By Stuart Diamond, Special To the New York Times
About 12:40 A.M. on Dec. 3, a worker at the Union Carbide pesticide plant here went to investigate a growing leak of methyl isocyanate. As he stood on a concrete slab above three large, partly buried storage tanks holding the chemical, the slab suddenly began to shake beneath him. ''There was a tremendous sound, a messy boiling sound, underneath the slab, like a caldron,'' the worker, Suman Dey, later recalled. ''The whole slab was vibrating.'' He said he started to run away, heard a loud noise behind him, turned to look and saw 60 feet of concrete at least 6 inches thick crack. ''The heat was like a blast furnace,'' he said. ''I couldn't get within six feet of it.'' Tungsten Carbide Strip
Gas Heads for Thousands He then heard a loud hissing sound, he said, and saw gas shoot out of a tall stack connected to the tank and form a white cloud that drifted over the plant and toward nearby neighborhoods where thousands of residents slept. ''I panicked,'' he said. ''Everybody panicked.'' Mr. Dey, a tall, soft-spoken man who has worked at the plant for five years, was among about a dozen workers and supervisors interviewed who were at the plant in the hours before and after a leak of methyl isocyanate, a chemical more toxic than cyanide that was used at the plant to make pesticide. The leak went out of control and caused the worst industrial accident in history, killing at least 2,000 people and injuring 200,000 in this central Indian city.
Toxicity Was Underestimated Nearly all the workers interviewed were making their first public statements since the disaster. They gave their accounts in Hindi through an interpreter; some declined to be identified. Virtually all the workers said they knew methyl isocyanate was dangerous and some said they knew it could be fatal, but the dozen workers said they underestimated its toxicity. No one said he knew it could kill many people quickly. The Union Carbide Corporation technical manual for methyl isocyanate is pointed on the hazards of the chemical and states that it ''may cause fatal pulmonary edema,'' which is an accumulation of fluid in the lungs. But although the manual was distributed to managers that handle methyl isocyanate at the Bhopal plant and was seen by some of the workers there, most of the factory's employees had not read or understood it, according to former technical officials at the factory.
Workers Expected Slow Day The hours before the accident, the workers said, unfolded this way: About 100 workers reported for duty on the eight-hour shift beginning at 2:45 P.M. on Sunday, Dec. 2. The production plant to make methyl isocyanate had been shut down since Oct. 22, the workers recalled, and they were not particularly busy. In addition, they said, most major maintenance was now being done on the day shift during the week. Outside the walled factory, it was a typical Sunday afternoon in the teeming old city of Bhopal, local residents later recalled. The open-air market was bustling; some of the vendors, as usual, were calling out their items for sale from carts and small booths on narrow streets. Across the street from the factory, children played in the dirt outside the slum huts crammed together in a community called Jai Prakash Nagar. Inside the factory, workers said, the pesticide Sevin was being produced. The Sevin plant, after having been shut down for some time, had been started up again about a week before but was still running at far below normal capacity, the workers said. To make the pesticide, carbon tetrachloride is mixed with methyl isocyanate and alpha-naphthol, a coffee-colored powder that smells like mothballs. The methyl isocyanate, or MIC, was stored in the three partly buried tanks, each with a 15,000-gallon capacity. Nitrogen Leaked Out Workers had not been able to use the methyl isocyanate in one of the tanks, No. 610, to make the pesticide for more than a week, they said, because they could not get the chemical out of the tank. Every time they tried to push it out and into the Sevin plant by pumping in nitrogen, they said, the nitrogen leaked out somewhere; they did not know where. The methyl isocyanate supervisor on duty during the second shift of Dec. 2, the workers said, was Gori Shankar, who had arrived two months before from a Calcutta battery factory owned by Union Carbide India Ltd., which also owns the Bhopal plant. The company is owned 50.9 percent by the Union Carbide Corporation of Danbury, Conn. About 9:15 P.M. Mr. Shankar telephoned one of the methyl isocyanate operators, Rahaman Khan, who was in the plant's canteen having tea, Mr. Khan recalled. Mr. Shankar could not be reached for comment on the workers' accounts, and plant officials declined to put The New York Times in touch with specific employees. Mr. Khan said Mr. Shankar asked him to come to the MIC area of the plant and clean a pipe. The pipe, about 25 feet long and 8 feet off the ground, led from a device that filtered crude methyl isocyanate before it went into the storage tanks, Mr. Khan said. Inside the pipe was a valve that had been closed.
