“SECRETARY OF LABOR,Complainant,v.FMC CORPORATION,Respondent.OSHRC Docket No. 83-0488 and 83-0489_DECISION_Before: BUCKLEY, Chairman; RADER and WALL, Commissioners.This case is before the Occupational Safety and Health Review Commissionunder 29 U.S.C. ? 661(j), section 12(j) of the Occupational Safety andHealth Act of 1970, 29 U.S.C. ?? 651-678 (\”the Act\”). The Commission isan adjudicatory agency, independent of the Department of Labor and theOccupational Safety and Health Administration. It was established toresolve disputes arising out of enforcement actions brought by theSecretary of Labor under the Act and has no regulatory functions. _See_section 10(c) of the Act, 29 U.S.C. ? 659(c).IFMC’s plant in South Charleston, West Virginia, produces chlorine byprocessing brine through diaphragm-type electrolytic cells. Nitrogentrichloride (NCl3), an explosive compound, is a by-product of theelectrolytic process and is contained in the waste material, which isknown as \”gunk\”. These cases concern the procedure by which FMCneutralizes the potentially hazardous waste material and then removes itfrom the system. First, the gunk, which accumulates at the bottom oftwo scrubbers, is transferred through pipes to three gunk tanks, alsoknown as reboilers. During this process, chloroform and carbontetrachloride are added to dilute the concentration of NCl3 Once a gunktank is filled, it is heated by either steam, hot water, or both so thatany chlorine present in the tank will be vaporized. This is called\”gasification.\” When this process is completed, the residue is pumpedto a reactor, where it is neutralized with hydrochloric acid and thentransferred to drums for disposal. There is a diked pit in the area ofthe gunk tank that contains a pump. The purpose of that pump is toprevent accumulations of the hot water that is used in gasifying the tanks.On March 1, 1983, an employee of FMC, Gary Gessel, was killed in anexplosion of NCl3. The explosion occurred in the transfer line from oneof the gunk tanks to the reactor, and secondarily in the gunk tankitself, due to an exothermic reaction of NCl3. Apparently, the explosionoccurred because Gessel failed to follow his supervisor’s instructionsto turn off the water that was heating the gunk tank during thegasification process. While the record does not establish conclusivelywhat caused the explosion, there is some evidence that Gessel placed thehot water hose on the pipe, thereby causing the hot water to contact thepipe and overheat the NCl3, starting the reaction that resulted in theexplosion.As a result of his investigation of the explosion, the Secretary issuedin Docket No. 83-0488[[1]] a citation alleging two violations of section5(a)(1), 29 U.S.C. ? 654(a)(1) of the Act.[[2]] The items focus on twoof the steps in the procedure described above: the gasification of thegunk tanks to drive off the chlorine; and the addition of chloroform todilute the concentration of NCl3. The citation alleges thatdeficiencies in those procedures exposed employees to the hazard ofexplosion.[[3]] A penalty of $720 was proposed for each item.Administrative Law Judge Paul Brady affirmed the citation and assessed asingle $700 penalty. Judge Brady did not differentiate between the twoitems, but rather treated the citation as an integrated whole. Havingapparently viewed the problem as stemming from the heating required byFMC’s gasification process, the judge defined the hazard as the dangerof explosion when heat is applied to gunk tanks and transfer linescontaining NCl3.The judge found that both FMC and the chlor-alkaline industry in generalwere aware of hazards involving concentrations of NCl3 in gunkoperations. More specifically, the judge found that the evidenceestablished that FMC’s methods of gasifying gunk tanks constituted ahazard. The judge noted that the employees were not aware of the hazardassociated with sudden heat buildup and that there was no ruleforbidding the application of steam or hot water to the transfer line. Judge Brady also noted that the transfer lines were placed in proximityto the water hose used to heat the gunk tank. With such a setup, theJudge concluded, it was incumbent on FMC to institute a safe andeffective program governing its work practices. The judge found thatFMC had failed to do so and therefore found that FMC had violatedsection 5(a)(1). FMC petitioned for review of the judge’s decision.For the reasons that follow, we find that the Judge erred in affirmingthe citation. With respect to each step in the process that theSecretary has challenged, we conclude that the Secretary either failedto prove that FMC’s safety measures were inadequate or failed to provethe likely utility of measures intended to reduce the hazard ofexplosion. Accordingly, we vacate the citation.IITo prove that an employer violated section 5(a)(1), the Act’s generalduty clause, the Secretary must prove that the cited employer failed tofree the workplace of a hazard that (1) was recognized by the citedemployer or its industry, (2) that was causing or likely to cause deathor serious physical harm, and (3) that could have been materiallyreduced or eliminated by feasible and useful means of abatement. _Pelron Corporation_, 12 BNA OSHC 1833, 1986 CCH OSHD ? 