University of Alberta

Total Mercury Analysis in Water

 Printer Friendly

Determination of Total Mercury in Water by Oxidation, Purge and Trap, and Cold Vapor Atomic Fluorescence Spectrometry.

Modified by Vincent St.Louis and Jane Kirk. Original version written by Mark L. Olson and John F. De Wild, Wisconsin District Mercury Laboratory, U.S. Geological Survey, Madison, Wisconsin. This protocol has been listed here with permission of the original authors.


1.0 Scope and Application
2.0 Summary of Methods
3.0 Safety Issues
4.0 Sample Preservation, Containers, and Holding
5.0 Reagents and Standards
6.0 Quality Control
7.0 Procedure
8.0 Archiving
9.0 References

1.0 Scope and Application:

1.1 Applicable Matrices: This method may be ustabed to determine mercury in filtered or unfiltered water samples.

1.2 Minimum Reporting Limit: The minimum reporting limit for this method is 0.04 ng/L (nanograms per Liter).

1.3 Dynamic Range: This method is designed for the measurement of total mercury (Hg) in the range of 0.04 - 100 ng/L. The upper range may be extended to higher levels with the selection of a smaller sample volume.

2.0 Summary of Method:

Bromine Monochloride (BrCl) is added to the sample container to oxidize all forms of Hg to HgII oxidation state. After a minimum of 12 hours the BrCl is neutralized by addition of Hydroxylamine Hydrochloride (NH2OH*HCl). Following neutralization, Stannous Chloride (SnCl2) is added to the sample to reduce the Hg from the HgII to the Hg0 oxidation state. The Hg0 is purged onto gold-coated glass bead traps (sample). The mercury vapor is thermally desorbed to a second gold trap (analytical) and from that detected by cold vapor atomic fluorescence spectrometry (CVAFS). Samples high in organic matter may require initial pretreatment in an ultra violet (UV) digester to remove the colour associated with the organic matter.

3.0 Safety Issues:

Specific safety concerns for each chemical can be found in the Material Safety Data Sheets for that chemical, all of which are located in the laboratory. Two extremely important areas of safety for this method are addressed below.

3.1 Chronic mercury exposure may cause kidney damage, muscle tremors, spasms, personality changes, depression, irritability and nervousness. Due to the toxicological and physical properties of Hg, only highly trained personnel using extremely cautionary procedures should handle high concentration standards. These cautionary measures include use of gloves and high volume hoods when preparing standards.

3.2 Strong acid solutions are employed in the cleaning of equipment, preparation of reagents and in sample preservation. Proper acid handling techniques should be employed whenever acids are being used. These techniques include the use of acid resistant clothing, self-contained breathing apparatus/respirators, and the utilization of high volume fume hoods.

4.0 Sample Preservative, Containers, and Holding Times:

4.1 Sample containers consist of Teflon bottles cleaned at the laboratory. New Teflon bottles are rinsed with tap water, and cleaned by immersing in concentrated reagent grade nitric acid heated to 65C for 6-8 hours. Bottles are then rinsed in Milli-Q water and immersed in 30% ultra trace HCl heated to 65C for 6-8 hours. Immediately following removal from the bath, the bottles are immersed in Milli-Q water and rinsed at least 3 times with Milli-Q water. Following the rinsing step, each bottle is filled with 1% Baker Instra-analysed HCl and capped. The exterior of the bottles is allowed to air dry in our mercury-free class 100 clean room. Dry equipment is double bagged in new zip-type bags with the unique identifier written on the outer bag. Used sample bottles are cleaned the same way.

4.2 Samples are preserved by the addition of 0.2% concentrated Baker Instra-analysed hydrochloric acid (HCl).

4.3 Properly preserved samples may be stored for up to 6 months before analysis is carried out.

5.0 Reagents and Standards:

5.1 Reagents: All reagents and/or dry chemicals used to make reagents must be of the highest purity available from the vendor and shown to be low in mercury. Upon receipt at the laboratory, containers will be marked with the date of receipt and stored in appropriate areas. When reagents are mixed for use in this method, the person who mixes them will initial and date the reagent container.

