In the #SheriSangji case: Sufficient training and oversight?

Is an experiment with an air sensitive catalyst an appropriate way to gauge experimental skill and technique to handle a pyrophoric reagent? That appeared to be one of the arguments that the defense attorney of University of California, Los Angeles, chemistry professor Patrick G. Harran was setting up last month in a court hearing.

Harran faces felony charges of labor code violations relating to the death of researcher Sheharbano (Sheri) Sangji. Sangji died from injuries sustained in a 2008 fire in Harran’s lab that started when she was handling tert-butyllithium.

C&EN and the Safety Zone covered the preliminary hearing in Harran’s case. One of the charges centers on failing to provide chemical safety training. In cross-examination of prosecution witnesses, Harran’s defense attorney, Thomas O’Brien, seemed to be building the assertion that Harran had provided sufficient training and oversight by watching Sangji do an earlier experiment involving a Grubbs II catalyst. From Sangji’s lab notebook, here are the experimental details:

Oct. 14, 2008, experiment that Harran observed

  • Air-sensitive reagent was a Grubbs II catalyst, which loses potency on exposure to air
  • Working in a glove bag, Sangji added 63 mg of the catalyst to a 50-mL flask. She then added 2.5 mL of 1,2-dichloromethane 1,2-dichloroethane to the flask, followed by 250 mg of vinyl glycine dissolved in 2.5 mL of 1,2-dichloromethane 1,2-dichloroethane and 256 mg of undecen-1-ol simultaneously over 20 minutes. She lowered the flask into an 80 ºC oil bath and stirred it under reflux for 20 hours. She sampled the reaction solution to run thin-layer chromatography at 16 and 20 hours. She filtered the solution and then purified it on a silica gel column.
  • Sangji’s notes aren’t clear whether this entire process was done in a glove bag or just the step of weighing the catalyst.

Dec. 28, 2008, experiment that started the fire

  • Air-sensitive reagent was tert-butyllithium (tBuLi), which ignites spontaneously in air
  • Sangji was scaling up an Oct. 17, 2008 experiment to produce 4-hydroxy-4-vinyl decane. The first step of the synthesis was to generate vinyllithium. In October, she added 28 mL anhydrous ether and 3.0 mL vinyl bromide to a 200-mL flask. After stirring the mixture for 15 min at -78 ºC, she added 54 mL of 1.67 M tBuLi. She stirred the mixture for two hours, moved it to a 0 ºC bath for 30 minutes, and took it back to -78 ºC. She then used a double-tipped needle to transfer 3.90 mL of 4-undecanone in 6 mL ether to the vinyllithium solution. She stirred the solution for two hours, then quenched it with sodium bicarbonate. She put the quenched mixture in a separatory funnel, collected the organic phase, dried it to remove residual water, and rotovapped it to remove the solvent from the product.
  • Sangji doesn’t say it in her notebook, but she was probably not working in a glove bag to do this reaction. Going by what she did in December, she was more likely working in a hood, running nitrogen lines to the tBuLi bottle and reagent flask, and using a syringe to transfer tBuLi from one to the other.
  • Sangji scaled up this experiment three-fold in December and used a 60-mL syringe for the tBuLi transfer. We know that she did not clamp the bottle, and so was likely holding it upside down in one hand while manipulating the syringe in the other. She was probably on her second or third transfer, reusing the needle and syringe, when the syringe plunger came out of the barrel, exposing the tBuLi to air and starting the fire. Sangji’s clothes caught fire and she was burned on her thighs, torso, arms, and neck.

What say you, Safety Zone readers? Was a 63-mg Grubbs II experiment an appropriate one by which to gauge Sangji’s skills and technique to handle tBuLi at the 54- or 160-mL scale?

Author: Jyllian Kemsley

Share This Post On


  1. Is the yield of the Grubbs II reaction probative as to her technique? As I recall, it’s quite low. (Don’t know if that’s because of the substrates, or air. Probably the substrates.)

  2. I don’t think you can equate the handling of 63mg of Grubs II with the dispensing of 160 mL of a pyrophoric solution.
    This especially as she used a plastic syringe, non gas tight? The plastic probably was swollen by the solvent making each successive transfer more difficult, thus requiring more force on the plunger to syringe the material, ultimately pulling the plunger from the barrel.

  3. @CJ – I haven’t done the math to figure out her yield for that reaction.

    @Quintus – Yes, plastic syringe.

  4. The techniques and the dangers are completely different. I was a pro at Schlenk techniques before I had used tBuLi for the first time. I still conferred with a post-doc and had him look over my shoulder the first time did the transfer. If I had to transfer 160 mL of the stuff, I would opt for a cannula. That is way too much tBuLi to transfer by syringe.