27,605 (No.82-388, 1986).Judge Brady defined the hazard as the danger of explosion when heat isapplied to gunk tanks and transfer lines containing NCl3. The evidencein this case clearly establishes that some danger of explosion alwaysexists when heating NCl3-FMC’s gasification process necessitates theheating of gunk containing NCl3. The Secretary has neither alleged noradduced any evidence to suggest that the gunk neutralization process canbe accomplished without heating the tanks to drive off excess chlorine. Therefore, if the hazard were to be defined as the danger of explosionwhen heat is applied to the gunk tanks, it would be impossible for FMCto rid its workplace of the hazard. The intent of section 5(a)(1) is,however, to reduce preventable hazards. _Pelron Corporation_, 12 BNAOSHC at 1835, 1986 CCH OSHD at p. 35,871. To further that intent,hazards must be defined in a way that apprises the employer of itsobligations and identifies conditions or practices over which theemployer can reasonably be expected to exercise control. _Pelron_, 12BNA OSHC at 1835, 1986 CCH OSHD at p. 35,872. Accordingly, we definethe hazards in this case as those practices, procedures or conditionsthat increase the likelihood of an explosion.Both FMC and the chlor-alkaline industry recognized that the likelihoodof an explosion is greatly increased when there are either inadequatemethods of temperature control during gasification, or unknownconcentrations of NCl3 due to improper methods of adding and monitoringthe addition of the diluent chloroform. This is not in dispute. Rather, the dispute centers on, and we must decide, whether FMC’sprocedures for preventing high temperatures and high concentrations ofNCl3 were inadequate. Also at issue is whether the Secretaryestablished that there were additional measures that would have feasiblyand materially reduced the risk of harm. _See_ _Cerro Metal ProductsDiv. Marmon Group, Inc_., 12 BNA 1821, 1823, 1986 CCH OSHD ? 27,579, p.35,829 (No. 78-5151,1986).IIIThe first item of the citation alleges that FMC’s method of monitoringand controlling temperatures was insufficient, thus increasing thedanger of an explosion due to an exothermic reaction of NCl3 in the gunktanks and transfer pipe.Before 1980, FMC heated the gunk tanks during gasification by applyinghot water to the top of the tanks with a hose. Because of a waterdisposal problem, FMC added a steam heating system in which steam wasapplied with a sploger (a pipe with holes) to the bottom of the tank. In colder weather, hot water would be applied by hose in addition to thesteam to speed gasification. When used to gasify the gunk, hot wateraccumulated in a pit. The pipes carrying the gunk from the gunk tanksto the reactor ran above this pit. To prevent the hot water from risingand contacting the pipes, FMC installed in the pit a sump pump thatstarted automatically when the water reached a pre-set level.To prevent the gunk from reaching a critical temperature, employees wereinstructed not to allow the temperatures to rise above 10- 20? C duringthe gasification process. The temperatures of the gunk tanks weredetected by a thermocouple on each of the tanks. Digital displays of thetemperatures detected by the thermocouples were located in a controlroom several feet away from the tanks. A continuous digital readout ofthe temperatures in a gunk tank could be obtained by punching up theassigned number on the buttons in tile control room.The Secretary first argues that FMC failed to adequately train itsemployees with respect to the hazards of overheating the gunk tanks. In support of this assertion, the Secretary argues that employees wereallowed to use hot water to speed gasification and were not instructedto gasify the tanks slowly. The Secretary also notes that FMC had nowritten procedures for gasifying the tanks, and new control roomoperators were trained only by other operators. Moreover, someemployees were not informed of the hazard of explosion caused byallowing temperatures to rise too high or too rapidly.We find that the Secretary failed to establish that FMC’s trainingmethods were inadequate. Although FMC did not have a formal trainingprogram, a formal training program is not necessarily required bysection 5(a)(1). _See_ _Jones & Laughlin Steel Corp_., 82 OSAHRC 34\/A2,10 BNA OSHC 1778, 1782, 1982 CCH OSHD ? 26,128, p. 32,887 (No. 76-2636,1982); _Pelron_, 12 BNA OSHC at 1838, 1986 CCH OSHD at p. 35,874. Thequestion is one of substance, not form. The testimony establishes thatall operators were aware that temperatures were not to rise above 10-20?C. Similarly, while several employees could not recall being advised ofthe danger inherent in heating the tanks too rapidly, all were awarethat the tanks should be heated gradually. Indeed there is no evidencethat the tanks were heated too rapidly. Although the record establishesthat gasifying the tanks in less than half an hour was hazardous, theevidence indicates that the shortest period of time taken to gasify thetanks was 45 minutes. We also note that the operators who testified hadbeen in their positions for periods ranging from four to thirteen yearsand either were experienced or were trained by highly experiencedemployees. Moreover, FMC operated its facility for some 25 yearswithout an accident, thereby evidencing that its program of on-the-jobtraining was successful in ensuring proper operating procedures.Although the Secretary