5.1.1Milli-Q water: Ultra pure Milli-Q water shown to be > 18 MW starting from pre-purified source (distilled, RO, etc.). The water is delivered through a 0.2 uM filter. All water is obtained from a Millipore ELIX 10/Milli-Q water purification system.

5.1.2 Hydrochloric Acid: Baker Instra-analysed HCl (containing less than 5 ng/L Hg) or equivalent.

5.1.3 Bromine monochloride (BrCl): Dissolve 27 g of reagent grade Potassium Bromide (KBr) in 2.5 L concentrated Baker Instra-analysed HCl. Place a Teflon coated stir bar into the bottle and stir for 1 hour in a fumehood. Slowly add 38 g reagent grade Potassium Bromate (KBrO3) to the solution while stirring. CAUTION: This needs to be done slowly and in a fume hood because large quantities of free halogens are produced. As you add the KBrO3 to the solution, the colour should change from yellow to red to orange. Cap bottle loosely and allow to stir for an additional hour. The BrCl is analyzed for Hg prior to adding to samples.

*To reduce the Hg content of the reagents, muffle both the KBr and the KBrO3 overnight at 250C. Take the reagents out of the muffle furnace while still hot, and dissolve KBr in HCl as soon as possible.

5.1.4 Hydroxylamine hydrochloride (NH2OH*HCl):

Dissolve 30 g of NH2OH*HCl in 100 mL of reagent water in a clean glass Erlenmeyer flask. Add 100 uL of SnCl2 and purge with Hg free N2 overnight at 300 ml/min. Prepare fresh every 6 months.

5.1.5 Stannous chloride (SnCl2) Add 500 mL of reagent water to 100 g SnCl2 in a clean Erlenmeyer flask. Add 50 mL concentrated Baker Instra-analysed HCl and purge overnight with Hg free N2 a 300mL/min. Prepare fresh at least every 6 months.

5.1.6 Nitrogen (N2). Ultra High Purity grade that is passed through a gold bead trap attached to the outlet of the tank to remove any Hg.

5.1.7 Argon (Ar). Ultra High Purity grade that is passed through a gold bead trap attached to the outlet of the tank to remove any Hg.

5.2 Standards: Upon receipt at the laboratory or on the day of preparation, containers should be labeled with the date received or made and the initials of the person preparing them. The stock and sub stock standards should by stored outside of the clean laboratory to prevent contamination of the entire lab.

5.2.1 Stock Standard (1000 mg/L): Commercially available Hg standard verified against a NIST standard reference material. All subsequent standards are prepared using the stock standard. Before preparing other standards, ensure the expiration date of the stock standard has not been exceeded.

5.2.2 Sub stock or Secondary Standard (1000 ng/mL): The stock standard (1000 mg/L) is diluted by a factor of 1000 with 1% Baker Instra-analysed HCl.

5.2.3 Working standard (1 ng/mL): The secondary standard (1000 ng/mL) is diluted by a factor of 1000 in a clean Teflon bottle. This working standard must be compared to the previous working standard and agree within 5%. Prepare fresh every 6 months or until it goes off.

6.0 Quality Control:

Each analyst hired in our laboratory must demonstrate the ability to generate acceptable accuracy and precision with this method. This includes the ability to reproduce standards, establish acceptable daily detection limits (DDL), produce acceptable relative percent differences between replicates, and produce spike recoveries that meet acceptance criteria.

6.1 Vapor Standards: Three vapor standards (morning, noon, and end of the day) are run throughout the day to ensure that the Tekran Model 2500 CVAFS Mercury Detector (see below) is functioning properly and that there are no leaks in the analytical or thermal desorbtion system. Using a syringe, 25 pg of Hg vapor is injected directly onto a sample trap on the analytical system. The liquid and vapor standards are compared and must be within 10% of each other. If they are not within 10%, then the problem must be resolved before analysis continues.

6.2 Bubbler blanks: A bubbler blank is prepared by adding 1 mL of SnCl2 to a bubbler containing approximately 150 mL of Milli-Q water. Blanks are critical to the reliable determination of Hg at low levels. Frequent analysis of bubbler blanks is required to demonstrate freedom of system contamination and the absence of carry over from one sample or standard to the next.

