Document RJBRLvz2w7e0YZgbYjpGLdeya

t\A|^M.'WJ'_-I-,: * W*MJB_' V "`^'.y^WjlKHWM enraicals SUMMARY REPORT OF REACTIVE CHEMICALS INCIDENTS fl'nnrnft<i JmifmflHeriirtfrTriflV-*- --'-^-`ri-'**** J*vv FOR THE MONTH OF DECEMBER ---- '-- January 14,1997 The following reactive chemicals incidents were reported during the month of December: A dilute sulfuric add was inadvertently mixed with 92% sulfuric add when a flow meter malfunctioned during the production of a batch of cation ion exchange resin. The mistake was discovered about an hour later, and fortunately the materials had not mixed due to stratification in the tank. The concentrated material was removed slowly avoiding a potential heat of mixing inddent. An acddent occurred during the startup of an oxygen mixer section in an ethylene oxide process. A mixture of hydrocarbons from the gas loop apparently backed into the piping section prior to the mixer, became mixed with oxygen and ignited. The resulting explosion in this section of the piping caused a significant fire involving the escaping hydrocarbons and oxygen. Analysis of this acddent is continuing and a final report will be completed in February. Monochloroacetic add backflowed into a caustic line containing 49% caustic when a reactor < *. recycle/neutralization loop became plugged. The blockage was apparently caused by the salt formed from the neutralization reaction, and the control system was not adequate to detect and shut down the gravity feed caustic system. An esterification reactor producing DURSBAN* insectidde was not in operation for a couple of days but was held at 60 C and under vacuum (a wait step, not a shut down step). A leaking valve allowed 600 kg of one of the reactants (dimethyl phosphoro-chloridothioate (DMPCT)) into the reactor. The reactor is on weigh cells and pressure monitored. Several reactor weight and vacuum alarms occurred during the period of the leakage, but there was no action taken. At re-start the alarms were all reset to allow the production to start. When the esterification reactor was being preheated to 120 C to allow the proper transfer of intermediate sodium trichloropyridinate, the DMPCT decomposed and exothermed up to 180 C, causing the reactor rupture disk to relieve. Fortunately, no transfer of intermediate took place - this could have added to the severity of the incident. A small fire occurred in the vent from an extruder in a styrenics compounding facility. The extruder was mixing styrene polymer, Acrawax and a rubber concentrate for a sample order. It was concluded that there was too short of an extruder mixing length in relation to the required die pressure. A fire occurred in a laboratory cubicle during a reactor wash with 155 C isopar solvent. When the solvent was automatically dumped to a glass dump pot, there was a spill that was ignited from an unknown source. The glass dump pot is being replaced with a stainless steel pot. "Restricted for Use Within Dow" DO 072811 CONFIDENTIAL  A brominated epoxy residue in a finishing tower decomposed during a standard clearing operation to replace the packing. An upset upstream in the process had allowed caustic to get into the resin feed to the tower. Caustic enhances the decomposition of the epoxy. The decomposition was exothermic, and hydrogen bromide gas was produced. The tower relieved through the pressure relief at 150 psig. The packing was later found to be fouled with a black solids. Water was inadvertently added to an epoxidizer reactor prior to startup after 3 weeks of being shutdown. An organic stream containing epichlorohydrin was added to the reactor, and the water layered out on top of the organic. Caustic carbonate solution was then fed to the top of the reactor. This initiated a hydrolysis reaction with the EPI, causing the reactor to increase in pressure and relieve through the relief valve. The water entered the reactor through purges on pump seals during the shutdown period. m Twenty feet of a 6-inch plastic drain line shattered during an unplugging step. The line had become plugged with magnesium hydrate and strontium carbonate. A 32% HC1 solution was being used to dissolve the plug. This procedure had been utilized successfully in the past; but the plugging this time was more severe, and the CO2 gas that was generated by the add decomposition of the carbonate caused the line to overpressure. A small fire occurred near the entrance of a drier of DRYTECH*. A significant amount of material had built up outside the drier. Ignition was caused by either a damaged heat shield baffle or oil contamination that may have lowered the autoignition temperature. The investigation is continuing. A plastic drum containing used absorbent pads containing Rando 68 oil and diluted organic peroxides was found to be smoldering in a waste storage area. The fire was quickly extinguished. Autoignition of organic containing absorbents continues to be a reoccurring inddent. The peroxides here make the inddent even more probable. Editor's note: A large number of inddents this month were caused by leaking valves and inadequate start-up procedures. We need to be espedally careful when restarting equipment that has been idle for a period of time, and make sure our procedures look for things that might have changed. Process Safety Services Trademark of The Dow Chemical Company For address change, contact Sharon Hartnagle, Process Safety Services, 2020 Dow