JRHS 2013; 13(1): 37-42

Copyright © Journal of Research in Health Sciences

Optimization of a Novel Setup for an On-Line Study of Elemental Mercury Adsorption by Cold-Vapor Atomic Absorption Spectrometry

Mohammad Javad Assari (MSc)a, Abbas Rezaee (PhD)b*, Ahmad Jonidi Jafari (PhD)b, Abdolrahman Bahrami(PhD)c

a Department of Occupational Health Engineering, Faculty of Medical Sciences, Tarbiat Modares University, Tehran, Iran

b Department of Environmental Health Engineering, Faculty of Medical Sciences, Tarbiat Modares University, Tehran, Iran

c Department of Occupational Health Engineering, School of Public Health, Hamadan University of Medical Sciences, Hamadan, Iran

* Correspondence:Abbas Rezaee (PhD):E-mail: rezaee@modares.ac.ir

Received: 31 October 2012, Revised: 25 December 2012, Accepted: 06 February 2013, Available online: 09 February 2013


Background: The objective of this work was developing a simple and stable time-based on-line setup for assessing the potential of mercury (Hg) vapor adsorption of the commercial sorbents used in air sampling and control operation followed by cold vapor atomic absorption spectrometry (CVAAS).

Methods: A special designed separation chamber was used where reduction of the injected Hg (II) solution took place. Purge gas passes through this chamber resulting to a prompt release of mercury vapor, purging into the adsorbent that regulated at the desired adsorption temperature. After sorbent saturation, in order to study the adsorption parameters of sorbents (activated carbon and bone char) such as breakthrough time (BTT), and adsorptive capacity, mercury gas stream was passed through the sorbents, directly transport to the CVAAS.

Results: Preliminary experiments concerning the reductant solution showed that SnCl2 offers higher stability than NaBH4. Around the loading range 0.125-2.5 ml min1 of 100 µg l-1 Hg(II) solution, a linear calibration curve with the equation P=0.134L-0.017 (p, peak area; L, Loading flow) and a correlation coefficient r=0.996 was obtained, and the detection limit was improved up to cL=1µg l-1. The relative standard deviation of five measurements of lowest flow loading of Hg (II) was RSD=2.8%. The significant differences were observed in the breakthrough time and mercury adsorptive capacity between activated carbon and bone char (P=0.010).

Conclusion: This novel setup is suitable for an on-line study of elemental mercury adsorption, determination of breakthrough time and adsorption capacity, and because of its stable performance during all experiments; it can be applied to the time based studies.

Keywords: Mercury, Vapor generation, Adsorption, Separation, Cold vapor atomic absorption, Spectrometry


Mercury and its compounds are recognized as potentially hazardous materials and are rated as the main category of environmental and occupational pollutants. Global emissions of mercury to the atmosphere have increased 3 times during the recently century year, due to primarily anthropogenic releases, associated with fossil fuel burning and waste incineration1,2. Mercury exists in several physical states and chemical forms depending on temperature and pressure. Because of its propensity to undergo biological interaction, mercury generally undergoes complex and difficult-to-predict changes in its concentration levels and chemical forms. As point of physical view, mercury species in the air, in the form of gaseous are more than particulate components. Mercury has been grouped into three categories: (1) elemental mercury; (2) inorganic mercury; and, (3) organic mercury3,4.

Adsorption in solid materials is a process that offers great potential for achieving high quality air, with respect to mercury emissions, because of the most of existing air pollution control technologies, such as electrostatic precipitators and bag houses, cannot remove effectively the gas phase mercury emissions, except for the particle-bound mercury atoms5-7.

Cold vapor atomic absorption spectrometry (CVAAS) is recognized as a routine procedure for mercury determination in various fields, due to the advantages of efficient matrix separation, high analyte transport efficiency, high selectivity, simple instrumentation and ease of automation8. There are two different ways for producing constant mercury vapor concentration in a gas stream: Application of permeable membranes and evaporating the liquid mercury, while each method has its advantages and limitations. Continuous flow (CF) or flow injection (FI) and recently sequential injection (SI) systems can overcome the existing draw backs because they offer automated handling of sample, improved precision, drastic diminishing of reagents and sample consumption and increased sampling frequency9-13.

