View Article Online / Journal Homepage / Table of Contents for this issue
Communication
Optimization of microwave digestion for determination of
selenium in human urine by flow injection-hydride
generation-atomic absorption spectrometry
Fangshi Li,* Walter Goessler and Kurt J. Irgolic
Institute for Analytical Chemistry, Karl-Franzens-University Graz, A-8010 Graz, Austria.
E-mail: njfangli@jlonline.com
Received 11th August 1998, Accepted 29th September 1998
A microwave digestion program, which completely decom-
poses and oxidizes selenium compounds in urine to selenate,
was developed by monitoring the pressure and the temperat-
ure during microwave digestion. The efficient decomposi-
Published on 01 January 1998. Downloaded on 06/06/2018 20:13:56.
tion and quantitative recovery of trimethylselenonium
iodide spiked into urine was achieved in 18 min using the
optimized microwave program reaching 200 °C and 8 bar.
The selenate in the digest was reduced to selenite by
hydrochloric acid with the aid of microwave energy. Urea
was found useful to eliminate NOx fumes, which might be
absorbed in the digest and interfere in the determination of
selenium by flow injection-hydride generation-atomic ab-
sorption spectrometry (FI-HG-AAS). The recovery of trime-
thylselenonium iodide, selenomethionine, and selenoethio-
nine added to urine was 96.5–105%. The whole procedure,
FI-HG-AAS determination following microwave digestion
of urine sample and microwave reduction of selenate in the
digests into selenite, was checked with two Standard
Reference Materials 2670 (toxic metals in human urine). The
results showed good agreement with the certified values
(normal level 30 ± 8 mg Se l21 and elevated level 460 ± 30 mg
Se l21. The detection limit of the whole procedure was 3 mg
Se l21 urine. Selenomethionine and selenoethionine were
found unstable during the microwave heating used to reduce
selenate to selenite. Such a microwave reduction procedure
should be cautioulsy used to distinguish selenate from
selenite in the matrices which might contain organic
selenium compounds.
1 Introduction
Selenium is both a toxic and an essential trace element for
humans and animals. The selenium concentration in urine is
used as an indicator of the selenium status of the human body.
Urine may contain inorganic selenium compounds, selenite and
selenate, and organic selenium compounds including trime-
thylselenonium ion (TMSe), selenomethionine (SeMet), sele-
noethionine (SeEt), etc.1–3 The presence of TMSe in urine is
believed to indicate that the daily dose of selenium is in excess
of the required intake.4,5 Accurate and precise methods are
needed for the determination of selenium in urine. Among the
analytical procedures for the determination of total selenium in
urine described in the literature,6 the most often used methods
are spectrofluorometry, atomic absorption spectrometry, and
neutron activation analysis.
Hydride generation-atomic absorption spectrometry (HG-
AAS) offers the advantages of low detection limits and
relatively simple instrumentation.7 However, the determination
of total selenium in urine by HG-AAS may produce inaccurate
results, because either organic selenium compounds, notably
TMSe, are not completely decomposed to selenate, or selenium
is lost during digestion.8 Several reports indicate that the
digestion procedures are largely responsible for the discrep-
ancies among concentrations for total selenium in the same
urine samples determined by HG-AAS.9,10
Microwave-assisted digestion in closed Teflon vessels can be
performed in a much shorter time than digestions in open
beakers on a hot plate. Closed vessels prevent losses of volatile
selenium compounds. However, the procedure for the micro-
wave-assisted digestion of organic selenium compounds must
be optimized6 to achieve complete conversion of the organic
selenium compounds present in urine to selenate.
Selenate needs to be reduced to selenite prior to the formation
of the hydride, because selenium hydride can only be formed
initially from selenite. Several reducing agents have been used
to perform this reduction. Boiling HCl solution is the preferred
reducing media and the optimum HCl concentration for a
quantitative reduction of selenate to selenite has been estab-
lished in the 5–6 mol l21 range.6 This reduction step has been
most often performed in open beakers on a hot plate. Recently,
microwave-assisted reduction procedures were used to over-
come some losses of Se by evaporation and to shorten the
reaction time.11–14 The microwave reduction was also used in
on-line inorganic selenium speciation.13–16 To our knowledge,
however, so far the stability of organic selenium compounds
during microwave reduction has not been studied.
In this study, urine samples were spiked with TMSe, which is
claimed to be rather resistant to mineralization under oxidizing
conditions, and the two selenoamino acids, SeMet and SeEt, as
representative examples of organic selenium compounds that
may be present in urine. The spiked urine samples were digested
with mixtures of nitric acid and hydrogen peroxide in a closed,
pressurized microwave system with the goal to find the
conditions for their complete, reliable, and quick conversion to
selenate. Selenate was microwave reduced to selenite prior to
the determination of selenium by FI-HG-AAS. The stability of
the organic selenium compounds during the microwave reduc-
tion was studied.
2 Experimental
2.1 Instrumentation
Digestion of urine samples and reduction of selenate to selenite
were carried out in a high-performance microwave digestion
system (MLS 1200 MEGA, Milestonl, Leutkirch, Germany)
with a rotor for ten Teflon digestion vessels designed for
pressures up to 30 bar and for temperatures up to 250 °C. The
vessels are equipped with a pressure release system to prevent
explosions. A specially designed vessel with a temperature and
a pressure sensor allows the progress of the digestion to be
monitored. The temperature and pressure data are displayed on
a screen and can be stored and visualized with a dedicated
computer and software.
The flow injection-hydride generation system was a com-
puter-controlled Perkin-Elmer (Norwalk, CT, USA) Model
FIAS-400 with an AS-90 autosampler. A Hitachi (San Jose, CA,
Anal. Commun., 1998, 35, 361–364 361
 View Article Online

