May 2008

System Integration

It's the law, now what?

Continuous monitoring of mercury emissions will be necessary for all stationary sources where annual Hg emissions exceed 29 pounds of Hg


  • Stack gas sampling for Hg analysis from coal-combustion has many challenges.
  • The important issue facing electric utility industry is a short compliance timeline.
  • The commercialization of the various mercury control technologies is afoot.
By Dr. Nenad Sarunac

With mercury regulations pending and control technologies in the full-scale demonstration stage, measurement of mercury (Hg) in combustion flue gas is of critical importance.

The ability to accurately and reliably measure mercury is fundamental to demonstrating compliance when regulations spread and, in the meantime, to ensuring adequate quantification of mercury removal during the demonstration and commercialization of the various mercury control technologies.

The important issue facing electric utility industry is a short compliance timeline: by 1 January 2009, certified continuous mercury monitors (CMM) need to be in place. Following certification, a certified CMM should collect 12 months of mercury emissions data. Reporting of data for compliance monitoring would start on 1 January 2010. This gives a year for CMM installation and certification, and two years until mandatory reporting for emissions compliance.

Collecting a representative flue-gas sample for Hg analysis from coal-combustion flue gas produces many challenges.

The complexity of flue gas chemistry, relatively high temperatures, reactivity of mercury species, and particulate loading are matters of attention to ensure the flue gas sample that reaches the mercury-measuring device is representative of the gas stream within the duct or stack.

In addition to measuring total mercury accurately, the identification and quantification of species of mercury is also very important.

Mercury emissions from anthropogenic sources occur in three forms: solid particulate-associated mercury, Hg(p); gaseous divalent mercury, Hg2+; and gaseous elemental mercury, Hg0.

Continuous monitoring of mercury emissions will be necessary for all stationary sources where annual Hg emissions exceed 29 pounds of Hg.

However, because mercury is present in three different forms, the measurement process and analysis is considerably more complicated.

Monitors and methods

The Ontario Hydro wet chemistry method and dry sorbent trap methods provide good results for total and speciated mercury measurements. However, they can fail to provide the real-time data often necessary for environmental compliance.

Hg CEMs are similar to other combustion system CEMs in that a sample leaves the gas stream, sees some conditioning, and travels to a remote analyzer for detection.

Although on-line emission analyzers are expensive to purchase, install, and maintain, they offer several benefits:

  • Real- or near-real-time emission data
  • Operational data for process control and environmental compliance
  • Evaluation of control strategies
  • Greater understanding of process variability and operation
  • Greater public assurance

As they are currently working, Hg CEMs possess several challenges to long-term, low-maintenance continuous operation for flue gas mercury monitoring.

The two main challenges include the areas of sample collection and flue gas conditioning. Collecting a representative flue-gas sample for mercury analysis from coal-combustion flue gas is very difficult.

One must address the complexity of flue gas chemistry, high temperatures, reactivity of mercury species, and particulate loading to ensure the flue gas sample that reaches the mercury-measuring device is representative of the gas stream within the duct.

To a great extent, heated sample lines, gas conditioning systems, and material of construction have addressed many of these issues; however, monitoring dirty locations remains difficult, especially when reactive ash is present and traditional probe filters are used.

The Environmental Protection Agency (EPA) has proposed a new performance standard that outlines the requirement for mercury measurement using continuous emission monitors-Performance Specification 12A (PS-12A).


Here is a summary as they are pertinent to defining the requirements of Hg CEMs.

  • The Hg CEMs must be capable of measuring the total concentration in µg/m3 (regardless of speciation) of vapor-phase Hg and recording that concentration on a dry basis, corrected to 20°C and 7% CO2.
  • Particulate-bound Hg is not included.
  • The CEMs must include a diluent (CO2) monitor and an automatic sampling system.
  • Calibration techniques and auxiliary procedures are not here.
  • Procedures for measuring Hg CEM relative accuracy, measurement error, and drift are here.
  • Hg CEM installation and measurement location specifications and data reduction procedures are included.
  • There is an outline of procedures for comparison with reference methods.

