As
species (MAsV, DMAsV, and TMAsVO) in acidic media under
analytical conditions and (ii) find how the conditions of HG
influence the formation of corresponding and noncorrespond-
ing arsanes. For separation and detection of generated gaseous
arsanes CT with atomic fluorescence detection was used.
Switzerland), a glass gas−liquid separator (GLS) with a forced
outlet of total volume of 7 mL, a chemifold between the pumps
and the GLS, the dryer (10 cm long polypropylene tube of 1.7
cm i.d. filled with solid NaOH pellets), and a cryogenic trap
(CT). The chemifold was built using polytetrafluorethylene
(PTFE) tubes (0.75 mm i.d.) and polyetheretherketone T-
junctions. The CT was described previously,15,24 and it
consisted of a 300 mm glass U-tube (2.4 mm i.d.) packed
with Chromosorb WAW-DMCS 45/60, 15% OV-3 (Supelco,
USA). The U-tube was wrapped with a resistance wire for
gradual heating by an appropriate voltage of 21 V (∼1 A).
Acid was pumped by the pump PP1 at the flow rate of 4 mL
min−1. As standard was introduced into the acid channel by a
six-port injection valve (Rheodyne, USA) with a 0.6 mL loop
volume. A carrier gas (argon or helium) merged the acid
channel downstream to the standard introduction point. The
mixture was then introduced to the bottom of the GLS so that
carrier gas could bubble through the liquid. The acid channel
was provided with an additional T-junction upstream with the
GLS that served for cleaning of the GLS with DIW manually by
a syringe. Another channel was used to remove liquid waste
from the GLS manually by a syringe. The solution of THB was
pumped by the pump PP2 at the flow rate of 1 mL min−1 (see
Figure 1a), and it was introduced to the bottom of the GLS
EXPERIMENTAL SECTION
■
Standards and Reagents. Deionized water (DIW; < 0.2
μS cm−1, Ultrapur, Watrex, USA) was used for the preparation
of all solutions. A 1000 μg mL−1 As standard solution (Merck,
Germany) was used as iAsV stock standard solution. Stock
solutions of 1000 μg mL−1 As were prepared for iAsIII, MAsV,
DMAsV, and TMAsVO species in DIW using the following
compounds: As2O3 (Lachema, Czech Republic); Na2CH3AsO3·
6H2O (Chem. Service, USA); (CH3)2As(O)OH (Strem
Chemicals, Inc., USA); and (CH3)3AsO (obtained by courtesy
of Dr. William Cullen, University of British Columbia, Canada).
The total As content of methylated arsenic species standards
was confirmed by liquid sampling graphite furnace-atomic
absorption spectrometry as described previously.24 The purity
of all species was confirmed by HPLC-ICPMS as >99.9% for
DMAsV and TMAsVO and 97.5% for MAsV that contained 2.5%
B
DOI: 10.1021/acs.analchem.6b00735
Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry
Article
below the liquid level. If explicitly stated the THB flow was
mixed with the other acid flow and pumped by the second
channel of the pump PP2, at the flow rate of 1 mL min−1
(Figure 1b). THB was thus partially decomposed (hydrolyzed)
in a hydrolysis coil (HC) downstream the point of mixing the
THB flow with the other acid flow. The output of the gases
from the GLS was connected to the dryer. The filling of the
dryer was changed approximately every 3 days.25
The generator was operated either in the direct transfer
mode or in the collection one. (i) In the direct transfer mode −
the carrier gas was argon, and the NaOH dryer was directly
connected to the T-junction serving to introduce hydrogen
(flame hydrogen) and argon (flame argon) for an atomizer (not
shown in Figure 1). (ii) The setup of the CT collection mode is
shown in Figure 1 − the carrier gas was helium.
or TMAsVO) were treated. In these cases, we present “arsane
fraction” which is defined as a peak area of the given arsane
divided by the sum of peak areas of all arsanes recorded in the
same chromatogram. Arsane fractions were corrected (sub-
tracted) to inorganic arsenic impurities in the MAsV standard.
When appropriate, we deal with sensitivity. In the direct
transfer mode it is expressed as the peak area of a given species
related to its concentration. In the CT collection mode, it
means sum of the peak areas (of the single chromatogram)
related to the concentration of the standard. All measurements
were performed at least in triplicate and error bars in the figures
mean standard deviations.
RESULTS AND DISCUSSION
■
Effect of HCl Concentration. Figure 2 shows dependence
of the normalized peak area of iAsIII, iAsV, MAsV, DMAsV, and
Atomizer. A miniature diffusion flame (MDF) was used as
the atomizer. The atomizer optimum observation height (7
mm), flow rates of flame hydrogen (200 mL min−1), and flame
argon (600 mL min−1) were optimized previously − see ref 26
for details.
