4.18  Analysis of mercury in precipitation and air

4.18.1  Analysis of mercury in precipitation

4.18.1.1  Instrumentation

The most common procedure for the analysis of mercury in precipitation is oxidation with BrCl, pre-reduction with NH2OH.HCl followed by reduction of the aqueous Hg to Hg0, purging onto gold traps and thermal desorption and analysis using Cold Vapour Atomic Fluorescence Spectroscopy (CVAFS) (Bloom and Crecelius, 1983; Bloom and Fitzgerald, 1988). The analysis procedure can be performed in manual or automated modes. A detection limit (defined as 3 times the standard deviation of the blank concentration) 2 ng/l is necessary for the accurate analysis of total mercury in precipitation samples from remote stations.

The most reliable technique for the analysis of mercury is atomic fluorescence spectrometry (AFS). Atomic absorption spectrometry (AAS) may be used but requires larger sample volumes due to higher detection limit. AFS and AAS instruments are available from a number of different manufacturers.

Borosilicate glass is the recommended material for the reaction and purging of flasks where mercury is reduced and volatilised for the pre-concentration step. Acid-washed Teflon tubing should be used. Ordinary polyethylene or rubber tubing is not suitable.

4.18.1.2  Sample storage and handling

All water samples for mercury analyses must be handled with care in order to avoid contamination. Sample bottles should only be handled in laboratories where mercury or any mercury compounds in pure or concentrated forms have not been handled. Samples for analysis of total mercury should be preserved with low blank HCl (5 ml 30% acid/l). Precipitation samples should be stored in the collection bottles, in double plastic bags in the dark in a refrigerator or cold room. A storage time up to 6 months can be acceptable but it is absolutely necessary to test this under the conditions employed in the individual laboratories. Shorter storage times are recommended if methylmercury is analysed. Plastic gloves must always be used when the plastic bags are opened. If possible, the plastic bags should be left around the bottles during the analysis. The bottles should not be placed on laboratory surfaces that may have been exposed to mercury or chemical reagents containing mercury.

4.18.1.3  Chemicals and glassware

Purging flasks for SnCl2-reduction:
Acid-cleaned borosilicate glass (Pyrex) wash-bottles are used.

Hg-free Nitrogen/Hg-free Argon:
The gas should go through a gold trap or coal filter

High purity water:
Purified water with >18 MW resistance and a low mercury blank.

Air in laboratory:
All glassware and samples should be handled in a laboratory containing low concentrations of mercury (not more that 10 ng/m3 if possible). A clean bench (or some other clean zone arrangement) of class 100 should be used for handling reagents, for some sample treatment and for the drying of glassware.

Hydrochloric acid:
30% HCl (Suprapur) from Merck is recommended. Other manufacturers may provide equally high quality hydrochloric acid. Regular blank checks should be made. For the preparation of SnCl2 solution, 37% HCl (P.A.) is necessary.

Brominemonochloride solution:
Must be prepared in a fume hood with great care. Use safety goggles.
Add 11.0 g KBrO3 and 15.0 g KBr to 200 ml high purity water. Stir the solution with a magnetic bar for 1 hour and add 800 ml 30% HCl very slowly. Large amounts of acid fumes and gaseous free halogens will form and will evaporate from the solution. The solution can be prepared in an empty HCl bottle.

Hydroxylammonium chloride:
Dissolve 120 g NH2OH.HCl in 1 l high purity water. This chemical reagent sometimes contains high mercury concentrations. Adding 1 g Chelex 100 ion exchange material can lower the mercury content. Blanks must be checked carefully.

Stannous Chloride solution:
Dissolve 200 g SnCl2.2H2O in 100 ml 37% HCl (p.a.) and dilute to 1 l with high purity water. Purge this solution with mercury-free N2 for 12 hours and then store it in the dark. Aliquots of 100 ml may be removed and used as working solutions for analysis. These aliquots should be purged continuously with mercury-free N2.

Mercury calibrating solution:
Standard solutions can be prepared from commercially available mercury standards. A parallel check using two standard solutions of different origin is recommended. One of these can be made from pure chemicals (e.g. Hg0 dissolved in concentrated HNO3 and diluted to the appropriate volume).