would require a formal training program, there isno evidence that such a program would have significantly elevated thelevel of safety at the facility, especially given the experience of theoperators who gave the on-the-job training.The Secretary next argues that FMC’s temperature control technology wasinadequate. He asserts that monitoring the temperatures on the gunktanks was difficult because the thermocouples on the tanks displayedtemperatures only after a series of buttons were punched in the controlroom. According to the Secretary, FMC also should have installed ahigh-temperature alarm that would immediately warn employees when thegunk tank temperatures had reached dangerous levels.At the hearing, W. Duane Colpous, a retired chemical engineer with 35years experience in the chlorine industry, was called by the Secretaryas an expert witness. It was his opinion that the temperatures at theplant were inadequately monitored. He criticized FMC’s failure torecord temperatures on a con tinuous basis so they could be followed ona chart and opined that FMC should have used a high temperature alarm towarn employees when temperatures approached the upper limit of safeoperation. Colpous’ testimony was contradicted by FMC’s expert witness,Dr. Chester Grelecki,[[4]] a Ph.D in chemistry with considerableexperience in the chlor-alkaline industry, and a consultant in the areaof chemical process design. Grelecki found nothing inadequate aboutFMC’s temperature detection system and testified persuasively that FMC’ssystem was up to industry standards.There is no evidence to support the Secretary’s claim that punchingbuttons to obtain a continuous readout of gunk tank temperatureinterferes with the operators’ ability to monitor temperature changes. Testimony of the control room operators Eleanor Garrett and Robert Jonesestablishes that they had no difficulty monitoring the temperature ofthe gunk tanks in the control room. Similarly, despite Colpous’opinion that FMC’s temperature detection system would have been improvedby installing a high temperature alarm and continuously recording gunktemperatures, the evidence establishes that the employees successfullymonitored temperatures with the system in use. Accordingly, we findthat although the evidence establishes that the use of high temperaturealarms and the continuous recording of gunk tank temperatures might havemarginally improved the ability of the employees to detect hightemperatures, it does not establish that such improvement would havematerially reduced the hazard of undetected high temperatures in thegunk tanks. While a means of abatement need not be perfectlyprotective, it must be shown by the evidence to promise a materialreduction of the hazard. _See_ _Chevron Oil Co._, 83 OSAHRC 19\/B2, 11BNA OSHC 1329, 1334, 1963-84 CCH OSHD 26,507, p. 33,724 (No. 10799,1983). This was not shown on this record.The Secretary next focuses on alleged deficiencies in FMC’s control oftemperatures on the transfer pipe. The Secretary observes that FMC didnot have thermocouples on the transfer lines, thus making it impossibleto determine the temperatures in those lines. Furthermore, the Secretaryargues that the transfer lines should have been relocated out of the pitto an area where they would not have been exposed to the hot water usedto heat the gunk tanks. Finally, the Secretary argues that FMC shouldhave adopted a system to \”dump\” large quantities of diluents into thepipes to reduce hot spots and to \”dump\” the contents of the gunk tanksinto diluents when the tanks reached critical temperatures.The evidence is undisputed that FMC was unable to determine thetemperature of the gunk as it passed through the transfer pipes. Thedanger of hot spots developing without warning due to contact with anexternal heat source, such as hot water or steam, and an ensuingexothermic reaction resulting in an explosion was therefore a dangerinherent in FMC’s system.We find, however, that the Secretary failed to establish either thefeasibility or likely utility of an abatement method with respect tothis danger. The first method of abatement suggested by the Secretarywas to place thermocouples on the transfer lines that would be able todetect hot spots. He points to the testimony of Mr. Colpous thatthermocouples on the line would improve the ability to detect hotspots. However both Dr. Grelecki and Mr. Colpous testified that due tothe unpredictability of hot spot formation, numerous thermocouples wouldhave to be placed along the transfer pipes. Mr. Colpous could not statehow many thermocouples would be necessary or how far apart they wouldhave to be placed to be effective but opined that even one would behelpful. Dr. Grelecki, on the other hand, testified that to thoroughlymonitor the pipes, thermocouples would have to be placed about a footapart along the entire length of the pipes. Lionel Updyke, a chemicalengineer for FMC, testified that the thermocouples might have to beplaced two to three inches apart along a hundred feet of piping in orderto pick up a localized heat source. Both Dr. Grelecki and Lionel Updyketestified that the use of thermocouples to detect hot spots on thetransfer pipes was not an accepted practice within the chlor-alkalineindustry to reduce the hazard of hot spot formation within the pipes. Moreover, Dr. Grelecki testified that