Spectrometer. The in-house assembled research grade
nondispersive AFS was employed. The experimental setup of
AFS is described elsewhere.24,26 Briefly, the instrument was
equipped with an As electrodeless discharge lamp (EDL system
II, PerkinElmer, USA) as a radiation source, an interference
filter (193 nm, full width at half-maximum 18.7 nm, CVI Melles
Griot, USA) to isolate fluorescence radiation from the atomizer,
and a solar blind photomultiplier (165−320 nm, PerkinElmer
Optoelectronics, USA) as the detector.
Procedure and Data Evaluation. Direct Transfer Mode.
A standard (concentration of 2 ng mL−1) was injected into the
flow of acid, and it was dosed (by PP1) into the GLS. The total
volume of liquid was 1.2 mL (0.6 mL of standard and 0.6 mL of
acid). Subsequently, recording of the fluorescence signal was
switched on (210 s read time), and 2 mL of THB was being
pumped into the GLS. After recording had been completed, the
remaining liquid was removed from the GLS, and the GLS was
cleaned with DIW.
Figure 2. Dependence of normalized peak area of As species on
III
V
■
●
▲
concentration of HCl; (black) − iAs , (red) − iAs , (blue)
−
V
V
V
⧫
▼
MAs , (green) − DMAs , (brown) − TMAs O; concentration of
As species was 2 ng mL−1, 2 mL of 1% THB addition.
CT Collection Mode. Dosing of standard (concentration of
0.5 or 2.0 ng mL−1 in the case of TMAsVO when the HG
efficiency was low) and acid (1.2 mL total) was the same as in
the direct transfer mode. Then approximately 3/4 of the U-tube
was manually immersed into liquid N2. Immediately thereafter
1 mL of THB was added. If explicitly stated, acid was pumped
and mixed with THB in the HC as illustrated in Figure 1b.
After introduction of THB had been accomplished, another 150
s was allowed to complete the reaction and for transport of
arsanes from the GLS to the CT. The volatilization step started
by removing the flask with liquid N2 and switching on heating
of the U-tube. The signal (chromatogram) was recorded during
the entire volatilization step (90 s read time). Finally, the
heating was switched off, the remaining liquid was removed,
and the GLS was cleaned in the same way as in the direct
transfer mode.
Peak area was invariably employed as the analytical quantity.
Since the AFS detector output yields a response in μV units
that are specific to the instrument, peak area values cannot be
simply related to those yielded by different instruments.
Therefore, in the direct transfer mode, we deal with normalized
peak areas − related to that of 2 ng mL−1 iAsIII that was
generated by 1% THB from 1 mol L−1 of the used acid (HCl,
HNO3, and CH3COOH). In the most experiments in the CT
collection mode, only single species standards (MAsV, DMAsV,
TMAsVO on concentration of HCl measured in the direct
transfer mode. Taking into account that the atomization
efficiency of all arsanes in the MDF is equal24 and that iAsIII can
be converted to the arsane with 100% efficiency under
optimized conditions of HG,27 it is evident that all the As
species except TMAsVO are completely converted to arsanes in
the range from 0.5 to 4 mol L−1 of HCl, while TMAsVO is
completely converted to arsanes when concentration 1 mol L−1
or lower is used. It should be highlighted that increasing HCl
concentration inevitably leads to broadening MAsV, DMAsV,
and TMAsVO peaks and narrowing of iAsIII and iAsV peaks. At
HCl concentration of 2 mol L−1 or higher, the peak broadening
resulted even to a decrease of normalized peak area of
TMAsVO (Figure 2). Although this could be compensated with
addition of a higher volume of 1% THB, it was not further
employed due to a pronounced increase of iAs content in the
blanks.
However, in the CT collection mode when arsanes were
generated from a mixed standard containing the same
concentrations of iAsIII/iAsV, MAsV, and DMAsV, a higher
peak area of AsH3 and lower peak areas for CH3AsH2 and
(CH3)2AsH were obtained. Namely, the use of 1 mol L−1 HCl
and 1% THB resulted in 95 3% and 77 4% peak area for
C
DOI: 10.1021/acs.analchem.6b00735
Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry
Article
MAsV and DMAsV, respectively, relative to the peak area of
AsH3. This should be related to exactly the same normalized
peak area of iAsIII+V, MAsV, and DMAsV in 1 mol L−1 HCl
(Figure 2). This discrepancy suggests a demethylation of MAsV
and DMAsV under the employed HG conditions.
ascribed to different concentrations of standards which were
used in this work (0.5 ng mL−1), while Talmi and Bostick
worked with concentrations 2 to 5 orders of magnitude higher.