  5. Well for smaller quantities of such materials I always used Hamilton syringe, gas tight and you have a hard time pulling the plunger out.
    For larger amounts I used a cylinder and pumped the material out of the cylinder using nitrogen pressure through the valves.
    By the way I worked in chemical development so I was used to such large transfers.

  6. The correct use of a glove bag is not the same as the correct use of a syringe for the transfer of air-sensitive materials. I would agree that both skills are needed to transfer air-sensitive materials but it would be inappropriate to assume that the trainee can successfully use a syringe to transfer an air-sensitive material just by observing their glove bag technique. However, if professor Harran observed the October 17th experiment and evaluated Sangji’s technique as perfect when transferring the 54mL of tert-butyllithium then I would say it would be appropriate to say she was trained in a syringe transfer.

  7. “1,2-dichloromethane” is not a chemical name, used twice there.

    No of course it is not in any way appropriate training for large quantity of t-BuLi. They are desperate and hoping that the general ignorance of most people will let them get away with this.

  8. She did more than a couple of things *really* wrong that that makes it clear that she either had very little understanding of how to do this safely, or she was taking terribly dangerous shortcuts. I suspect more of the former, but it’s just a hunch.

  9. @swihart – Whoops, yes, of course 1,2-dichloromethane is wrong. I was transcribing from the notebook and clearly not paying enough attention to what made chemical sense. Post updated!

  10. @silane – Harran did not observe the Oct. 17 tBuLi experiment. According to what he told investigators, he did not know at the time that Sangji was even doing it, although he learned of it later. Regarding the Dec. 29 experiment, Harran told investigators that he generally knew what Sangji was doing but didn’t discuss it with her in detail.

  11. As regards the safety training. I worked in industry where safety was a big issue. Safety audits once a month, regular training mandatory for all lab workers, critical evaluation of experiments especially on scale and so on. Form what I remember of my university time this did not happen. Safety was handed down form year to year in the post-grad. lab. That is, it was never formalised.
    Sadly I don’t think much has changed in the intervening years unless, of course, a tragic accident happens, which makes everyone sit up and take stock. But even then it will slip back into the background as the pressures of an academic lab. mount to publish, get results, bring in money etc.
    So safety in the academic environment tends to take a back seat and it’s up to the individual to look after his/her own safety.

  12. “So safety in the academic environment tends to take a back seat and it’s up to the individual to look after his/her own safety.”

    This is pretty effective and concise summary, but are we just accepting that as THE WAY IT IS, or is there some hope to change it? It’s not the way the so-called real world operates, and so why should it be the way the ivory tower operates?

  13. Swihart has already said it – is this a joke with the Grubbs II? It better be! *shakehead*

  14. It should not be “accepted as the way it is”. If one thing comes out of this I hope there will be a change in the academic perception of lab. safety issues.
    But, usually the experiments are small scale, even to bringing through more starting materials. So should safety considerations be included? I think they should. This was around a 250 mmol reaction scale, at around a molar concentration of reagents. Not to concentrated, indeed quite dilute. but large for a research lab. Perhaps they did not see the requirement for a risk analysis. Wrong in my view, especially as I would regard the ether to have a greater hazard potential, especially in combination with pentane. (tBuLi in pentane?).
    Risk analysis is a vital part of any experiment and should be documented with the lab notebook.

  15. Not even close. In a corporate lab, any use of something as dangerous as t-butyl lithium, in any amount, would require a safety review and written procedure, even for people with dozens of times Sanji’s experience. At the scale she was working at, a second review would almost certainly be required, involving several managers, a process chemist or engineer, etc, likely with a combined lab experience of over a century.

    Academic lab safety is terrible. Harran’s lab was even worse than normal, and the buck stops at his desk.

  16. This one is easy – NO!

    And I will repeat what I wrote here once before – the Aldrich bulletin is woefully inadequate and should not be used in place of training or as a primary training document.

  17. “No” doesn’t even begin to cover how non-comparable those two experiments would be (Note on the Aldrich bulletin: even the post-incident UCLA training video is atrocious.) An idividual of her level of chemistry experience should never have been handling that chemical in the first place. Any suggestion to the contrary is either coming from someone woefully ignorant or someone with a vested interest. Whenever I have worked with that chemical (about 30 times in 20+yrs …along with trimethylaluminum)I have two well trained people flanking me – one with a fire extinguisher.

  18. Jyllian,
    Please feel free to remove this post as it does not respond to the topic. But I would appreciate it if you took a look and perhaps commented in your blog as to why the ACS is propagating this. Thanks.


  1. My Website – Time for a Better Safety Culture - [...] wake of the recent happenings in the Sheri Sangji/Patrick Harran case, I wanted to take a few minutes to …