Issue Date: October 5, 2009
Lessons Of The Tragedy At UCLA
In reading the tragic tale of Sheharbano (Sheri) Sangji's death resulting from a series of events culminating in the ignition of the pyrophoric reagent tert-butyllithium, I saw several valid points raised regarding the need for better control of laboratory safety in academic environments (C&EN, Aug. 3, page 29). Nowhere, however, did I see anyone ask why an extremely dangerous reagent was used to generate an organometallic compound (vinylmagnesium bromide) that can be purchased.
I have supervised synthetic chemists for more than 20 years. Some of them have had no practical experience, and others have had several decades' worth. Whatever their level, I have always enforced one common rule: Buy your bonds rather than make them whenever possible.
Sometimes this isn't feasible, usually for budgetary reasons. Sometimes synthesis is undertaken to teach inexperienced chemists good techniques. Both of these are valid reasons for making compounds in the laboratory, but there must be limits. If money is the problem, one would do well to consider the legal costs and fines for violations of safety rules versus the cost of a premade reagent.
Researchers should focus on the goal of the synthesis and ask if there is a safer way to achieve it. There are enough methods out there, and just because a synthesis method is published in the literature does not make it the best. Anyone who has attempted to reproduce a literature procedure knows how mercurial (no pun intended) the art of organic synthesis can be. If a reagent is simply too dangerous or toxic to use, find another way.
James T. Palmer
Templestowe, Victoria, Australia
The article "Learning from UCLA" tells me that many chemists need to better understand the importance of laboratory safety showers. Immediate use of a safety shower is the best option in a clothes-on-fire situation. A safety shower not only puts fire out, it also removes heat the fire has deposited in tissues, preventing further damage after the fire stops. Personal protective equipment (PPE) goes only so far. Hand burns occurred in spite of glove use during the procedure described. Even the model researcher shown in the photograph accompanying the C&EN article should locate her safety shower before she finds herself in a hair-on-fire situation. Chemists need to be aware of proper emergency response procedures as well as PPE.
Unfortunately, safety shower use has not been prominent in discussing this accident. While the text of the California Fatality Assessment & Control Evaluation (CA/FACE) Program report on the UCLA accident cites the need for better emergency response procedures, including shower drills, no reference to safety showers appears in the summary box of that report. The CA/FACE text discussing hypothetical shower use seems poorly written. Standard operating procedures for pyrophorics, even those cited on the ACS website, make no mention of proper emergency response procedures.
Perhaps there is confusion about the relative risk posed by a small amount of water-reactive material in a shower compared with the risk posed by being entangled in a much larger amount of material burning uncontrolled. If there is any question that the standard chemist's response to clothes-on-fire—that is immediate use of the safety shower—did not apply in this accident, data need to be produced to justify that position.
James W. Lewis
Santa Cruz, Calif.
The articles on the recent terrible accident at UCLA, have yet to explain exactly why tert-butyllithium is so dangerous. Is it because it is supplied in pentane? Rapid evaporation is said to lead to a pyrophoric solid. In contrast, n-butyl- and sec-butyllithiums are more usually supplied in hexane or cyclohexane and are apparently safer. Should tert-butyllithium be supplied in these solvents instead of pentane?
H. K. Hall
Jyllian Kemsley's story hit at all relevant points about safety practices and mind-set in university laboratories; however, the working environment in university laboratories is never discussed. Many faculty members expect (and in some cases require) their students, especially graduate researchers, to work extremely long hours, often on weekends and university holidays. In such a setting, safe practices can very easily take secondary priority.
Personally, I have friends at two well-known universities in Texas whose faculty advisers have mandated a workweek of 60 hours or more. Some advisers schedule regular weekly research group meetings late at night or on weekends. As a result, students are left with no choice but to work seven days a week. In many instances, the situation is worse for international students who typically depend more on their faculty advisers.
Imagine a researcher working on Saturday and following all the safety precautions, but an instrument failure leads to an accident. There will be no university official present and, in many cases, university gates or the department will be closed as well. Sangji was carrying out a potentially risky reaction on Dec. 29, the Monday following the Christmas holiday and two days before New Year's Eve. Although there can be no compromise on safe practices, faculty also need to ensure that their group members enjoy the work atmosphere and take responsibility for their actions.