Effect of THB Concentration. The effect of THB
concentration on demethylation was investigated in the
concentration range from 0.25 to 4% at three different
concentrations of HCl − 0.25, 1.0, and 6.0 mol L−1. Figure 5
Fractions of all arsanes arising from MAsV, DMAsV, and
TMAsVO were determined in the concentration range from
0.05 to 6.0 mol L−1 of HCl. Dependence for MAsV is shown in
Figure 3. (The dependences for DMAsV and TMAsVO are in
Figure 5. Dependence of arsane fractions formed from MAsV on
●
▲
concentration of THB; (black)
− AsH3, (blue)
− CH3AsH2;
concentration of MAsV was 0.5 ng mL−1, 0.25 mol L−1 HCl, 1 mL of
THB addition.
Figure 3. Dependence of arsane fractions formed from MAsV on
●
▲
concentration of HCl; (black)
− AsH3, (blue)
− CH3AsH2;
concentration of MAsV was 0.5 ng mL−1, 1 mL of 1% THB addition.
shows the dependence of volatile products formed from MAsV
at 0.25 mol L−1 HCl. The maximum conversion to AsH3 occurs
at 2% of THB; an increasing or decreasing concentration of
THB leads to a lower fraction of AsH3. The dependences were
similar for 1.0 and 6.0 mol L−1 of HCl − just the extent of
conversion to noncorresponding hydrides was lower − in
agreement with Figure 3. Similar influence of THB at 0.25 mol
L−1 HCl was observed for DMAsV and TMAsVO. Maximum
conversion to CH3AsH2 from DMAsV occurred at 2% of THB
with a small fraction of AsH3 as well, and TMAsVO had the
maximum conversion to noncorresponding hydride at 1% (see
the Supporting Information for Figures S-2a,b). Maximum
conversion to noncorresponding hydride was always observed
at 1−2% of THB. Since the demethylation dependences on the
concentration of HCl and THB were similar for MAsV, DMAsV,
and TMAsVO, we suppose that the demethylation is caused by
the influence of hydrolytic products of THB.
the Supporting Information − Figure S-1a,b). Typical
chromatograms of volatile products from DMAsV using 0.25
and 1.0 mol L−1 of HCl are shown in Figure 4a,b. In the case of
MAsV at a lower concentration of HCl a higher fraction of
noncorresponding hydride (AsH3) is formed at the expense of
CH3AsH2: almost 28% of MAsV is converted to AsH3 at 0.05
mol L−1 HCl (Figure 3). The dependence for DMAsV and
TMAsVO showed a similar trend as MAsV. A minimum for the
fraction of (CH3)2AsH formed from DMAsV was found in the
region 0.1−0.25 mol L−1 HCl with the value of 63 1%. The
fraction of (CH3)3As from TMAsVO had a minimum at 0.1 mol
L−1 HCl with the value of 50 3%. Higher HCl concentration
is more suitable for generation of the corresponding hydride. In
contrast to Talmi and Bostick,20 neither formation of
(CH3)2AsH from MAsV nor (CH3)3As from DMAsV was
observed in the whole concentration range. This could be
Figure 4. HG−CT−AFS chromatograms of volatile products generated from DMAsV using (a) 0.25 mol L−1 HCl, (b) 1.0 mol L−1 HCl, (c) 1.0 mol
L−1 HNO3; red solid line − DMAsV standard, black dashed line − blank; concentration of DMAsV was 0.5 ng mL−1, 1 mL of 1% THB addition.
D
DOI: 10.1021/acs.analchem.6b00735
Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry
Article
The above observations indicate that in the presence of HCl,
THB (or some of its hydrolysis intermediates) can cleave the
As−C bond. The extent of demethylation for As species was
always in the order of TMAsVO > DMAsV > MAsV; this can be
attributed to the number of As−C bonds in the molecule.
Effect of Dissolved Oxygen and Day-to-Day Repeat-
ability. Talmi and Bostick20 reported that removing dissolved
oxygen from all reagents strongly suppressed the demethylation
extent. To check the potential oxygen role, all solutions were
purged by Ar to remove dissolved oxygen. Demethylation for
both species (MAsV and DMAsV) decreased by one-third (at
0.25 mol L−1 HCl with 1% THB) indicating that dissolved
oxygen plays some role, but it is not a dominant factor.
Demethylation was quite repeatable during the day (relative
standard deviation of the arsane fraction was typically below
1%); however, day-to-day repeatability was significantly worse
(e.g., demethylation of MAsV in 1 mol L−1 HCl and 1% THB
ranged between 8−17% during the days). This could be
ascribed to differing oxygen concentrations in reagents.