USA) Model Z-6100 flame atomic absorption spectrometer
equipped with an electrically heated quartz tube17 served as a
selenium-specific detector. The absorbance signals from the
spectrometer were transferred via the analog port of the
converter to a PC and processed with software written in-
house.18
2.2 Reagents and chemicals
NANOpure water (18.0 MW cm) was obtained by double
distillation in a quartz still (Destamat, Heraeus) and subsequent
passage through an all-quartz Barnstead NANOpure cartridge
system (Barnstead NANOpure, Boston, USA). All solutions
were prepared with NANOpure water. A solution of sodium
borohydride was prepared by dissolving 3.0 g sodium borohy-
dride (Merck, Rahway, NJ, USA) and 0.5 g sodium hydroxide
(Merck) in 1.0 l NANOpure water and filtered before use. A
solution of 40% urea was obtained by dissolving 40 g urea
(Loba-Chemie, Vienna, Austria) in 100 ml NANOpure water.
Hydrochloric acid (32%) and nitric acid (65%) were purified
Published on 01 January 1998. Downloaded on 06/06/2018 20:13:56.
The flasks were filled to the mark with NANOpure water.
Selenium was determined by FI-HG-AAS under the conditions
specified in Table 2.20
3 Results and discussion
3.1 Optimization of microwave digestion procedure
To elucidate the influence of the composition procedure on the
determination of total selenium in urine with FI-HG-AAS, the
recoveries of TMSe, SeMet, and SeEt from aqueous solutions
and from spiked urine samples were tested.
Without digestion, no signal was obtained from the aqueous
solution of TMSe, SeMet or SeEt (10 mg Se l21 for each
compound) by FI-HG-AAS.
The first digestions were performed for 1.0 ml aqueous
solution of TMSe (160 mg Se l21), SeMet (95 mg Se l21), or
SeEt (85 mg Se l21) with the standard digestion program (Table
1) recommended by the manufacturer of the microwave
digestion system for digestion of urine.21 Then, 2.0 ml of 32%

