Lithium Aluminum Hydride/ Tetrahydrofuran Explosion
A researcher at X was seriously injured last December when reducing a substrate using lithium aluminum hydride (LAH) in tetrahydrofuran (THF). Within the last year, at least two other accidents involving procedures using LAH and THF have been documented. Due to the inherent hazards of LAH and THF, researchers must thoroughly plan out experimental protocols and incorporate safety measure to mitigate the hazards of this procedure. We have consulted with an outside expert in these issues, and he has made a number of important safety recommendations for this procedure.
Following a typical protocol, an experimenter would:
- Heat and flush a 3-neck, glass flask with nitrogen to drive off all moisture.
- Remove heat source and cool the flask, but continue to flush with nitrogen.
- Add a stir bar, THF (freshly distilled), and LAH
- Flush with nitrogen for the rest of the procedure surround the flask with an ice bath.
- Turn on the stirrer.
- Start water running through the closed loop of the condenser.
- Start drop-wise addition of the substrate (which is dissolved in freshly distilled THF)
However, this 'typical' setup is not necessarily the best setup.
Listed below are several recommendations pertaining to this procedure:
- Use enough solvent to dissolve all LAH. Adding substrate to a slurry of undissolved LAH and solvent is almost as dangerous as adding it to dry LAH. The solubility of LAH in THF is 13g LAH/100g THF at 25oC, and in diethyl ether, 35g LAH/100g diethyl ether. Aldrich does not make solutions more concentrated than 1M (38g/liter). It is recommended to make solutions no more concentrated than 1M.
- Add LAH to THF rather than adding THF to LAH when preparing solutions. Dissolving LAH in THF is very exothermic! If THF is added to dry LAH, the LAH can easily overheat and decompose exothermically, especially on larger-scale reactions.
- Keep the ratio of LAH to substrate low. If the reaction goes awry, it's safer to have only a 2-fold excess LAH rather than a 10-fold excess to deal with.
- Ensure that the stopcock on the substrate dropping funnel works smoothly. If the stopcock sticks, too much substrate may be delivered, creating excess heat in the reaction flask.
- Ensure that the reaction flask is under a nitrogen blanket. Double check that the nitrogen inlet tube is securely fastened and all air is excluded from the reaction vessel.
- Prepare the substrate carefully to exclude any residual solvents that might react with LAH. This way, you won't have to use as much excess LAH.
- Ensure that the substrate/THF solution is free of peroxides. Any added THF should be freshly distilled.
- Use chilled silicone oil instead of ice and water as a cooling medium. This is now current industrial practice for large-scale reactions. If the flask breaks for any reason, LAH will not react with silicone like it does with water.
- Use an explosion shield when working with large-scale reactions. Lowering the fume hood sash and wearing protective eyewear is adequate with smaller scale reactions.
- Quench the reaction mixture, by addition of water or other quenching agents, using extreme caution. Add the quenching agents slowly!
If you have any questions, call your safety office.
Explosive Decomposition of an Organic Azide
Key Instruction Points:
- Review risk assessment when scaling up reactions.
- Use engineering safeguards for containment and remote handling when using reactive materials.
A chemistry graduate student was isolating an organic azide (benzyltriethylammonium azide) as an intermediate in a process to synthesize a complex organic molecule to be used in a cancer treatment. (She was trying to prepare a 5-deoxy-5-azido nucleoside by azide displacement of the 5-tosyloxy derivative). Several days earlier she had isolated a small amount of this organic azide intermediate by using a rotary evaporator to drive of the reaction solvent. Approximately 0.5 grams of material were initially isolated and used to run analytical tests to demonstrate the purity of the isolated intermediate. Now that she had demonstrated that the initial steps in her synthesis process were successful she scaled up the process 20 fold in order to isolate enough organic azide to continue her synthesis.
At approximately 9:00 on a Sunday night, while working in the lab with two other graduate students, she completed the isolation of approximately 7-8 grams of organic azide in the rotary evaporator. The rotary evaporator was set upon the open bench in the middle of the laboratory.
After isolating the organic azide from a 1:1 solution of acetone and methylene chloride in the Buchi rotary evaporator, she lifted the 250 ml round bottom flask containing the organic azide from a water bath, with the handle provided for this purpose, using her left hand, while her right reached out for the flask.