Figure 6. Dependence of arsane fractions formed from MAsV on
●
concentration of HCl used for prereaction of THB; (black) − AsH3,
V
▲
(blue) − CH3AsH2; HC volume was 40 μL, concentration of MAs
Prereaction of THB with HCl. THB hydrolysis can be, to a
certain extent, controlled when THB is online mixed with HCl
solution in the HC before the mixture is pumped into the GLS.
The degree of THB hydrolysis is thus influenced by HCl
concentration and by the reaction time which is proportional to
the volume of the HC. Unfortunately, there are no kinetic data
on THB hydrolysis in acidic aqueous media at ambient
temperature. The reactions are too fast to measure; therefore, it
is quite speculative to determine which hydrolytic product is a
dominant species after partial hydrolysis, but it is certainly not
BH4− because hydrolysis of BH4− is too fast (Theoretically 99%
was 0.5 ng mL−1, 1 mL of 1% THB addition, 0.6 mL of 0.25 mol L−1
HCl added with standard.
When the HC was enlarged to 825 μL (2 mol L−1 HCl was
mixed with THB in the HC), the demethylation decreased to
0.0 0.7%, 1.2 0.3%, and 1.7 0.2% for MAsV, DMAsV, and
TMAsVO, respectively, and no decrease of sensitivity was
observed for MAsV and DMAsV. The sensitivity for TMAsVO
decreased to 72%. The decrease in demethylation with the
larger HC could be caused by hydrolysis of an “acid-resistant”
n
−
hydridoboron species − L2BH2 . The next and last hydrolytic
of BH4 is hydrolyzed in less than 3 μs at pH 0, but the usual
product which can react with analyte is L3BHn. These results
mixing time in the HG reaction system is approximately 20−30
ms.).7,28 Due to the design of the hydride generator the
minimal volume of the HC was approximately 40 μL; this
corresponded to the residence time between 0.12 and 0.17 s
(calculated on the assumption that two or three molecular
hydrogens arise − reactions 3−5).
suggest that L3BHn species does not cleave the As−C bond.
n
L2BH2 species probably can cleave the As−C bond but is not
mainly responsible for demethylation. Since BH4− is hydrolyzed
too fast, it is unlikely that this species can act as a derivatization
agent for HG at all.3 Therefore, it is even more unlikely that it
n
cleaves the As−C bond. Consequently, LBH3 seems to be
Using the 40 μL of HC, 1% THB with 0.1% KOH (0.28 mol
L−1 together) was mixed online with HCl at a concentration
from 0 to 6 mol L−1. Sensitivity during these measurements was
constant and equal for both MAsV and DMAsV with the
exception of 6 mol L−1 HCl when the sensitivity for DMAsV
decreased to approximately 54% of the maximum. In the case of
TMAsVO the sensitivity decreased exponentially from a
concentration of 0.5 mol L−1 (at the 6 mol L−1 the sensitivity
was only 7% of the maximum). Figure 6 demonstrates that
demethylation decreased for MAsV with a higher concentration
of HCl reaching a minimum in the range from 0.25 to 1 mol
L−1. These minima correspond to 0.3 0.2% of AsH3 formed
from MAsV. HCl concentration of 0.1 mol L−1 of HCl is
obviously not high enough to prereact THB sufficiently. For 2.0
mol L−1 HCl and higher there is a slight but significant increase
in demethylation: 5.0 2.0% of AsH3 from MAsV. This can be
caused by hydridoboron intermediates which are more “acid-
responsible for the demethylation process. This is supported by
−
the fact that (HO)BH3 species has an even higher reduction
2
−
power than BH4 .
Influence of identity of a ligand (H2O, OH−, and Cl−) in
LBH3 on the demethylation effect is not considered here,
n
because there are no data on which one of those species should
be present in the majority.
Other Acids. The use of other acids was also investigated
with respect to demethylation during HG − HClO4, HNO3,
H2SO4, and CH3COOH. HG from TRIS buffer after treatment
with 2% L-cysteine, which was found suitable for speciation
studies,15,24 was also tested and compared.
