CLIN.CHEM. 30/2, 243-245
(1984)
Reduction of Organic Mercury in Water, Urine, and Blood by Sodium
Borohydride for Direct Determination of Total Mercury Content
Shiomo Margel and Joseph
Hlrsh
We haveestablished
theoptimal
conditions
for use of sodium
borohydride as the reducing agent before the direct determination of mercury in water, urine, and blood by atomic
absorption
spectroscopy.
We evaluated
the effects
of pH,
temperature, and cupric sulfate concentration on the direct
determinationof both organic and inorganic compounds of
mercury. Accurate and precise quantification of mercury
requiresthatthe pH be between 9.3 and 9.5, the reaction
temperature above 25 #{176}C,
the reactiontime longerthan 1
mm, and, for urine samples only,the cupricsulfate
concentration 10 &mol/L. The detection limit of the method is 1 to 2
ng and the precision
(CV) is3.8% forbloodand 4.0% for
urine.
Additional Keyphrases: atomic absorption spectroscopy
trace
elements
toxicology
conditions for complete reduction
In the efforts to develop a simple,accurate,and rapid
forthe quantification
of mercury, especially
of
organic mercury compounds, in biological liquids such as
blood and urine, the main difficulty
istheextreme volatility
of mercury and many ofits compounds. The current methods used formercury determinationsare based on neutron
activation
(1), colorimetry
(2), and atomic absorptionspectechnique
simple,
rapid, and accurate
methods
currently
method is an improvement
over
used tomeasure mercury in body fluids.
Materials and Methods
Reagents.
All reagents
used were of analytical
grade.
Mercuric
chloride, molecular
sieve 4A, and the anti-foam
agent, tri-n-butyl phosphate, were from British Drug House,
Bucks., U.K.; methylmercury(II)
chloride was from Columbia Organic Chemicals, Columbia, SC 29290; and phenylmercuric acetate was from Aldrich Chemical
Co., Milwaukee, WI 53201. All glassware was soaked in concentrated
nitric acid for several hours, washed exhaustively
with
doubly distilled water, and dried in an oven. Heparinized or
citrated normal human blood samples and 24-h urine samples were collected in polyethylene bottles.
We prepared the followingaqueous solutions:
sodium
borohydride,100 g/L in NaOH (1 molIL);NaOH, 50 gIL;
H2S04, 50 mL/L; NaHCO3, 1 molIL;and cupricsulfate,
0.83
mmoL’L.
Aqueous standard
stock solutionsof mercuric chloride
(Hg,250 mg/L),methylmercury(ll)chloride(Hg, 100 mg/L),
and phenylmercuricacetate(Hg, 250 mgfL) were prepared
by adding 33.82 mg of mercuric chloride,12.52 mg of
methylmercury(H) chloride,and 41.96 mg of phenylmercuricacetate,
respectively,
to 100-mL volumetricflasksand
trophotometry
(3-6). Neutron
activation
is expensive
and
dilutingto
volume.
Aqueous
standards of the mercury
not readilyavailablein most laboratories.
The colorimetric
compounds
(10
pg
of
Hg
per
milliliter)
were prepared by
measurements
are slow, require
several
extraction
steps,
appropriatedilutions
ofthe above standards.The 10 .tg/mL
and are subjectto many interferences.
Atomic absorption
bestmeets the essential
requirementsformercury analysis, standardswere furtherdilutedwith water to obtain the
but the flame atomic absorptionmethod is insufficiently working standards.
Apparatus.
To measure mercury,we used the Model 50A
sensitive for quantification
of mercury in such samples.
mercury
analyzer
(Perkin-Elmer Corp., Norwalk, CT
Several methods have been reported for furnace
atomic
06856), assembled
as illustrated
in Figure 1.The reaction
absorption
spectroscopy,
but theyarecontroversial
and have
tubes
were
2.5
x
20
cm
glass
tubes. The drying
tubes
not proved themselvestobe applicable
toactual samples or
consisted
of a 1-L Erlenmeyer
flask, containing
about 600
certified standards (3, 4). Currently, cold-vapor
atomic abmL of doubly distilled
water, and an 0.8 x 22 cm glass pipe
sorptionspectrophotometry,
based on thework ofHatch and
filled
with
molecular
sieve
4A. Air flow, monitored with a
Ott (6), is the method of choice for determining mercury in
flowmeter, was directed through
the reaction tube and the
biological
fluids.
Most ofthepublishedreportsinwhich this
drying
tubes
into
the
absorption
cell.
techniqueisused requirewet digestionofthe samples as a
Procedure,
Prepare aqueous standards of mercury (10 to
necessarypreliminarystep,with the concomitant major
drawbacks ofprolongedtime (several
hours forcompletion), 500 ng) from a standard 1 pg/mL solution of the appropriate
mercury compound. Add standard
to the reactiontube,
greatermercury loss,
and highreagentblanks.Few publicafollowedby distilled water up to 2 mL and 6 ml..of the
tionsdescribea directmethod for the determinationof
NaHCO3 solution,
asbuffer.
Beforetheanalysis,
run several
mercury in biological
liquids.
