Poison Gas Concerns

CORRALES COMMENT VOL.XXVIII, No. 3 March 21, 2009

Federal Toxics Agency Reviews Poison Gas Concerns Over Intel’s Waste Chemical Emissions

By Jeff Radford

Fourth in a series

The deadly chemical warfare gas phosgene is not being released into the air as a byproduct of Intel’s microchip manufacturing, say federal investigators for the Agency for Toxic Substances and Disease Registry (ATSDR).

Their finding is based on Intel’s assertion that the waste acid gases that could form phosgene are isolated in the factories’ exhaust streams.

Even so, Intel has insisted on being allowed to discharge to the open air over Corrales 5.3 tons of phosgene as part of its State-issued air pollution permit. That is, Intel officials say they neither use nor release phosgene in any manufacturing process, but in case it was somehow formed and released, Intel would still comply with its permit as long as the amount was under 5.3 tons in a year’s time.

But retired Los Alamos National Laboratories chemist Fred Marsh has contended for years that Intel does release phosgene, despite its denials. He says he has seen phosgene’s chemical signature in the spectrographs produced at a Corrales home downwind from Intel…and so have analysts studying results of other Fourier transform infrared spectrometers (FTIRs) positioned around Intel.

“Although Intel has always denied using phosgene, their permit addresses what is released, not what is used,” Marsh said after reviewing the community health consultation draft report issued last month.

“And the ATSDR report confirms that phosgene is readily formed in the Intel exhaust from carbon tetrachloride that Intel has used and released in multi-ton quantities, as we have been saying for many years.

“Because Intel claims that they neither use nor release phosgene, I have repeatedly asked Intel representatives to delete the 5.3 tons of phosgene their permit allows them to release,” Marsh noted. “But they have always insisted on keeping their ‘right’ to release a quantity of phosgene that could kill tens of thousands of people.”

A leader of Corrales Residents for Clean Air and Water (CRCAW) until he moved out of state a few years ago, Marsh was an active participant in the earlier Corrales Air Toxics Study funded by the U.S. Environmental Protection Agency (EPA) and administered by the Air Quality Bureau of the N.M. Environment Department (NMED).

His expertise was key to selection of the FTIR purchased with community donations in 2003. The device detects and displays chemicals that modify the spectrum of a reflected infrared light beam.

Two other FTIRs were operating at the same time when air monitoring was conducted for the 2002-04 Corrales Air Toxic Study. One was on site under contract to the Air Quality Bureau; another was operated by an Intel consultant.

All three detected a chemical identified as phosgene, March recalled.

Although the voluminous data collected for the Air Toxics Study comprised much of the material ATSDR studied for its four-year health consultation, the agency discounted those detections of phosgene, apparently accepting Intel’s interpretation that the FTIR readings were “false positive.”

“ATSDR disregards the FTIR measurements made by [Intel consultant] TRC on Intel property that conclusively showed many toxic chemicals at orders of magnitude above safe levels,” March contends. “ATSDR also refused to look into the dismissal of phosgene and other toxic chemicals as ‘false positives’ even though they were found by all three FTIRs during the same period.”

This is how ATSDR addressed the phosgene issue verbatim in the recently released community health consultation. “Community Concern – Phosgene and False Positives. Four compounds detected by the open path FTIR, including nitric acid and phosgene, were determined to be ‘false positives’ by the N.M. Environment Department, yet phosgene was reported by Intel-New Mexico.

“ATSDR Response: Phosgene can be formed when chlorinated hydrocarbon compounds (e.g. chloroform) are exposed to high temperatures, such as what occurs in thermal oxidizers. Intel-New Mexico reports, however, that its exhaust ventilation system is designed to prevent corrosive chlorine-containing gases from being vented to the thermal oxidizers.

“Intel-New Mexico and NMED open path FTIR used different methods for analyzing their data. ATSDR did not attempt to re-evaluate false positive determinations made by NMED’s consultant-spectroscopists. ATSDR is recommending that Intel-New Mexico, in partnership with New Mexico, educate the public on the process controls and safety measures that prevent the formation of phosgene during Intel-New Mexico plant operations.”

That phrasing suggests that ATSDR accepts Intel’s assurance that phosgene cannot be produced. The ATSDR team does not recommend that further testing for phosgene be conducted, but rather that Intel officials should persuade nearby residents that the lethal gas can’t physically be released.

ATSDR tem member Debra Gable, an environmental health scientist, explained their position this way: “We’re saying if that information [that Intel is not releasing phosgene] is available, then it should be provided to the community.”

But Marsh says the ATSDR response on phosgene release apparently misunderstands how the chemical would be formed and routed to the air. “The carbon tetrachloride also reacts with water vapor at slightly elevated temperatures to form phosgene. And these are the conditions in the scrubbers through which carbon tetrachloride is released.

“I made a major fuss about Intel actually measuring 1.4 tons of carbon tetrachloride being released from just one of their many scrubbers in the fourth quarter of 2003. Yet they reported zero release, based on their phony emission factor.

“When I complained to Robert Samaniego [then Intel’s air permit compliance officer] about this emission factor obviously being too low, he agreed, but told me only Intel could initiate emission factor changes.”

Therefore, Marsh pointed out, ATSDR’s dismissal of possible phosgene releases “totally ignores the fact that there is an obvious and plausible mechanism by which phosgene can be formed in Intel scrubbers, and it ignores the fact that three different FTIRs using three different operators and three different software packages all found phosgene near Intel during the same period.”

“Moreover, phosgene was found hundreds of times by TRC, Intel’s own contractor, with the highest frequencies and levels found at Intel’s eastern boundary adjacent to Corrales residential areas.”