Perchloric and Sulfuric Acid. The volatile products of MAsV,
DMAsV, and TMAsVO formed from the reaction of THB with
HClO4, H2SO4, and HCl (for comparison) are shown in Table
1. One mol L−1 HClO4 shows a similar demethylation effect on
MAsV compared to HCl, and the effect is slightly higher in the
case of DMAsV and TMAsVO. In the case of 0.5 mol L−1 H2SO4
the demethylation of MAsV, DMAsV, and TMAsVO was even
higher. Sensitivities were almost identical with those achieved in
HCl medium. The mechanism of demethylation in the case of
HClO4 and H2SO4 is probably similar as for HCl. However, in
the presence of these acids only hydridoboron species with
ligand H2O and OH− are formed because perchlorate, sulfate,
and hydrogensulfate are weakly coordinating anions. Since the
n
resistant” like L2BH2 (eq 7). The similar behavior was also
observed for DMAsV and TMAsVO which had a minimum at
0.5 mol L−1 yielding 1.3 0.2% of CH3AsH2 and 1.4 0.0% of
(CH3)2AsH, respectively. The increase in fractions of non-
corresponding arsanes at 2.0 mol L−1 to values 3.8 0.1% of
CH3AsH2 from DMAs and 4.7 0.3% (CH3)2AsH from TMAs
VO was also evident (see the Supporting Information for
Figures S-3a,b).
E
DOI: 10.1021/acs.analchem.6b00735
Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry
Article
Table 1. Volatile Products Formed from Methylated Arsenic Species by Reaction of THB with Different Acids
MAsV
DMAsV
TMAsVO
acid (mol L−1
)
AsH3
CH3AsH2
AsH3
1%
CH3AsH2
(CH3)2AsH
AsH3
0%
CH3AsH2 (CH3)2AsH
(CH3)3As
HCl (1)
12 1%
12 1%
28 1%
88 3%
88 1%
72 2%
4
3
5
15 1%
20 0%
27 3%
81 1%
77 1%
67 3%
1
1
3
4
7
0%
0%
19 2%
24 1%
26 1%
76 5%
67 3%
60 0%
HClO4 (1)
H2SO4 (0.5)
0%
3%
0%
0%
12 0%
(CH3COO)3BH− + CH3COOH → (CH3COO)4B− + H2
demethylation effect of HCl is the lowest from these acids (at
the same acidity), the hydridoboron species with chloride
ligand probably suppresses the demethylation.
(11)
Our results indicate that none of these intermediates can
cleave the As−C bond. To test whether demethylation caused
by HCl could be prevented in the presence of CH3COOH its
addition to 1 mol L−1 HCl was investigated. A significant but
not complete suppression of demethylation was found. Both
0.25 and 1 mol L−1 of CH3COOH addition to HCl suppressed
demethylation of MAsV by 50%. As for demethylation of
DMAsV the addition of 0.25 mol L−1 and 1 mol L−1 of
CH3COOH suppressed demethylation by 60% and 80%,
respectively. This can be explained by the competing reactions
during THB hydrolysis giving rise to formation of (CH3COO)-
Acetic Acid. First, sensitivity of iAsIII in 1 mol L−1 of acetic
acid did not significantly differ from that in 1 mol L−1 HCl.
Figure 7 shows dependence of the normalized peak area
−
n
BH3 and LBH3 (from HCl).
To conclude, acetic acid from the demethylation point of
view could be more useful for speciation analysis of arsenic;
however, an incomplete generation efficiency from MAsV and
mainly from iAsV should be expected. The addition of acetic
acid to HCl solves the problem partially.
Nitric Acid. Nitric acid appears to be convenient for HG
since all the investigated arsenic species yielded a uniform
normalized peak area at the HNO3 concentration of 1 mol L−1
(Figure 8). Moreover, the corresponding sensitivity did not
Figure 7. Dependence of normalized peak area of As species on
− iAsIII, (red)
− iAsV,
■
●
concentration of CH3COOH; (black)
(blue)
− MAsV, (green)
− DMAsV, (brown) − TMAsVO;
⧫
▲
▼
concentration of As species was 2 ng mL−1, 2 mL of 1% THB addition.
obtained in the direct transfer mode of iAsIII, iAsV, MAsV,
DMAsV, and TMAsVO on the concentration of acetic acid.
iAsIII, DMAsV, and TMAsVO are easily converted to the arsanes
in the whole range. MAsV is completely converted to arsane at
concentrations 2 mol L−1 and higher. The most important is
that even 8 mol L−1 is not enough to generate iAsV
quantitatively. This could be explained by the fact that acetic
acid is a weak acid (pKa = 4.8) in comparison to iAsV (pKa =
2.3). This correlates well with the assumption that As species
must be present in an undissociated form to be efficiently
converted to hydrides by the reaction with THB and/or its
hydrolyzed intermediates.29−31
In the CT collection mode experiments, no demethylation of
MAsV, DMAsV, and TMAsVO using acetic acid was observed in
the whole concentration range from 0.5 to 6.0 mol L−1.
