Air Pollution Control Equipment Services, Emissions

Monitoring Madness

Issue 1 and Volume 110.

High quality continuous emission monitoring capability can be as essential as high quality emission control equipment. Future mercury monitoring and control requirements add to the justification for better CEMS.

By Steve Blankinship, Associate Editor

The U.S. Environmental Protection Agency requires utilities to record and report emissions of nitrogen oxide (NOx), sulfur dioxide (SO2), carbon dioxide (CO2), carbon monoxide (CO) and oxygen (O2) on a continuous basis. Continuous emissions monitoring systems (CEMS) allow them to do it.

CEMS that were primitive by today’s standards started becoming standard equipment in coal-fired power plants in the 1970s following passage of the Clean Air Act. In the early 1990s another 1,500 CEMS were placed into service to comply with amendments to the Act to control acid rain. Now, a decade and a half later, the U.S. coal-fired fleet is once again facing the prospect of replacing or upgrading CEM systems.

Major components of existing systems are wearing out, becoming outdated, or both. New reporting requirements promulgated by EPA and local regulators – coupled with opportunities related to the cap-and-trade approach to emissions control – are also incentives to acquire better CEMS. And looming mercury (Hg) control requirements mean utilities will not only have to reduce mercury emissions, but measure, record and quantify those reductions using the best technology available.

“If you have a good SO2 system installed, you can reduce your lime cost on an FGD,” says Rich Hovan, environmental products marketing manager for Forney Corporation, which has about 500 CEM systems installed at power plants in the United States.

“If you have a good NOx monitoring system on an SCR, you can control your ammonia injection and control temperatures better, which will extend the life of your SCR. That’s important,” says Hovan, noting that many state regulations have put an end to the ‘old NOx season’ (starting the SCR in April and shutting it down in October). Beginning this year, several states will require utilities to operate their SCRs 24/7, 12 months a year. “That can cut the lives of SCRs in half,” he says. “So tuning is now more important. It makes sense to install the best CEMS. Their cost relative to other plant components is very little. And the replacements today – especially the analyzers – are of much better quality than the ones installed 10 or 15 years ago.”

How much better? Some of the systems placed into service when SO2 limits were 500 ppm can’t reliably and accurately measure current SO2 emission limits, which can be as low as 50 ppm. NOx detection limits are falling as well. “Ten years ago you were looking at 30 to 100 ppm NOx levels,” says Dave Vojtko, national sales manager for Horiba Instruments. “Now people are concerned with sub-10 ppm levels – more like 2 to 3 ppm for NOx. New analyzers are now fabricated with much more inert materials with increased detection capabilities and employ surface mount electronics.”


This CEMplicity continuous monitoring system from Forney is a stand-alone CEMS that measures SOx, NOx, CO and CO2. The system is designed to be located near the probe in order to minimize sample line length and allow for short response time. Photo courtesy of Forney Corp.
Click here to enlarge image

Frank Duckett, project manager for Thermo Electron, one of the world’s largest suppliers of continuous emission monitoring analyzers, says that while most companies are not presently installing new systems per se, they are replacing old analyzers at a fairy robust rate. “The analyzers installed in the early 1990s are 15 years old and that’s about the lifetime for an analyzer,” he says. “We’ve reached a period where technology advancements in analyzer sensitivity combined with lack of spare parts for old ones are driving a need for replacements. Operators can’t maintain what they have so we see a big push to get newer instrumentation.”

Maintenance A Big Part of the Cost

Maintenance represents a significant amount of the expense associated with owning and operating a monitoring system. One of the most time-consuming tasks involved in maintaining a CEMS is reviewing and recording daily CEMS checks. An estimated 50 to 70 percent of manpower requirements for maintaining a CEMS are consumed recording daily system readings such as dilution air pressure, sample flow rates and analyzer lamp voltages. In addition, pumps must be repaired and filters replaced. Probes must be maintained and absorbent chemicals replenished. Calibration checks must be performed daily, which requires the use of cylinders containing pure concentrations of whatever gases an analyzer must measure. Depending on the gases being measured, cylinders cost from $150 to as much as $3,000 for mercury calibration gas, and last about three months.

Many plants employ a maintenance staff dedicated to CEMS because they can’t afford the downtime if its CEMS is not working. Regulations forbid plant operation without fully functional continuous monitoring equipment. “It’s essential to keep them in top shape so you can keep the plant running,” says Mike Corvese, also a project manager at Thermo Electron. “Furthermore, the cap-and-trade programs make monitoring a bit more critical because it can determine how many credits and allowances a plant produces for its owners. That can equate to many thousands of dollars.”

Mercury: Tricky and Sticky

The likely requirement for power plants to someday capture CO2 does not affect CEMS because carbon dioxide is already measured. Mercury, on the other hand, is an entirely different animal, requiring completely separate, unique and far more expensive CEMS.

Hg is much harder to measure accurately than SO2, NOx or other traditional gases. Mercury can only be measured in its elemental form and not when it has been oxidized, which is unfortunately the form it tends to take through the combustion of bituminous coals. Therefore, all monitoring systems must convert oxidized mercury to elemental form for it to be measured. One way to convert mercury back to its elemental form is to heat it, then quickly quench it before it reverts to oxidized mercury. Keeping it oxidized is difficult since, like other emissions, mercury is extracted from some point beyond the boiler and transported through sample lines, often called umbilicals, to analyzers located in an enclosure at the base of the stack. Oxidized Hg is sticky and can adhere to the sides of the sample line.