6.2.1 Acceptance criteria: No bubbler blank must contain more than 25 pg of Hg. A daily detection limit (DDL) may be calculated prior to analysis of samples. The DDL is computed using the following formula:

DDL= ((3 x s) x RFm) / 0.125 L

s = standard deviation among peak areas found for a set of bubbler blanks purged simultaneously on all bubblers RFm = ratio of concentration to peak area from the standard curve (sec. 6.2.1) The acceptable value for the DDL must be 0.04 ng/L or less.

6.2.2 Corrective Actions: If a bubbler blank is found to contain more than 25 pg Hg at the beginning of the day, another set of bubbler blanks should be run to ensure the entire system has been purged and that the value is true. If this second set blanks is also high, the analyst must isolate and correct the problem before continuing. If a bubbler blank is found to contain more than 25 pg Hg, during the course of the day's analyses, the data produced on that bubbler should be rerun or carefully evaluated and flagged as being suspect. If the DDL exceeds 0.04 ng/L, another set of bubbler blanks should be run to ensure the entire system has been purged and that the value is true. If this second set blanks is also high, the analyst must isolate and correct the problem before continuing.

6.3 Standards: A standard is prepared by adding a known volume of working standard and 1 mL of SnCl2 to a bubbler containing approximately 150 mL of mercury free water.

6.3.1 Liquid standards must be within 10% of each other and of the vapor standard2.

6.3.2 Corrective action: If the standards fail to meet acceptance criteria, an additional set of standards must be run to ensure no operator error exists. If the second set of standards still does not meet acceptance criteria, the analyst must isolate and correct the problem before continuing.

6.4 Duplicates: At minimum, 20% of samples are run in duplicate.

6.4.1 Acceptance criteria: Duplicate mercury concentrations must be within 10% of each other.

6.4.2 Corrective action: If the duplicate concentrations are not within 10% then the sample must be run a third time. If the relative standard deviation between the three replicates is greater than 10%, the sample is flagged and/or analyzed and fourth time.

RSD (%) = (s/mean) X 100

s = standard deviation among the three replicates

6.5 Matrix spike: A matrix spike is prepared by adding a known concentration of working standard to a sample. A matrix spike must be analyzed each run or every tenth sample whichever is greater.

6.5.1 Acceptance criteria: Percent recovery for a matrix spike must fall between 90 and 110%.

% Recovery = ((Concentration of Spiked Sample Concentration That The Spike Added)/Concentration of Unspiked Sample) X 100

6.5.2 Corrective actions: If the percent recovery falls beyond the range of 90 to 110%, a second spike on that sample must be run or another sample from the set must be spiked. If percent recovery for the second spike falls beyond the range of 90 to 110%, the batch of samples analyzed for that day are flagged identifying either high or low recovery.

7.0 Procedure

7.1 Comments: The samples are collected using ultra clean sampling techniques.

7.1.1 Interferences: Free halogens: The destruction of the gold traps exists if they are exposed to free halogens resulting in low mercury values. This is avoided with the addition of a soda lime trap directly upstream of the sample traps during purging. Water vapor: Water vapor may collect on the gold traps during the purging step. If water vapor is present on the traps this will give a false peak during analysis. This is avoided with the addition of a soda lime directly upstream of the sample traps during purging and with a 10 min drying step after purging. Carrier gas: The fluorescent intensity of the detector is strongly dependent on the inertness of the carrier gas. The dual amalgamation step virtually eliminates quenching due to impurities in the carrier gas, but it is the analyst's responsibility to ensure high purity inert carrier gas and a leak free analytical train.

7.1.2 Helpful hints: Working with detection limits in the parts per trillion range, protecting these samples from contamination cannot be over emphasized. The greatest difficulty in low level mercury analysis is preventing the samples from becoming contaminated. Extreme caution must be used throughout the preparation, collection and analysis procedures to avoid contamination. It is very important that the laboratory air be low in both particulate and gaseous Hg. If a bubbler shows signs of contamination, it can be cleaned by dissolving approximately 2 g of Potassium Hydroxide in 250 mL of reagent water and soaking for 1 hour or more. If the peak area for Hg on a trap exceeds 3 ng, reburn the trap. If the peak exceeds 5 pg, reburn the trap, if a peak is still present, discard the trap and incorporate a new trap in its place.