When acetic acid reacts with THB, hydrolytic products are
different compared to HCl2
Figure 8. Dependence of normalized peak area of As species on
III
V
■
●
▲
concentration of HNO3; (black) − iAs , (red) − iAs , (blue)
V
V
V
⧫
▼
− MAs , (green) − DMAs , (brown) − TMAs O; concentration
of As species 2 ng mL−1, 2 mL of 1% THB addition.
significantly differ from that in 1 mol L−1 HCl. This
dependence is quite similar to that for HCl, but in contrast
to HCl there is no marked demethylation for MAsV and
DMAsV: the fraction of noncorresponding arsanes formed from
DMAsV in the whole studied concentration range from 0.1 to
6.0 mol L−1 was lower than 1%. In the case of MAsV it was also
under 1% with the only exception of concentration of 0.1 mol
−
−
CH3COOH + BH4 → (CH3COO)BH3 + H2
(8)
−
−
(CH3COO)BH3 + CH3COOH → (CH3COO)2BH2 + H2
(9)
−
(CH3COO)2BH2 + CH3COOH → (CH3COO)3BH− + H2
(10)
L−1 when it was 3
1%. Typical chromatogram of volatile
F
DOI: 10.1021/acs.analchem.6b00735
Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry
Article
products generated from DMAsV standard using a 1.0 mol L−1
of HNO3 is shown in Figure 4c.
HNO3 appears to be a suitable acid for HG of arsenic (when
TMAsVO is not present in the samples) because (i)
methylation of As species is preserved and (ii) equal generation
efficiencies are obtained for all species. The HG system with L-
cysteine prereduction and TRIS buffer medium leads only to
formation of corresponding hydrides for all As species.
Nevertheless, the situation for TMAsVO was completely
different. At the concentration 0.1 mol L−1 the fraction of
(CH3)2AsH was 1.4
concentration caused a gradual increase of (CH3)2AsH fraction.
At the concentration of 6 mol L−1 it was 37
5% of
0.4%. A further increase of HNO3
ASSOCIATED CONTENT
* Supporting Information
(CH3)2AsH (the dependence of demethylation on the acid
concentration was linear in the range 0.1−6 mol L−1 − see
Supporting Information Figure S-4). The formation of other
noncorresponding hydrides (AsH3, CH3AsH2) was not
observed, which is in agreement with the measurements for
DMAsV where only the corresponding arsane was formed. The
demethylation of TMAsVO probably proceeds via a different
mechanism and may be associated with the reaction rate of
(CH3)3As formation. As indicated by the low yield at a higher
HNO3 concentration the reaction is slower. Therefore,
TMAsVO remains for a longer time in the liquid phase, and
the demethylation becomes more efficient. The difference
between HNO3 and other acids is not easy to explain, because
nitrate ions are weakly coordinating anions compared to
chloride and in the postulated reaction scheme of the analyte-
borane complex23 there is no influence of acid (or anions)
considered.
■
S
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.anal-
chem.6b00735.
Details on demethylation of DMAsV and TMAsVO
(PDF)
AUTHOR INFORMATION
Corresponding Author
*E-mail: karel.marschner@biomed.cas.cz.
■
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This work was supported by GA CR (grant no. P206/14-
23532S), by the Institute of Analytical Chemistry of the CAS, v.
v. i. (project no. RVO: 68081715) and by Charles University in
Prague (project SVV260317).
L-Cysteine Prereduction. Analogously, no demethylation was
also observed when As species (with the exception of
TMAsVO) were prereduced with L-cysteine to generate
corresponding hydrides from 0.75 mol L−1 TRIS buffer
medium (pH adjusted to 6.0) which is in agreement with our
previous work using these generation conditions for speciation
analysis by HG-CT coupled to AAS, AFS, or ICPMS
detectors.15,24,32 The L-cysteine changes the nature of the
substrate (reduces to trivalent arsenic species and forms thiolate
complexes) as well as changes the nature of hydridoboron
REFERENCES
■
(1) Ded
̌
ina, J.; Tsalev, D. L. Hydride Generation Atomic Absorption
Spectrometry; John Wiley & Sons: Chichester, UK, 1995.
(2) D’Ulivo, A. Spectrochim. Acta, Part B 2004, 59, 793−825.
(3) D’Ulivo, A. Spectrochim. Acta, Part B 2010, 65, 360−375.
(4) Wang, F. T.; Jolly, W. L. Inorg. Chem. 1972, 11, 1933−1941.
(5) D’Ulivo, A.; Onor, M.; Pitzalis, E. Anal. Chem. 2004, 76, 6342−
6352.
33,34
−
species (e.g., RS−BH3 ).
