Hacker News Reader

Comment by dnautics | Hacker News Reader

I do think derek lowe is a bit too paranoid about HF. I've worked with HF before, and with a proper setup (negative pressure, calcium carbonate filter, teflon equiment) it is really quite reasonable to handle. I would entrust a conscientious first year grad student to it (as I was entrusted). I only screwed up once [1], and the consequence was we lost about 1/8 of our calcium carbonate filter column. You can tell because the section that has most recently been expended is warm on the outside of heavily shielded carbonate vessel).

[1] the screw up was I had left the HF open (but not pulling) for 20 minutes instead of 5.

View on HN

I think you're not paranoid enough about HF. It doesn't wash off, and is small enough to pass through skin to form insoluble mineral precipitate with calcium in bone and blood, which can result in amputations from dead tissue. It's one of the few acids capable of eating through silicates in glass or rock.

Even at lower concentrations, you still need specialized gloves, apron, and ideally a face shield, in addition to standard PPE. Source: worked with HF for 3 years.

Agreed.

If I had to choose splashing my arm with a concentrated acid, I would choose sulfuric or nitric acid over HF.

Unlike the other acids that just burn the surface of the skin, HF rapidly penetrated the skin and can mess up your potassium levels. Plus tissue destruction is much deeper.

That's why you keep calcium carbonate salve around. Also you're more vigilant around hf. I've spilled conc sulfuric and nitric acid on my lab coats. HF will never get that close, as safe equipment is in a closed loop but not that complex or byrdensome (it's not like explosives research for example)

There are some idiots that use open air HF for etching. That is a very bad idea.

I should clarify by open in the description I meant valve open. The system was still closed.

But who had to carry the jar of HF into the lab and attach it to the closed system?

It's a gas canister, there are extremely secure valves. I've done it myself.

As in, a highly pressurized gas canister? In a container that is presumably 100% impervious to HF corrosion, yet somehow also made out of a material that can withstand those pressures?

How is that safer than neutral-pressure, aqueous HF? I'm perfectly comfortable with bottled HF, to the point that I literally handled it in the dark (light-sensitive samples), but I'd never set foot in the same room as a pressurized HF gas canister. A spilled HF bottle is a major incident, but ultimately straight forward to contain, whereas an HF canister leak would kill off an entire lab.

I've done enough manifold setups and worked with enough gas cylinders to understand that they can leak in all kinds of ways, even with a wealth of experience. As I said initially: you are definitely not paranoid enough about this setup.

This also begs the question of whom exactly has an unobtainium pump that can withstand HF to put it into a cylinder.

I'm just going to say that your ignorance about HF is exactly my point. 100% HF is obtainable commercially in labs and only requires shipping precautions on the order of what you would expect for toxic volatiles.

Derek covers this in one of his articles. [1]

>It’s sold compressed in metal cylinders, like other gases, but it needs some care in packaging. The stuff is so corrosive that special alloys need to be used, usually ones high in nickel. If you stick an ordinary gas regulator on top of an HF cylinder, you’re in for trouble, and the complete destruction of the regulator is the least of your worries.

>HF has actually been used right out of the cylinder for a long time in Merrifield peptide synthesizers.

[1]: https://blogs.sciencemag.org/pipeline/archives/2004/03/03/th...

Fair enough. You’re using it in a certain way that makes accidental exposure less likely.

I was speaking more generally. It’s not uncommon for labs to handle concentrated HF in a similar manner as other acids - I.e. pouring it into a beaker in a fume hood

Yes this is a mistake but to categorically say that HF is scary and dangerous actually creates, in my opinion, a mindset that is detrimental to operator's of safe HF equipment and also encourages use of more wasteful (e.g. FMOC) techniques which also have tradeoffs (nonsynthesiazbility is more of a problem with FMOC than BOC).

And I was clear that setup matters in my gp.

The irony is that HF is probably worse in lower concentrations partly due to it's dissociative properties, which I won't get into a technical discussion about, and general human stupidity. If you're using 100% HF it's safer, because you're using not very much (we're talking 5-10 mL at a time), the equipment you're using is by nature safer, and you're more vigilant. It's also much easier to mechanically handle 10 mL than 100 (or more mL).