7.2 General Description: Refer to section 2 of this procedure for a summary of this method.

7.3 Labware:

7.3.1 All-plastic pneumatic fixed-volume and variable pipettors in the range of 5 L to 10 mL.

7.3.2 Analytical balance capable of measuring to the nearest 0.1 g.

7.3 Sample preparation

7.3.1 Ultraviolet light (UV) oxidation: Samples with a high degree of organic colour are placed in a UV digestion box until colourless. The UV digestion box consists of a large box lined with aluminum foil, shiny side out, and contains 2 UV lights. This step may be skipped if the samples are clear water samples where the organic colour in the sample will not interfere with the colour produced from the addition of BrCl (Olson et al 1997).

7.3.2 BrCl oxidation: After the UV oxidation or a determination that UV treatment is not necessary, the sample is oxidized with the addition of BrCl. Enough BrCl is added to give the sample a distinct yellow colour. If the yellow colour remains after 12 hours, this indicates excess BrCl in the sample. If the yellow colour is absent, additional BrCl is added. Additions of BrCl must be repeated until the sample colour indicates the presence of excess BrCl. The amount of BrCl that is necessary to oxidize the samples may vary from < 1% up to 5% in some cases. The BrCl needs to be analyzed prior to each addition to document the amount of Hg added to the sample.

7.3.3 Neutralization: The excess BrCl in the sample needs to be reduced to avoid destruction of the gold coated glass bead traps by the presence of free halogens. To reduce the BrCl add 200 L of NH2OH*HCl for every 1 mL of BrCl added to the sample. Swirl the sample. The yellow colour will disappear, indicating the destruction of the BrCl. Allow the sample to react another 5 min before analysis. Only neutralize samples immediately prior to analysis. Samples should not be stored neutralized. The volume of NH2OH*HCl added to the sample must be recorded.

7.4 Instrumentation:

7.4.1 Purging and preconcentration: Regulator capable of supplying 30 psi of pressure. Various sizes of Teflon tubing and fittings. Flow meter(s) capable of maintaining a N2 flow of 300-400 mL/min. Needle valve to shut off N2 flow to bubblers. Gold coated glass bead traps: The gold coated glass bead traps are constructed of a 7 mm quartz tube, 4" long and with a constriction at 1" from the outlet end. A quartz plug is placed into the inlet end, about 0.7 g (3.5 cm in the tube) of gold coated beads are added and the inlet end is plugged with another piece of quartz wool. Fittings for gold traps are made from small pieces of 0.320 PTFE heat shrink tubing. End plugs for gold traps are made from Teflon rod. Bubblers: The bubblers are 250 mL borosilicate glass Erlenmeyer flasks with the standard 24/40 tapered neck. The sparging stopper has a coarse glass frit that extends to the bottom of the flask. Soda lime traps: The soda lime traps are constructed of 0.320 PTFE heat shrink tubing with custom machined Teflon end plugs. The traps are filled with 4-8 mesh soda lime. New soda lime traps are pre-purged for 40 min before collection of a sample onto a gold trap. Soda lime traps that have been used before are pre-purged for 20 min. Soda lime traps should be repacked every 2-3 days.

7.4.2 Desorbtion and analysis: Analytical train: The Analytical train consists of 2 variable current transformers, 2 cooling fans, and a timer. The transformers are connected to Nichrome coils that are wrapped to fit around the sample traps and the analytical trap. First the sample trap is heated to at least 320C for 4 min and 30 sec. Then the analytical trap is heated for the same amount of time or until the Hg peak has come back down to the baseline on the computer. After the heating of each coil, the fans are activated to cool the traps. Detector: The detector is a commercially available Model 2500 CVAFS Mercury Detector from Tekran (Toronto, ON) equipped with a mass flow controller capable of measuring 40 mL/min. Integrator: The detector analog output is connected to a personal computer with Varian STAR integration software. Peak areas are recorded.