In the case of TMAsVO, HG is
performed only from TRIS buffer medium, and no L-cysteine is
used for prereduction because TMAsVO is thus gradually
reduced to (CH3)3As and consequently lost from the
solution.35 No demethylation was also observed at these
conditions as well.
̌
(6) Dedina, J. In Encyclopedia of Analytical Chemistry, Supplementary
Vols. S1-S3; Meyers, R. A., Ed.; John Wiley & Sons: Chichester, UK,
2011; pp 897−936.
̌
(7) D’Ulivo, A.; Dedina, J.; Mester, Z.; Sturgeon, R. E.; Wang, Q. Q.;
Welz, B. Pure Appl. Chem. 2011, 83, 1283−1340.
(8) Braman, R. S.; Foreback, C. C. Science 1973, 182, 1247−1249.
(9) Howard, A. G.; Salou, C. Anal. Chim. Acta 1996, 333, 89−96.
(10) Cabon, J. Y.; Cabon, N. Fresenius' J. Anal. Chem. 2000, 368,
484−489.
CONCLUSION
■
Demethylation of MAsV, DMAsV, and TMAsVO during the
reaction with THB under analytical conditions in the presence
of HCl, HClO4, and H2SO4 was characterized. It should be
highlighted that the demethylation does not bring any risk of
incorrect results of the postcolumn HG since the resulting
sensitivity does not depend on the degree of methylation of the
resulting hydride.36 However, in the case of As speciation
analysis based on generation of substituted hydrides, the
demethylation can endanger accuracy of the resulting
speciation namely in the case when HCl medium is employed
for HG.37−47
(11) Ellwood, M. J.; Maher, W. A. J. Anal. At. Spectrom. 2002, 17,
197−203.
(12) Hsiung, T. M.; Wang, J. M. J. Anal. At. Spectrom. 2004, 19, 923−
928.
(13) Devesa, V.; Del Razo, L. M.; Adair, B.; Drobna,
́
Z.; Waters, S. B.;
Hughes, M. F.; Styblo, M.; Thomas, D. J. J. Anal. At. Spectrom. 2004,
́
19, 1460−1467.
́
(14) Narvaez, J.; Richter, P.; Toral, M. I. Anal. Bioanal. Chem. 2005,
381, 1483−1487.
(15) Matousek, T.; Currier, J. M.; Trojan
Ishida, M. C.; Gonzalez-Horta, C.; Musil, S.; Mester, Z.; Styblo, M.;
Ded
̌
́ ́
kova, N.; Saunders, R. J.;
́
́
In principle, even in the case of HCl medium, the
demethylation can be almost completely eliminated when
using the mixing coil to partially hydrolyze THB (to minimize
̌
ina, J. J. Anal. At. Spectrom. 2013, 28, 1456−1465.
(16) Tseng, C. M.; Amouroux, D.; Brindle, I. D.; Donard, O. F. X. J.
Environ. Monit. 2000, 2, 603−612.
n
concentrations of LBH3n and L2BH2 ). However, this requires a
(17) Prohaska, T.; Pfeffer, M.; Tulipan, M.; Stingeder, G.; Mentler,
A.; Wenzel, W. W. Fresenius' J. Anal. Chem. 1999, 364, 467−470.
more complicated setup of the hydride generator. In the
CH3COOH medium the degree of methylation is preserved,
but generation efficiency from iAsV and MAsV is unsatisfactory.
In the case of HNO3 the methylation (i.e., speciation
information) is preserved for MAsV and DMAsV, but TMAsVO
is partially demethylated to form (CH3)2AsH. Therefore,
(18) Feldmann, J.; Grumping, R.; Hirner, A. V. Fresenius' J. Anal.
̈
Chem. 1994, 350, 228−234.
(19) D’Ulivo, A.; Meija, J.; Mester, Z.; Pagliano, E.; Sturgeon, R. E.
Anal. Bioanal. Chem. 2012, 402, 921−933.
(20) Talmi, Y.; Bostick, D. T. Anal. Chem. 1975, 47, 2145−2150.
G
DOI: 10.1021/acs.analchem.6b00735
Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry
Article
(21) Van Elteren, J. T.; Das, H. A.; De Ligny, C. L.; Agterdenbos, J.;
Bax, D. J. Radioanal. Nucl. Chem. 1994, 179, 211−219.
(22) Schmeisser, E.; Goessler, W.; Kienzl, N.; Francesconi, K. A.
Anal. Chem. 2004, 76, 418−423.
(23) Regmi, R.; Milne, B. F.; Feldmann, J. Anal. Bioanal. Chem. 2007,
388, 775−782.
(24) Musil, S.; Matouse
Anal. Chem. 2014, 86, 10422−10428.