I hope that universities will care for the well-being of their students, especially graduate students, whose work-life balance is often overlooked. They are the workhorses of academic research and are often unknowingly ill-treated.
"Learning from UCLA" is one of the most important articles published in recent years. When coupled with her previous articles in C&EN on the death of Sangji, Kemsley has struck a chord with me.
As a private investigator of fires, explosions, and chemical releases, the focus of this investigation should be on the lessons to be learned. The investigations may take more than one year. Although the family and loved ones wanted the answer immediately, the painstaking collecting of evidence, review of the facts, testing the facts, and then conclusions drawn (based on these facts) take time. Patience is a virtue, especially when the goal is that this needless death not be repeated.
Many things can be done to enhance chemical safety, especially in the university laboratory setting. While matters have changed dramatically since I was Sangji's age, a basic understanding of both what you want to do and the safest way to successfully do a procedure are critical. The young, relatively inexperienced scientist must not be afraid to ask questions. This is how we learn.
As an investigator, I often find that someone can be at fault for not having read the instructions or warning labels. Just because you've done something "umpteen times" doesn't eliminate an unexpected outcome.
In working with professionals in the fire service, maintaining open dialogue, especially where insurance claims or litigation could be in the offing, is essential. While many firefighters do not have the formal education C&EN readers have, they are the ones who enter hazardous areas. We have an obligation to listen to them and to understand what they are telling us.
Those who work in an environmental health and safety regulatory capacity are our partners, not adversaries, in ensuring that the workplace is safe. The "Right To Know" laws and the Occupational Safety & Health Administration (OSHA) have been part of safe industrial practice for more than 25 years. I am pleased to see easily accessible Material Safety Data Sheets in many sites that I visit.
Mitigating potential chemical hazards ought to be part of a student's training. Proper handling of chemicals based on respective flash point, explosion/flammability limits, auto-ignition temperature, etc., can be just as important as the percent yield. The system Kemsley describes for student research in the U.K. meets this criterion.
David M. Manuta
I appreciate C&EN's in-depth coverage of the UCLA tragedy. I have a recommendation from my 30 years of experience in chemical safety at research universities. Before working with radionuclides in an academic lab, the director (principal investigator) must submit a project-specific plan to the institutional Radiation Safety Committee for review. The committee ensures that the plan assesses risks, is complete, and includes safety procedures. Committee approval requires that lab staff be trained and rules be followed. The plans are updated and reviewed again every few years as projects change.
Similarly, before working with recombinant DNA in an academic lab, an experiment-specific plan must be approved of by the Institutional Biosafety Committee, following a nearly identical system. At most institutions, work with pathogens requires the same review, approval, and oversight.
This safety system works in academia. It makes people think before handling hazardous materials, so they take extra, appropriate precautions. It reduces accidents and injuries. It makes sure that minimum standards are met. These president-appointed, mostly faculty committees make and enforce institutional safety rules. Top-to-bottom accountability for safety is clear from the start. For decades, this safety system has fit well with the collegial, self-governing culture of higher education.
OSHA requires that laboratories have a chemical hygiene plan. A chemical hygiene plan is an excellent tool for risk assessment, documenting safety procedures, and training. Although the American Chemical Society supported this performance-based requirement in 1989, it was never embraced by chemists. The C&EN article reviewing the UCLA accident doesn't even mention it.
At most academic institutions, there is one generic chemical hygiene plan for the entire institution. It is rarely updated. As was the case at UCLA, the chemical hygiene plan is not laboratory- or project-specific. At most academic institutions, there is no institutional chemical safety committee. Some state universities are not subject to OSHA and have no chemical hygiene plan at all.
I'm not suggesting more regulations or OSHA enforcement. I'm suggesting that we chemists take the lead in using the chemical hygiene plan to its full potential, in a system that works in academia.
Peter A. Reinhardt
New Haven, Conn.
Congratulations to Kemsley for superb reporting on Sheri Sangji's tragic accident and death. One issue not addressed was why Sangji died. One might suppose legal issues are involved.
tert-Butyllithium is a spectacularly pyrophoric reagent, but it was probably only the trigger in this incident. The "bullet" that caused her death was the ignited spilled hexane. The safety shower that could have lessened her injuries was used neither by Sangji nor by either of the two fellow-chemists who responded to her cries.
