A recent question on poisons was answered in part by bringing up the concept of $\pu{LD_{50}}$ and animal testing and so on; none of that was new to me, but it did bring to mind a question I don’t know the answer to:

Has there ever been a substance that scientists have actively tried to determine $\pu{LD_{50}}$ for, and could not because they were unable to determine a dosage low enough to kill “just” half of the subjects?

Please note that this question is not restricted, strictly, to any technical definition of “poison.” Rather, answers must meet these requirements:

  1. There must be research (considering that I am looking for something that might be a negative result and those do not get published as much as they should be, I won’t demand peer-reviewed publication, but the goal should have been to attempt peer-review publication if it had worked out—effectively, I mean serious work), and

  2. that research must use the term “$\pu{LD_{50}}$” to characterize the substance.

Basically, I don’t want to get into debates about what is, or is not, a poison here. If a researcher is willing to call something $\pu{LD_{50}}$, then I am willing to accept it as a “poison” for the purposes of this question.

I also don’t want speculation, or for a user here to characterize something as $\pu{LD_{50}}$ when the underlying research doesn’t call it that. It is not enough to say, for example, “well I’m sure even a single atom of antimatter inside your body would be pretty bad,” you need to cite a particular researcher who has performed experiments with the goal of determining what they themselves called $\pu{LD_{50}}$ for that substance.

I suspect the answer is no, but I have no idea how to research something like this.

  • 1
    $\begingroup$ Comments are not for extended discussion; this conversation has been moved to chat. $\endgroup$ Commented Jan 16, 2018 at 17:56

2 Answers 2


Of course no. Botulotoxin is probably the strongest known, and still its $\rm LD_{50}$ is counted in nanograms per kilogram, which is pretty manageable. Sure, working with such tiny amounts requires some special measures, but still, it is way greater than one molecule. You can divide it again, and again, and again.

So it goes.

  • $\begingroup$ Thanks; I’d heard botulotoxin referred to as the strongest, but I wasn’t sure that there might not be caveats to that, strongest with the $\pu{LD_{50}}$ actually measured or something. $\endgroup$
    – KRyan
    Commented Jan 16, 2018 at 16:08
  • 6
    $\begingroup$ "one prion molecule makes two - those two go on and make four, those four make eight, and so on, until the whole brain is just one liquifiied spongy mass. " blogs.scientificamerican.com/a-blog-around-the-clock/… I wouldn't rule out one prion molecule delivered in a particular way killing (slowly) more then 50% of the people. $\endgroup$
    – DavePhD
    Commented Jan 16, 2018 at 16:57
  • 4
    $\begingroup$ @vaxquis that kind of "if" does not an LD50 make. $\endgroup$
    – hobbs
    Commented Jan 16, 2018 at 18:05
  • 1
    $\begingroup$ Seeing as I’m the one who asked the question, I think I’m entitled to determine what it’s asking about, not you. Yes, I understand that $\pu{LD_{50}}$ is typically used for toxicants, would not generally be used for pathogens, etc. But the definition of $\pu{LD_{50}}$ really is compatible with anything that might, at some dosage, kill 50% of subjects. And while scientists would almost never use it that way, almost never is not the same as never. If a researcher decided to use $\pu{LD_{50}}$ as a measure of a substance’s danger, that is good enough for me here. $\endgroup$
    – KRyan
    Commented Jan 16, 2018 at 18:10
  • 2
    $\begingroup$ Botulotoxin doesn't reproduce. I think that's quite a solid line. $\endgroup$ Commented Jan 16, 2018 at 20:44

Some options that do not work, but have been suggested:

  1. Prions: looks like LD50 has been successfully determined for the most common Sc237 and 263K prions: Prions: A Challenge for Science, Medicine, and the Public Health System. Though it is also referred to as "Infectious dose" as opposed to "Lethal dose".

  2. Polonium: While some sources claim that polonium cannot be characterized by LD50, this is not due to high lethality but just because the effects vary depending on what kind of cancers or acute radiation disease the radiation causes. Actual LD50 is around 90 ng for humans for acute radiation disease.

  3. Antimatter: looks like around 1 ng would be plenty to kill a person. LD50 would be much lower, but probably easy enough to measure anyway. Real problem in the experiment would be delivering the antimatter without premature annihilation.

But in reality, it just depends on the means available to the scientist trying to determine the LD50. For example in this study, "For the remaining solvents LD50 could not be determined due to volume limitations, and have been given a value of <1.0 ml/kg."

The limit of current science would have to be a poison that would kill with a single molecule, as it is quite possible nowadays to manipulate and separate single molecules. Or otherwise single indivisible unit, such as a single cell.

  • $\begingroup$ Please limit your answer to only those substances that have had actual research into determining their “$\pu{LD_{50}}$.” So that book on prions might work (though it’s unclear to me since only one page appears to use the term and I don’t get to see enough context to see if they’re actually discussing the $\pu{LD_{50}}$ of the prions), maybe polonium. Also, things that do have an $\pu{LD_{50}}$ aren’t actually answers to the question; they could be useful if you can prove they are as deadly as things can possibly get (suggesting that the answer to the question is no). $\endgroup$
    – KRyan
    Commented Jan 16, 2018 at 20:59
  • $\begingroup$ @KRyan, well, if you want to be nitpicky like that, I'll just point to the second last paragraph that fits the nitpicky interpretation of your question by having a group of scientists who were actively trying to determine LD50 and couldn't. $\endgroup$
    – jpa
    Commented Jan 16, 2018 at 21:20
  • $\begingroup$ I don’t consider it nit-picky at all; a question open-ended enough to allow random speculation like antimatter here should have been closed as being overly-broad and/or primarily opinion-based. Anyway, yes, the second paragraph could work, except I have absolutely no way of confirming your assertions here. A link to the work would improve matters tremendously. Note the updates to the question clarifying matters. $\endgroup$
    – KRyan
    Commented Jan 16, 2018 at 21:22
  • $\begingroup$ Yeah, paywalls suck. If you want, searching for that phrase on google scholar should (I think) show that it does exist there even if you don't have access to the full text. $\endgroup$
    – jpa
    Commented Jan 16, 2018 at 21:26
  • $\begingroup$ This is about #1 again? Yeah, I tried that, only got one hit (and one false-positive for a citation abbreviated as LD), which did talk about $\pu{LD_{50}}$ but it wasn’t clear what substance was being discussed in that section, and I couldn’t get at the surrounding context. Certainly not enough to see that they attempted to determine an $\pu{LD_{50}}$ but could not due to being unable to isolate a small enough dosage. $\endgroup$
    – KRyan
    Commented Jan 16, 2018 at 21:32

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.