Magos and Clarkson (7), the
blanks
through
the
system,
to
remove
tracemercury
confirst to reporta method of analysis
for mercury
in blood,
taminants and to ensure a stableblank signal.
used SnCl2 and the reagentsSnCl2 plusCdC12 fordetermiDetermine mercury in blood samples as follows: Pipet 1
nations of inorganic and organic mercury,
respectively.
A
mL ofwhole bloodintothe reactiontube.Add the mercury
second method for the directdeterminationof mercury in
standard,0.1mL ofthe antifoamagent,distilled
water (up
urine (8) and inblood (9) involveduse ofthe reducingagent
to
2
mL),
and
6
mL
ofthe
bicarbonate
buffer
solution.
When
NaBH4. The proceduresforurine and blood samples were
the temperature
of the solution in the reaction tube reaches
notinterchangeable.
25 #{176}C,
rapidly add 0.35 mL of the sodium
borohydride
We studied the use of sodium borohydride
as a reducing
solution,
cap
the
tube,
and
shake
or
stir
the
solution
for 1
agent for the direct determination
of mercury. The resulting
mm. Turn on the air pump and read the absorbancepeak.
Turn off the air pump only when the absorption reading
returns
to zero or to the background
level. Flushing the
Department of Plastics Research, Weizmann Institute of Science,
Rehovot, Israel.
Received September
1, 1983; accepted October 28, 1983.
system
thoroughly
residual
mercury
between
samples
efficiently
eliminates
contamination.
CLINICAL CHEMISTRY, Vol. 30, No. 2, 1984
243
00
HgCI2
CLAMP
I
60
3HgCl
oniNG
THERMOSTAT
Fig. 1.
Diagram
oftheapparatus
used
Determine mercury in urine and water samples by substituting 0,1 mL of the cupric sulfate solution or 0.1 mL of
distilled water, respectively,
for the 0.1 mL of anti-foam
agent.
To study the effect of pH on the mercury analysis, we
substituted
6 mL of distilled water containing
various
amounts of H2S04 solution for the 6 mL of the NaHCO3
buffer. We also studied the effect of temperature
on the
mercury determination.
The following two factors must be especially considered
during determination
of mercury:
(a) The peak absorbance
depends on the air flow, the
length of the absorption cell, the final volume, and the dead
volume of the apparatus.
All these must therefore be kept
constant.
(b) The mean analytical
recovery of mercury standards
added to water, blood, and urine is not precisely the same (7,
9). Therefore, the calibration standard
must be prepared in
the same matrix as the sample.
Results and Discussion
0
2.0
6.0
0.0
4.0
pH
Fig. 2. Analytical recovery of mercury compounds added to blood as a
function of pH
0.3tg of the mercurycompounds was determinedat25C andreactiontime of 1
mm. Each pointof the graph is based on the mean of triplicatedeterminations
&
I0
20
30
40
Sodium borohydride,
a mild specific reducing agent, reT(#{176}C)
duces groups such as aldehydes
and disulfides but not
Fig. 3. Analytical recovery of mercury compounds
added to blood asa
organic cids or sulfoxides. It reacts violently with water to
function of temperature
liberate hydrogen. Under alkaline aqueous conditions, how0.3 g ofthe mercury compoundswas determined at pH 9.3 and reaction time of
ever-e.g.,
1 molJL NaOH-the
solution is more stable and
1 mm. Each point on the graph isbased on the mean of thplucate
determinations
can be kept for several weeks. Sodium borohydride reacts
water and blood. For urine, however, the recovery of methylwith mercury compounds to form elemental mercury. We
determined the optimal conditions for the measurement
of
mercury chloride is highly dependent on the cupric sulfate
mercury by an intensive study of the effects of temperature
concentration,
being optimum at 10 mol of Cu504 per liter
and pH on the analytical recovery of mercury compounds.
(Figure 4) (8).
Effect of pH and temperature.
Figures
2 and 3 illustrate
Precision.
We prepared blood and urine samples containthe analytical recovery of inorganic and organic mercury
ing 100 ng of mercury per milliliter by adding a standard
compounds from blood samples as a function of pH and
containing equal amounts of mercury as mercuric chloride
temperature.
The determination
of mercuric chloride is
and methylmercury
chloride. Twenty replicate analyses,
independent of both pH and temperature;
e.g., the recovery
under the optimum conditions described above, of 1-mL
of 0.3 pg of mercuric chloride from blood at both pH 13.3 and
aliquots of these samples yielded CVs of 3.8% for blood and
pH 8.8 is about the same, 97%, which is the same as for
4.0% for urine.
mercuric chloride at temperatures
of 32 #{176}C
and 4 #{176}C.
On the
Analytical
recovery. Samples of urine and citrated plasma
other hand, analytical recovery of methylmercury
chloride
from five normal volunteers were collected in polypropylene
depends highly on pH and temperature,
being optimum
containers. To each sample we added either mercuric chlochloride, or phenylmercuric
acetate
between pH 9.3 and 9.5 and at temperatures
above 25 #{176}C.ride, methylmercury
These same comments apply to recovery of mercuric chlo(final concentration
100 ng/mL). Analytical
recoveries of
these compounds from blood and urine consistently exceedride and methylmercury
chloride from water and urine.