The flask exploded in her hand, shattering all of the glass associated with the rotary evaporator and glass containers close by on the lab bench. Parts of the condenser were found in a hallway approximately 15 feet away.
Her recollection of the incident and the nature of her injuries, indicate that she did not have the opportunity to break the vacuum on the system or stop the rotation of the flask. It is believed that the raising of the flask alone from the warm water bath initiated the decomposition of the shock sensitive organic azide, perhaps by creating a movement in a contaminated ground glass joint. However, the graduate student does not feel that solvent "bumping" occurred in this case. This could have caused the azide compound to contaminate the glass joints.
Injuries and property damage caused by the incident:
The glass fragments from the exploding flask severely lacerated the graduate student's right hand and cut her cheek and forehead. The force of the explosion blew her to the floor where she lay stunned and bleeding. The safety glasses she was wearing protected her eyes from glass fragments; otherwise she may have been blinded. The two students with her immediately came to her aid and called an ambulance that transported her to the hospital five minutes away. That night a four-hour surgical operation removed the glass from her face and hand and subsequent surgery restored most, but not all, of the functionality of her hand. She lost the ability to move her thumb. She also underwent multiple plastic surgery operations to improved her appearance.
Resources spent responding to the incident: The local fire department responded, and because the incident involved an explosion, the State Fire Marshall's office was also called in. Three University EHS employees took part in the six hour investigation with the three state inspectors and two representatives of the local fire department. The building was closed until the investigation was complete. Upon completion of the State Fire Marshall's investigation, EHS employees cleaned up the spilled materials and blood.
Cause of the incident:
The explosion was caused by the rapid decomposition of the organic azide which it is believed had worked its way into the ground glass joints between the product flask and the glass column on the rotary evaporator. However, after interviewing the graduate student it was apparent that several factors lead up to the incident including:
- The graduate student had underestimated the risks associated with the material she was isolating. Although she was aware generally of the decomposition potential of azides she did not know just how shock sensitive the organic azide she was isolating was – even though this information was available in the literature.
- Due to the underestimation of risk, she isolated the azide on open bench with out adequate containment such as a laboratory hood and shielding, personal protective equipment, or procedures.
- She did not reassess the risk when scaling up her reaction. If she had, she would have realized that the material being handled had significant explosive power and due to its inherent instability required substantial shielding and remote handling.
To prevent future accidents of this type the following steps were taken:
- The types of "high-risk" reactions that were being conducted in the Chemistry Department were identified. Based on the type of reaction and the scale (quantity of material), the appropriate safety precautions (both engineering controls and personal protective equipment) were identified and placed in a matrix. This safety precaution matrix table was distributed throughout the Chemistry Department and required to be followed.
- A formal peer safety review process was established that required the following steps be completed before graduate students were allowed to beginning research: (1) a comprehensive literature review must be conducted (safety and chemistry); (2) a protocol safety review form summarizing the hazards and precautions to be taken is completed; and (3) The planned research, information uncovered in the literature review, and safety review form, is reviewed with a peer.
- A shared use facility was established in which high-risk reactions could be performed and special procedures for performing these reactions established.
Stirred Reaction Flask Explosion
Key Instruction Points:
- Don't leave reaction unattended.
- Use proper PPE.
- Control sources of contamination.
- Set chemical hood sash to lowest height possible.
At 10:11:44am, Wednesday, 9 February xxxx, the Fire Department received an alarm from the Chemistry Building, and responded with fire and EMS personnel. County Sheriff officers also responded. At about 10:35am, EH&S personnel arrived at the incident site.
At about 10:10am, an explosion occurred within the Chemistry Laboratory. A Ph.D. research student, performing an experiment inside a fumehood, was injured by flying glass shards, which were generated from an explosion that occurred in a reaction flask (see photo below). Although the fumehood sash was partially down (about half way), the researcher received injuries mostly to the right side of his face (see photo below) and to his left hand and arm. No injuries were associated with the eyes since the researcher was wearing safety glasses with side shields.
The researcher was de-conned in the laboratory emergency shower and received first aid from laboratory personnel, who are also safety representatives for the laboratory. After the first aid treatment, the researcher was escorted from the building to meet the arriving Fire Department EMS personnel. The EMS personnel then transported the researcher to a near by building for further deconning in a hot water shower. Afterwards, the researcher received additional first aid before being transported to the Hospital ER for treatment and observation. Late in the afternoon, the researcher was release from the ER. On the following Monday, the researcher returned to his laboratory at the Chemistry Building.