from the viewpoint of engineeringprinciples and industry standards, such a method of monitoring is notpractical.Having considered the testimony of both experts, we find the Secretaryfailed to establish the feasibility of placing thermocouples on thepipes. Although Mr. Colpous testified that even one thermocouple wouldbe helpful, his own testimony establishes that to detect a hot spot thethermocouple would have to be located in the immediate vicinity of thehot spot. A thermocouple might enable FMC to monitor temperatures inthe pipe as a whole, but it would be of little if any benefit indetecting hazardous hot spots.The Secretary next argues that the hazard of hot spot formation couldhave been reduced by relocating the transfer lines out of the pit wherewater, used to heat the gunk tanks, would accumulate. He notes that,after the accident, the transfer lines were relocated, thus establishingthe feasibility of the measure.[[5]]Although we agree that the Secretary established the feasibility ofrelocating the lines, we find that the evidence fails to establish thelikely utility of this abatement measure. We do not believe the recordestablishes that the lines were located in an area that endangeredemployees, a sump pump was used in the pit to prevent the accumulationof water. It activated automatically when the water reached one inch indepth. Superintendent Updyke testified that, with the pump operating,it was impossible for water to accumulate to the level of the pipes,which were more than 3-1\/2 inches above the bottom of the pit. Heexplained that while the flow rate of the hot water hose was fourgallons per minute, the pump was able to handle seven gallons perminute. Several operators testified that the pumps had worn out andwere subject to corrosion. However, FMC checked the pump three times aweek and required an inspection sheet, indicating whether the sump pumpwas in proper condition, to be completed after each inspection. Thesereports establish that during the period January 5 through February 28,1983, a problem occurred with the sump pump only once and that the pumpwas repaired by the next inspection. Both Robert Jones and James Jones,control room operators, testified that during the month prior to theaccident the pump was working well. Gunk neutralization operatorCassell stated that normally a PUMP is kept in good working condition.[[6]]Despite the efforts by FMC to keep the sump pump in good repair, FMCoperators testified that water did reach the lines. Their testimony,however, is in conclusive regarding the time when this occurred. BothRobert Jones and James Jones testified that they had seen the linescovered with water but they did not identify when this happened. Gunkneutralization operator Cassell said he had observed water over thepipes during the two months before the accident but weekly inspectionlogs for the period of January 7 through February 25, 1983, namingCassell as operator, all state that there was no water standing in thediked area during that period. In addition control room operatorEleanor Garrett, who recalled occasions when water accumulated to pipelevel, stated this occurred when FMC was using a different system ofwater disposal. She did not testify that water had ever accumulated topipe level since the institution of the system used at the time of thecitation.In any event, even if water did reach the lines, there is no evidencethat this water created or was likely to create a hazard. As previouslynoted, superintendent Updyke testified that the flow rate of the hotwater hose was four gallons per minute. He also testified thatapproximately 300 gallons of water would be required in the pit beforeit would back up to the lines. Considering these facts together withother characteristics of FMC’s operation, Dr. Grelecki opined that bythe time the water would reach the pipes its temperature would be toolow to heat the contents of the pipe to any significant temperature. Based on the evidence it would take at least one hour and fifteenminutes for enough water to accumulate to reach the lines. The waterwould, of course, be cooling during this time period. At some point intime the water would actually act as a conductor of heat away from thepipes much like the cooling system in an automobile engine. There issimply no evidence to support the view that the pipes were or could everbe \”submerged in hot water\” as the dissent postulates. Dr. Greleckialso testified that FMC’s facility design was both appropriate andconsistent with industry practice. Given this evidence, we are notconvinced that there was a problem with the water disposal system in useat the time of the inspection.The fact that FMC relocated the lines after the accident does not alterour conclusion. As we recently noted in another case, \”Employers maydecide . . . to take other precautions against injury or illness out ofan abundance of caution rather than a belief that the absence of suchprecautions would expose employees to a significant risk of harm.\” _Kastalon, Inc. & Conap, Inc_., ___ OSAHRC ___, 12 BNA OSHC 1932, 1986CCH OSHD ? 