trends in analytical chemistry, vol. 76, no.

2, 1997 73

any number of samples. In such a case, the experi- [21 R.W. Kennard and L.A. Stone, Technometrics, 11
menter can reach a compromise between the infor- (1969) 137.
mation content that is required of the calibration set [31 I. Ruisanchez, J. Lozano, M. S. Larrechi, F.X.
and the cost he/she is willing to pay for it. Rius and J. Zupan, Anal. Chim. Acta, in press.
[41 P.F. De Aguiar, B. Bourguignon, M.S. Khots,
D.L. Massart and R. Phan-Than-Luu, Chemom.
Intell. Lab. Syst., 30 ( 1995) 199.
Acknowledgements D. Mathieu, PhD Thesis, Marseille, 198 1.
[51
[61 V.V. Fedorov, Theory of Optimal Experiments,
J. Ferre thanks the Comissionat per a Universi- translated and edited by W.J. Studden and E.M.
tats i Recerca of Generalitat de Catalunya, for pro- Klimko, Academic Press, New York, 1972.
viding a doctoral fellowship (FI 194-700 1). Finan- 171 J. Fern? and F.X. Rius, Trends Anal. Chem., sub-
cial support from the Spanish Ministry of mitted for publication.
Education and Science (DGICyT project BP93- [81 Matlab, The Mathworks, South Natick, MA.
0366) is acknowledged.
J. Ferrk and F.X. Rius are respectively PhD student
and Professor of Analytical Chemistry at the
References Department of Chemistry, Universitat Rovira i
Virgili, PI. Imperial Tcirraco, 1, E-43005 Tarragona,
[ 1 ] G.E.P. Box, W.G. Hunter and J.S. Hunter, Statis- Spain. Tel: 34-77-558 155, Fax: 34-77-559563,
tics for Experimenters, Wiley, New York, 1978. e-mail: ferre @@mica. urv. es.

Speciation of methylarsenic, methyl-

and butyltin, and methylmercury
compounds and their inorganic analogues
by hydride derivatization

James H. Weber centrations in environmental samples is often rou-

Durham, NH, USA tine, speciation is much more difficult. Speciation is
the identification and quantitation of forms of
Derivatization by sodium borohydride con- chemicals. This review describes hydride derivati-
verts low environmental concentrations of zation by sodium borohydride (NaBH4) and spe-
non-volatile, ionic analytes into volatile com- ciation of inorganic arsenic, tin, and mercury, their
pounds that can be transported to and sepa- methylmetal derivatives, and butyltin compounds.
rated on chromatographic columns. This Hydride derivatization volatilizes all these com-
review describes hydride derivatization of pounds allowing their trapping on an analytical col-
selected organometallic compounds and umn, elution and separation, atomization, and
their inorganic analogues in complex envi- detection (Fig. 1). Atomization is necessary
ronmental media. for the common detection methods atomic
absorption spectrometry (AAS ) and atomic
fluorescence spectrometry ( AFS ). The water
1. Introduction trap before the packed column protects it from
water and prevents water interference in AFS.
Although the determination of total metal con- Scientists have speciated organometallic com-
0165-9936/97/$17.00
PMSO165-9936(96)00085-4 0 1997 Elsevier Science B.V. All rights reserved.
74 trends in analytical chemistry, vol. 16, no. 2, 1997

Batch This brief review describes hydride derivatiza-

Continous Flow
Flour Injection
tion for environmental determinations of methylar-
senic, methyl- and butyltin, and methylmercury
compounds and their inorganic analogues. It
NaBH4
HYDRIDE DERIU. includes a discussion of interferences with hydride
derivatization in environmental samples and how
to minimize them. The book by Dedina and Tsalev
r&d [ 1 ] is an excellent source of details on hydride
derivatization. Also useful is a review by Hamasaki

A/
CRYOGENIC 1 e.g. MethylsiloHane
et al. [ 3 1, who described environmental processes
involving methyl compounds of arsenic, tin, and
on solid support
mercury.
1 PRCKED CLMN.1