Valve Had Not Been Sealed Mr. Khan and several other workers said Mr. Shankar told him to open a nozzle on the pipe and put a water hose in to clean the inside. ''He came along with me and stayed there while I did the procedure,'' Mr. Khan said. ''I connected a hose to the pipe to be cleaned and opened a drain.'' It was about 9:30 P.M., he recalled. Mr. Khan said he noticed that the closed valve had not been sealed with a slip blind, a metal disc that is inserted into pipes to make sure that water does not leak through the valve. Valves were notorious for leaking at Bhopal, the workers said. They knew, they said, that water reacted violently with methyl isocyanate. Page 67 of the MIC operating manual for Bhopal says: ''Isolate the equipment positively by inserting suitable blinds. Isolation by valve or valves is not to be relied upon.'' Mr. Khan said he and Mr. Shankar left the area while the pipe was being cleaned. Unattended, water flowed into the pipe, out pipe drains and onto the floor, where it entered a floor drain, Mr. Khan and other workers recalled. The water was to continue to flow, the workers said, for about three hours. According to Mr. Khan, it was the only pipe in the methyl isocyanate unit being washed during the second shift that day. 'I Didn't Check to See' ''I knew that valves leaked,'' said Mr. Khan, who has a high school education and less training than the amount originally established for methyl isocyanate plant operators. ''I didn't check to see if that one was leaking,'' he said. ''It was not my job.'' Many workers said they believed water from the unsealed valve was the most likely trigger for the accident that came hours later. Shakil Qureshi, the methyl isocyanate supervisor on duty at the time of the accident, said later: ''If water caused this accident, it is the fault of the management of Union Carbide. In fact, it is all our faults.'' He said it would have been extremely difficult to check whether the valve was leaking; the plant, he said, did not have the proper instruments. He said he found out later that there was no indication in notes on the daily maintenance log to insert a slip blind, although there was a note to wash the pipe. ''But the daily notes are always vague,'' he said. Pressure of Gas Rises But Causes No Alarm About 10:30 P.M., the workers on duty prepared for the change in shifts that was to occur about 15 minutes later, some of them recalled. Among other things, they said, they logged the pressure indicated on the gauge in the control room for MIC tank No. 610. It was two pounds per square inch - normal, they said. Operators said they did not record the temperature of the tank. ''For a very long time we have not watched the temperature,'' one worker said. ''There was no column to record it in the log books.'' Operators said the temperature of the methyl isocyanate was usually nearly 20 degrees centagrade, or 68 degrees Fahrenheit, although the plant's operating manual specifies that it be kept below 5 degees centigrade, or 41 derees Fahrenheit. Sometimes in the summer, operators said, the methyl isocyanat storag tank temperature indicator went off off the scale, which was 25 degrees centigrade at the maximum, or 77 degrees Fahrenheit. About 11 P.M., Mr. Dey later recalled, he noticed that the pressure gauge for Tank 610 read 10 pounds per square inch - five times what it had been half an hour earlier. Mr. Dey, who was in the control room and was the senior operator on duty, said he had thought nothing of it; it was still a relatively normal pressure. Half an hour later, Mr. Qureshi, whose office is across the hall from the control room, had the same reaction, he later recalled - he thought one of the two readings was faulty. ''Instruments often didn't work,'' he said. ''They got corroded. Crystals would form on them.'' About 11:30 P.M., workers in the methyl isocyanate area said, they realized that there was a methyl isocyanate leak somewhere: Their eyes began to tear. Detecting leaks by the effect of the gas on the eyes was standard procedure at the plant, and at least one leak a month was detected this way at the plant, they said. ''We were human leak detectors,'' Mr. Dey said. The practice violated procedures laid out in the technical manual. The workers began to look for the source of the leak they recalled.
V.N. Singh, one of the workers, said he and the others walded around the MIC structure, which looks like a small refinery, and spotted a drip of liquid about 50 feet off the ground and some yellowish-white gas accompanying the drip. ''It was a small but continuous drip,'' he said. While the other workers kept looking at it - they later said they thought they had found the source but were not certain - Mr. Singh said he went by himself to inform Mr. Qureshi.
We are having trouble retrieving the article content.
Please enable JavaScript in your browser settings.
Thank you for your patience while we verify access. If you are in Reader mode please exit and log into your Times account, or subscribe for all of The Times.
Thank you for your patience while we verify access.
Already a subscriber? Log in.
Tungsten Mechanical Seal Want all of The Times? Subscribe.