Measuring particles through experiments

[_PJ_]

How is it reliably ascertained that the "particles" 'measured' can be "proven" to have "traversed the experiment" and are not just conveniently (or inconveniently depending on perspective) incident / external?

Is this obtained through statistical analysis from a large number of repetitions, because I find it hard to accept that any individual case can be 'measured'.

 

[_PJ_]

Sorry but the moderator that moved this post to its own thread.The question was very specifically in context to "delayed choice experiments" with regards to determination that the detection can be reliably assured to be of something that has actually traversed the experiment overall.

I know that usually measurements are repeated and the statistical results are analysed across all sciences in order to enforce reliability and discard erroneous anomalies. This question was not in regards to the nature of reliable data collection as may be interpreted without context, but with regards to the implications given in descriptions of delayed choice experiments that a "particle" detected can be known to have actually participated in the experiment at all.

 

[PeterDonis]

How is it reliably ascertained that the "particles" 'measured' can be "proven" to have "traversed the experiment" and are not just conveniently (or inconveniently depending on perspective) incident / external?

Although you raised this question in the context of delayed choice experiments, the answer is the same for any quantum experiment: you record particles being emitted from one or more sources, and you record particles being detected by one or more detectors. The correlation between the source and detector records is what tells you that the particles you detected originally came from the sources, not from somewhere outside the experiment. The experimental arrangement itself is what tells you that particles emitted from the sources will traverse various other devices like beam splitters, etc.

 

[DrChinese]

How is it reliably ascertained that ... because I find it hard to accept that any individual case can be 'measured'.

The scientific team designs an experiment and executes it. If suitable, they publish the results. It is up to the reader to accept or reject those results.

Generally speaking, scientific papers are oriented towards other experts in the field. So they use techniques and standards that are well-accepted by expert consensus. However, there is no single authority on scientific matters.

In delayed choice experiments, the setup is described in detail and usually relies on the work of previous experimental teams (who are referenced). So the assumption is that you are familiar with, or will become familiar with, those references.

 

[bhobba]

Just to make 100% clear what's going on. Peter gave the correct answer - correlation. There is a 100% correlation in every experiment done so far between emitter and detector. So its assumed they traveled from emitter to detector. Strictly speaking in QM you can't say that - but when it is said what is really meant is correlation - its just understood.

Thanks
Bill

 

[vanhees71]

Well, yes, but one should also say that of course, any experiment needs both statistics (i.e., a lot of carefully equally prepared setups of the particles and measurement devices) and systematic error analysis. In HEP physics of course you have always some "misidentified particles", need for "background subtraction" to filter out the signal of the process of interest (e.g., to observe the Higgs boson peak above the background of other processes leading to similar signals) etc. etc. Indeed, at the end it has all to do with correlations as said above in #3.

 

[_PJ_]

Thanks Bhobba. I think you had a better grasp of what I was actually asking (again, without the context of the previous replies in the thread this was actually posted, it seems to apply to a more general query as to "how are measurements taken" which is not at all what I was referring to.

I suspect PeterDonis is referring to more general cases as I already noted the statistical rationale. Also DrChinese seems to be even more general in terms of the "validity of scientific results" overall...
Neither of which are at all my actual question, which I do appreciate was not necessarily evident given it is no longer in the appropriate context of the thread following comments in which it was posted in response to.

Of course, if the mathematical formality predicts that a certain outcome will occur with a given probability, and sufficient repetition of experiments not only helps verify the theory/model, but also thereby reinforces that the experiment is conducted appropriately.

HOWEVER - my real point at heart was more in regards to the comments that allude to making measurements of AN ELECTRON (or whatever 'particle') at various stages throughout the experiment - there is, in my opinion, no reasonable means to identify that such an electron is indeed the exact same one (what does that even mean?)
EXAMPLE

If A theory of whatever suggests (for example) that 'particles' go from A through B and 50% result at C and 50% result at D and that separately one can identify the particle emission at A, then they can also be detected if required at B, then detectors at C and D agree precisely (when sufficient numbers have been involved), that there is a 50/50 disttribution between thee two --- Then any detection at C or D one can then claim, rightfully of course, that it has agreed with every step of the experiment and theory --- That's not what I was questioning.