J/
1 ATOMIZATION 1 Wart2 furnace
2. Speciation
compounds
of methylarsenic

NaBH4 volatilizes inorganic As( III) and

As(V) by formation of arsine (AsHJ), and
ARS, AFS, AES, ICPIMS, Me,,AsO(OH)s-. (y1= 1-3 ) by formation of
etc. methylarsenic( III) hydrides and trimethylarsen-
Fig. 1. Schematic representation of hydride derivati- ic( III) (Table 2) [ 11. The general (Eq. 1) and sam-
zation experiment including trapping, atomization ple (Eq. 2) reactions for derivatization of methyl-
and detection. Abbreviations: AAS = atomic absorp- arsenic compounds are not balanced because the
tion spectrometry, AFS = atomic fluorescence spec- fate of NaBH4 is unknown.
trometry, AES = atomic emission spectrometry, and
Me,AsO(OH),_, + NaBH4 + Me,AsHs-,
ICP-MS = inductively coupled plasma mass spec-
trometry. (n = l-3) (1)
arsenic(V), tin(IV), and mercury(I1) as well as
their inorganic analogues in environmental sam- MesAsO + NaBH4 + MesAs (2)
ples by hydride derivatization (Table 1) [ 1 1.
Three methods commonly used to derivatize MeJAs+, arsenosugars, and arsenoamino acids are
not directly derivatized by NaBH4 and must be
organometallic compounds are alkylation by
Grignard reagents and sodium tetraethylborate separated and decomposed into inorganic arsenic
(NaBEtb) [ 21, and hydride formation by NaBHa. before hydride derivatization [ 4 1. Cullen and

Advantages of derivatization by NaBH4 relative to Reimer [ 5 ] have reviewed many aspects of inor-
the two alkylation reagents are that NaBH4 is less ganic arsenic and methylarsenic compounds in the
expensive and its solutions are more stable. environment.
Grignard reagents are particularly difficult to use
because they are unstable to water. The high stabil- Table 1
ity of NaBH4 in water means that hydride deriva- Speciation of inorganic species and organometallic com-
tization is experimentally easy for any method of pounds in the environment by hydride derivatizatior?
introducing samples and reagents (Fig. 1). One
Inorganic Organometallic species
disadvantage of hydride derivatization relative to
species
alkylation methods is that hydride derivatives of
methyl- and ethyllead compounds are difficult to Ge( IV) MeGe3+, Me2Ge2+
form and/or are not very stable. A problem com- As( Ill), As(V) MeAsO(OH)z, MezAsO(OH), MesAsO
mon to all derivatizing reagents is that chemicals Sn( IV) Me,Sn(4-“)+, BunSn(4P7L)+ (n= l-3)
Hg(lf) MeHg +b, Me2HgC
found in environmental samples commonly include
metals, ions, ligands, and hydrophobic compounds “[I].
that sometimes interfere with derivatization reac- bNot in above reference.
tions as discussed below. ‘No derivatization needed.
trends in analytical chemistry, vol. 16, no. 2, 1997 75