But hey I worked with 100% HF for seven years (BOC protected amino acid synthesis), so what do I know...

Only if we ignore that HF is volatile, meaning that you're dealing with much more (not dissociated) HF evolving from your reagents. Even with a fume hood, you've got much more HF in your face versus a 5% concentration. Calcium gluconate gel will do nothing for HF in your lungs.

I disagree. If something happens with an apparatus, you fucking shut the hood quick. There is very fast positive pressure air blowing in through the sash opening, the apparatus is bolted into the wall, sashed above the nonglass barrier, and there is negative pressure inside the apparatus, so the tendency is to not leak out but leak in.

If you have a 5% aqueous solution most of that HF is still binary undissociated HF, which will leak into the atmosphere so you could very easily unawaredly be breathing it in if you are working with it in any open air situation below the sash line (which is what the tendency is when you're working with aqueous concoctions).

I fail to see how quickly shutting the hood in cases of catastrophic failure address my concerns. When you're dealing with pressure differentials and compressed gasses, explosions and implosions are a very real risk, and shutting the fume hood, even quickly, will not help in the fraction of a second these failures would take to fill your lab with aerosolized HF and shrapnel.

Even working at negative pressures, you can easily end up with a positive-pressure manifold if your sample is unexpectedly evolving gas at a higher than usual rate, or your valves aren't configured as intended. To be clear, these aren't academic exercises; I've actually seen all of these happen. With closed systems, you have to anticipate explosive pressures and plan accordingly. HF is incompatible with that kind of failure.

https://www.chem.purdue.edu/chemsafety/chem/HFfatality.html

I thought of the same case.

Guy spills a beaker (100-250 mL) of HF on his thighs. They end up amputating his leg but he still dies anyways.

Yeah you should never use open air HF. If you are in a closed system it is very safe.

You're saying that as long as you do everything perfectly all the time you're (mostly) safe. Real life proved again and again that there no such thing as "all perfect all the time". Mistakes will happen and when they do HF is far less forgiving than other instruments of the devil.

It's not a matter of perfection, it's a matter of tooling. It's not like saying "you're safe from segfaults in c if you program correctly", it's like saying "you're safe from segfaults in erlang".

Yes, mistakes will happen. Yes, you can segfault erlang. Yes there are tradeoffs. But, I can have a junior deploy to prod confidently in a BEAM language in a few months, I can have a first year grad student use HF confidently in a few weeks.

A segfault is a segfault and for all intents and purposes it's the same regardless of the language you used to make the mistake. If a segfault in Erlang guaranteed that you lost a hand would you want to avoid it?

On the other hand (pun intended) chemicals intrinsically make a burn more or less serious. It's the difference between getting a rash or dying a horrible death.

In a risk matrix the probability has less weight than the severity. So a "catastrophic" severity will make the risk shoot up faster than even "certain" probability. Regardless of tooling and precautions taken the impact of the mistake makes a difference in any field and will make people want to completely steer away from some.

I think it can be summed it up like this: If you were tied to a chair and someone was about to drip something on you what would you rather it be: HCl or ClF3?

The severity of a plane crash is rather high. Do you prefer walking?

Well... yes. Plenty of people do (and actual alternatives to flying) specifically because they consider flying too dangerous, uncomfortable, unwieldy, etc. and everyone agrees a crash is unlikely but catastrophic.

But you're being pedantic now. Nobody takes the plane to a place they can walk to. Many people don't take the plane even to a place where they can go by train or car. I understand that you may have to work with something like ClF3 but what everyone is telling you is that they'd rather avoid it if they have a choice. Just like most people would avoid working with unstable explosives, dangerous pathogens, extreme sports, and any very dangerous job in general even if all of these have their own set of precautions.

As the danger level increases people justifiably would rather avoid that specific task. Precautions are a mitigation but overall the sentiment stays, the job becomes much more stressful and the outcome of a mistake more severe.

The more dangerous the task and more severe the consequences of a mistake, the fewer the people that would want to deal with it. Take from this what you will.

Reply on news.ycombinator.com