7.5 Initial start-up, calibration and sample nalysis:

7.5.1 Check pressure in Argon tank to verify adequate volume for the days analyses.

7.5.2 Adjust mass flow controller at detector to read 40.0 mL/min.

7.5.3 Check baseline at detector, acceptable baseline readings are from 0.0050 and 0.0250. If the baseline is out of that range, adjust by turning the offset knob. Record date, initial reading and adjusted reading in the instrument notebook.

7.5.4 Start the day by burning the set of sample traps to clean (if they have been sitting unused for more than 2 days) and running a vapor standard. While these traps are being burned, the soda lime traps are purged. To purge the soda lime traps, attach them to the outlet of the bubblers, rinse and then fill the bubblers with approximately 150 mL of reagent water, and attach the nitrogen line to the inlet of the bubbler. Purge with nitrogen at 300 mL/min for 20 or 40 min, depending upon if the traps are new or have been used before. Remove the plugs from the ends of the first trap and place it into the analytical train by threading it through the center of the Nichrome wire coil. Ensure that the Nichrome wire completely covers all of the gold beads and glass wool of the trap. Start the timer for 4 min and 30 sec. Once this timer goes off, switch the fans, turn off the heat on the first Nichrome coil, and switch on the heat for the second. Start the timer again and turn off the heat once the Hg peak comes off on the computer and is back down to baseline. After the 9-minute cycle is complete repeat the steps in for each of the remaining traps. The reduction of samples, purging onto gold traps, and drying of the traps takes 30 minutes, as does the burning of 4 traps, so you will be trapping samples on gold traps while burning (analyzing) the previous round. This is the cycle you will follow throughout the day.

7.5.5 Remove the caps from the inlet and outlet of the bubblers and thoroughly rinse the bubblers with Milli-Q water.

7.5.6 Dispense approximately 150 mL of Milli-Q water into each of the bubblers and add 1.0 mL of SnCl2. Ensure that the caps are tightened on the bubblers and allow to react for 10 min.

7.5.7 Attach the nitrogen line to the inlet of the bubbler, the cleaned sample traps to the soda lime traps, and begin purging at 300 mL/min with nitrogen. Purge for 10 min.

7.5.8 Attach the sample traps to the nitrogen lines with teflon drying plugs, reduce the nitrogen flow to 150 ml/min, and allow the sample traps to dry for 10 min. This represents a bubbler blank.

7.5.9 When the above 30-minute cycle for the bubbler blanks has elapsed (sec. 7.5.5 7.5.8), remove the gold traps from the end of the soda lime traps, cap both ends of the gold traps, and shut-off the flow of the N2.

7.5.10 Analyze the gold traps as in sec. to While these traps are being burned, proceed with step 7.6.11.

7.5.11 Rinse the bubblers and refill with approximately 150 mL Milli-Q water. First add 1 mL SnCl2 to the bubblers, and then 100 L of the working standard. Allow the standards to react for 10 min and then proceed with steps 7.5.7 and 7.5.8. This represents your standards.

7.5.12 Analyze the gold traps as in sec. to While these traps are being burned proceed with step 7.6.13.

7.5.13 Rinse the bubblers and refill with approximately 100 mL Milli-Q water. Add 1 mL SnCl2 and then add approximately 50 mL of neutralized water sample. You must weigh the water sample before and after addition to the bubbler so that you know exactly how much you have added to the bubbler for analysis. Allow the samples to react for 10 min and then proceed with steps 7.5.7 and 7.5.8 as before.

7.5.14 This chart summarizes a day of sample analysis.


Bubbler 1

Bubbler 2

Bubbler 3

Bubbler 4







0.100 ng

0.100 ng

0.100 ng

0.100 ng

Sample set A





Sample set B





Sample set C





Sample set D


DUP e.g. S4



Sample set E

DUP e.g. S1




Sample set F


DUP e.g. S9



BB = bubbler blank
SX = sample
SPIKE = matrix spike

7.5.16 All samples need to be bracketed by standards, if the sample peak area is greater than the highest standard, either a higher standard is analyzed or the sample is analyzed using a smaller sample volume. 7.6 Calibration and performance documentation: During the analysis run, the analyst must evaluate the calibration data, bubbler blank values, and standard values to ensure acceptance criteria (sec. 6.0) are being met. The following information must be recorded in the mercury logbook.