Description of Experimental Procedure
The experiment being performed was a modification of the Simmons-Smith cyclopropanation procedure for the synthesis of species for reacting with olefins. Very simply, in a stirred reaction vessel under a dry argon atmosphere containing an ultra-low water and oxygen solvent (250 mL dichloromethane; CAS# 75-09-2) cooled to -10oC by dry ice in acetone, two reactants (diethyl zinc; CAS# 557-20-0 and methylene iodide; CAS# 75-11-6) are sequentially introduced via a fill funnel under dry argon pressure. Photos of the experimental apparatus and equipment are shown in the Incident Investigation section below. In between fillings, the funnel is rinsed with the solvent fed from a one-use, sterile plastic syringe. First, diethyl zinc (13 mL) is added from its container with a double-ended needle to the solvent at about 2.5 mL/min. Next, methylene iodide (22 mL) is added with a glass syringe to the solvent mixture at about 1 mL/min. After the final addition, the solvent mixture is allowed to continue to stir for 20 min producing chemical species for reacting with olefins. As the reaction goes to completion, the solvent mixture turns milky with the formation of a fine precipitate, which is normal.
Description of Incident
The experiment was performed as stated in the SOP, which was recorded in the researcher's lab notebook, up to and including the addition of the methylene iodide to the dichloromethane and diethyl zinc mixture under an inert argon atmosphere. After the methylene iodide (22 mL) was added at a rate of about 1 mL per min over a time period of about 25 min, the researcher noted that the experiment was proceeding normally. At this point, he left the experiment in the fumehood for the reaction to continue for about 20 min. However, he decided to return after about 10 min to check on the experiment. Upon returning, he noted the stir bar was not rotating due to the formation of an unusual amount of precipitate in the bottom of the reaction flask. The reactant mixture was clear, but with no liquid phase separation. The appearance of the reactant mixture was unusual; normally the mixture would appear milky white due to the suspension of a fine precipitate. Although the stir bar was not rotating, the researcher did not perceive the risk of an explosion. So he immediately proceeded to restart the stir bar. During this process, when he had the reaction flask in his left hand, the contents of the flask detonated. From the flying glass shards, the researcher sustained serious injuries to the left hand, right neck, and right side of the face (see above photo) from flying glass shards.
NOTE: Terms such as explosion and detonation will be used through out this report with the realization that a very rapid release of energy may have occurred without an actual detonation. Regardless, the energy release and the subsequent pressures were so rapid and great that the neck of the flask could not vent the pressure buildup. After the incident, there was no evidence of fire/smoke or other combustion products.
After an interview with the injured researcher, a reenactment of the experiment was performed substituting water for the chemicals: dichloromethane, diethyl zinc, and methylene iodide. The experimental setup and equipment are shown in the photos below (clockwise: experimental apparatus, diethyl zinc container, and rinse syringe).
The experimental apparatus, under a positive-pressure argon atmosphere, is continuously fed from an argon cylinder through a drying column. Setting atop the round bottom flask, which is the reaction vessel, is a septum-fitted funnel for feeding the reactants. The reactants are fed via double-ended needles (diethyl zinc), plastic syringes (dichloromethane), or glass syringes (methylene iodide) into the funnel, and then fed into the flask through a stopcock. The only difference in the mock setup and the experimental setup was the use of an open bath rather than a half-sphere Dewar. This difference was not judged to be a factor in the incident.
Primary physical hazards associated with the chemicals components were the flammability of dichloromethane (LFL 13%, UFO 23%), methylene iodide and diethyl zinc incompatibilities (see http://xxxx.edu/~msds/
), and the pyrophoricity, water reactivity, and explosive heat-sensitivity of diethyl zinc. Of particular concern are the incompatibilities of the two reactants with other chemicals such as alkenes, oxidizers, copper-zinc alloys, potassium-sodium alloys, and potassium. For example, alkenes in the presence of the reactant mixture could result in an explosive reaction and in the presence of potassium form a shock-sensitive mixture. During the reenactment of the experiment, several possible causes for the incident were identified and are addressed in the following table.
Hazard Assessment Table
Introduction of water via the glassware and reusable syringes
Since there is a SOP for washing glassware and reusable syringes, it is not likely. If it did happen, there might be no sign to small amounts of visible emissions in the flask; no signs were noted.