Table 2 mercury because stannane (SnHb), methyltin com-

Overview of speciation of arsenic compounds by hydride
pounds (Me71Sn4-71, y1= l-3), and butyltin com-
derivatizatior?
pounds ( BullSnH4-71, y1= 1, 2) occur naturally
Analyte PH Product without need of derivatization by NaBH4 (Table
3 ) [ 7- 10 1. These hydrides form in the environment
AsOH )3 -1 to7 AsH3 by completely unknown processes.
AsO(OH)~ > 0.3 AsH3 Derivatization of inorganic tin and methyl- and
MeAsO(OH)z 0.3-l (HCI) MeAsHz
Me*AsO(OH) 0.3-l (HCI) Me*AsH
butyltin compounds by NaBH4 is commonly done
Me3As0 0.3-l (HCI) Me3As at about pH 3 using HNOs or acetic acid [ 11,121.
The general reaction of R,Sn(4-n)+ (R= Me and
“[ 1I. Bu; y1= O-3) (Eq. 6) and that of the Me2Sn2+ exam-
ple (Eq. 7) illustrate derivatization reactions.
A problem with the methods summarized in
Table 2 is that different arsenic compounds have R nSn(4P”)++NaBH4 -+ R,SnH4_, (y1= l-3) (6)
different optimum pH values for derivatization and
Me2Sn2+ + NaBH 4 + Me2SnH2 (7)
that at constant pH (e.g., pH OS) they have differ-
ent sensitivities. This is a common problem in
simultaneously derivatizing several compounds. The limits of detection (background +30) with an
Le et al. [ 61 solved this problem and obtained AAS detector for methyl- and butyltin compounds
excellent reproducibility and sensitivity for arsenic range from 11 to 45 pg [ 13,141.
compounds by adding 2% aqueous L-cysteine Methyltin compounds in the environment form
(HSCH&HNH$02H) along with NaBH+ The by natural methylation of inorganic tin while butyl-
reasons for the improvements are that L-cysteine tin compounds result from anti-fouling marine
(RSH) reduces As(V) compounds to As(II1) paints that release tributyltin. An interesting study
analogues and that its anion (RF ) binds the result- [ 15 ] compared methyl- and butyltin concentra-
ing arsenic( III) compounds before hydride deriva- tions in oysters from the unpolluted Great Bay
tization (Eqs. 3-5). However, this procedure does Estuary (New Hampshire, USA) and polluted
not distinguish inorganic As(II1) and As(V) Loire Estuary (France). Oysters from the Great
(Eq. 3). Bay Estuary contained (dry weight) inorganic tin
(65OrtlOO “g/g), MeSn3+ (ca. 100 “g/g), and
AsO(O As(OH), + As(SR), + AsI& (3) less than the limit of detection (6 rig/g)) of any
butyltin compound. Loire Estuary oysters, in con-
MeAsO(OH),! + MeAs( -+ MeAsH (4) trast, contained only a low 40 rig/g of inorganic tin
and no detectable methyltin compound. However,
they contained considerable BusSn+ (2200 ng /g)
MezAsO(OH) --) MezAs(SR) + MezAsH (5) and its degradation products Bu2Sn2+ (840 rig/g))
and BuSn3+ (580 ng/g). This example clearly
shows the effects of anti-fouling paints containing
3. Speciation of methyl- and butyltin tributyltin in the marine environment. An excellent
compounds and detailed review by Fent [ 161 describes many
aspects of organotin compounds in the environ-
ment with an emphasis on toxicological implica-
Tin compounds differ from those of arsenic and tions.

Table 3
Natural formation of stannane and methyl- and butyltin hydrides

Compound observed Sample Reference

Me2SnH2, Me3SnH Chesapeake Bay water 181

SnH4 Decaying macroalgae [71
SnH4, MeSnH3 Wastewater 191
MenSnH~4_,I (n=O-3) BunSnH4-71 (n= 1,2) Domestic waste dump 1101
76 trends in analytical chemistry, vol. 76, no. 2, 7997