27,643 at p. 35,970 (Nos. 79-3561 and 79-5543, 1986). TheSecretary introduced no evidence to show the reason for FMC’s relocationand we decline to speculate as to why the lines were moved.Finally, the Secretary argues that FMC should have installed a dumpsystem that would have utilized large quantities of diluent or coolentsto neutralize the development of an exothermic reaction. Although theSecretary presented evidence to establish that the introduction ofdiluents into the gunk would reduce the hazard of an explosion, we findthat the record fails to establish the feasibility of such a system.The use of a \”dump system\” was recommended by the Secretary’s expertwitness Duane Colpous. Mr. Colpous testified that either carbontetrachloride could be quickly added to the pipes as a coolant when ahot spot developed, or the material in the tank could be dumped into acoolant. He stated that the latter type of dump system is used in TNTplants to reduce the danger of explosion. However, Colpous did nottestify that he had ever seen such a \”dump system\” used in a chlorineplant, or present any details that would indicate how such a systemwould work in the FMC facility. He also did not identify anysimilarities ties between a TNT plant and FMC’s facility that wouldindicate that a \”dump system\” could be used in a chlorine plant.[[7]]Dr. Grelecki, on the other hand, testified that a dump system would notbe feasible in reducing the hazard of an explosion. According to Dr.Grelecki, because the pipes would be full, there would be no way to\”dump\” quantities of diluent into the lines in a manner that wouldeffectively neutralize a hot spot. According to Dr. Grelecki, anattempt to \”dump\” diluents into a full pipe would only displace the gunkalready in the line and move the hot, spot to another location withinthe pipe. Moreover, Dr. Grelecki testified that because of the rapidityof an exothermic reaction of NCl3, once one was detected there would notbe sufficient time to \”dump\” the material in the tanks into a coolant.We find Dr. Grelecki’s testimony to be more persuasive than that of Mr.Colpous. While Mr. Colpous discussed a dump system in only generalterms, Dr. Grelecki specifically identified technical problems thatwould have rendered such a system infeasible in a chlor-alkaline plant. The Secretary introduced no evidence to show how FMC might overcomethose technical problems. We therefore conclude that the Secretaryfailed to establish the feasibility of a dump system.We find that the evidence is insufficient to establish either that FMCfailed to adequately train its employees in the hazards of overheatingthe tanks or that FMC failed to institute adequate methods oftemperature control, and vacate item 1 of the citation.VIThe second item of the citation alleges that FMC increased the hazard ofexplosion by using improper methods of adding and monitoring theaddition of chloroform to its system, resulting in unknown quantities ofNCl3 in the system.The upper limit for a safe concentration of NCl3 in the gunk tank is5%. To maintain a safe concentration FMC adds chloroform to thescrubbers and carbon tetrachloride to the gunk tanks. Besides being adiluent in the scrubbers, the chloroform acts as an antifreeze for thecarbon tetrachloride in the gunk tanks, which freezes at -23? C. Thetemperature of the chlorine in the gunk tanks is -33? C. By itself, thecarbon tetrachloride would freeze and separate, leaving a concentrationof pure NCl3. The chloroform lowers the freezing point of the carbontetrachloride, allowing it to be an effective diluent. A device calleda rotometer controls the addition of chloroform and can be adjusted tovary the chloroform flow.The evidence establishes that both FMC and the chlor-alkaline industryrecognized that the addition of insufficient quantities of chloroform tothe scrubbers would greatly increase the danger of explosion fromexcessively pure concentrations of NCl3. The Secretary does not arguethat FMC’s formal procedures and requirements for the addition ofchloroform and testing for NCl3 concentrations were inadequate. Rather,the issue is whether FMC’s failure to follow those procedures increasedthe hazard of explosion. The Secretary points to three specificinstances where FMC’s failure to follow its written procedures allegedlyincreased the hazard. First, the Secretary contends that FMC failed tosample the gunk to determine the concentration of NCl3 as often asrequired by its own _Chlorine Operating Bulletin_. Second, theSecretary contends that, during a couple of weeks before the accident,problems with the rotometer resulted in quantities of chloroform beingadded to the scrubbers that were below the amounts required by FMC’s ownoperating procedures. Third, the Secretary argues that on the daybefore the accident FMC began an experiment that resulted in the flow ofchloroform to the scrubbers being interrupted for three hours.We find that the Secretary failed to establish that FMC’s failure tofollow the sampling requirements in its own _Chlorine Operating__Bulletin_ increased the possibility of an explosion due to undetectedhigh concentrations of NCl3. The Bulletin, written in 1973, requiresthat the gunk be sampled at least three times a week to ensure that theconcentration of NCl3 does not exceed 5%. Duane Colpous testified that,in his opinion, the gunk should be sampled daily, but he agreed thatthree times a week would be adequate. Terry Fontalbert, FMC’s seniorprocess engineer, testified that the gunk was not sampled for NCl3concentrations as often as recommended by the Bulletin. However, hefurther