Effect of reaction time. Analytical
recovery of mercuric
ed 96% (Table 1).
chloride in water, blood, or urine is independent
of the
Detection limit. The detection limit, defined as the quantireaction time. Therefore, mercuric chloride can either be
ty of the element that gives a reading equal to twice the
standard deviation of a series of at least 10 determinations
determined immediately after the sodium borohydride soluof a near-blank concentration
(10), was 1 to 2 ng of Hg in
tion is added or later. Methylmercury
chloride, however,
should not be analyzed until at least 1 mm after the sodium
blood and urine.
borohydride solution is added.
Notes: In general, inorganic mercury compounds such as
Effect of cupric sulfate. The concentration
of cupric sulfate
mercuric nitrate and organic mercury compounds such as
does not influence the determination
of mercuric chloride in
phenylmercuric
acetate, behave like mercuric chloride and
urine or of mercuric chloride and methylmercury
chloride in
methylmercury
chloride, respectively.
244
CLINICAL CHEMISTRY, Vol. 30, No. 2, 1984
The method described is also applicable to sea water.
Mercury compounds, particularly
organic mercury compounds, should be measured under optimum conditions, pH
9.3-9.5, temperature
25 #{176}C
(or above), reaction time 1 mm
(or more), and cupric sulfate concentration
(for urine samples only) of 10 pmolIL. Under these conditions we observed
a linear relationship
between analytical
recoveries and
concentration
of added mercury in the range 0.05 to 0.7 pg
per sample. If the optimal conditions
are not followed
closely, recovery of organic mercury will be incomplete and
the total mercury concentration
calculated for a sample will
be low.
C.)
1.
0)
This research was supported by the Committee on Prevention
and Research
CuSO4(M)X 106
Fig. 4. Effect of cupric sulfate concentration (in 1zmol/L) on the analytical
recovery of methylmercury chloride
The cupnc sulfate was added to aqueous solutions containing0.25 g of the
mercury compounds. The mercury was determined at 25 “C. reaction time of 1
mm, and pH 9.3
Table 1. Analytical Recovery of Mercury in Three
Compounds Added to Blood and Urine8
Recovery.
Subject
H9CI2
%
CH3HgCI
C0H5H9CO2CM3
Blood
1
97
97
96
2
3
101
99
96
96
97
96
4
98
97
97
98
101
96
97.2 ± 2.17
5
Mean
98.4
±
1.67
96.8
±
2.64
Urine
1
98
96
2
102
95
3
4
5
Mean
97
98
97
98.4 ± 2.07
97
98
96
asamples
containing 0.1 g
96.4
± 1.14
96
97
102
97
97
97.8 ± 2.39
of mercury compounds were analyzed at
25 #{176}C,
pH 9.3. with a 1-mm reaction time; CuSO4, 10 Mmol/L, was added to the
urine samples. Results are the mean of triplicate determinations.
Compounds of alkaline and alkaline earth metals such as
NaCl, KC1, CaCl2, and MgC12 (1 mol/L) do not interfere in
the analysis for mercury.
in Occupational
Health,
The Government
of Israel.
References
1. Sjostrand B. Simultaneous
determination
of mercury and arsenic in biological and organic materials by activation
analysis.
Anal Chem 36, 814-819 (1964).
2. Nobel S. Nobel D. Determination of mercury in urine. Clin Chem
4, 150-158 (1958).
3. Issag HJ, Zielinski WL. Hot atomic absorption spectrometry
method for the determination of mercury at the nanogram and
subnanogram level. Anal Chem 46, 1436-1438 (1974).
4. Owens JW, Gladney ES, Issaq HJ, Zielinski WL. Loss of mercury
during flameless atomic absorption spectrometry. Anal Chem 48,
787-788 (1976).
5. Kuldvere A, Andreassen BT. Determination of mercury in
seaweed by atomic absorption spectrophotometry
using the PerkinElmer MHS-1. At Absorpt Newsi 18, 106-107 (1979).
6. Hatch WR, Ott WL. Determination
of sub-microgram quantities
of mercury by atomic absorption spectrophotometry.
Anal Chem 40,
2085-2087 (1968).
7. Magos L, Clarkson TW. Atomic absorption determination
of
total inorganic and organic mercury in blood. J Assoc Off Agric
Chem 55, 966-971 (1972).
8. Toffaletti T, Savory J. Use of sodium borohydride for determination of total mercury in urine by atomic absorption spectrometry.
Anal Chem 47, 2091-2095 (1975).
9. Sharma DC, Davis PS. Direct determination of mercury in blood
by use of sodium borohydride reduction and atomic absorption
spectrophotometry.
Clin Chem 25, 769-772 (1979).
10. Basic Atomic Absorption Spectroscopy, A Modern Introduction,
Varian-Techtron Pty, Ltd., Springvale, Victoria, 3171 Australia,
1975, pp 79-80.
CLINICAL CHEMISTRY,Vol. 30, No. 2, 1984 245