under subboiling conditions in a quartz distillation unit. HCl were added to the digest. The solution in the closed vessels
Selenium stock solutions were prepared with NANOpure water was heated with 600 W microwave power for 5 min to reduce
from sodium selenite pentahydrate (Merck), anhydrous sodium selenate to selenite. This microwave reduction program was
selenate (Fluka, Buchs, Switzerland), seleno-dl-methionine proved11 to quantitatively reduce selenate to selenite in 5–6
(Sigma, St. Louis, MO, USA), selenoethionine (Sigma), and mol l21 HCl media. The remaining treatment was the same as
trimethylselenonium iodide synthesized according to the lit- that in Procedure (section 2.3).
erature procedure.19 Selenium stock solutions were stored in a Fig. 1 shows the recoveries from aqueous solutions of TMSe,
refrigerator at 220 °C before use. Dilute solutions for analysis SeMet, and SeEt determined by FI-HG-AAS following stan-
were prepared daily with NANOpure water. dard microwave digestion and reduction, in comparison to the
The urine samples were filtered through 0.2 mm cellulose
nitrate filters, stored in pre-cleaned polyethylene containers at Table 2 FI-HG-AAS instrumental parameters for the determination of
220 °C and analyzed within 48 h. No preservative was selenium in urine20
added.
Wavelength 196.0 nm
Current of hollow cathode lamp 12 mA
2.3 Procedure Slit 1.3 nm
FIAS fill time 10 s
Urine (1.0 ml) was pipetted into the Teflon digestion vessels. FIAS inject time 15 s
65% nitric acid (2.0 ml) and 30% hydrogen peroxide (1.0 ml) Reductant 0.3% NaBH4 in 0.05% NaOH
were added to each vessel. Ten loaded vessels were covered HCl 1.0 mol l21
Quartz tube temperature 900 °C
securely and fastened into the rotor. The rotor was placed into Sample loop 500 ml
the microwave oven. The optimized digestion procedure (Table Argon gas flow rate 110 ml min21
1) was performed. After digestion, the rotor with the vessels was Reductant flow rate 5.3 ml min21
transferred into a water bath and cooled down to room HCl flow rate 7.3 ml min21
temperature. To each of the above digestion vessels, 3.0 ml of Waste flow rate 15 ml min21
32% HCl were added. The covered vessels were heated in the
microwave oven for 5 min at 600 W. When the vessels were
cooled down, 100 ml of 40% urea solution was added to each
vessel and the vessels were shaken for 5 min. The solutions
were quantitatively transferred into 10 ml volumetric flasks.

Table 1 Microwave digesion programs (standard and optimized) for the

conversion of organic selenium compounds in 1 ml urine to selenate

(Time/min)/(Power/W)

Step Standarda Optimizedb

1 3/250 2/450
2 1/0 0.5/0
3 3/250 3/600
4 1/0 0.5/0
5 4/250 10/700
6 5/400 2/0 (cool)
7 3/600
8 3/250
9 2/0 (cool)

Total time/
min 25 18
a Recommended by the manufacturer: digestion mixture 1.0 ml 65% Fig. 1 Recoveries of TMSe, SeMet, and SeEt from aqueous solutions
HNO3 and 0.5 ml 30% H2O2. b Optimized digestion program: digestion determined by FI-HG-AAS following standard microwave digestion and
mixture 2.0 ml 65% HNO3 and 1.0 ml 30% H2O2. microwave reduction of selenate to selenite in comparison to selenite
solution.