testified that it was not necessary to sample the gunk thisoften. He noted that the Bulletin was prepared early in the history ofthe system, and speculated that the sampling requirement was includedbecause the company was still learning the dynamics of the process. Fontalbert’s testimony was corroborated by Dr. Grelecki. Dr. Greleckitestified that industry practice is to frequently monitor the NCl3levels when first establishing a gunk neutralization process. Afterenough control over the process is established and confidence in theprocess grows, monitoring can occur less frequently. At some point,Dr. Grelecki testified, monitoring of NCl3 levels can stop altogether.Our conclusion that FMC’s failure to monitor NCl3 concentrations asrequired by the _Bulletin_ was not hazardous is strengthened by evidenceof other FMC work practices. Both employee testimony and work sheetsintroduced at the hearing establish that employees were required tocheck and record chloroform levels every two hours. Moreover, LionelUpdyke testified that FMC tested their production process five days aweek to determine the amount of NCl3 being produced. Therefore, FMCknew the amount of NCl3 entering the neutralization process. Given FMC’sregular monitoring of the chloroform being added to the scrubbers andthe amount of NCl3 being introduced into the neutralization process, andFMC ‘s long experience with the process, we conclude that FMC adequatelycontrolled the concentration of NCl3 in the system. Accordingly, itsfailure to monitor NCl3 concentrations three times a week did notexacerbate the hazard of explosion.The Secretary next argues that a pluggage in the line feeding chloroforminto the scrubber that existed for a couple of weeks before theaccident, together with problems with the rotometer which regulates theflow of chloroform, resulted in inadequate amounts of chloroform beingadded to the scrubber. We find that, despite the technical problemsdescribed by the Secretary, the evidence establishes that the requiredmonitoring of the chloroform by employees ensured that adequatequantities of chloroform were being added.[[8]]Operator Robert Jones testified that, during the problem with thechloroform line and rotometer, his regular monitoring of chloroformlevels disclosed that an inadequate amount had entered the scrubber. Accordingly, he made adjustments in the rotometer until the properamount entered the system. Operator James Jones also stated that whenthere was trouble with the chloroform flow he would adjust the flow andkeep checking the level until the proper amount entered the scrubber. He also testified that whenever the rotometer or line was plugged,management would send an instrument man to correct the problem.Duane Colpous opined that, during the days before the accident,chloroform was added at one-third to one-half of the rates called for inFMC’s _Operating Bulletin_ or about five gallons in a 24-hour period.Lionel Updyke, however, noted that even if chloroform was flowing belownormal levels, the amount in the system before the accident was adequatebecause the system was operating at only 65% capacity. He also statedthat the minimum- required flow of chloroform was 1\/4 gallon per hour orsix gallons in a 24- hour period. Updyke’s conclusion was alsosupported by Dr. Grelecki, who noted that the fact that the explosiontook place in the lines and not in the tank indicates that the flow ofchloroform was sufficient.Analysis of the gunk made after the accident revealed a chloroformconcentration of 24.6%. In comparison, FMC’S _Operating_ _Bulletin_calls for a 20% concentration, while Lionel Updyke testified that 10%was sufficient. The accuracy of the post-accident analysis wasquestioned by Duane Colpous, who testified that the instability of NCl3and its ability to react with moisture might have resulted in thesamples showing a lower level of NCl3 than actually existed at the timeof the accident. Dr. Grelecki agreed that the gunk samples wereunreliable, but stated that \”very reliable\” samples from the reactorindicated an NCl3 concentration of 3-1\/2 – 4%, well-within the 5% limitfor safe operation. Thus, the NCl3 was adequately diluted.We also find that, while FMC did have occasional problems with therotometer and chloroform lines, it promptly repaired the defects. Moreover, its experienced and competent operators readily recognized andcompensated for the technical problems to ensure that the flow ofchloroform remained at safe levels. The post-accident analysis of thegunk, which provided the only concrete evidence of diluent levels at thetime of the accident, indicated more than sufficient levels ofchloroform. Those results were supported by other tests that showedthat the NCl3 was adequately diluted. We therefore find that theproblems with the rotometer and chloroform lines did not create a dangerof explosion due to overly pure concentrations of NCl3, and thatemployees were not exposed to a hazard of explosion due to unsafeconcentrations of NCl3. We therefore conclude that FMC took adequatemeasures to maintain NCl3 concentrations at a safe level and vacate item2 of the citation.Accordingly, the judge’s decision in No. 83-0488 is reversed and items 1and 2 of the citation are vacated. The direction for review as to No.83-0489 is vacated.FOR THE COMMISSIONRay H. Darling, Jr.Executive SecretaryDATED: AUG 28 1986BUCKLEY, Chairman, dissenting in part:I respectfully dissent from my colleagues’ disposition of item 1 of thecitation. In my view, the Secretary established that both FMC and thechloralkaline industry in general were aware of the hazard of locatingpipes containing NCl3 in an area susceptible to external sources ofheat. By placing its transfer pipes in an area where it was inproximity to hot water, FMC failed to free its workplace of the hazard,and therefore was in violation of section 5(a)(1) of the Act.