Why? Possibly because chemists are never trained to use the shower because it produces a huge volume of water of uncertain quality without a drain and is, thus, too messy to demonstrate. Also, the behavior of tert-butyllithium and other instantly pyrophoric compounds can't be appreciated until it is seen. Potential users should be trained by deliberately allowing a small volume to be exposed to air and ignite. Also, we all should see a small volume of hexane deliberately ignited to appreciate the intensity of the fire. It is important to prepare the mind.
Stephen T. Ross
Upon reading Jyllian Kelmsley's article on the death of research assistant Sheharbano (Sheri) Sangji in the UCLA chemistry department, I thought immediately about the risks to students in my own laboratory.
For any reactions that are sensitive to water or oxygen, we use an argon glove box with a recirculating vapor trap. It is possible to transfer tert-butyllithium and other pyrophoric liquids in a fume hood; however, it is also possible for small deviations from best practices to result in tragedy.
It seems to me that an engineering solution is called for to minimize the risk of laboratory fires due to spills of pyrophoric liquids. No amount of training can ensure that pyrophoric liquids will never leak during transfer. I suggest that organic labs handling pyrophoric liquids be equipped with at least one inert-atmosphere glove box to preclude accidental exposure to air; otherwise, such a tragedy could be repeated.
I was moved by the comment, "Lab workers have to have the mind-set that something can always go wrong." That mind-set starts with a culture of safety awareness set by the lab's management. When safety inspections are done only one time per year, and deficiencies of a given inspection are not corrected promptly, what message is being sent to those lab workers?
Of particular concern in the UCLA accident is what appears to be copious amounts of flammable solvents cluttering a lab that is using pyrophoric reagents! It's not difficult to clean up lab clutter and store solvents in appropriate cabinets. When immediate steps are not taken to correct such simple and obvious deficiencies, it is easy to see how the mind-set of the lab worker can shift from "What can go wrong" to "It won't happen to me."
It was stated that for "unknown reasons, the syringe plunger came out of the barrel." I know the reason, because the same accident happened to me. Extreme force is needed to fill a 50-ml syringe with (sometimes viscous) liquid through a very small bore needle, like the 20-gauge one described. Even the elaborate setup shown in the picture on page 31 is not foolproof.
Use of the gas-tight syringe shown in the photo will not always prevent accidents; with reuse, I have had the Teflon head of the aluminum plunger come off because liquids have gotten inside the head and corroded it. I generally use the simple polypropylene/polyethylene syringes with the offset Luer tips, and I have had other accidents peculiar to that design. The head of one reused plunger broke off while I was pulling on it, causing my hand to jerk upward and for me to lose control.
During another syringing, the needle separated from the Luer tip of the barrel, because the needle was held to the Luer tip only by friction (there was no threaded connection). Pulling the plunger so far up that it is at the head of the barrel is risky. Tightly grasping the barrel of plastic syringes in the palm of one hand while pulling upward with the other actually makes pulling more difficult because the plunger won't move easily. Applying downward pressure with the thumb and forefinger on the lip of the barrel with one hand while pulling upward with the other hand doesn't allow for good control of the barrel.
To avoid using extreme force, a larger diameter needle must be used. Also, there should be an attachment to the syringe (I don't know of the existence of any as of yet) that prevents the plunger from being unintentionally pulled out of the barrel. No one ever showed me how to syringe properly and safely, nor was I taught how to choose a needle. Was this unfortunate student taught, or did she pick it up like I did?
After reading the detailed investigation on the deadly ULCA accident involving tert-butyllithium, I feel sad that there has not been "enough information available to say for certain" the direct cause of this tragedy. From my personal experience with n- and tert-butyllithium, the two most widely employed organolithium reagents, I would like to point out that tert-butyllithium is commonly manufactured as a pentane solution, compared with hexanes for n-butyllithium.
While tert-butyllithium is more chemically reactive than the n-isomer, pentane is a solvent more volatile and flammable than hexanes, with a boiling point (36 ºC) slightly below the normal human body temperature. This means that holding a syringe barrel firmly for long enough could warm up the tert-butyllithium solution inside to reach a considerable pentane vapor pressure, which could have contributed to the unexpected ejection of the syringe plunger in the accident.