4. Speciation of mercury compounds hydride that is trapped on the cryogenic column.

The second type of interference is any chemical
NaBH4 volatilizes inorganic mercury( II) by such as a metal-binding ligand or hydrophobic sub-
reducing it to Hg” (Eq. 8) and monomethylmercury stance that decreases the ability of NaBH4 to react
cation (MeHg+ ) by forming monomethylmercury with the analyte.
hydride (MeHgH) (Eq. 9) [ 17-191.
5.2. interferences from metal ions
Hg(I1) + NaBH4 + Hg” (8)
An excellent method of analyzing metal ion
MeHg+ + NaBH4 4 MeHgH (9) interferences is to distinguish solution phase and
gas phase contributions. Martin and Donard [ 281
Dimethylmercury (MezHg), a volatile liquid, is in their study of interferences by various metal ions
unchanged during hydride derivatization. Limits on hydride derivatization of MeSn3+ and
of detection (30) in our unpublished work using Bu,Sn(4-“)+ ( yt = l-3) used the following method.
hydride derivatization in 0.3 M NaCl at pH 7 are 0.3 First they obtained the total loss in sensitivity
pg for MeHg+ and MezHg. Two reviews [ 20,211 (area/ng ) by comparing peak areas in the absence
describe methods for speciating mercury com- and presence of metal ions. Then they divided the
pounds in environmental samples. A review by loss due to interference into solution and gas phase
Stein et al. [ 221 described the environmental dis- effects in the following manner. They reacted
tribution and fate of mercury compounds. NaBH4 with metal ions in the absence of analytes
Most researchers derivatize Hg( II) to diethyl- and cryogenically trapped potential volatile metal-
mercury (Et2Hg) and MeHg+ to methylethylmer- containing compounds (probably metal hydrides)
cury (MeEtHg) using NaBEt4 [ 231. One reason on the column. Then without warming the column
that derivatization by ethylation is more commonly they derivatized the organotin analytes in the
used than hydride derivatization is that MeEtHg absence of metal ions and trapped the organotin
has been isolated as a pure compound, but hydrides. Next they heated the column to co-elute
MeHgH has not. However, there is much evidence organotin hydrides and volatile interfering com-
that MeHgH, which has a 1.5 to 2 h half-life in pounds, and determined the peak areas of the or-
the gas phase [ 17,18 1, is sufficiently stable for its ganotin analytes. This two-step experiment gives
determination according to the methods of Fig. 1. In the loss in sensitivity due to gas phase interference.
addition scientists have characterized it by several Subtraction of this sensitivity loss from the total
methods including Fourier transform infrared spec- sensitivity loss gives the solution phase contribu-
trometry (FT-IR) [ 171, mass spectrometry (MS) tion. Different organotin compounds show differ-
[ 17,18, 241, and ‘H and 199Hg nuclear magnetic ent effects from metal ion interferences. For exam-
resonance (NMR) [ 25-271. FT-IR and MS studies ple, sensitivity loss for Bu2Sn2+ is primarily due
confirmed the identity of MeHgH by also studying to gas phase interferences while that for Bus&r+
MeHgD. is nearly equally due to solution and gas phase
effects.
Martin and Donard [ 28 ] then tested the ligands
5. Detection and elimination of EDTA and L-cysteine to see how effectively they
interferences could bind metal ions including Al(III), Fe( II),
Ni( II), Cd( II), and Pb( II) and prevent their reac-
5.1. introduction tions with NaBHA. EDTA is undesirable because it
complexes MeSn3+ and BuSn3+ decreasing their
Hydride derivatization, like other derivatization sensitivity. In contrast L-cysteine is effective
methods, is plagued by interferences that decrease because it strongly binds metal ions, but not organ-
sensitivity and reproducibility [ 11. These fall into otin compounds. The result is improvement in both
two classes. The first type of interference is due to the sensitivity and reproducibility of the determina-
metal ions that are reduced by NaBH4 to active tions.
metals. The metals catalytically decompose There has been much speculation on why metal
NaBH4 or, more likely, decompose organometallic ions interfere with hydride derivatization in solu-
hydrides. Alternatively, the reaction of the metal tion and gas phases [ 11, but there is no consensus.
ion and NaBHa may result in a volatile metal Some researchers have proposed solution interfer-
trends in analytical chemistry, vol. 16, no. 2, 1997 77