Although mercury in metallic form has a considerable vapor pressure (0.0016 mbar at 20 °C corresponding to a concentration of ca. 14 mgm3 in air), always needs to be separated efficiently from the liquid phase, in order to ensure that Hg0 is transferred to the atomic absorption cell and consequently to improve the sensitivity. This is achieved by the gasliquid separator device. Generally, there are three approaches for the separation of the gaseous phase: (a) hydrostatic (U-type), (b) with forced outlet, (c) micro porous poly-tetra-fluoro-ethylene (PTFE) membrane. In batch CVAAS methods, the Gas Liquid Separator (GLS) is employed also as reaction chamber for mercury reduction. On the other hand, various GLS configurations for on-line systems have been presented in the literature, in order to increase peak height of the transient signal14-15.

The objective of this work was developing a simple and stable time-based on-line setup for assessing the potential of mercury vapor adsorption of the commercial sorbents used in air sampling and control operation followed by CVAAS. The proposed setup involves a new separation chamber, which in parallel operates as reactor. This was tested by activated carbon and bone char in order to determine the breakthrough and total capacity of this sorbent.



Ultrapure water (TKA; Smart 2 Pure; Germany) was used. All chemicals including HgCl2, SnCl2·2H2O, NaBH4, NaOH, 65%w/w HNO3, and 37%w/w HCl were analytical-reagent grade provided from Merck, Darmstadt, Germany. Argon (Ar) and N2 gases were research grade and supplied from Roham, Tehran, Iran which used as the purge gas. A stock solution (1000 mg l1) of Hg (II) was prepared by dissolution of mercury chloride salt. The mercury stock solution was stored at 4°C prior to use. Working standard solutions of Hg (II) were prepared by appropriate stepwise dilution of stock standard solution in 0.5 mol l1 HNO3 to the required levels just before use. The reducing agents (SnCl2 in 0.6 mol l1HCl or NaBH4 in 0.1 mol l1 NaOH) was prepared fresh daily. The reductant solutions were purified from possible traces of elemental mercury by 30 min degassing with argon prior to use.


An atomic absorption spectrometer model AAnalyst 700 Perkin Elmer (Precisely) equipped with quartz cell. An Hg hollow cathode lamp (Lumina™ Perkin Elmer) was used as the radiation source. Peak absorbance of the transient signal was chosen for quantification. Instrumental parameters were set up as follows: lamp intensity: 6 mA; wavelength: 253.7 nm; spectral band pass: 0.7 nm; quartz cell temperature: 25 °C; and measurement mode: peak height.

The proposed setup was consisted of two peristaltic pumps (JMS model SP-510 and JAZ-1800) and adjustable valves. The whole system was controlled via WinLab32TM software. A silicon tube (4.9 mm i.d.) was used for separation chamber and CVAAS connection, in order to keep the dead volume of the system at minimum level and thus the vapor dispersion at lower level. Ar gas was used for the mercury vapor release from liquid mixture. The separation chamber was constructed of a cylindrical polyethylene container (164.3 mm length, 72.7 mm i.d.) and the push-fit connections made of poly-tetra-fluoro-ethylene. The channels in upper cell managed two tubes (4.9 mm i.d.) made of silicon, one tube for the inlet purge gas and another for outlet mixture (Hg0 + Ar). The channels in lower cell managed two parallel outlets, one for HgCl2 solution and the other for reducing agent. The above system is shown schematically in Figure 1.


Mercury vapor generation; the proposed mercury vapor generator unit was based on reduction of Hg (II) solution and bubbling purge gas through on it (Figure 1). The efficiency of mercury reduction and discharging of it was checked by pumping the Hg (II) and reducing agents (SnCl2 or NaBH4) into the separation chamber for long and short period. HgCl2 solution and reducing agent were loaded sequentially to the container by activating peristaltic pumps. The reduction of Hg (II) causes mercury drift that it provides a well detectable mercury concentration in the gas stream. The reducing agent and HgCl2 solution continuously stirred by using a magnetic stirrer in order to assure better mixing and sufficient reduction time. Ar passes through separation chamber resulting to a prompt release and transportation of mercury vapor to the CVAAS (Figure 1). The peak height of the absorbance signal that was proportional to mercury vapor concentration in the solution was used for all measurements. Each instance repeated three times in all cases, and repeatability was shown with relative standard deviation (RSD %).