Center, Midland, MI - , _ '"C * FOR THE MONTH OF MAY SUMMARY REPORT OF REACTIVE CHEMICALS INCIDENTS *- .. -"r, ! , i - S] \ -. * ^ ~1 ^ ' ,-'* i* 3 ~. June 12,1996 The following reactive chemicals incidents were reported during the month of May: Two operators were working to clear a plug in the magnesium chloride cell feed handling system. Due to several events linked together, a significant quantity of water had been introduced into the elevator/conveyor system and into this hopper. The nature of the magnesium chloride cell feed is such that the right amount of water in combination with the magnesium chloride cell feed can produce a liquid/solid mixture with around 225 F. In this case, a sufficient quantity of water did reach the hopper creating a hot mixture that spilled out, resulting in bums to the two operators. One individual passed away 18 hours after the event; the other should have no long term impairment. A significant chlorine release occurred following start up of a new system in a chlorine liquefaction area. A red rubber shipping gasket on the bottom 8-inch nozzle of a chlorine tank C was accidentally left in place during construction. The pipe was bolted together with both the shipping gasket and the correct gasket installed on top of it. The process chlorine eventually reacted with the red rubber causing the release. The FiAA plant has successfully implemented a new MOD control scheme designed to control reactor runaway conditions that have caused process problems in the past. In the HAA process, EO, PO and acrylic acid are reacted to make hydroxyalkyl acrylate. Recently a reactor was experiencing the initial indications of an exotherm. The new control system caught the excursion quickly enough to successfully avert a loss. A 2-liter glass lab reactor exploded during a chlorination reaction. Chlorine gas and 1-hexene were being slowly added to a solvent consisting of a 10% by volume mixture of ethylene glycol 0 dimethyl ether in water (glyme). After about one minute there was a loud pop accompanied by a flash of light. The upper portion of the reactor baffle was propelled off the reactor. Apparently the reaction occurred in the head space of the vessel between chlorine and the glyme solvent. A tank trailer containing chlorpyrifos tank bottoms was in the process of being prepared for cleaning. Water and industrial strength bleach were added to the material for odor control. It was well understood that bleach and chlorpyrifos will react, so the materials were added slowly. Next, one gallon of a deodorizer material NI-712 was added to the truck. Black smoke started coming from the manway due to the unexpected reaction of the bleach with the deodorizer. The operator attempted to close the hatch on the trailer to contain the smoke, but was unable to, due to the pressure being generated within the trailer by the reaction. DO 07?813 CONF TOFNTIAL "Restricted for Use Within Dow'  Pyrophoric tungsten carbide was leaked from a process and was sucked up into a dust collector system that also contained carbon dust. The materials ignited, causing flames to come out the vent. The fire department extinguished the fire, minimizing associated damage. A laboratory reagent bottle of 1:1 nitric add and water was prepared and stored in a plasticcoated glass bottle. Five days later, the volumetric pipette on the top of the bottle was propelled upward into the ceiling tile due to internal pressure being generated in the bottle. The remaining liquid was found to be dark green in color that later turned brown. It was found to contain a residue of oxalic add and acetic add. The source of the contamination is being investigated. For address change, contact Sharon Hartnagle, Health & Safety Services, 2020 Don? Center, Midland, MI, E-mail U07663S DO 072814 CONFIDFNTTAl SUMMARY REPORT OF REACTIVE CHEMICALS INCIDENTS FOR THE MONTH OF APRIL rtih'li nitiifi li it Tli' liliiaii^JrtniiifTilfri^WfciiiiiBi^fil i , , ,in,ri ________________________________ May 9,1996 The following reactive chemicals incidents were reported during the month of April: (Thanks to thefolks who submitted a number ofvaluable "Learning Experiences" this month.) A white vapor cloud was observed coming from a vent hose in a centrifuge bay. The building was evacuated, and emergency personnel were called. The white vapor was later discovered to be ammonium chloride that had apparently formed in the vent header. An open cross tie between two vent headers was later discovered. One header contained HC1, and another con tained ammonia from a leaking relief valve on a refrigeration system. This probably allowed HC1 and ammonia to come together during an upset with the vent header pressures, produc ing the ammonium chloride. (Leveragable Learning - Do your vent header systems operate in such a way that inadvertent mixing might occur, producing an uncontrolled chemical reaction?) ,1. A tank truck that was supposed to be loaded with diethanolamine was inadvertently loaded with 50% caustic. The empty truck (with diethanolamine loading paperwork) arrived at Dow, bdC along with a large number of caustic tank trucks. Security thought that this truck was for the caustic plant and sent it there. The truck was loaded with 50% caustic; however, the mistake was caught before it left the site when it was weighed and found to be overweight. (An incident like this, if not caught, could easily cause a reactive