MERCURY 
Mercury concentration was determined on a 
separate portion of the sample. Two hundred milli- 
grams of sediment (100 mg, if sample concentra- 
tions were expected to be >50 parts per billion) 
were decomposed in a 1-oz teflon screw-top vial 
with 2 mL of concentrated HNO3 (J. T. Baker 
Chemical Co.) and 2 mL of HC1O, (G. Frederick 
Smith Chemical Co. (GFS)) double distilled from 
Vycor, a pure silica glass. The mixture was heated 
in a capped vial until the solution reached 200 °C. 
The solution was then heated with the cap off for 
about 45 minutes, after which the samples were 
removed from the heat source. Immediately, 1 mL 
of concentrated HNO, was added; the vial was 
filled with H,O and capped tightly until used. The 
sample solution then was added to a flask contain- 
ing 125 mL of HO and 4 mL of 10-percent 
(weight: volume) SnCl2 in 20-percent HCl. Nitrogen 
was passed through the solution to remove elemen- 
tal Hg, which was collected on gold foil located in 
the center of the coils of an induction furnace. 
Activation of the furnace released the Hg, which 
was measured by a cold-vapor AA technique. 
Blanks, standard rocks, and internal sediment 
standards were analyzed for each set of samples. A 
series of solutions was prepared that had the same 
Hg-concentration range expected in the samples. 
The concentrations of Hg in bottom sediments 
determined during the first year of the monitor- 
ing program were typically less than the detection 
limit of 0.01 ppm. During the second and third year 
of monitoring, we tested new procedures designed 
to lower the detection limit. 
The contribution of Hg from various brands of 
nitric acid was determined. Baker ‘analyzed re- 
agent grade” contained less than 0.5 ppb Hg, the 
lowest concentration of the acids tested. Baker 
“ultrex’”’ contained 2 ppb Hg, and Mallinckrodt 
nitric acid contained 1.3 ppb Hg. During the 
checks of HC1O,, we found that some bottles of 
GFS double-distilled HClO, contained 5 ppb Hg. 
We ultimately selected GFS HClO, double dis- 
tilled from Vycor, which was found to contain less 
than 0.5 ppb Hg. The Hg concentration of each 
new bottle of acid and of every other reagent was 
determined before the reagent was used for analy- 
sis. The Hg contribution from the combined 
reagents was reduced to 0.5 ng+0.1 ng. 
We tried to lower the detection limit by increas- 
ing the sample size. Subsamples weighing 1 g were 
analyzed with various combinations of nitric and 
perchloric acids. The results were not encouraging 
because digestion was incomplete when small acid 
volumes were used or because blanks were too high 
when large acid volumes were used. The high sedi- 
ment concentration in suspension during the gas- 
stripping procedure may have adsorbed some of 
the Hg and accounted for the lower concentration 
measured for large samples. 
Another method of increasing sample size was 
successive plating of Hg vapor from three 200-mg 
aliquots onto the gold foil of the induction furnace. 
This technique yielded poor reproducibility among 
replicates and decreased the number of samples 
that could be analyzed in a day by a factor of 3. 
The selection of reagents having the lower Hg 
concentration, the addition of a digital readout 
voltmeter, and the optimization of the optical 
system in the cold-vapor AA detection system 
(manufactured by Laboratory Data Control, Inc.) 
reduced the detection limit of our procedure from 
0.01 ppm to 0.005 ppm. 
The magnitude of Hg lost while oven drying sedi- 
ment samples also was evaluated. Aliquots of bulk 
sediments from station M06-13A and aliquots of 
the fine fraction from station M05-16 were ana- 
lyzed wet, and the results were compared to sam- 
ples that were oven dried at different temperatures. 
We found no evidence of Hg loss as a result of dry- 
ing bulk sediments at temperatures between 40 
and 100 °C, but we observed some loss (about 
42 percent) when drying fine-fraction samples at 
100 °C. 
ADDITIONAL METHODS 
Results of Ba and Cr analyses on selected 
Georges Bank samples were cross-checked by an 
energy-dispersive X-ray fluorescence technique 
(Johnson, 1984). The determination of Ba concen- 
tration was made with a Kevex 0700 energy- 
dispersive X-ray fluorescence spectrometer. 
Powdered samples of about 1 g were analyzed with 
a gadolinium secondary target for excitation of the 
K-alpha line. The ratio of Ba intensity to the 
gadolinium Compton scatter intensity was used to 
correct for absorption effects. This ratio then was 
compared to a standard calibration curve to deter- 
mine the concentration of Ba. 
The X-ray fluorescence technique was used on 
all samples found to have more than 500 ppm 
Ba during the first analysis by acid decomposition 
and ICP spectrometry. The X-ray fluorescence 
technique is highly accurate in samples enriched 
with BaSO,, which is a difficult mineral to dis- 
solve completely if present in high concentration. 