The two most prominent mercury measurement methods are cold-vapor atomic absorptive spectrometer (CVAAS) and atomic fluorescence (AFS). Atomic absorption measures elemental mercury by shining a light through it and measuring how much of a particular frequency of light is dimmed as the Hg absorbs it. Fluorescence illuminates the mercury vapor and measures how it glows in the test chamber. Both fluorescence and absorption require all oxidized mercury be converted to elemental mercury so it can be seen by the detector. Flue gas contains a number of “bad actors” including SO2 and HC1 that make it impossible to measure mercury. Hg reacts with HCl to form HgCl2, which is not seen by the detector, and SO2 absorbs at the same wavelength interfering with the measurement. Gold amalgam traps may be used to separate the Hg from these effects but must be protected from the acid gases.

The atomic fluorescence technique is inherently more sensitive than atomic absorption for mercury. Both Tekran and Thermo dilute the sample to mitigate the problems of acid gases. Some fluorescent technologies use gold amalgam traps, then carry the samples to the analyzer in argon – an inert gas. Thermo does not use gold amalgam, injecting the sample directly into the analyzer while still in the inert environment.

“The move toward dilution has helped the technology a lot,” says Chuck Dene, air emissions monitoring and control manager for EPRI. He notes how in the 1990s, diluting samples rather than trying to transport the fully extracted gas with full concentrations of SO2, water vapor and chlorides greatly improved the reliability of conventional CEMS. He explains that the reason the industry didn’t adopt dilution for mercury monitoring at first was the concern that mercury levels were so low that diluted samples couldn’t be measured. “But atomic fluorescence, which is several orders of magnitude more sensitive than atomic absorption, lets you do it,” he says. He adds that with atomic absorption, it is necessary to use gold amalgam for pre-concentration. Dene notes that atomic absorption works best when acid gas removal and conversion to elemental mercury take place as close to the stack as possible.

Conventional CEMS sample lines are heated to about 200 F, but mercury sample lines are heated to about 350 F. That minimizes oxidation and also helps the samples flow because mercury mixed with chloride becomes sticky and clings to the inside walls of the lines. Maintaining mercury sample line temperatures at significantly higher temperatures than those needed for conventional CEMS adds dramatically to the power consumption of mercury CEMS.

Mercury probes are also different than those used for other power plant emissions, due in part to mercury’s volatility and ability to be absorbed by other components. As a result, mercury probes require special conditioning and treatment. Mercury probes are typically designed to remove moisture from the mercury sample. “Most mercury probes use the inertial filter probe design, which minimizes the particulate build-up that mercury will react with,” says Dene.

A consensus within the industry is that the big year for mercury CEMS will be 2007. “There aren’t too many commitments yet for mercury CEMS, because the compliance date is still several years away and because there are still too many things that can change before mercury CEMS are needed,” says Vojtko, whose company installed the first mercury CEMS to operate continuously at a coal-fired power plant in the United States at PSEG’s Hudson Generating Station in Jersey City, New Jersey. Horiba’s DM-6 system has also been demonstrated in short-term evaluations at Kansas City Power & Light’s Hawthorn Plant and EPA test sites. “I would say none of the systems out there are quite in their final state,” says Vojtko. “There are still lots of tweaks and modifications that must be made.”

EPRI has been working on a manual approach to mercury measurement, whereby mercury is absorbed on a carbon trap that is then taken to a laboratory where it is measured using atomic absorption or fluorescence. The QuickSEM method (quicksilver emissions monitoring system) is based on modifying a technique for capturing both elemental and oxidized mercury. The technique was simplified by eliminating the portion that captures oxidized mercury and trapping both oxidized and elemental Hg on a single carbon trap. The method is currently allowed for mercury measurement under Appendix K of the Clean Air Act, and while it is labor intensive and does not provide real-time data, it represents a far less capital intensive approach to mercury monitoring.

“We envision sampling over a week and you would get an integrated average of what your plant’s mercury emissions have been,” says Dene. In accepting the method and adapting it to Appendix K, EPA requires dual train sampling, and a spiked third section. The spiked third section is intended to demonstrate that the analysis records all the mercury.

All told, the capital cost for non-mercury CEMS (often referred to as “acid rain” CEMS), including data acquisition, analyzers, conditioning, probes, and sample lines, will come to at least $100,000 per unit. Mercury CEMS cost much more. Analyzers alone can cost four times what conventional CEMS analyzers cost. In addition, special software programming needed to keep abreast of specific situations must also be considered. Under service agreements, data companies typically charge about $200,000 per year to keep systems current.

Also, a major change in how CEMS data is reported and formatted is coming over the next few years.

The move will take data collection out of the mainframe computer realm and put it into formats more easily assimilated and processed by modern computers by conversion to XML format – a Web-based and database-friendly format. That will require new software for CEMS owners and operators.

“This requires software providers to produce new CEMS programs and also opens the door for EPA to very easily request additional data, maybe even source by source,” says Mark Shell, who is vice president of ESC, whose CEMS software is presently operating in about 500 U.S. power plants. “For example, they could have a source that is of specific concern and might require that source to report something a little bit different than another source. That was almost impossible before because the old reporting format was very fixed.

Beta testing of the new CEMS EPA XML format starts in 2007, and 2008 will be a transition year. The new reporting protocol will become required in 2009.