362 Anal. Commun., 1998, 35, 361–364


standard solutions prepared from selenite. The selenium
recoveries were 22.7% for TMSe and close to 100% for SeMet
and SeEt.
When the standard microwave digestion program was
applied to 1.0 ml of urine, clear and colorless digests were
obtained. However, the recovery of TMSe from the spiked urine
samples was about zero.
The development of a microwave digestion procedure, by
which TMSe in urine would be completely decomposed, and by
which the closed digestion vessels would not open during
digestion, was desired. To optimize the microwave digestion
procedure for TMSe in urine, a specially designed vessel, which
allowed a sensor of temperature and pressure to be connected
and the progress of the digestion to be monitored, was
substituted for one of the ten digestion vessels. Urine (1.0 ml)
and TMSe solution (100 ml of 1.60 mg Se l21) were pipetted into
this vessel. The digestion mixture of 65% nitric acid (1.0, 1.5 or
2.0 ml) and 30% hydrogen peroxide (0.5, 1.0 or 1.5 ml) were
added. To each of the other nine digestion vessels, 3 ml of
NANOpure water was added. The rotor with the ten loaded
vessels was placed into the microwave oven. The pressure limit
Published on 01 January 1998. Downloaded on 06/06/2018 20:13:56.
was set at 20 bar and the temperature limit at 200 °C. The
digestion power and time were gradually increased until a
power level and time combination produced complete diges-
tion. During the digestion procedure, the pressure and the
temperature in the specially designed vessel were monitored.
After microwave digestion, 32% HCl (3.0 ml) was added to the
digest. Selenate was microwave reduced to selenite. The
remaining treatment was the same as in Procedure (section
2.3).
Fig. 2 shows the temperature and pressure curve for
microwave digestion of 1.0 ml spiked urine with the standard
and the optimized microwave heating program (Table 1).
During the standard microwave digestion, the temperature (Ts)
reached up to 130 °C and the pressure (Ps) up to about 0.8 bar.
During the optimized microwave digestion, the temperature
(To) reached up to 200 °C and the pressure (Po) up to about 8
bar. With the optimized program, the recovery of TMSe iodide
spiked in urine was 105% (Table 3), and that of SeMet and SeEt
96.5–101%. These results demonstrate the efficiency and the
validity of the optimized microwave digestion program.
3.2 Elimination of the nitrogen oxide interference
When determining selenium in urine by FI-HG-AAS following
microwave digestion of urine samples and microwave reduction
Fig. 2 Temperature (T) and pressure (P) curves for microwave digestion
of 1.0 ml spiked urine. Ts and Ps: standard digestion program; To and Po:
optimized digestion program.
View Article Online
of selenate to selenite, the blank solution had a high absorbance
signal, because of the NOx fumes absorbed in the digest. Similar
absorbance signals from the blank solution were observed,
when determining arsenic in urine by FI-HG-AAS following
digestion.22 The NOx fumes could lead to erroneously high
results for determinations by HG-AAS. The removal of these
fumes from the digest required open-vessel heating with
concentrated sulfuric acid in a microwave oven, which was both
potentially dangerous and labor intensive. In this study, adding
50–300 ml of 40% urea solution to the digests before
determination of selenium by FI-HG-AAS was found useful in
ridding the sample of interfering NOx fumes that might be
present after digestion.
3.3 Detection limit and accuracy
The detection limit of the determination of selenium in urine by
FI-HG-AAS following the optimized microwave digestion and
reduction was 3 mg Se l21 urine. The precision obtained
expressed as the relative deviation of five replicates was 4.7%.
The method was validated by two reference urine samples
(NIST 2670), the normal level 30 ± 8 and the elevated level 460
± 30 mg Se l21. The determined results (28.2 ± 2.4 and 477 ± 30
mg Se l21) showed good agreement with the certified values.
Therefore the proposed procedure can be used to determine total
selenium in urine.
3.4 Efficiency of the microwave reduction procedure
In this study, selenium in the urine samples could not be
detected directly by FI-HG-AAS without digestion of urine
sample and reduction of selenate to selenite, because most of the
Table 3 Recovery of TMSe from aqueous solution, or spiked to 1.0 ml
urine, after optimized microwave digestion and microwave reduction (n =
3)
TMSe added/ Se found/ Se recovery
Sample ng Se ng Se (%)
NANOpure water 160 161 ± 9 101 ± 6
Urine 0 38.3 ± 3.0 —
Urine 160 206 ± 7 105 ± 7
Fig. 3 Comparative studies of the influence of the sample treatment on the
determination of selenium in urine by FI-HG-AAS. (1) Microwave heating
used to reduce selenate to selenite; (2) optimized microwave digestion and
then microwave reduction.
Anal. Commun., 1998, 35, 361–364 363
 View Article Online

Acknowledgement
F. Li thanks the Austrian Academic Exchange Services for
awarding a scholarship to study at the University of Graz.

References
1 G. Koelbl, PhD Thesis, Institute for Analytical Chemistry, Karl-
Franzens-University, Graz, Austria, 1994, pp. 157–159.
2 J. M. Marchante-Gayon, J. M. Gonzalez, M. L. Ferandez, E. Blanco
and A. Sanz-Medel, Fresenius’ J. Anal. Chem., 1996, 355, 615.
3 K. L. Yang and S. J. Jiang, Anal. Chim. Acta, 1995, 307, 109.
4 X. F. Sun, B. T. G. Ting and M. Janghorbani, Anal. Biochem., 1987,
167, 304.
5 R. Hasunuma, M. Tsuda, T. Ogawa and Y. Kawanishi, Bull. Environ.
Contam. Toxicol., 1993, 51, 756.
6 M. Sanz Alaejos and C. Diaz Romero, Chem. Rev., 1995, 95, 227.
7 J. Dedin and D. L. Tsalev, Selenium, Hydride Generation Atomic
Absorption Spectrometry, John Wiley and Sons Ltd., UK, 1995, ch.
13.
Published on 01 January 1998. Downloaded on 06/06/2018 20:13:56.