Injection of water with the dichloromethane or methylene iodide
The dichloromethane is dried in the purification process, which is under a dry nitrogen atmosphere; not likely contaminated. The methylene iodide is transferred from a glass bottle, which is used by several researchers. Probably some water could be introduced resulting in no sign to visible emissions in the flask; no signs were noted.
Injection of oxygen with the dichloromethane or methylene iodide
Oxygen is removed from the dichloromethane in a purification process, which is under a dry nitrogen atmosphere; not likely contaminated. The methylene iodide is transferred from a glass bottle, used by several researchers. Probably some oxygen could be introduced resulting in no sign to visible emissions in the flask; no signs were noted.
Loss of argon atmosphere
The argon cylinder was still under pressure, and argon was still flowing after the incident.
Increase temperature of the reactant mixture
The only heat source is the heat of reaction. The magnetic stirrer did not have a heating element. The final reactant was added over a 25-min time period to prevent large temperature increases. In addition, the reactant mixture was cooled in a Dewar to -10oC by a dry ice/acetone solution. Nevertheless, if a large temperature increase had occurred, detonation could result.
Introduction of a fourth chemical component via the glassware, syringes, or contamination in other components
Because of the very strict cleaning, rinsing, and drying procedure used on the glassware and reusable syringes, it unlikely that amounts of contamination could be introduced that could result in an explosion. Evaluation of the solvent and diethyl zinc sources indicates it is unlikely that these are sources of contamination. However, the methylene iodide (stabilized with copper or silver mesh) is purchased, stored, and used out of a glass bottle by many different researchers. It is judged to be a possible source of contamination.
A definitive conclusion could not be made as to the specific cause of the detonation of the reactant mixture. However, based on the above hazard assessment, two likely causes of the explosion detonation were a very rapid increase in temperature of the reactant mixture and the introduction of a fourth chemical, as a contaminant, into the reactant mixture. The introduction of a third reactant could possible explain the formation of unusual amounts of precipitate, which settled to the bottom of the reaction vessel stopping the magnetic stirrer from rotating. The stopping of the magnetic stirrer was judged to be the abnormal occurrence that preceded the explosion and perhaps was the causality for the explosion.
The use of safety glasses probably saved the researcher from serious eye injures. Nevertheless, additional protection to the face and body would have reduced the number and severity of the injuries received.
- Do not leave this experiment unattended; ensure that the reactant solution is continuously stirred.
- Use additional personal protective equipment such as full faceshield and blast shield when conducting experiments with highly reactive components.
- Assess the operation of the dichloromethane purification unit to ensure high purity.
- Review the glassware cleaning procedure to ensure no contaminates are present.
- Discontinue the use of sterile syringes for introducing chemicals into the experimental apparatus. The term "sterile" is no indication of "chemical contamination levels.
- Always set the fumehood sash at the lowest usable height.
- Assess the use of the methylene iodide to reduce the likelihood of contamination and implement the following controls:
- Date the container as to when received.
- Date the container as to when opened.
- Implement procedures to ensure minimum container open time.
- Set criteria for methylene iodide use, considering such parameters as color and age.
Chemical Solution Preparation Explosion
Key Instruction Points:
- Review hazards of chemical combinations prior to start.
- Use proper PPE for the task.
- Lower hood sash to the lowest possible height.
Description of Incident
A teaching assistant was preparing a 30 L solution 0.04 M KMnO4 in 0.5 M H2SO4 for use in chemistry instructional laboratories. She was working by herself. This is a laboratory where all solutions for use in the general chemistry laboratories are prepared. Her supervisor was working in his office which is adjacent to the preparation lab and connected by a doorway. A chemistry class was underway in a lab across the hallway. This teaching assistant had a BS and MS in chemistry and working at the university for two years. Her responsibilities included preparing solutions for the chemistry program since the start of her employment.
She indicated she was following a procedure she has followed without incident in previous semesters (estimated 10-12 previous preparations). She indicated the procedure was contained in a notebook containing standard procedures for preparing all of the reagents used in the two classes. On the day of solution preparation, was not referring directly to the written procedure since she had used it sufficiently in the past that she did not need to refer to it. The quantity prepared is typically 10-20 L; this was the first instance she could recall in which 30L was to be prepared.
The notebook containing the written procedure has not been located.