When someone tries to create a scenario, then such as "what if we measure ... of the electron at B and then measure if that electron is at C or D" for example, WHEN there is some attempt to identify a specific single, individual electron - that is what I am having difficulty with and I suspect Bhobba appreciated that point, given the mention of being able to state it strictly.

The correlation mentioned by PeterDonis is exactly as I mentioned that statistical results.

 

[vanhees71]

I'm again not sure what your problem really is, but QT tells you that indeed you can not individualize indistinguishable particles. In QT indistinguishable really means indistinguishable, i.e., if you exchange two electrons all that happens is that the corresponding many-electron state vector $|\psi \rangle$ changes its sign, i.e., the state itself, which is given by the projection operator $\hat{\rho}=|\psi \rangle \langle \psi|$ doesn't change at all. I.e., you cannot distinguish in any possible way the situation where just two electrons are interchanged. That the state ket changes sign means that the electrons are fermions (as all half-integer-spin particles are). Other kinds of particles (e.g., all mesons) are bosons. Here the multi-particle state ket doesn't change at all under permutations of indistinguishable particles.

 

[Cthugha]

I am also not quite sure what you are aiming at, but let me discuss this from the experimental point of view for the example of experiments using single photons. First, an experimentalist would determine the dark count rate of his setup. So he will check how often detection events will occur even when the single photon source he intends to use is not present. This may happen due to stray light or other unwanted effects of the detection electronics. However, for really modern detectors such as superconducting nanowire detectors these numbers can be pretty low. Next, the experimentalist has to make sure that his single photon source indeed emits single photons. This is typically checked using a Hanbury Brown-Twiss type of experiment. You take the emission put it on a beam splitter and place two detectors at the exit ports. If the source is a true single photon source, these detectors will never click simultaneously, otherwise you can deduce the error rate of the source (how often it will emit more than one photon compared to how often it will emit one photon) by comparing the mean coincidence count rate to the mean detection rate. Using these numbers, you can determine the relative probability that in any single repetition of the experiment detection events occur due to the intended preparation of a single photon or due to unwanted effects and can think about whether this number is sufficient for your experiment. If this rate is not sufficient you need to optimize your source or experiment. For example in some experiments single photon pulses may be pretty short,e.g. few picosecond pulses which are repeated every about 13 ns. Then you could in principle use time gating to reduce background counts.

 

[_PJ_]

"I'm again not sure what your problem really is, but QT tells you that indeed you can not individualize indistinguishable particles. "

This is precisely my point.

So if a comment claims to measure "THE particle" at one part, then measure THE same particle again at another part of an experiment, I would question how they might do so reliably.

As cthuga explains, careful preparation can help ensure a level of integrity of an experiment, that reduces possobility of 'external' anomalous particles etc. but not that the same particle is measured

 

[PeterDonis]

if a comment claims to measure "THE particle" at one part, then measure THE same particle again at another part of an experiment, I would question how they might do so reliably

So can you give us a textbook or peer-reviewed paper that actually makes such a claim? In other words, that actually claims to measure "the same particle" when there is more than one particle in the experiment?

 

[_PJ_]

So can you give us a textbook or peer-reviewed paper that actually makes such a claim? In other words, that actually claims to measure "the same particle" when there is more than one particle in the experiment?

No. Of course I cannot provide that.

Whyever would you expect that I might provide such a thing?!

I think it extremely unlikely that such a claim would be made in such, nor would that pass publication or review. If ever it did, it would be highly suspect and extremely 'fringe'.

The question posed off thecuff response to earlier comment, if that poster had some such source (which I doubt) it would be best to ask them to provide it, surely? Not me who questions it with incredulity in the first place.

Maybe thete's some translation or language barrier, but I tend npot to conflate PF user thread 'comment' with 'texbook or peer-reviewed paper'.

I made repeated efforts to clarify that my question was resåonding to the original comment and with particular (no pun) context, but you seem to utterly neglect this and persist with your idealised version.

 

[_PJ_]

At the end of the day, either i ought to have included quotation from the commentor to which my question was in response to (I did not since in its original location, I assumed the context was appropriate and sufficient), or the moderator ought not to have made this its own thread with such a misleading title because as it stands now, this thread appears to be a generalised, naive question regarding "how experiments are conducted" in broad sense, rather than the (potentially flippant) semi-rhetorical position to highlight the flaw in reasoning that it was to address.

This has clearly caused some confusion.

 

[PeterDonis]

Thread closed for moderation.