7.6.1 Date of analysis.

7.6.2 Type and date prepared for reagents and standards used.

7.6.3 Name of analyst.

7.6.4 Identification of bubbler contents, volume analyzed, instrument response, and sample trap identification for each analysis performed.

7.6.5 Comments pertaining to special samples run, problem samples, corrective actions taken, and results of any calculations performed to ensure acceptance criteria are being met.

7.7 Shut-down:

7.7.1 After the last sample has been run, the following steps must be performed to properly store the bubblers until the next analysis run. Shut off N2 flow at the needle valve and at the tank regulator. Remove the N2 line from the inlet and the soda lime trap from the outlet of the bubblers. Rinse the bubblers with reagent grade water. Fill the bubblers to approximately 95% volume with reagent grade water.

7.7.2 After the last sample trap has been burned, leave the trap in the analytical train to avoid contamination from room air. Reduce flow at the mass flow controller to the minimum flow allowable (approximately 6 mL/min).

7.8 Maintenance records are kept for each machine.

7.8.1 Nichrome wire temperature should be checked regularly.

7.9 Calculations

7.9.1 Final (corrected) concentrations: Raw data is entered into an Microsoft EXCEL spreadsheet that calculates final concentrations. The concentration of a sample is calculated by comparing the sample peak area to the standard peak area to gain a value in ng of mercury. This ng value is then divided by the volume of sample analysed to get a concentration value. The concentration value must than be adjusted or corrected by subtracting additional Hg that has been added to the sample during preservation, bromination and analysis.

7.10 Data validation and evaluation: After the data has been entered into the Microsoft EXCEL spreadsheet, the analyst verifies that no values have been incorrectly entered into the spreadsheet. Data is evaluated as to reasonability if historical data from a site exists.

7.11 Reporting:

7.11.1 Reporting units: Total mercury as ng/L Hg.

7.11.2 Reporting levels and significant figures: Report to the nearest 0.01 ng/L for values less than 10 ng/L. Report to three significant figures for values exceeding 10 ng/L.

7.11.3 Data transfer: After the data has been verified in the Microsoft EXCEL spread sheet it may be transferred to the customer via e-mail, hard copy, or the internet.

8.0 Archiving:

All raw data produced in the laboratory is archived in binders and lab books. All electronic data is archived on the laboratory manager's computer, and regularly backed up onto CD.

9.0 References:

9.1 Method source:

U.S. Environmental Protection Agency, 1996, Draft Method 1631: Mercury in Water by Oxidation, Purge and Trap, and Cold Vapor Atomic Fluorescence Spectrometry: EPA 821-R-96-012, Office of Water, 32 p.

Olson, M.L. Cleckner, L.B., Hurley, J.P., Krabbenhoft, D.P., Heelan, T.W. 1997, Resolution of matrix effects on analysis of total and methyl mercury in aqueous samples from the Florida Everglades: Fresenius Journal Analytical Chemistry. 358: 392-396

9.2 Deviations from source method and rationale:

9.2.1 Method 1631 states that a 100 mL aliquot of sample should be poured into a clean 125 mL bottle before bromination and neutralization. To avoid the potential contamination associated with this transfer, this procedure recommends bromination and neutralization in the original sample bottle. Bromination in the original sample container also has the benefit of stripping any mercury that may have adhered to the container wall.

9.2.2 Gold coated glass bead traps are used in place of sand traps. There are two advantages to glass beads: 1) less back pressure, and 2) the ability to handle higher temperatures during analysis which provides cleaner burning traps (Olson et al 1997).

9.2.3 To increase the confidence in the analytical result obtained using this procedure, all samples are analyzed in duplicate, which is not required in Method 1631.

9.2.4 Method 1631 establishes bubbler blank control limits at 50 pg. This method sets the control limits for bubbler blanks at 25 pg.

Last Modified: 2002-10-25