4.18.1.4  Pre-treatment

The collected samples are preserved with HCl prior to storage or during sampling. Before analysis a chemical oxidation step is performed using BrCl. This reagent efficiently converts stable mercury forms to water soluble species that can be easily reduced by SnCl2. Before analysing the sample, excess BrCl is removed using a mild reducing agent such as NH2OH.HCl or ascorbic acid.

4.18.1.5  Preparation of reducing vessels

Fill the wash bottles with about 50 ml water containing 2.5 ml of the SnCl2.solution and 2 ml 30% HCl. Purge the solution with N2 for 20 minutes before checking the bubbler blank value.

At the end of each day, the bottles should be rinsed thoroughly with de-ionised water and then filled (at least covering the glass frit) with Aqua Regia until use. Before starting the next set of analyses, the Aqua Regia should be transferred to a storage bottle (Aqua Regia can be re-used for up to a month) and the reduction vessel rinsed, first with de-ionised water and then with high purity water (e.g. Milli-Q).

4.18.1.6  Reduction step

The bubbler blank value should be checked by connecting a gold trap to the bubbler and purging the solution with N2 for 20 minutes, then analysing the mercury collected. The mercury collected on the gold trap is the bubbler blank and should not exceed a few picograms.

In all collection and purging steps, a glass tube containing baked quartz wool should be connected between the bottle and the gold trap to avoid exposing the gold surface to droplets of acid solution.

After the bubbler blank has been checked, a clean gold trap should be connected to the outlet and an aliquot of the pre-treated precipitation sample added to the bubbler flask. The bubbler flask should then be placed on an electronic balance and the amount of sample added weighed. The reduction and purging should be allowed to proceed for 10 to 20 minutes.

4.18.1.7  Detection

The traps should be dried at about 40°C in a mercury-free N2 flow for 5 minutes prior to analysis. They should then be connected to the AFS detector on line with the helium gas flow. The mercury is then thermally desorbed either directly into the detector or onto an analytical trap. If an analytical trap is used, a second heating step should be performed before the detection. The advantage of the dual amalgamation is that the influence of any interfering substances adsorbed on the first trap may be reduced and also that the mercury adsorbed onto the second analytical trap will be more easily desorbed and a sharper peak obtained.

After the analytical step the gold trap should be allowed to cool. It should then be removed from the gas stream and stoppered with Teflon plugs. It should be stored in a plastic bag if not immediately used again.

4.18.1.8  Calibration

Standard solutions can be prepared from commercially available mercury standards. Calibration should be performed by using 4 standards in each run.

4.18.1.9  Quality control - Quality assurance

The calibration step is critical. In general, the basic principle is always to use two independent calibrant solutions One of these can be made from pure chemicals (e.g. Hg0 dissolved in concentrated HNO3 and diluted to the appropriate volume). For mercury, commercially available standard solutions can be used but regular checks against a reference standard must be made. Certified reference materials should be used if available, but reference standards can also be prepared from pure mercury compounds. Traceability is an important step and all standard solutions must be regularly checked against a reference material. In the absence of aqueous phase reference standards, solid materials may be used.

As an independent check on the analytical results, a Hg0 vapour source can be used consisting of liquid mercury in an enclosed vessel from which vapour samples can be drawn with a gas tight syringe.

4.18.1.10  Special problems

The analysis of low level mercury concentrations in aqueous samples is associated with a number of potential errors mainly emanating from blank problems and poor recovery.

Blank values usually arise from the use of reagents of poor quality or from glass vessels or tubing. Careful checking and documentation of all steps in the analytical procedure is necessary in order to identify the source of the blank.

4.18.1.11  Summary

 

Recommendation

Acceptable alternatives

Sample pre-treatment

BrCl oxidation, NH2OH.HCl pre-reduction

Ascorbic acid

Preconcentration

SnCl2 reduction, purging, collection on gold traps

 

Detection

AFS

AAS

Detection limit

< 2 ng/l

 

QA/QC

Blank determinations, use of traceable reference materials

 

 

4.18.2  Analysis of mercury in air

Mercury collected on gold traps is analysed after desorption of the mercury.