To make 30 L of the solution, 833 ml of 18 M sulfuric acid was needed. This was measured out, transferred to a 1 L beaker which was placed on a stir plate in the hood. The volume of acid was obtained from two sources: the last 200-300 mL of sulfuric acid remaining in a bottle that had already been opened and the balance from a new unopened bottle. Both sulfuric acid bottles were 2.5 L in size. A 40 L Nalgene tank was filled with approximately 22 L of deionized water and placed in the hood. Solid reagent grade potassium permanganate (189.648 grams) from a bottle dated 9/10/96 was weighed out into a clean beaker on a top loading balance on the lab desk. This was slowly added to the acid on the stir plate (estimated over a period of 30 seconds). The solution heated quickly and began to boil in 1-2 minutes. Because it was spattering, the solution was picked up with gloved hands so that it could be transferred into the DI water and diluted before any more was lost. Before any of the solution could be transferred to the DI container, the beaker broke, spraying acid and permanganate everywhere.
The teaching assistant sustained chemical burns to the upper body, any uncovered areas. She was wearing safety glasses, not goggles, gloves, and a short sleeved shirt.
The supervisor later indicated that the calculated amounts described above for preparation of the 30 L solution were correct. If the material had not broken, the contents would have been added to the 22 L DI container over the course of a few minutes, the container would have been filled 30 L with DI, stirred, and then dispensed into three 2 liter containers. The remaining solution would be placed into a 20 L carboy for subsequent refills of the 2 liter containers for lab work later in the week.
Probable Cause of the Incident:
Several pertinent references in the literature regarding the mixtures of potassium permanganate and sulfuric acid indicate caution. Bretherick's Handbook of Reactive Chemical Hazards states:
"Addition of concentrated sulfuric acid to the slightly damp permanganate caused an explosion. This was attributed to formation of permanganic acid, dehydration to dimanganese heptoxide and explosion of the latter, caused by heat liberated from interaction of sulfuric acid and moisture. A similar incident was reported previously, when a solution of potassium permanganate in sulfuric acid, prepared as a cleaning agent, exploded violently.
Manganese heptoxide is formed as a dense green-brown oil by reaction between potassium permanganate and concentrated sulfuric acid. Kleinberg, Argersinger and Griswold, Inorganic Chemistry pg 534 states that the reaction between potassium permanganate and sulfuric acid is:
2 KMnO4 + H2SO4 = Mn2O7 + K2SO4 + H2O
AManganese heptoxide begins to lose oxygen at 0B and decomposes with explosive violence when warmed.
Durrant and Durrant, Introduction to Advanced Inorganic Chemistry pg 1014 states ...
A Manganese heptoxide, exists as dark green, explosive crystals. It is made by adding powdered potassium permanganate to cooled concentrated sulfuric acid. A dark green solution is formed which is explosive. Manganese heptoxide is stable at -5B, but it begins to give off oxygen at 0B, and at about 10B it explodes yielding manganese dioxide.
The most probable cause of this incident was the explosion of manganese heptoxide formed by the reaction of potassium permanganate and concentrated sulfuric acid.
- An alternate procedure for preparing the solution will be developed. Procedures for preparation of all instructional lab reagents will be reviewed.
- Use of protective equipment will be re-emphasized. The teaching assistant wore glasses and gloves, but no lab coat and faceshield, sustaining burns to the face, arms, and upper torso.
- Re-emphasize the availability and requirement to review health and safety information resources provided to supervisors and employees. This information (ACS booklet) indicates caution in mixing of sulfuric acid and potassium permanganate.
Phenyl Azide Compound Erupts During a Vacuum Distillation
A Post-Doc was purifying a fluorinated phenyl azide compound via vacuum distillation over a heating/stirring mantle. The resulting explosion caused the ceramic mantle fragments to cut and embed themselves in the experimenter's face. Fortunately, she was not seriously hurt and she was wearing her safety glasses.
What can be done to prevent this from occurring again?
There is no substitute for pre-planning your experiment and to discuss various techniques with your supervisor. Heating mantles are not good choices for vacuum distillation if the materials used are heat sensitive or unstable (such as most azides). This is because it is difficult to regulate precise temperature control with a heating mantle. A better choice would have been to use a hot oil bath or use chromatographic techniques to isolate the substances.