[[9]]In my view, FMC’s expert, Dr. Chester Grelecki, established thatnitrogen trichloride in a pipe is highly susceptible to exothermicheating and explosion if exposed to external heat. This is due to theconfinement of the NCl3 in the pipe, which allows no space for the gasesgenerated during heating to expand, and to the very low autoignitiontemperature for NCl3. His testimony also confirms industry knowledge ofsuch susceptibility.The record generally substantiates that the explosion which occurred atFMC’s chlorine purification facility was initiated in a pipe carryingNCl3 from a reboiler (where a chlorinated organic mixture containingNCl3 was heated to vaporize and remove chlorine) to a reactor forneutralization of the NCl3. In the heating process, water at atemperature of 80?C was applied externally to the reboiler in thevicinity of the pipe, and the pipe was in an area where hot waterfrequently accumulated. As the majority opinion points out, theevidence indicates that hot water which was being applied externally tothe reactor flowed directly onto the transfer pipe, overheating theNCl3, and commencing an exothermal reaction within the pipe. Theevidence, however, is not clear as to precisely how the hot water camein contact with the transfer pipe. Either of two hypotheses can bemade: that the water level in the diked pit rose due to a faulty sumppump, reaching the transfer pipe which traversed the diked pit, or thatthe hot water from the hose which was being used on the reboiler wasdirected onto the transfer pipe.The record establishes the substantial likelihood that hot water couldaccumulate in the pit to a level where it came in actual contact withthe pipe. Testimony of the operators established that the sump pumpsfrequently broke down. Problems with the pumps were especially acuteduring the winter when they would freeze up. Significantly, it wasduring such cold weather when the operators were most likely to use hotwater to gasify the tanks. While FMC did act promptly to repair thepumps, there is no evidence that gasification was stopped during sumppump malfunction. Indeed, according to FMC’s daily inspection report,on January 21, 1983, the day the sump pump was reported to be not inproper condition, two drums of nitrogen trichloride byproduct weregenerated.Moreover, unlike my colleagues, I find no inconsistency in the evidenceconcerning whether, during the two months before the accident, hot wateraccumulated to the level of the transfer pipes. The majority finds itsignificant that, despite operator Robert Jones’ testimony that heobserved the water level reach the pipes during the two months beforethe accident, weekly logs kept during that period show that there was nostanding water in the diked area. They fail to consider, however, thepossibility that the accumulation of water observed by Jones was removedby the time of the weekly inspection.I am also not persuaded by Dr. Grelecki’s testimony that even if thewater did reach the level of the pipes, it would have cooled off toomuch to constitute a hazard. First, Dr. Grelecki’s opinion was notshared by Duane Colpous who testified that the line could have beenheated up by the hot water in the pit. Second, Dr. Grelecki’s opinion isundercut by other aspects of his testimony. When discussing thetemperature dynamics in the tank, Dr. Grelecki recited a principle ofphysics that is particularly applicable to the accumulation of water inthe pit. Dr. Grelecki stated that \”a hot zone in a liquid . . . tends torise because it is less dense than the surrounding cold fluid. Sothat–if you had a big vessel and a hot spot would develop in thatvessel it would tend to rise and sort of stir itself, like if you heat apot on the bottom the hot water gets from the bottom to the top.\” Applying this principle to the pit area, it is apparent that while theoverall temperature would have cooled, hot water continuing to flow intothe pit would rise to the top where the pipes are located.The evidence also indicates that at the time that the instruments in thecontrol room showed an increase from -30?C to -5?C in the contents ofthe reboiler, the operator went out to the reactor, removed the hosefrom the reboiler, and then went to the shut-off valve, some 40-50 feetaway, to close the valve. It is not improbable that he directed theflow of water into the diked pit, and it is possible that the hosesomehow became aimed at the transfer pipe.FMC should have anticipated and avoided the danger that external heatsources would come into proximity to the pipe. FMC knew that hot waterwas applied externally to assist in heating the contents of the reboilerand that hot water was in proximity to the transfer pipe. It also knewthat heat applied to the transfer pipe could initiate an exothermicreaction of the NCl3. Its failure to prevent such contact increased therisk inherent in the chlorine purification process which FMC was using. The Secretary established the feasibility of relocating the pipes. Indeed, after the accident, FMC did relocate the pipe to an area awayfrom the hazards presented by the pit.Accordingly, I find that the record establishes that FMC failed to freeits workplace of a recognized hazard that was likely to cause death orserious injury, and that the Secretary established a feasible means ofabatement. Therefore, I would affirm item 1 of the citation.