4.18.2.1  Sample pre-treatment

Before analysing the mercury content of the gold trap, a drying step is recommended. Small amounts of water vapour may have condensed on the gold surface and may interfere in the analysis step. The gold traps can be heated to 40-50°C for 5-10 min in a stream of dry N2 without any measurable loss of mercury.

4.18.2.2  Analysis

The analysis of mercury in air samples is generally made using double amalgamation CVAFS (Fitzgerald and Gill, 1979; Bloom and Fitzgerald, 1988). In this procedure, the gold trap is mounted in series with a second analytical trap in a gas stream (Hg-free argon) leading to the CVAFS detector. Heating is achieved with a heating wire (e.g. NiCr). In the first step the mercury is thermally desorbed from the first sampling trap onto the second analytical trap. The second trap is then rapidly heated and the mercury is transported into the CVAFS with an integrator.

The analytical steps are as follows:

  1. thermal desorption from the field trap to the analytical trap: 500ºC for 4 minutes, with 30 ml/min flow rate
  2. thermal desorption from the analytical trap to the AFS: 800ºC for 25 seconds, with 30 ml/min flow rate
  3. total gaseous mercury calculation: Peak Area of the integrated AFS Signal.

4.18.2.3  Calibration

Mercury-saturated air is supplied from a closed flask (ca 350 ml), containing 30-40 ml of mercury (Dumarey et al., 1985). The inner pressure is kept at atmospheric pressure by means of a side-arm, which has access to ambient conditions via a capillary. The flask is placed in a thermostat (20ºC ± 0.1ºC). 0.1 ml saturated air is removed via a septum by using a gas-tight syringe (Hamilton #1810). 0.1 ml of air at 20ºC and 101,325 Pa contains 1.316 ng Hg according to Ideal Gas Law (Table 4.8). Its accuracy depends mainly on the temperature of the mercury-saturated air, which must be lower than the ambient temperature to prevent condensation of mercury in the syringe. By preconditioning the syringe, initial irreproducible measurements caused by sorption are avoided. Under optimal conditions the standard deviation of the injection with 0.1 ml (n=10) should be better than 10%. The re-establishment of the equilibrium between liquid and gaseous mercury depends on the cleanness of the pool surface. After some time, mercury at the surface becomes oxidised by atmospheric oxygen and the upper layer must be removed.


Table 4.18.1: Concentrations of Hg calibration gas as function of temperature.

Temp °C

ng Hg/ml

Temp °C

Ng Hg/ml

Temp °C

ng Hg/ml

Temp °C

ng Hg/ml

17.0

10.22

20.5

13.72

22.2

15.79

23.9

18.14

18.0

11.12

20.6

13.83

22.3

15.92

24.0

18.29

19.0

12.10

20.7

13.95

22.4

16.05

24.1

18.44

19.1

12.20

20.8

14.06

22.5

16.18

24.2

18.59

19.2

12.31

20.9

14.18

22.6

16.31

24.3

18.74

19.3

12.41

21.0

14.30

22.7

16.45

24.4

18.89

19.4

12.52

21.1

14.42

22.8

16.58

24.5

19.04

19.5

12.62

21.2

14.54

22.9

16.72

24.6

19.20

19.6

12.73

21.3

14.66

23.0

16.86

24.7

19.35

19.7

12.83

21.4

14.78

23.1

17.00

24.8

19.51

19.8

12.94

21.5

14.90

23.2

17.13

24.9

19.67

19.9

13.05

21.6

15.03

23.3

17.28

25.0

19.83

20.0

13.16

21.7

15.15

23.4

17.42

26.0

21.49

20.1

13.27

21.8

15.28

23.5

17.56

27.0

23.27

20.2

13.38

21.9

15.40

23.6

17.70

 

 

20.3

13.49

22.0

15.53

23.7

17.85

 

 

20.4

13.60

22.1

15.66

23.8

17.99

 

 

 

4.18.2.4  Quality assurance

The necessary quality control steps are primarily associated with gold trap collection and analytical instrument reliability.

All gold traps must be individually calibrated at regular intervals. This is most conveniently done using a source of gaseous mercury, i.e. a thermostated vessel containing liquid mercury from which gaseous samples can be drawn with a gas tight syringe. Gold traps with low recovery must be discarded.