While the Post-Doc was wearing eye protection, the fume hood sash was in the wide-open position. This allowed the fragments to strike her face. If the sash must be open during the experiment, a portable blast shield should be used. If you know that the materials are unstable, safety glasses with a full face shield would be appropriate choices for PPE.
Bursting Chemical Container
The chemical and physical properties of liquids may change when gases are dissolved in solution. Supplier MSDSs may lack precautionary information addressing the hazards associated with gases dissolved in solution. This incident reinforces the importance of understanding the properties of chemicals before acquisition, use and storage.
Effects of incident:
The incident was an injury free event that could have resulted in shrapnel wounds from glass projectiles, and eye / skin contact with a chemical solution.
During a routine chemical inventory, containers stored in a small refrigerator rated for flammable storage were inspected and scanned with a barcode reader.
An original factory 800 mL glass bottle containing 1,4-dioxane with 0.5 M ammonia dissolved in solution was taken out of the refrigerator, and set on a lab bench top temporarily to affix a barcode and collect the specific information needed from its label. Approximately 1-2 minutes after the container was set on the bench top, it burst. The glass bottle was under enough pressure to cause it to burst into 4 large sections, scattering three of them across the lab bench top. The bottom section of the bottle remained in place and was still "holding" the frozen (but now melting) contents. Fortunately, no injuries or damage were caused and proper response / clean-up procedures were followed after the incident.
The incident investigation revealed that the container as received from the chemical supplier had been unopened. The outer cap had been sealed and the inner Sure-Seal cap had been in place. Furthermore it was determined that neither the container label nor the MSDS had explicit warnings about the refrigeration hazard. However, the label did state that the solution contained a 0.5 M concentration of dissolved ammonia gas. Further investigation revealed that 1,4-dioxane freezes at 10-12 degrees C. When ammonia is added to 1,4-dioxane, and the solution is allowed to freeze / thaw, pressure is generated inside the container due to the ammonia coming out of the solution.
Corrective actions to prevent reoccurrence:
- Do not store 1,4-dioxane containing ammonia in a refrigerator. This chemical should be consigned to a flammable storage cabinet at room temperature.
- The chemical supplier was contacted regarding the incident and the lack of specific precautionary information in the MSDS.
- Seek assistance as needed from site industrial hygiene professionals.
Unintended Overpressurization of Sealed Vial Results in Rupture
- Heating chemical compounds in excess of their boiling points within a sealed system may create over-pressurization hazards.
- When planning chemical synthesis work, the presence of a sealed system should be viewed as a trigger for additional safety review.
Effects of incident
- Shattered glass and chemicals escaped the chemical fume hood and could have injured nearby workers
- Potential for chemical exposure and burns
- Other equipment and experiments within the hood could have been compromised.
- Clean up efforts were required by hazmat trained personnel.
A researcher working in a chemical laboratory placed 5 ml of methyl iodide, a hazardous compound with a boiling point of 42 degrees C, in a 15 ml labeled closed glass vial along with another non-hazardous compound. The researcher set the apparatus in a lab fume hood and pulled down the sash. The researcher then heated the glass vial containing the two compounds to 150 degrees C in a Pyrex oil bath containing silicone oil, and left for the night.
In the morning, a different lab member discovered the broken oil bath and contacted safety personnel and the researcher. Industrial Hygiene and Emergency Services personnel determined there was no exposure hazard. The hood and adjoining area were then cleaned by appropriately trained personnel.
The glass vial was not rated for the pressure generated when the compound was heated and converted from a liquid to a gas. The vial was not outfitted with a pressure relief valve.
The researcher did not recognize that heating the liquid to a temperature well in excess of its boiling point could lead to pressures sufficient to break the glass vial and, therefore, did not realize that this experiment was outside the scope of the approved work plan.
The safety training for pressurized systems did not address the hazards of over-pressurization resulting from heating liquids above their boiling point in sealed systems.
The use of Pyrex containers for oil baths heated to temperatures above 100 degrees C is discouraged, as written in various safety documents.
Corrective actions to prevent reoccurrence
Safety documentation and training in chemical synthesis work should include a discussion of pressurization that occurs when heating sealed systems and associated mitigation strategies.
Unattended experiments should only be allowed with a posting at the laboratory entrance describing the contents and conditions of the experiment.
Oil baths heated in excess of 100 degrees C should be carried out in either a porcelain or metal bath container with silicone oil that is changed frequently. Repetitive heating of silicone oil has been known to lower its flash point.