When pressed by Corrales Comment whether ATSDR itself had been convinced that phosgene is not being released, team leader Peter Kowalski replied, “They said the waste gas streams are kept separated; essentially keeping the chlorinated compounds out of the gas stream that’s designated for volatile organic compounds… non-chlorinated compounds. That’s what we based our opinion on.”

And as to the acceptability of declaring readings to be false positives, Kowalski said “It’s not our role to issue a second opinion on the appropriateness of false positives. But we do feel like the FTIR methodology has a kind of science and an art to it.”

At any rate, he said, “We would want an up-front methodology on how to deal with possible ‘false positives.’”

Corrales Comment pressed Kowalski for a definitive statement on whether his team is convinced phosgene can’t be released. He replied, “Based on information provided from Intel, yes, we believe there is no possibility that phosgene is being released.

“But we will look further into whether hexaflouroethane is released.”

Although Marsh had identified the compound hexaflouroethane as a likely cause of Corrales residents’ illnesses back in 2004, ATSDR’s study did not examine that in the draft report.

The chemical, often described as a “semiconductor gas,” has been detected and measured in the air at times when villagers living near Intel have reported industrial odors and sickness, Marsh pointed out.

Hexaflouroethane is a chemical precursor to hydrogen fluoride which Intel admits using and releasing. It is much heavier than air and therefore might be expected to sink to ground level after leaving the factories’ stacks.

It was, in fact, repeatedly detected at ground level by FTIRs around Intel during the Corrales Air Toxics Study 2002-04. It was found more often than any other industrial chemical, 36 times, by the community-purchased FTIR during the period September 13-29, 2003, Marsh recalled.

The chemist said health effects hexaflouroethane produces on humans are strikingly similar to complaints by Corrales residents over the past decade and a half: respiratory problems, headaches, nausea, throat, skin and eye irritations, coughing, dizziness, fatigue and loss of consciousness.

It is also odorless and invisible.

When he found hexaflouroethane’s signature in the spectrometer’s displays, Marsh thought he might have solved a big part of the mystery. He noted that the chemical reacts with water vapor (such as within the lungs or in the airborne water vapor pouring from Intel’s cooling towers and acid gas scrubbers) to form the toxic hydrogen fluoride and carbonyl fluoride.

“Because carbonyl fluoride is so similar to phosgene (carbonyl chloride) in structure and toxicity, it is also known as flourophosgene,” Marsh pointed out.

The retired chemist made that point to Intel’s corporate-level physician, Don Fisher, in November 2003. “We found tetraflouromethane and hexafluoroethane more often than any other compounds, and these were always present when people got sick,” he wrote. “Because the New Jersey Department of Health Hazardous Facts Sheets included a list of symptoms for hexaflouroethane that were nearly identical to those reported by some of Intel’s nearest neighbors, we felt we had identified the culprit.

“I stand behind my statement that hexaflouroethane can react with water to form hydrogen fluoride and carbonyl fluoride (also known as flourophosgene). The fact that carbonyl fluoride was found by neither FTIR [operated by the Air Quality Bureau and by Intel] is not surprising, as neither instrument has this compound in its reference library.”

In a presentation Marsh made to the NMED-assembled task force for the Corrales Air Toxics Study, he noted that “Dr. Don Fisher, a physician on Intel’s payroll, claims that carbonyl fluoride cannot be formed or exist long enough to be measured. Yet as a scientist, I am swayed by experimental evidence to the contrary. Intel’s own third quarter report for 2003 includes TRC stack testing data in which carbonyl fluoride was measured on Scrubber 8s.4.2ab at 2,280 parts per billion, and again later at the same scrubber at more than 1,600 parts per billion.

“This is the same Central Utility Building scrubber that is closest to Corrales residential area, whose malfunctions (according to Intel whistle-blowers) have been the source of many of the toxic pollutants that reach Corrales.”

Marsh’s paper for the task force then referenced a published article, “Using a Catalytic Technique to Abate PFC Emissions in a 300-mm Etch Tool,” which he said “provides additional proof that carbonyl fluoride is formed in semiconductor manufacturing processes.

“In fact, that article discusses how the toxic carbonyl fluoride byproduct of the semiconductor manufacturing process can be removed by a pre-treatment scrubber.”

Marsh took Intel to task for denying it releases such toxins. “Intel continues to deny that it is the source of the phosgene that was found hundreds of times by its own contractor on Intel property and in nearby neighborhoods. Yet there can be no question that Intel releases flourophosgene, which is as toxic as phosgene or perhaps even worse, as its production is accompanied by hydrogen fluoride.

“Moreover, published symptoms of exposure to flourophosgene are identical to those reported in the [CRCAW-distributed] Corrales Health Survey by hundreds of residents who live near Intel,” Marsh reported in his 2004 presentation.

“So whether it’s phosgene or flourophosgene, there can be no doubt that Intel is the source of the toxic compounds” which he thinks have caused Corrales residents’ illnesses.

After reviewing the ATSDR draft report last month, Marsh reflected on the situation in which Corrales residents find themselves. “A terrible mistake was made when Intel was allowed to build one of the world’s largest semiconductor manufacturing plants within several hundred yards of existing residential areas.

“That mistake was compounded when NMED was given the political mission of protecting Intel as its client, even at the expense of public health and the environment. NMED’s approval of Intel’s ‘minor source’ permit, and its automatic granting of any and all permit changes requested by Intel, led us to seek ATSDR’s help,” Marsh observed.

“Although we’re pleased with many of ATSDR’s initial findings, we hope our input to their draft report will help them to identify Intel as a source of our pollution problem, and then persuade NMED and Intel to work with us to restore the clean air we enjoyed before Intel’s arrival.”