ence from metal borides formed by reactions Acknowledgements

between metal ions and BHd-, but this hypothesis
is unlikely due to their required high-temperature This research was partially supported by NSF
synthesis [ 29 1. It is more likely that NaBH4 grant BCS 9224717. It was presented in a different
reduces metal ions to finely divided, very reactive form at the 4th Symposium on Analytical Sciences
metals that rapidly decompose BHb- and/or on June 4-6, 1996 in Brussels, Belgium.
organometal hydride derivatives in the solu-
tion phase. BHd- may also form volatile
metal hydrides that cause gas phase interfer-
References
ences.
[ l] J. DZdina and D.L. Tsalev, Hydride Generation
5.3. interferences from hydrophobic substances
Atomic Absorption Spectrometry, Wiley, New
York, 1995.
Francois and Weber [ 111 found that hydropho- [ 21 S. Rapsomanikis, Analyst, 119 ( 1994) 1429-
bic Bus%+ extracted from eelgrass by a dichloro- 1439.
methane-methanol mixture gives poor sensitivity [ 3 ] T. Hamasaki, H. Nagase, Y. Yoshioka and T. Sato,
and reproducibility during derivatization with CRC Crit. Rev. Environ. Sci. Technol., 25 ( 1995)
NaBH4. In contrast, when BusSnH is added to the 45-91.
eelgrass extract, excellent sensitivity and reprodu- [ 4 ] R. Rubio, A. Padr6, J. Alberti and G. Rouret, Anal.
cibility result. This experiment demonstrates that Chim. Acta, 283 (1993) 160-166.
the problem is not mass transport of BusSnH, but [51 W.R. Cullen and K.J. Reimer, Chem. Rev., 89
lack of its formation in the extract containing (1989) 713-764.
hydrophobic compounds. Our solution was to add 161 X.-C. Le, W.R. Cullen and K.J. Reimer, Anal.
Chim. Acta, 285 (1994) 277-285.
tetrabutylammonium borohydride (Bu4NBH4 ) to
[71 O.F.X. Donard and J.H. Weber, Nature, 332
the reaction because it solubilizes BHd- in hydro- (1988) 339-341.
phobic phases. An easier method, in retrospect, 181 J.-A. Jackson, W.R. Blair, F.E. Brinckman and
would be to add BubNCl along with NaBH4 in the W.P. Iverson, Environ. Sci. Technol., 16 (1982)
derivatization reactions. By a process called phase 110-119.
transfer catalysis Bu~NBH~ would go into the [91 O.F.X. Donard, P. Quevauviller and A. Bruchet,
hydrophobic phase, the BHd- would react, and Water Res., 27 (1993) 1085-1089.
the Bu4Nf would return to the aqueous phase. 1101 J. Feldman, R. Grtimping and A.V. Himer, Fres. J.
This method might also solve problems when Anal. Chem., 350 (1994) 228-234.
doing derivatization with NaBEt4 in the presence [Ill R. Francois and J.H. Weber, Mar. Chem., 25
of hydrophobic substances. (1988) 279-289.
[121 R. Francois, F.T. Short and J.H. Weber, Environ.
Sci. Technol., 23 (1989) 191-196.
1131 O.F.X. Donard, S. Rapsomanikis and J.H. Weber,
6. Summary Anal. Chem., 58 (1986) 772-777.
[I41 L. Randall, O.F.X. Donard and J.H. Weber, Anal.
This review describes derivatization of methyl- Chim. Acta, 184 ( 1986) 197-203.
arsenic, methyl- and butyltin, and methylmer- [I51 J.S. Han and J.H. Weber, Anal. Chem., 60 ( 1988)
cury compounds and their inorganic analogues by 316-319.
reactions with sodium borohydride. Derivatization iI61 K. Fent, CRC Crit. Rev. Toxicol., 26 (1996) l-
of nanogram and sub-nanogram amounts of non- 117.
volatile, ionic analytes in complex environmental [ 171 M. Filippelli, F. Baldi, F.E. Brinckman and G.J.
media into volatile hydrides allows their transport Olson, Environ. Sci. Technol., 26 (1992) 1457-
1460.
to and separation on chromatographic columns.
The separation step allows speciation, i.e. determi- iI81 R. Puk and J.H. Weber, Anal. Chim. Acta, 292
(1994) 175-183.
nation of specific forms of the elements studied.
1191 R. Ritsema and O.F.X. Donard, Appl. Organomet.
Hydride derivatization, like other derivatization Chem., 8 (1994) 571-575.
methods, has interferences due to metal ions, com- 1201 R. Puk and J.H. Weber, Appl. Organomet. Chem.,
plexing agents, and hydrophobic compounds. The 8 (1994) 293-302.
final topic discusses how to overcome these inter- [211 J.H. Weber, in H. Sigel and A. Sigel (Editors),
ferences. Mercury and Its Effects on Environment and Biol-
78 trends in analytical chemistry, vol. 16, no. 2, 1997