Fig. 4 Calibration curves for determination of selenite, SeMet and SeEt by 8 S. J. Hill, J. B. Dawson, W. J. Price, I. Shuttler and J. F. Tyson,
FI-HG-AAS following microwave heating used to reduce selenate to
selenite. 9 J. Neve, M. Hanocq and L. Molle, Microchim. Acta, I, 1980, 259.
10
selenium existing in the urine is selenate, organic selenium 11
compounds, or bound on proteins.
12
Two sample treatment procedures were studied for deter-
13
mination of selenium in urine by FI-HG-AAS. One involved the
microwave heating used to reduce selenate to selenite. The other 14
included the optimized microwave digestion and reduction 15
steps. The two procedures provided different results (Fig. 3).
The difference should have corresponded to trimethylsele- 16
nonium ions and selenoamino acids, which could not be reduced
to H2Se by sodium borohydride so they could not be detected by 17
FI-HG-AAS. But the following experiment contradicted this
18
assumption.
SeMet and SeEt in 6 mol l21 HCl media could be detected by 19
FI-HG-AAS following the microwave reduction step (600 W, 5 20
min). Fig. 4 shows the calibration curves of the determination.
This result means that SeMet and SeEt were unstable in 6 21
mol l21 HCl during microwave heating. According to the
literature,23,24 selenoamino acid is very easily oxidized and thus 22
may decompose during acid hydrolysis.
Reduction of selenate to selenite is often used to distinguish 23
selenate from selenite, especially in the on-line determination of
24
inorganic selenium compounds by FI-HG-AAS with micro-
wave reduction. However, according to this study, such a
microwave reduction procedure should be cautiously used to
distinguish selenate from selenite in the matrices which might
contain organic selenium compounds.
J. Anal. At. Spectrom., 1996, 11, 281R.
V. W. Bunker, M. S. Lawson, M. E. Stansfield and M. F. Clayton, Br.
J. Nutr., 1988, 59, 171.
F. Li, Ph.D Thesis, Institute for Analytical Chemistry, Karl-Franzens-
University, Graz, Austria, 1998, pp. 125–126.
R. Munoz Olivas and O. F. X. Donard, Talanta, 1998, 45, 1023.
M. G. Cobo-Fernandez, M. A. Palacios, D. Chakraborti, P. Quevau-
viller and C. Camara, Fresenius’ J. Anal. Chem., 1995, 351, 438.
L. Pitts, P. Worsfold and S. J. Hill, Analyst, 1994, 119, 2785.
L. Pitts, A. Fisher, P. Worsfold and S. J. Hill, J. Anal. At. Spectrom.,
1995, 10, 519.
J. L. Burguera, P. Carrero, M. Burguera, C. Rondon, M. R. Brunetto
and M. Gallignani, Spectrochim. Acta, 1996, 51B, 1837.
D. Mayer, S. Haubenwallner and W. Kosmus, Anal. Chim. Acta,
1992, 268, 315.
G. Koelbl, K. Kalcher and K. J. Irgolic, J. Autom. Chem., 1993, 15,
37.
J. L. Hoffman, J. Chromatogr., 1991, 588, 211.
F. Li, PhD Thesis, Institute for Analytical Chemistry, Karl-Franzens-
University, Graz, Austria, 1998, p. 86.
MLS Application Report 09.07.93., Mikrowellen-Labor-System
GmbH.
C. P. Hanna, J. F. Tyson and S. McIntosh, Clin. Chem., 1993, 39,
1662.
C. Hammel, A. Kyriakopoulos, U. Roesick and D. Behne, Analyst,
1997, 122, 1359.
G. Zdansky, Seleno amino acids and peptides, in Organic Selenium
Compounds: Their Chemistry and Biology, ed. D. L. Klayman and
W. H. H. Guenther, John Wiley and Sons, New York, 1973, Ch. XII,
p. 596.
Paper 8/06342G

364 Anal. Commun., 1998, 35, 361–364