A Bench Scale Chemical Reaction Results in an Explosion
This incident illustrates the importance of careful planning and verification of reagent and solvent mixtures when scaling chemical reactions.
Effects of incident
An employee suffered glass shrapnel injuries as a result of an explosion involving a reaction in a bench scale chemistry laboratory. The explosion shattered glassware and lab fume hood sash panes.
A laboratory employee suffered multiple glass shrapnel injuries to the face, arms, chest, and eye as a result of an explosion involving a reaction with tetrahydrofuran, water, 2,6 lutidine, and mercury perchlorate hydrate inside a lab fume hood. While the reaction mixture was stirring, the lab worker weighed out the desired amount of mercury (II) perchlorate hydrate (84 g) and then began adding the solid to the reaction flask in 5 g portions over a 10 minute period. Once the addition of the solid was complete, the individual closed the hood sashes and returned to the balance area to clean residue from the weighing operation. As soon as he was done cleaning, he returned to the hood and noticed that the reaction mixture (which was sitting in an ice bath) was boiling and had turned black. In the next instant, a powerful explosion occurred which shattered the reaction flask, stirrer, lab jack and two hood sash panes.
The lab worker was wearing safety glasses and a lab coat. A colleague assisted the individual under the safety shower for several minutes. Upon exiting the shower, the individual was led to the eye wash due to continued vision problems. Police, paramedics, firefighters, and EH&S responded to the incident. The injured worker was transported to a nearby hospital E.R. for treatment. Surgery was required to address an eye laceration. The individual was released from the hospital two days following the surgery, and made a full recovery with no vision impairment.
The incident investigation indicated that the reaction mixture was too concentrated. The reaction chemicals were scaled up, but the amount of solvent was not scaled up to an equivalent proportion. The amount of material used in the reaction was approximately 20 times more concentrated than the quantity specified in the literature. Electronic software was used to calculate the quantities of all reactants but not the quantity of solvent. Additionally, the lesser amount of solvent used in this reaction made it difficult for the magnetic stirrer bar to effectively mix the relatively-insoluble perchlorate into the reaction mixture, thus leading to localized overheating and detonation.
Corrective actions to prevent reoccurrence
Laboratory staff were re-trained to recognize that solvent proportions are critical to the safe conduct of this and other experiments. Additionally laboratory staff were advised that electronic software can be utilized to calculate the proportions of reactants, but verification of all constituents and the associated hazards with each should occur before proceeding with revision of any documented experimental components, including proportions, concentrations and temperature.
Gas Cylinder Gasket Melted Due To Reaction with Anhydrous Hydrogen Chloride
Watch for incompatible materials in all chemical processes.
Effects of incident
Injury-free event: Anhydrous hydrogen chloride under high pressure could have been released into the laboratory
A researcher used a re-configured stainless steel regulator for an upcoming experiment. The reconfiguration consisted of changing out a 350 CGA fitting with a 330 CGA fitting for anhydrous hydrogen chloride (HCl) service. The Regulator Shop performed the change-out of the CGA and supplied a "white plastic" gasket to seal the high-pressure fitting. After the pressure safety check, the researcher used the regulator to fill a small chamber with a small amount of HCl. Following the filling, the regulator and the filling system were evacuated.
When the HCl lecture bottle was taken off-line, it was noted that the "white plastic" gasket had reacted with the anhydrous hydrogen chloride and melted (see photo). Apparently, the white gasket was not an inert material such as Teflon, but actually a plastic that de-polymerized when in contact with HCl.
Inattention to compatibility of materials in contact with a toxic and corrosive gas.
Corrective actions to prevent reoccurrence
- Teflon is compatible with anhydrous hydrogen chloride, and Teflon CGA gaskets were made available to staff and users.
- The Regulator Shop verified material compatibility for all gases in service.
- Research project safety documents were revised to specify the person responsible for reviewing the compatibility between gases and materials of construction, including the gasket.
- The incident report was shared with all researchers who use compressed gas cylinders.
Chemical Container Failure Due to Over-pressurization
- Cooled organic liquids should be allowed to return to room temperature before re-bottling.
- Organic liquids used in cold-baths and recovered for reuse should be allowed to off-gas and then stored in self–venting containers.
Effects of incident
This incident was an injury-free event, however exposure to flying debris and chemicals could have occurred had the container been left to cool on a laboratory workbench during normal work hours. Mixed chemical reactions could have increased severity of incident.
