Environmental Engineering Reference
In-Depth Information
cleaned with strong acid prior to use. These materials have
been found to be free from Hg contamination and therefore
suitable for work at low, ambient levels. However, Tefl on
shows the best performance regarding both contamination
and loss-free storage of aqueous samples. Sampling devices
(such as Go-Flo water-sampling bottles) should be Tefl on-
coated if possible and the surface in contact with the sample
acid should be cleaned with dilute acid (e.g., 1N HCl for
8 hours or more).
Approximately 250 mL of water is typically needed for
a total Hg analysis. Personnel handling sample containers
should do so with gloved hands (clean, nonpowdered latex
or vinyl gloves), and samples should be double bagged in
clean plastic bags for shipping as outlined in EPA Method
1669,
Sampling Ambient Water for Trace Metals at EPA Water
Quality Criteria Levels
(EPA, 1996)
.
Logar et al., 2004). This method has been promulgated as
EPA Method 1631 (EPA, 2002).
Sample decomposition of all organo-, organic-, and par-
ticulate Hg species in a water sample is necessary if total Hg
is to be measured. This is generally achieved using the oxi-
dizing agent BrCl in an HCl solution (Bloom and Crecelius,
1983) or techniques based on UV oxidation in HCl solution
(Ahmed et al., 1987; May et al., 1987). The use of other wet
chemical oxidative mixtures is limited because of relatively
high reagent blanks. In the case of humic-rich water sam-
ples, a combination of BrCl and UV oxidation is very effec-
tive and results in complete recovery.
After the decomposition step, mercury is usually isolated
from the aqueous sample matrix for gas-phase detection by
reducing the Hg in a sparging vessel and purging it from
solution w ith an iner t, Hg-f ree, gas fl ow. The most common
reducing agent used is SnCl
2
, but NaBH
4
may also be used
(Iverfeldt, 1988; Heraldsson et al., 1989; Gill and Bruland,
1990). The evolved Hg
0
is then swept from the purging cell
either onto an amalgamation media or directly into the
detector. Preconcentration of the Hg
0
in the gas stream
onto a gold trap by amalgamation is applied in nearly all
analytical procedures used for the most sensitive measure-
ment of Hg in natural waters (Bloom and Crecelius, 1983;
Gill and Fitzgerald, 1987). Like the CVAAS methods, the
fl uorescence signal is proportional to the mass of Hg col-
lected, which is quantifi ed using a standard curve.
Most of the trace-level analytical methods for water
samples, such as EPA Method 1631, were written for ambi-
ent environmental water samples. The sample-preparation
techniques described in these methods may not be rigor-
ous or adequate for all types of water samples. For waters
with complex matrices, such as highly organic industrial
wastes and sewage infl uent, it is important to modify the
techniques to ensure that the digestion of the sample is
complete. Verifi cation of digestion can be determined by
performing matrix spikes on the most complex matrices
to be analyzed. Interferences in the method can result
with unusual samples. For example, high concentrations
of iron can result in the formation of iron chloride during
the oxidation step that will precipitate and scavenge Hg
out of the sample. Method modifi cations often used to
overcome matrix interferences include sample dilution,
the addition of extra oxidizer, and heating the sample
(e.g., ~50°C for several hours) (Lytle et al., 2007).
For the analysis of elemental and reactive Hg in water,
CVAFS is recommended, as it is the only method with a
suffi ciently low detection limit to accurately quantify the
relatively low concentration of these forms of Hg in water
samples. For the analysis of elemental Hg, no oxidation or
reduction step is performed; the elemental Hg is simply
purged directly from the sample, preconcentrated onto an
amalgamation media (generally gold) and then quantifi ed
using CVAFS. Similarly, for the analysis of reactive Hg, no
oxidation step is performed; the sample is reduced with a
Determination of Total Mercury
in Natural Waters
Analytical techniques suitable for total Hg determination
in natural waters at the picogram level are either based on
cold-vapor atomic absorption spectrometry (CVAAS), cold-
vapor atomic fl uorescence spectrometry (CVAFS), ICP-MS,
inductively coupled plasma atomic emission spectrometry
(ICP-AES), or atomic absorption spectrometry (AAS). Of
these, CVAAS is the most widespread method, although
CVAFS is rapidly replacing CVAAS because of its superior
detection limits.
Mercury detection by CVAAS is based on the gas phase
absorption of 254 nm radiation by elemental Hg atoms in
an inert gas stream. Mercuric ions in the digested sample
are reduced to Hg
0
with SnCl
2
or a similar reductant and
then are purged out of the sample and directly into a gas
cell, where absorption is determined. The absorption sig-
nal (peak height or area) is proportional to the concentra-
tion of Hg in the gas cell. The mass of Hg in a sample is
quantifi ed using a standard curve, which is a function of
the volume of sample purged. This method has been pro-
mulgated as EPA Method 245.1,
Determination of Mercury in
Water by Cold Vapor Atomic Absorption Spectrometry
(in EPA,
1994).
However, the achievable detection limits for CVAAS
methods are not low enough to be useful for most ambient
water samples.
In recent years, CVAFS techniques have become
increasingly important, since the instrumental detection
limit is often less than 1 pg, which is an order of mag-
nitude better than ICP-MS, ICP-AES, or CVAAS. Since
the development of a simple, very sensitive (~0.3 pg)
and inexpensive CVAFS detector (Bloom and Fitzgerald,
1988; Kvietkus et al., 1983) many research groups have
used this instrumental approach for Hg measurements
of low-level air and natural water samples (Bloom and
Fitzgerald, 1988; Gill and Bruland, 1990; Liang and
Bloom 1993; Mason and Fitzgerald, 1993; Cai, 2000;