Mercury adsorption experiments were conducted by two type sorbents, activated carbon made of coconut charcoal and bone char. The mass and mesh of sorbents were 150 mg and 20/40, respectively. The reactor adsorber was a quartz tube (185.6 mm in length with an outer diameter of 7.9 mm and inner diameter of 4.9 mm) that held in a vertical position. The variety of adsorption units was described with more details by Skodras et al. 16. Quartz has good chemical resistance and is inert to elemental mercury. A furnace equipped with temperature controller was employed to keep the adsorbent reactor bed at a desired temperature. When the thermal stability was provided, the pure Ar was directed to the empty reactor. Then the mercury vapor generator unit that was considered to generate the desired mercury concentration was placed in line. At the same time, the reactor outlet was absorbed into the absorbent that was set the desired adsorption temperature, and then determined the Hg0 concentration by CVAAS (Figure 1). In order to study of adsorption parameters such as breakthrough time (BTT), and adsorptive capacity, the experiment was performed by using of 100 µg l-1 Hg (II) solution in both loading flow 1.25 and 2.5 ml min1. The purge gas stream and the adsorption temperature were adjusted at 200 ml min-1 and 50 °C respectively.

It should be noted that the mercury vapor released from the reaction vessel discharge of the lab by the efficient ventilation system and wastes disposal after finishing the experiment managed in an appropriate and approved waste disposal facility.

Figure 1: The schematic of a novel setup used for testing the elemental mercury adsorption. (1) Purge gas, (2) Dryer (3) Adjusted valve (4) Flow meter (5) Separation chamber (6) Stirrer (7) Hg (II) peristaltic pump (8) Reductant peristaltic pump (9) Waste (10) Temperature controller (11) Quartz tube (12) Sorbent (13) Oven (14) CVAAS


Optimization of proposed system parameters

The optimization of the system parameters was proceeded under the above procedure (Figure 1), using of Hg (II) standard solution in the concentration range on 10-1000 µg ll and loading flow rate in the range on 0.1-5 ml min1 respectively. The results were indicated that the flow rate and concentrations of mercury influenced the peak height. The optimization of system performance data are presented in Table 1. A linear calibration curve was obtained in the concentration on 100 µg ll and the loading range of 0.125-2.5 ml min1 of Hg (II) with the equation p=0.134L-0.017 (p, peak area; L, loading flow) and a correlation coefficient r=0.996, and the detection limit up to cL= 1µg l1 was improved (Figure 2). The relative standard deviation of five measurements of the lowest loading of Hg (II) was RSD=2.8%.

The effect of reducing agent concentration on the generation of mercury vapor was studied from 1 to 5% (w/v). The results clearly indicated that as the reducing agent concentration increased, signal intensity of mercury significantly increased and reached to a maximum when the SnCl2 and NaBH4 concentrations were 2.5% and 1% respectively (Figure 3). Furthermore, the effect of the reducing agent flow rate on the mercury vapor generation was also tested from 0.01 to 5 ml min1 keeping the concentration of the reducing agents at above values. That is, considering the sensitivity and the consumption of Hg (II) solution, the SnCl2 and NaBH4 flow rates were optimized in the ranges of 0.0250.125 and 0.010.1 ml min1 respectively keeping mercury concentration and flow rate at 100 µg l-1 and 1 ml min1. The results from preliminary experiments concerning the reducing agent showed that SnCl2 offers higher stability than NaBH4. When 1% NaBH4 was used it was observed that reproducibility reduced, despite the signal intensity increased. In addition, the peak height of mercury decreased when NaBH4 concentration was increased further to 3%. Thus, SnCl2 solution 2.5% (W/V) was selected as adaptable concentration of the reductant in all the subsequent experiments. In all cases, the solution was prepared in basic media (HCl 0.6 mol l-1), in order to stabilize this reagent. The optimal concentration and flow rate ranges of mercury and SnCl2 solutions were used throughout.