(25) Taurkova, P.; Svoboda, M.; Musil, S.; Matouse
̌
k, T.; Currier, J. M.; Styblo, M.; Ded
̌
ina, J.
́
́
̌
k, T. J. Anal. At.
Spectrom. 2011, 26, 220−223.
̌
(26) Marschner, K.; Musil, S.; Dedina, J. Spectrochim. Acta, Part B
2015, 109, 16−23.
(27) Van Elteren, J. T.; Das, H. A.; Bax, D. J. Radioanal. Nucl. Chem.
1993, 174, 133−144.
(28) D’Ulivo, A.; Baiocchi, C.; Pitzalis, E.; Onor, M.; Zamboni, R.
Spectrochim. Acta, Part B 2004, 59, 471−786.
(29) Braman, R. S.; Johnson, D. L.; Foreback, C. C.; Ammons, J. M.;
Bricker, J. L. Anal. Chem. 1977, 49, 621−625.
(30) Anderson, R. K.; Thompson, M.; Culbard, E. Analyst 1986, 111,
1143−1152.
(31) Howard, A. G. J. Anal. At. Spectrom. 1997, 12, 267−272.
(32) Matouse
Adair, B. M.; Drobna,
Spectrochim. Acta, Part B 2008, 63, 396−406.
̌
k, T.; Hernan
́
dez-Zavala, A.; Svoboda, M.; Langrova,
́
L.;
́
Z.; Thomas, D. J.; Styblo, M.; Dedina, J.
́
̌
(33) Brindle, I. D.; Le, X. C. Anal. Chim. Acta 1990, 229, 239−247.
(34) Pitzalis, E.; Ajala, D.; Onor, M.; Zamboni, R.; D’Ulivo, A. Anal.
Chem. 2007, 79, 6324−6333.
(35) Musil, S.; Matouse
691.
(36) Marschner, K.; Musil, S.; Ded
̌
k, T. Spectrochim. Acta, Part B 2008, 63, 685−
̌
ina, J. Anal. Chem. 2016, 88,
4041−4047.
(37) Drobna,
́
Z.; Waters, S. B.; Devesa, V.; Harmon, A. W.; Thomas,
D. J.; Styblo, M. Toxicol. Appl. Pharmacol. 2005, 207, 147−159.
́
(38) Hughes, M. F.; Devesa, V.; Adair, B. M.; Styblo, M.; Kenyon, E.
M.; Thomas, D. J. Toxicol. Appl. Pharmacol. 2005, 208, 186−197.
(39) Wen, W. H.; Wen, J. H.; Lu, L.; Liu, H.; Yang, J.; Cheng, H. R.;
Che, W. J.; Li, L.; Zhang, G. Toxicol. Appl. Pharmacol. 2011, 254, 41−
47.
(40) Devesa, V.; Adair, B. M.; Liu, J.; Waalkes, M. P.; Diwan, B. A.;
Styblo, M.; Thomas, D. J. Toxicology 2006, 224, 147−155.
(41) Valenzuela, O. L.; Borja-Aburto, V. H.; Garcia-Vargas, G. G.;
Cruz-Gonzalez, M. B.; Garcia-Montalvo, E. A.; Calderon-Aranda, E. S.;
Del Razo, L. M. Environ. Health Perspect. 2005, 113, 250−254.
(42) Hernandez
́ ̌ ́
-Zavala, A.; Valenzuela, O. L.; Matousek, T.; Drobna,
Z.; Dedina, J.; Garcia-Vargas, G. G.; Thomas, D. J.; Del Razo, L. M.;
̌
Styblo, M. Environ. Health Perspect. 2008, 116, 1656−1660.
́
(43) Sun, G. F.; Xu, Y. Y.; Li, X.; Jin, Y. P.; Li, B.; Sun, X. C. Environ.
Health Perspect. 2007, 115, 648−652.
(44) Calatayud, M.; Bralatei, E.; Feldmann, J.; Devesa, V.; Vel
́
ez, D. J.
Agric. Food Chem. 2013, 61, 12164−12170.
(45) Calatayud, M.; Vazquez, M.; Devesa, V.; Vel
Toxicol. 2012, 25, 2654−2662.
(46) Calatayud, M.; Velez, D.; Devesa, V. Chem. Res. Toxicol. 2012,
25, 2402−2411.
́
́
ez, D. Chem. Res.
́
(47) Zhao, R.; Zhao, M. X.; Wang, H.; Taneike, Y.; Zhang, X. R. Sci.
Total Environ. 2006, 371, 293−303.
H
DOI: 10.1021/acs.analchem.6b00735
Anal. Chem. XXXX, XXX, XXX−XXX