————————————————————————The Administrative Law Judge decision in this matter is unavailable inthis format. To obtain a copy of this document, please request one fromour Public Information Office by e-mail ([email protected]), telephone (202-606-5398), fax(202-606-5050), TTY (202-606-5386).FOOTNOTES:[[1]] Another citation issued to FMC was docketed as No. 83-0489. Although never officially consolidated, both Nos. 83-0488 and 83-0489were heard and decided together by the judge. Both docket numbers werelisted on the direction for review. However, neither party has takenexception to any part of the judge’s decision in No. 83-0489. Accordingly, the direction for review as to No. 83-0489 will be vacated.[[2]] Section 5(a)(1) states:Each employer–(1) shall furnish to each of his employees employment and a place ofemployment which are free from recognized hazards that are causing orare likely to cause death or serious physical harm to his employees.[[3]] Item 1 alleged a violation of section 5(a)(1) in that:Employees were exposed to the hazard of an explosion due to criticaltemperature change which could initiate a decomposition of NCl3(nitrogen trichloride) during the chlorine vaporization and gunkneutralization process at Cell Room #1, on or about March 1, 1983.Item 2 alleged a violation of section 5(a)(1) on the grounds that: Employees were exposed to the hazard of an explosion due to theinstability of NCl3 (Nitrogen trichloride) created by unknown amounts ofchloroform being added to the scrubber; and unknown levels Of NCl3 inthe chlorine vaporization and gunk neutralization process, Cell Room #1,on or about March 1, 1983.[[4]] Dr. Grelecki’s name appears misspelled throughout the record. Ourspelling of his name is derived from 3 _American Men &_ _Women ofScience_ (15th ed. 1982).[[5]] Under the Fed. R. of Evid. 407, evidence of post-accident measuresare admissible to establish feasibility. The Federal Rules of Evidenceare generally applicable to Commission proceedings. See Commission Rule72, 29 C.F.R. ? 2200.72.[[6]] The Secretary argues that another method of reducing the danger ofhot spots would have been to keep the sump pump in good repair to keepwater from accumulating in the pit. However, the evidence indicatesthat FMC properly maintained the pump and that the pump was replacedpromptly whenever a problem arose.[[7]] Similarly, the Secretary introduced an exhibit establishing that achlorine plant operated by PPG uses a mandatory dump system requiringthat the reactive material be dumped into a coolant at 60? F. As withthe TNT plant, however, the Secretary failed to introduce any evidenceto indicate that the techniques and processes used by PPG weresufficiently similar to those at FMC to warrant a conclusion that such adump system would be feasible at the FMC facility.[[8]] The Secretary also argues that an experiment conducted shortlybefore the explosion where the chloroform flow was cut-off establishesthat FMC failed to maintain the concentration of NCl3 at a properlevel. We do not agree. The test to which the Secretary refers wasconducted under the direction of FMC’s engineers. There was no showingthat FMC’s engineers were not skilled to perform this test. Nor isthere any evidence that cutting-off the chloroform flow for severalhours in any way increased the likelihood of an explosion or endangeredthe employees.[[9]] FMC contends that the citation did not contain any suggestion thatrelocation of the transfer lines was a proper method of abatement, andtherefore should not be considered by the Commission. I disagree. Thecitation expressly addressed itself to hazards caused by criticaltemperature changes during the nitrogen trichloride byproductneutralization process. Also, the citation stated that, \”among othermethods,\” a feasible and useful method of abatement would be to\”initiate the use of instrumentation to control temperatures on thepiping system.\” In my view, the citation put FMC on notice that if theSecretary failed to establish the feasibility of instrumentation todetect hot spots on the pipes, he would attempt establish other feasiblemethods of abating the hazard. To the extent the Secretary’s failureto specifically list the \”other\” methods of abatement may have confusedFMC, such confusion could have been remedied by moving for a moreparticular statement pursuant to Fed. R. Civ. P. 12(e). Similarly, tothe extent the Secretary’s raising of the issue of moving the pipes atthe hearing may have surprised FMC, any prejudice could have beenremedied by a motion for a continuance pursuant to Fed. R. Civ. P.15(b). Indeed, through its own expert witness, FMC introduced testimonyconcerning the cause of the explosion. Therefore, I conclude that FMCwas not prejudiced by having to address the relocation of the pipes.”