To check the influence of purge gas type and flow rate on peak height was tested within the range of 100-600 ml min1 keeping the concentration and flow rate of Hg (II) and SnCl2 solution at optimized values. The obtained results were shown that the peak height was increased by increasing of the flow rate, up to 200 ml min1 and over this value, it would be decreased. In addition, by increasing the purge gas flow rate higher than 400 ml min1 the peak height was significantly decreased. Thus for all subsequent experiments 200 ml min1 argon flow rate was used.

The results indicate a significant relationship between the stirring of separation chamber and peak height (P=0.010), this relationship especially was more when the air was used as purge gas. The relationship between type of purge gas (Ar, N2 or Air) and the peak height at the optimized experimental conditions was not significant. Furthermore, at these conditions, the relationship between quartz cell types (open or closed) and the peak height was not significant.

Using of novel setup for the determination of mercury adsorption

This experiment was performed by using initial concentration at 100 µg ll in both loading flow 1.25 and 2.5 ml min1. The mercury breakthrough curves, during 90 min adsorption experiment were determined by using of continuous collected data. Then the total adsorptive capacity of sorbents was calculated by comparison of the shapes and the bases of these curves. As shown in Table 2 in above loading flows, the obtained breakpoints from activated carbon were 45 and 12 min and from bone char were 6 and 1 min. As observed in Figure 3, the breakthrough curves of activated carbon followed a much more efficient profile than in bone char. Thus, for activated carbon, the mercury adsorbed quantity was higher than bone char. The significant differences were observed in the breakthrough time and mercury adsorptive capacity between activated carbon and bone char (P=0.010).

Table 1: The optimization of setup performance data

Table 2: Compression of breakpoints from activated carbon and bone char

Figure 2: Calibration curve of Hg vapor production and loading range of Hg (II); 100 µg l-1 Hg(II)  initial concentration, 200 ml min-1 Ar flow rate

Figure 3: Effect of reducing agents concentration on peak area; 100 µg l-1 Hg(II) initial concentration, 1ml min-1 Hg (II) flow rate


The simultaneous investigation of loading flow rate and concentration of Hg (II) solution, the type and loading flow rate of reducing agents (SnCl2 or NaBH4), the type and flow rate of purge gases (Ar, N2 or Air), the type of quartz cell used (open or closed), and stirring of the separation chamber during vapor generation, which has been less attention in literature, were the unique properties of this study.

As it is shown in Figure 2, by increasing the flow rate of Hg(II) solution up to 2.5 ml min1, the absorbance of the fixed amount of Hg almost linear increased and then out of linear mode, shown that larger concentration make Hg0 release more difficult. Thus, according to the initial mercury concentration (100 µg l1), the use of proposed system was a linear range between 0.125-2.5 ml min-1 and had good sensitivity (Table 1). The effect of mercury solvent volume on the sensitivity of the sequential injection system measured input has been studied in other researches 9, 12, 17. The study of Anthemidis et al. (2004) established that, despite larger sample volumes offer increased overall sensitivity, with developing a sequential injection system for trace mercury determination by CVAAS9. This is close to the results of the present study.

On the other hand, the vapor generation initiates into the separation chamber, which functions also as reaction chamber. Thus, in present system, it is more convenient and less time consuming, to load mercury and the reducing agents directly into the separation chamber, especially when large volumes are used. Consequently, the liberated mercury vapor can be delivered with the lower dispersion into the absorption cell. Based on these considerations, in the proposed system, in order to start the reaction with Hg (II) solution loading the reductant solution excess there was already inside the chamber (Figure 1). The efficiency of Hg reduction and purging was checked by pumping the Hg (II) for long and short period into the reaction vessel containing Sn (II) ions. After stopping the pump which delivers the mercury solution, the signal decrease since mercury free gas passes through the system. The difference of the tendency decreasing of the signal gives information about the accumulation of the mercury in the generator. Since a marginal difference found, it can be stated that, all of the delivered Hg (II) has been changed to mercury vapor. 