To provide an internal control on the field results, one set of 3 or 4 gold traps is routinely kept in the glass container during a field study. The set is then analysed along with the other field samples. In almost all cases, a typical mercury blank result of 5-30 pg Hg was observed.

4.18.2.5  Detection limit

The detection limit of the gold trap analysis is defined as 3 times the standard deviation of the trap blank provided that the trap blank is subtracted from the analysed amount. This detection limit, expressed in units of ng Hg, can be translated into an air concentration using the typical air volume sampled in this application. The detection limit can also be based on the requirement that the blank content of mercury on the traps should not exceed 10% of the total content of mercury collected during a normal sample period. The absolute value will depend on sampling time and air flow rate. As a guideline the following example can be used: in air containing 2 ng Hg/m3 a sample collected for six hours at an air flow rate of 0.5 l/min contains 0.36 ng. In this case, the blank content on the trap should not exceed 10% of 0.36, i.e. 0.036 ng. If the blank content is higher than this value, then the detection limit exceeds 2 ng/m3 under the conditions employed, and a larger sample volume is required.

4.18.2.6  Special problems

The analysis of mercury collected on gold traps is generally straightforward provided that the collection efficiency of the gold traps is checked regularly.

4.18.2.7  Summary

 

Recommendation

Acceptable alternative

Sample pre-treatment

Drying at 30-50oC if necessary.

 

Analysis

Dual amalgamation CVAFS.

CVAAS.

Detection limit

3s of trap blank and/or trap blank <10% of sample.

 

QA/QC

Check gold trap collection efficiency and recovery.

 

 

4.18.3  References

Bloom, N.S. and Crecelius, E.A. (1983) Determination of mercury in seawater at subnanogram per litre levels. Mar. Chem., 14, 49-59.

Bloom, N.S. and Fitzgerald, W.F. (1988) Determination of volatile mercury species at the picogram level by low temperature gas chromatography with cold-vapour atomic fluorescence detection. Anal. Chim. Acta, 208, 151-161.

Bloom, N.S., Prestbo, E.M., Hall, B. and Von der Geest, E.J. (1995) Determination of atmospheric mercury by collection on iodated carbon, acid digestion and CVAFS detection. Water, Air, Soil Pollut. 80, 1315-1318.

Dumarey, R., Temmerman, E., Dams, R. and Hoste, J. (1985) The accuracy of the vapour-injection calibration method for the determination of mercury by amalgamation/cold-vapour atomic absorption spectrometry. Anal. Chim. Acta, 170, 337-340.

Ebinghaus, R., Jennings, S.G., Schroeder, W.H., Berg, T., Donaghy, T., Ferrara, R., Guentzel, J., Kenny, D., Kock. H.H., Kvietkus, K., Landing, W., Mazzolai, B., Mühleck, Munthe, J., Prestbo, E.M., Schneeberger, D. Slemr. F., Sommar, J., Urba, A. Wallschläger, D. and Xiao, Z. (1999) International field intercomparison measurements of atmospheric mercury species at Mace Head, Ireland. Atmos. Environ., 33, 3063-3073.

Fitzgerald, W.F. and Gill, G.A. (1979) Subnanogram determination of mercury by two-stage gold amalgamation and gas-phase detection applied to atmospheric analysis. Anal. Chem., 51, 1714-1720.

Iverfeldt, Å. and Sjöberg, K. (1992) Intercomparison of methods for the determination of mercury deposition to convention waters. Göteborg, Swedish Environmental Research Institute (IVL Report B 1082).

Munthe, J. (1996) Guidelines for the sampling and analysis of mercury in air and precipitation. Göteborg, Swedish Environmental Research Institute (IVL-report L 96/204).

OSPAR (1997) JAMP Guidelines for the sampling and analysis of mercury in air and precipitation. London.

Schroeder, W.H., Keeler, G., Kock, H., Roussel, P., Schneeberger, D. and Schaedlion, F. (1995) International field intercomparison of atmospheric mercury measurement methods. Water, Air, Soil Pollut., 80, 611-620.


Last revision: November 2001