The basic steps all Hg CEMs must accomplish in order to effectively measure mercury in a flue gas stream include

  • Filter particulate matter from the sample gas while minimizing flue gas fly ash contact.
  • Transport the sample gas to a conditioning system, or condition the sample at the sampling port and transport the conditioned sample to the instrument.
  • Condition the sample by reducing all forms of mercury in the sample gas to Hg0 and remove moisture form the sample gas.
  • Measure the mercury in the flue gas sample.

Mercury analyzers are distinguishable by their measurement detection principle. Methods used include pre-concentration by gold amalgamation with Cold Vapor Atomic Absorption Spectrometry (CVAAS) detection, Zeeman modulated CVAAS, pre-concentration, and gold amalgamation with CVAFS detection. Instruments using AES and laser technologies are also in development.


CVAAS method: The Cold Vapor Atomic Absorption Spectrometry (CVAAS) method determines the mercury concentration in the gas by measuring the attenuation of the light produced by a mercury vapor lamp as it passes through a cell that contains the sample gas. The mercury atoms in the cell absorb mercury at their characteristic wavelength of 253.7 nm.

Zeeman modulated CVAAS removes the interference of other flue gas components by using a powerful magnet to shift slightly the wavelength of the mercury vapor lamp. The broadband absorbers will attenuate the signal at both wavelengths, and the difference between the signals is because of the mercury concentration.

CVAFS method:  Typically, the Cold Vapor Atomic Florescence Spectroscopy method uses gold amalgamation to pre-concentrate the mercury. After the mercury desorbs from the trap, it travels into the detection cell by an argon carrier gas. The mercury atoms in the cell are excited to fluorescence by a pulsed mercury discharge lamp, which a photomultiplier tube then measures.

AES method: The AES method is currently being developed by Envimetrics for mercury measurement. The method is based on the emission of light from mercury atoms induced by a high-energy source such as plasma. The light is emitted at the characteristic 253.7-nm wavelength.

Flue gas conditioning: CVAFS- and CVAAS-type mercury analyzers can only measure elemental mercury. Therefore, to measure the total mercury concentration in a sample gas stream, the oxidized forms must be reduced to elemental mercury in a conversion system. The most common method of reducing oxidized forms of mercury to elemental is using a liquid reducing agent such as stannous chloride. This method is used extensively, has proven to be problematic, but is receiving much technical attention toward improvement.

Particulate removal: Particulate-bound mercury captured on a filter can translate to Hg0, but because of particulate matter transport issues, it is impractical. In addition, EPA Draft Performance Specification 12A only requires CEMs to measure "the total concentration (regardless of speciation) of vapor phase mercury." Therefore, it is important to remove any particulate matter from the sample gas stream in a manner that ensures it does not interfere with the operation of the analyzer or impart a bias to the mercury data. To eliminate this problem, most CEM systems are either equipped with an inertial separation probe or a blowback filter.


Calibration: Regardless of the measurement technique or conversion system, all instruments must be calibrated. All of the instruments available are easily zeroed by passing a filtered mercury-free sample gas through the analyzer.

For information on the technologies and the companies that are developing them, see the report at


Dr. Nenad Sarunac ( is principal research engineer at the Energy Research Center at Lehigh University in Pennsylvania. This article comes from his report Armstrong Project: Evaluation and comparison of U.S. and EU reference methods for measurement of mercury, heavy metals, PM2.5, and PM10 emissions from fossil-fired power plants.

Stack emissions: Keeping them real

A comparison of surface adsorption effects in mercury and sulfur analyzer systems shows "just any" material can ruin sample, invalidate results

By Martin Higgins, Gary Barone, David Smith, and Ted Neeme

Mercury emissions have a significant impact on human health and the environment. Doctors and researchers have determined that mercury exposure causes nerve and brain tissue development problems in children and can harm the brain, heart, kidneys, lungs, and immune system of people regardless of age or sex.

The primary sources of mercury emissions are coal-fired power plants with 33% of total U.S. emissions, municipal waste incinerators 19%, commercial boilers, and medical waste incinerators.

Every year, 48 tons of mercury goes into the environment through coal powered power plant stack emissions. As a result, the U.S. Environmental Protection Agency (EPA) established mercury monitoring and emissions standards for coal-fired power plants and other point source mercury emitters.

U.S. EPA 40CFR parts 60, 63, 72, and 75 require coal-fired power plants to be compliant with mercury emission standards beginning 1 January 2009.