ogy. Vol. 34. Metal Ions in Biological Systems, University (7963). In 7963 he became an assistant
Marcel Dekker, New York, 1997, in press. professor of chemistry at the University of New
[ 221 E.D. Stein, Y. Cohen and A.M. Winer, CRC Crit. Hampshire and became a full professor in 1977.
Rev. Environ. Sci. Technol., 26 (1996) l-43. During his most recent sabbatical leave in 1991-
[ 231 L. Liang, M. Horvat and N.S. Bloom, Talanta, 41 92 at ENEA (La Spezia, Italy) he studied abiotic
(1994) 371-379. methylation of mercury(ll). He was an invited
[ 241 P.J. Craig, D. Mennie, N. Ostah, O.F.X. Donard participant in the European Community workshop
and F. Martin, Analyst, 117 ( 1992) 823-824. “Trends in Speciation Analysis in Environmental
[ 25 J P.J. Craig, D. Mennie, M. Needham, N. Oshah, and Food Matrices” (Rome, 1994). Recent invited
O.F.X. Donard and F. Martin, J. Organomet. lectures include ones at the international
Chem., 447 (1993) 5-8. Conference on Environmental and Biological
[ 261 K. Kwetkat and W. Kitching, J. Chem. Sot. Chem. Aspects of Main-Group Organometals (Bordeaux,
Commun., (1994) 345-347. 1994) and the 4th Symposium on Analytical
[ 27 ] P.J. Craig, M.L. Needham, N. Ostah, G.H. Stojak, Sciences (Brussels, 1996). He is on the editorial
M. Symons and P. Teesdale-Spittle, J. Chem. Sot. board of Applied Organometallic Chemistry.
Dalton Trans., (1996) 153-156. His research has been supported by the
[ 28 ] F.M. Martin and O.F.X. Donard, J. Anal. Atom. USA agencies NSF, EPA, and NOAA. He is
Spectrom., 9 (1994) 1143-1151. now at the University of New Hampshire,
[ 291 F.A. Cotton and G. Wilkinson, Advanced Inor- Chemistry Department, Parsons Hall, Durham,
ganic Chemistry, Wiley, New York, 5th edn., NH 03824, where he directs research of
1988, p. 165ff. graduate students on the biogeochemistry of
methylmercury compounds in the estuarine
James Weber’s education includes a PhD degree environment. He published more than 90 papers
in inorganic chemistry from the Ohio State in refereedjournals.

Molecular emission cavity analysis:

principles and applications
Antony C. Calokerinos ered in the early 1970s by Bogdanski, Townshend
Athens, Greece and (the late) Belcher at the Department of Chem-
istry of the University of Birmingham, UK. At that
Molecular emission cavity analysis (MECA) time, Belcher was interested in unusual flame tests
is a chemiluminescence technique in which which were mentioned in Feigl’s Spot Tests in Znor-
molecular emissions are generated within a ganic Chemistry. He asked Bogdanski to work for
cavity that is introduced into a hydrogen dif- his PhD on what was later called candolumines-
fusion flame. The hydrogen flame acts as the cence, in which the luminescent emission is gener-
environment of radicals, atoms and mole- ated on the surface of a solid in contact with a
cules which promote the chemical reactions hydrogen flame [ 11. Bogdanski used the cavity in
producing excited species and the cavity the head of an ordinary Allen screw for introducing
acts as the environment for stabilizing the the solid into the flame environment. The screw was
emitting species. The principles, develop- then cleaned with a mixture of phosphoric and sul-
ment and applications of MECA are outlined phuric acids and washed with water. When the
and future possibilities are discussed. screw was brought into contact with a hydrogen
flame, an intense blue emission followed by a
green emission was noticed. It was soon realized
1. Introduction: the genesis of a new that both emissions were arising from the extremely
analytical technique low residual amounts of the acids.
When the experiment was repeated by injecting a
Molecular emission cavity analysis was discov- solution of sulphuric acid into the Allen screw
0165-9936/97/$17.00
P//SO 165.9936 (96 )00096-9 8 1997 Elsevier Science B.V. All rights reserved.