chemicals incident at our customer. Good work by the employees at the Dow scales on catching this.) ;* An inbound truck was sent to a plant to be offloaded into an organic peroxide catalyst tank. In the process of checking the paperwork and the truck prior to unloading, the operator discovered that the truck contained acrylonitrile. Not only was a serious potential reactive chemicals incident averted, but a serious industrial hygiene exposure could have occurred in just sampling the truck to verify contents. (This excellent learning experience highlights the need to continue to be 100% accurate with our receipt ofraw materials operating procedures.) A minor injury occurred when a glass sample bottle containing hot chlorinated hydrocarbon bottoms material ruptured and splashed an employee. The bottle had just been removed from a hot sand bath as part of an analytical procedure. The bottle was broken due to either over pressure resulting from decomposition of the sample material to HC1 and chlorine, or due to thermal shock from setting the hot glass container onto a cold bench. In any case, the use of a glass sample bottle for this operation was deemed inadequate. E DO 07?8;j S donfiofniial  A fire occurred in a dryer for METHOCEL*. This followed an unplanned shutdown of the reactor of METHOCEL*. A failure of the reactor seal pump necessitated the shutdown of the agitator, causing the reactor to vent. Several other undesirable conditions and reactions occurred in the control sequence during restart. The shutdown apparently caused the produc tion of a product with significantly lower oxidation stability. During a later drying operation of the less stable product, the dust collector on the transfer system to the silo plugged and the dryer was shut down. This resulted in higher residence time in the dryer. These combined events in addition to the residual heat in the dryer resulted in the product catching fire. There was minimal loss. An operator at a chlor-alkali plant discovered that Silastic 732 RTV had been used on 2 PSVs | that were in chlorine service. Due to the high chlorine sealing pressure (140 psig), these valves i had proven difficult to seal in the past. It has been the valve shop's practice for some time to use Silastic 732 RTV on the threaded plugs and the bellows gaskets. There was no incident; however, the PSVs were removed from service and replaced with PSVs that contained Fluorolube grease rather than Silastic. Even though reactive chemical tests show the Silastic 732 RTV reacts with chlorine at a higher temperature than other Silastic RTV formulations, there are some credible scenarios that could have resulted in an incident with these 2 PSVs. - Using Silastic RTV sealants in chlorine service is normally not recommended. ^ 1 Trademark of The Dow Chemical Company For address change, contact Sharon Hartnagle, Process Safety Services, 2020 Dow Center, Midland, MI, or via e-mail. 00 07P816 CONFIDENTIAL ' ,, " * - 'v . t v.- - . m- ;, ;-i ,:-v -fc * .v-^,-. v, FOR THE MONTH OF JANUARY February 14,1996 The following reactive chemicals incidents were reported during the montit of January. January's indication of our performance shows we have started the year with a very good month. This report includes several recently reported incidents from 1995 that we are sharing for the learning value. An uncontrolled exothermic polymerization occurred in a packed tower containing HEA (hydroxyethyl acrylate). The reaction was caused by the loss of inhibitor activity, resulting from the inadvertent shutting off of the system purge air for several hours. (January, 1996) A hydrogen header coming off a chlorine cell series experienced a small internal explosion after the series was taken offline. An automatic nitrogen purge valve failed to open, allowing air to enter the hydrogen header. (January, 1996) Hot lubricating oil leaked onto insulation at 300 C and autoignited. (February, 1995) A faulty ethylene flow transmitter signal caused excess oxygen to flow into an oxychlorination reactor, causing a rapid temperature rise and a process shut down. (March, 1995) A failure in a PLC controller on an EO reactor control system allowed ethylene from the recycle gas loop to back flow into the oxygen line through a mixer, causing a 6-inch monel elbow to rupture. The rupture was a one foot long split on the inside radius seam. (July, 1995) An oxychlorination process experienced a lightoff in the head of a reactor, blowing the rupture disk. It is believed a contributing cause was the previous discontinuation of the use of "Denstone" in the reactor head cavity as a suppressor. (August, 1995) A rail tank car used for caustic shipments to a customer gave off a strong hydrogen sulfide odor when hot water was added to wash the car. Upon investigation, it was found that the car was being used by the customer to contain sodium hydrosulfide, without informing Dow. (August, 1995) A chlorine reaction occurred with the external bolting lubricant on a liquid chlorine pump, causing a small chlorine release. The release was detected within fifteen seconds, and the pump was automatically shut down. A copper based "Neverseez" lubricant was used, which is incompatible with chlorine. (October, 1995) For address change, contact Sharon Hartnagle, Health & Safety Services, 2020 Dow Center, Midland, Ml, E-mail U076635 DO 072817 T? CONFIDENTIAL "Restricted for Use Within Dow"