As it has been reported in the literature, the concentration and flow rate of the reductant are critical in optimizing the mercury signal in CVAAS 13, 14. In the present study peak height of the Hg0 as a function of the concentration and flow rate of both SnCl2 and NaBH4 were obtained, keeping mercury concentration and flow rate at 100 µg l-1 and 1 ml min1. As it is shown in Figure 3, low reductant concentration results in incomplete vapor generation of Hg0. As the SnCl2 and NaBH4 concentration increased up to 2.5% and 1% respectively, all of inorganic mercury was reduced to Hg0. This difference may be attributed to the variation of the vapor generating efficiency of the reducing agents, which other studies have referred to in12. Also, it was observed that reproducibility was decreased. Balarama Krishna et al. (2010), in study with the aim of speciation of inorganic and methyl mercury showed that peak area of mercury species decreases when NaBH4 concentration was increased further to 3% 18, which confirms the results of present study. 

The purge gas flow rate will affect mainly the determination by CVAAS and final mercury vapor concentration and consequence the detection limit and sensitivity of sequential injection system8, 9. The obtained results clearly indicate that flow rate increases to 200 ml min-1, peak height increases and then remains constant. Therefore, it is concluded that saturation of mercury was completed and mixing time for analyte dilution was sufficient. While the observed deterioration of reproducibility in flow rate higher than 400 ml min1 probably was related to leak fitting.

The significant relationship between the stirring of separation chamber and peak height especially when the air was used as purge gas, can be linked to form an oxide layer on the undisturbed surface, which causes a slow shifting of the peak height. Stirring of the reductant and mercury in the container has restored the initial mercury signal, and by using the renewable surface, constant mercury vapor concentration can be achieved, and if the flow rate of gas is not extremely high, the saturation can be achieved. Constant mercury vapor concentration could be achieved by using of the renewable surface.

As it is observed in the results, the increase in flow rate of Hg (II) solvent from 1.25 to 2.5 ml min-1 resulted in a significant decrease in the breakthrough time and mercury adsorptive capacity (Figure 4, Table 2). This can be attributed to the effective pore diffusivity decrease which in other studies referred to in5,16. Skodras et al. (2008), by study of two types of carbon active with different pore indicated that increase of the initial concentration from 0.1 to 0.35 ng/cm3 causes the decrease in breakpoint from 52 to 22 min19. It seems that carbon active particle produces a delay of the breakthrough, compared with bone char. That is, in the carbon active particle, adsorption breakthrough curves follow a much more efficient profile than that in bone char (Figure 4).

Figure 4: Compression of mercury vapor adsorption on both activated carbon and bone char; 100 (µg l-1) Hg(II) initial concentration, 185.6 mm column length, 4.9 mm column diameter, 500C adsorption temperature, 150 mg sorbent mass, 200 ml min-1 Ar flow rate. (a) With 1.25 (ml min1) Hg (II) flow rate. (b) With 2.5 (ml min1) Hg (II) flow rate


A novel mercury vapor generation unit based on CVAAS, undertaken to provide the advantages such as simplicity, use facilitating, small dead volume, quickly of mercury ions reduction, and the use of new separation chamber resulting thus to minimum vapor dispersion, it facilitates the successful manipulation of mercury loading in sorbent types is the ideal system for mercury vapor generation and as well as appropriate alternative for pre-concentration method. The above setup can successfully manage a wide range of mercury solution loading for varied sensitivity while its performance was stable during all experiments. This setup is adaptable to an on-line study of elemental mercury adsorption and the determination of breakthrough time and adsorption capacity. It could be applied to the time based studies.


The authors would like to thank for some instrumental and technical supports from Hamadan University of Medical Sciences.

Conflict of interest statement

The authors have no conflict of interests to declare.


This study was funded by Tarbiat Modares University.