As a result, coal plants must install mercury emissions monitoring equipment on approximately 1,300 coal units within the next two years, with a first phase emissions cap of 38 tons per year by 2010 and a second phase cap of 15 tons per year by 2018.

To meet compliance regulations, U.S. coal-fired utility plants will invest approximately $6 billion in capital projects targeted at mercury reduction.

Ongoing monitoring will cost the industry another $948 million per year. This does not include costs related to corrosive damage, as many stack environments common to coal-fired power generators emit sulfur compounds, sulfuric acid, hydrochloric acid, and other corrosives that cause stack monitoring and testing equipment failures.

Further complicating coal-fired plant stack test reliability is the complexity of mercury species present in the plume. Stack mercury emissions may exist in three forms: elemental mercury (Hg); the 2+ oxidation state (Hg++); and as attached to particulate matter.

In many stack emission streams, Hg++ will react with sulfur compounds, nitrogen, chlorine, and/or oxygen to produce sulfurous, nitrous, chloride and oxide mercury species.

Additionally, elemental and oxidized mercury can be lost to reactions and adsorptions on the inner surfaces of monitoring equipment.

Transport, retention of Hg

The combined effect is inaccurate mercury readings, which can result in costly retesting or may result in broad financial, environmental, and regulatory repercussions as a result of non-compliance issues.

Analytical testing costs alone can be substantial. Recent studies estimate a per-test sampling cost ranging from $100-$640 for a typical mercury analysis.

Because of the significant costs associated with inaccurate mercury sampling, we studied various common materials used in stack emission monitoring for their impact on test accuracy of mercury and sulfur samples.

Regulations developed for the characterization of mercury and sulfur emissions demand sampling and analysis systems that are capable of reliably and reproducibly transferring sample streams containing these active compounds.

Sampling system surfaces in the transport and retention of mercury and sulfur compounds were evaluated.

Mercury and sulfur streams are difficult to transport due to activity with and adsorption to ferrous surfaces.

Surfaces were tested, including 316L grade stainless steel, 304-grade stainless steel, and functionalized amorphous silicon-coated 316L grade and 304-grade stainless steel.

Industries benefiting from this study include stack gas sampling, environmental quality testing, refining streams, oil and gas exploration and transport, or any industry transporting or retaining active compounds.

Previous studies by Restek Corporation and O'Brien Corporation have focused on sulfur compound adsorption in static and dynamic sampling systems. Results of those studies have demonstrated significant adsorption in non-treated stainless steel storage and transfer systems.

This study's experimental data showed a 70% greater loss of mercury when stored in 304SS sample cylinders over a period of 50 days. Significant and rapid mercury loss when exposed to a 304SS surface begins upon sample charging and continues throughout the test duration.

Silicon coated surfaces show an initial mercury loss of 5% with sample loss stabilizing within seven days.

The test data demonstrate significant Hg/Sulfur or Hg/Hg oxide adsorption due to active stainless steel surfaces. Amorphous silicon coated surfaces exhibit 70% less mercury loss compared to bare 304SS surfaces.

Based on this data, analysts charged with monitoring mercury levels in coal-fired emission streams can significantly improve analytical performance by using functionalized amorphous silicon treated system components.


Martin Higgins ( is an ISA member and an engineer at Restek Corporation in Pennsylvania. David Smith ( is senior scientist at Restek. Gary Barone ( is business development manager at Restek. Ted Neeme works at Spectra Gases in New Jersey. See full complement of data for this study at


Lost strength, lost sample, lost data

Results from a comparison in which a gas containing 17ppbv of hydrogen sulfide was stored for seven days in untreated or in amorphous silicon treated stainless steel sample cylinders:

The response ratio for hydrogen sulfide, relative to a stable reference material, dimethyl sulfide (DMS), is steady at approximately 1:1 for at least seven days in amorphous silicon treated cylinders.

The response ratio appears in this case to look at loss or change over the life of the experiment. The data show an amorphous silicon treated system will reliably store ppb levels of the active sulfur-containing compound during transport from the sampling site to the analytical laboratory.

In contrast, hy-drogen sulfide de-graded rapidly in the untreated cylinder and was lost totally within 24 hours.

Sulfur compounds in treated stainless steel containers.