In addition, when tert-butyllithium ignites upon exposure to air, pentane serves as a best fuel to stoke the fire. To help minimize possible repetitions of such a disaster in the future, I suggest that both n- and tert-butyllithium be prepared and sold in solvents such as hexanes, cyclohexane, or even heptane, which have much higher boiling and flash points. I don't see an intuitive reason why n-butyllithium can be made in hexanes but tert-butyllithium cannot.
I just finished reading Kemsley's article "Learning from UCLA." Thorough reporting of the investigation will certainly help in fostering better safety awareness regarding highly pyrophoric materials. I feel compelled to write, however, because although there was a lengthy discussion of many of the conditions that contributed to the tragic outcome, the justification of the use of vinyllithium appeared rather tenuous.
I inferred from the article that the purpose of the experiment was purely preparative, so, although the use of that reagent was, according to the consultant Kemsley cited, "an acceptable way to approach the synthesis," I cannot but wonder why 1-M vinylmagnesium bromide in THF, a much less hazardous reagent than the 1.7-M tert-butyllithium reagent used to prepare the vinyllithium, had not been used instead.
Vinylmagnesium bromide, along with the corresponding chloride, are readily available from various vendors, are probably less expensive than the reagents used in the incident, and are known to add to simple ketones in nearly quantitative yield under the proper conditions, as could be confirmed by a 10-minute search of the literature. The consultant's opinion in apparent response to the suggestion of using a vinyl Grignard reagent that "side reactions would reduce the yield," would therefore seem to be irrelevant if not actually false. Even if true, I seriously doubt that a slightly higher yield could justify the risks inherent in preparing that reagent for the purposes of this work.
I believe the primary lesson from this tragedy is that the most important decisions regarding safety are made prior to entering the laboratory. In particular, before attempting to perform a synthesis using an extremely hazardous reagent—especially on a scale involving serious risk of fire, explosion, or exposure to toxic substances—one should always seek an inherently safer alternative.
Nothing is mentioned in Kemsley's article about the location of the nearest fire extinguisher, nor is information given about the proper method to fight a tert-butyllithium fire.
Since another person was in the lab and a second person in a nearby lab, why did neither of them try to use a fire extinguisher rather than pouring water on the victim? The MSDS says not to use water, but to use a fire extinguisher. Would a fire extinguisher have made a difference? The article does not make this clear.
I was sad to read the article on the recent fatality of Sheharbano Sangji, a research assistant at a UCLA chemistry laboratory, due to an accidental fire while she was performing a synthesis reaction.
I have spent most of my 35-plus-year career as a chemical engineer in the specialty gas/chemicals industry in a wide variety of positions (engineering, R&D, environmental health and safety, analytical, marketing, operations, and technical service). This has given me a unique perspective on materials such as arsine, phosphine, silane, chlorine trifluoride, trimethylaluminum, and more. Besides the typical chemical hazards of toxicity, reactivity, and flammability, the gases have the additional hazard of pressure.
One of my passions for the past 20 years has been to teach others how to handle these materials safely and what to do in the event of an incident. I have had the opportunity to experiment with many of these materials to better understand their behavior in a release and to design procedures and equipment to handle these. I have taught more than 6,000 people from private industry, fire departments, and government agencies in one- to three-day classes and have presented at numerous conferences.
Unfortunately, few attendees (less than 1%) have been from universities, yet these are the locations where I believe this training is most needed. In these laboratories, a wide variety of materials are being handled mostly by inexperienced graduate students who have yet to learn or remember the key safety rules of a laboratory. Some universities I have visited have excellent safety programs, but these are in the minority. Since the mid-1990s, I have been teaching a one-day MOCVD gas safety and ER course every two years at a major midwestern university, but these are the exception.
As the father of a daughter who will be a junior in a biochemistry program next year, this incident hit home. I have a concern with her safety as she proceeds to graduate school. As an undergraduate at a major state university with an excellent safety program, she is now regularly trained and her teaching assistants provide safety support during every lab. I will become concerned if she enters graduate school at a university that might not have the same level of safety and support.
Graduate school is also when she will start to experiment in areas that have not been explored and where safeguards have not been established. To do this safely, she will need to rely on the basic safety skills she has been taught as an undergraduate.
She can also be affected by the inexperience of others sharing the laboratory. I can't help but worry until she finishes school.
Eugene Y. Ngai
Whitehouse Station, N.J.
- Chemical & Engineering News
- ISSN 0009-2347
- Copyright © American Chemical Society