A researcher employed acetone in conjunction with dry-ice as a cooling bath in the distillation of toluene / tetrahydrafuran with aluminum hydride. When the distillation activity was complete and no dry-ice remnants were visible in the liquid, the acetone was recovered for reuse and re-bottled in the original glass container from which it was dispensed. The capped bottle was then placed in the lab's flammable liquids storage cabinet and the lab was closed for the day. Due to the late time of day (approximately 6:00 PM) the acetone was not allowed to return to room temperature before being re-bottled, capped, and stored.
Upon initial morning entry into the laboratory, the researcher discovered broken glass and liquid in the flammable storage cabinet. The researcher immediately notified the lab custodian who barricaded the room and made appropriate notifications to EHS personnel. Industrial hygiene staff conducted room air monitoring and once the scene was considered stable, personnel with appropriate PPE verified the absence of chemical incompatibility risks in the flammable storage cabinet. Five (5) bottles of solvent were broken inside the storage cabinet. The broken bottles contained acetonitrile, acetone, pentane and triethylamine. The chemicals were cleaned by the laboratory’s spill response team.
Due to the increased solubility of gas in cold solvents, the acetone, as a result of contact with the dry ice, became saturated with CO2 gas. As the acetone warmed in the capped bottle, CO2 was liberated causing the capped bottle to become over-pressurized and ultimately fail. Glass projectiles from the acetone bottle failure caused the breakage of the four (4) additional bottles of solvent.
Failure to allow acetone saturated with CO2 to return to room temperature and off-gas prior to re-bottling.
- Assumption - Acetone solvent when absent of dry-ice crystals is safe to rebottle.
- Unexpected condition - CO2 is highly soluble in acetone at low temperatures.
- Time Pressure - Acetone was not allowed to return to room temperature because of the late time of day.
Inadequate hazards analysis - Failure to identify CO2 solubility in dry ice/acetone bath resulted in failure to realize pressure hazard when acetone was rebottled and capped.
Corrective actions to prevent reoccurrence
The incident report will be shared with management and laboratory staff.
Texas Tech Laboratory Explosion
Chemical Safety Board Case Study - October 2011
On January 7, 2010, a graduate student within the Chemistry and Biochemistry Department at Texas Tech University (Texas Tech) lost three fingers, his hands and face were burned, and one of his eyes was injured after the chemical he was working with detonated. The Chemical Safety Board (CSB) investigated and found systemic deficiencies within Texas Tech that contributed to the incident: the physical hazard risks inherent in the research were not effectively assessed, planned for, or mitigated; the university lacked safety management accountability and oversight; and previous incidents with preventative lessons were not documented, tracked, and formally communicated. The lessons learned from the incident provide all academic institutions with an important opportunity to compare their own policies and practices to that which existed at Texas Tech leading up to the incident.
Looking beyond Texas Tech, the CSB identified a lack of good practice guidance recognized by the academic community; limitations in using the Occupational Safety and Health Administration’s (OSHA) Occupational Exposure to Hazardous Chemicals in Laboratories Standard (29 CFR 1910.1450) as guidance for mitigating physical hazards in the laboratory; and a missed opportunity for a granting agency to infl uence safety practices. While a vast number of references, standards and guidelines have been developed to describe and promote different types of hazard evaluation methodologies in an industrial setting, similar resources that address the unique cultural and dynamic nature of an academic laboratory setting have not been generated. Good-practice guidelines would provide universities a metric to evaluate their current hazard evaluation procedures against, or for schools with none in place, would enable a more rapid process for their development. If universities choose to use OSHA’s Laboratory Standard as guidance for developing a plan to mitigate chemical hazards, they need to understand that the standard was not created to address physical hazards of chemicals, but rather health hazards as a result of chemical exposures. Physical hazards though, as evidenced in the Texas Tech incident and the 2008 laboratory fire that resulted in the death of a staff research associate at University of California, Los Angeles (UCLA), are deserving of similar attention to that given to health hazards. Finally, the granting agency, which provides funding for the research and thus maintains a level of control and authority over the researchers, did not prescribe any safety provisions specific to the research work being conducted at Texas Tech until after the incident occurred.
The full report
addressing key learnings, effects of the incident, a description, causation, and corrective actions to prevent reoccurrence.