  1. World Health Organization. Environmental Health Criteria 118-Inorganic Mercury. Geneva: WHO; 1991.
  2. Hanisch C. Wheres Mercury Deposition Coming From? Environ. Sci. Technol. 1998;32 (7):176A-179A.
  3. Lu JY, Schroeder WH. Annual time-series of total filterable atmospheric mercury concentrations in the Arctic. Tellus. 2004;56:213-222.
  4. Agency for Toxic Substances and Disease Registry (ATSDR). Toxicological profile for Mercury. Atlanta: Department of Health and Human Services, Public Health Service; 1999.
  5. Lopez-Anton MA, Tascon J MD, Martinez-Tarazona MR. Retention of mercury in activated carbons in coal combustion and gasification flue gas. Fuel Process Technol. 2002;77-78:353-358.
  6. Laumb JD, Benson SA, Olson EA. X-ray photoelectron spectroscopy analysis of mercury sorbents surface chemistry. Fuel Process Technol. 2004;85(6-7):577-585.
  7. Galbreath KC, Zygarlicke CJ. Mercury transformations in coal combustion flue gas. Fuel Process Technol. 2000;65-66:289-310.
  8. Sturgeon RE, Guo X, Mester Z. Chemical vapor generation: are further advances yet possible? Anal Bioanal Chem. 2005;382(4):881-883.
  9. Anthemidis AN, Zachariadis GA, Stratis JA. Development of a sequential injection system for trace mercury determination by cold vapour atomic absorption spectrometry utilizing an integrated gasliquid separator/reactor, Talanta. 2004;64(4)1053-1057.
  10. Menendez-Garcia, Fernandez Sanchez ML, Sanchez Uria JE, Sanz-Medel A. Speciation of mercury by continuous flow liquid-liquid extraction and inductively coupled plasma atomic emission spectrometry detection. Microchimica Acta. 1996;122(3-4):157-166.
  11. Segade SR, Tyson JF. Determination of inorganic mercury and total mercury in biological and environmental samples by flow injection-cold vapor-atomic absorption spectrometry using sodium borohydride as the sole reducing agent. Spectrochimica Acta. 2003;58:797-807.
  12. Doering WE, James RR, Echols RT, Fresenius J. A sequential injection cold-vapor atomic absorption method for the determination of total mercury. J Anal Chem. 2000;368:475-479.
  13. Fang Z. Flow Injection Atomic Absorption Spectrometry, John Wiley and Sons Ltd.: West Sussex; 1995.
  14. Sanz-Medel A. Flow analysis with atomic spectrometric detectors, Analytical Spectroscopy Library; Elsevier: Amsterdam; 1999.
  15. Wuilloud JC, Wuilloud RG, Salonia JA, Olsina RA, Martinez LD. Modified gas-liquid separator for the determination of mercury with cold vapor flow injection-inductively coupled plasma optical emission spectrometry. Journal of Analytical Chemistry. 2002;57(9):799-801.
  16. Skodras G, Diamantopoulou Ir, Zabaniotou A, Stavropoulos G, Sakellaropoulos GP. Enhanced mercury adsorption in activated carbons from biomass materials and waste tires. Fuel Processing Technology. 2007;88:749-758.
  17. Raquel BR, Mesquita, Antonio OSS Rangel. A review on sequential injection methods for water analysis. Analytica Chimica Acta. 2009;648(1):7-22.
  18. Balarama Krishna MV. Chandrasekaran K, Karunasagar D. On-line speciation of inorganic and methyl mercury in waters and fish tissues using polyaniline micro-column and flow injection-chemical vapour generation-inductively coupled plasma mass spectrometry (FI-CVG-ICPMS). Talanta. 2010;81:462-472.
  19. Skodras G, Diamantopoulou Ir, Pantoleontos G, Sakellaropoulos GP. Kinetic studies of elemental mercury adsorption in activated carbon fixed bed reactor. J Hazard Mater. 2008;158:1-13.

  20. JRHS Office:

    School of Public Health, Hamadan University of Medical Sciences, Shaheed Fahmideh Ave. Hamadan, Islamic Republic of Iran

    Postal code: 6517838695, PO box: 65175-4171

    Tel: +98 81 38380292, Fax: +98 81 38380509

    E-mail: jrhs@umsha.ac.ir