# Quantum spin superposition interpretation

• phoenix3
In summary: Trying to think classically about quantum mechanics leads to many confusions and misunderstandings. Embrace QM for what it is, without trying to make too many classical pictures of what is going on. That is not to say that classical pictures are never useful, but rather that they are limited in their usefulness.In summary, the conversation discusses the probability of a particle's spin being measured as "spin up" along a certain direction when prepared along a different direction. The probability is given by ##P(up)=cos^2(\frac{\theta}{2})##, but the conversation highlights some incorrect assumptions and interpretations made in trying to understand this concept from a classical point of view. The expert summarizer suggests
phoenix3
Homework Statement
How to interpret superposition
Relevant Equations
##cos^2(\frac{\theta}{2})##
Here is my workings out:

If a particle's spin of magnitude ##\frac {\hbar}{2}## is prepared along direction ##\vec r_1## and subsequently its spin is measured along direction ##\vec r_2 ## at an angle ##\vec \theta ## to ##\vec r_1##, the probability of its being found "spin up" along is ##\vec r_2## is ##P(up)=cos^2(\frac{\theta}{2})## again with full magnitude ##\frac {\hbar}{2}##.

Classically, the spin might easily have been incorrectly predicted to be of magnitude ##\frac {\hbar}{2} . cos(\theta)## along ##\vec r_2 ##


in units of ##\frac {\hbar}{2}##, ##\frac {\hbar}{2} = 1##.

Starting with the seemingly 'incorrect' assumption that the component along ##\vec r_2 ## is ##\frac {\hbar}{2} . cos(\theta)##



##\frac {\hbar}{2} . cos(\theta) = 1.cos(\theta) ## in units of ##\frac {\hbar}{2}##
\begin{align}
1.cos(\theta) & = cos(\frac{\theta}{2}+\frac{\theta}{2}) \nonumber \\
& = cos^2(\frac{\theta}{2})-sin^2(\frac{\theta}{2}) \nonumber \\
& = (+1).cos^2(\frac{\theta}{2}) + (-1).sin^2(\frac{\theta}{2}) \nonumber \\
\end{align}

Multiplying accross by a time interval, ##\delta t##,

\begin{align}
1.cos(\theta).\delta t = (+1).cos^2(\frac{\theta}{2}).\delta t + (-1).sin^2(\frac{\theta}{2}).\delta t \nonumber \\
\end{align}

it seems that this can be interpreted to mean either:

(A) a combination of two vectors of magnitudes ## cos^2(\frac{\theta}{2}) ## and ##-sin^2(\frac{\theta}{2})## respectively, acting for duration ##\delta t## or


(B) a combination of two vectors of magnitudes 1 and -1 respectively, the first acting for a duration ##cos^2(\frac{\theta}{2}).\delta t## and the second acting for the remaining duration ##sin^2(\frac{\theta}{2}).\delta t##


with ##( cos^2(\frac{\theta}{2}).\delta t + sin^2(\frac{\theta}{2}).\delta t = \delta t )##

This latter interpretation, (B), would seem to justify quite nicely why the spin is always measured to have the same absolute unit magnitude and also why the probability of getting "spin up" along ##\vec r_2 ## is equal to ##cos^2(\frac{\theta}{2})## since according to this interpretation it is in fact "spin up" along ##\vec r_2## , ## cos^2(\frac{\theta}{2})## of the time.

Obviously once measured along ##\vec r_2 ## it is then in an eigenstate along ##\vec r_2 ## and the probability of a measurement along ##\vec r_1 ## being "spin up" along ##\vec r_1 ## is ##cos^2(\frac{\theta}{2})## for the same reason.

Is there an error in the above and if so where?

Last edited:
phoenix3 said:
Is there an error in the above and if so where?
The error seems to be that you are trying to think classically about quantum mechanics. My advice is to stop doing this and fully embrace QM without trying to make too many classical pictures of what is going on (although some classical pictures can help some of the time).

topsquark
Respectfully, DrClaude, that still doesn't explain where any error is, either in my math or in my interpretation of it. Besides, as long as the math correctly predicts the probability, where's the harm in having an interpretation of it that is non-mysterious?

phoenix3 said:
Respectfully, DrClaude, that still doesn't explain where any error is, either in my math or in my interpretation of it.
I'm tempted to answer that it's "not even wrong," but I'll humor you.

phoenix3 said:
If a particle's spin of magnitude ##\frac {\hbar}{2}## is prepared along direction ##\vec r_1## and subsequently its spin is measured along direction ##\vec r_2 ## at an angle ##\vec \theta ## to ##\vec r_1##, the probability of its being found "spin up" along is ##\vec r_2## is ##P(up)=cos^2(\frac{\theta}{2})## again with full magnitude ##\frac {\hbar}{2}##.
This is technically incorrect. A spin-1/2 system has a spin magnitude of ##\frac{\sqrt{3}}{2} \hbar##, with its projection on any axis being ##\pm \frac{1}{2} \hbar##.

phoenix3 said:
Classically, the spin might easily have been incorrectly predicted to be of magnitude ##\frac {\hbar}{2} . cos(\theta)## along ##\vec r_2 ## in units of ##\frac {\hbar}{2}##, ##\frac {\hbar}{2} = 1##.
Strange choice of units (one would usually choose ##\hbar = 1##), but not incorrect.

phoenix3 said:
Multiplying accross by a time interval, ##\delta t##,
What is the justification in doing this?

phoenix3 said:
\begin{align}
1.cos(\theta).\delta t = (+1).cos^2(\frac{\theta}{2}).\delta t + (-1).sin^2(\frac{\theta}{2}).\delta t \nonumber \\
\end{align}

it seems that this can be interpreted to mean either:

(A) a combination of two vectors of magnitudes ## cos^2(\frac{\theta}{2}) ## and ##-sin^2(\frac{\theta}{2})## respectively, acting for duration ##\delta t## or

(B) a combination of two vectors of magnitudes 1 and -1 respectively, the first acting for a duration ##cos^2(\frac{\theta}{2}).\delta t## and the second acting for the remaining duration ##sin^2(\frac{\theta}{2}).\delta t##
No. Just because it "looks like" the original ##cos^2(\frac{\theta}{2})## doesn't mean it is the same thing. The entire expression should be attached to measuring spin-up, even though a minus sign appears.

Try the same exercise with the probability of measuring spin-down and you will see that it doesn't work. What you did is close to what we call "numerology," where people manipulate equations and numbers until they look like something we already know, but this doesn't give any deeper understanding because this "game" can be played with any set of numbers.

phoenix3 said:
Besides, as long as the math correctly predicts the probability, where's the harm in having an interpretation of it that is non-mysterious?
The harm is in having an understanding that is actually incorrect.

PeroK and topsquark
This is technically incorrect. A spin-1/2 system has a spin magnitude of , with its projection on any axis being .
DrClaude said:
This is technically incorrect. A spin-1/2 system has a spin magnitude of , with its projection on any axis being .
DrClaude said:
This is technically incorrect. A spin-1/2 system has a spin magnitude of , with its projection on any axis being .
yes i know, but it only ever 'measured' to be ##\frac {\hbar}{2}##

DrClaude said:
What is the justification in doing this?
to include 'time' into consideration. One can't pretend time is irrelevant.

DrClaude said:
Try the same exercise with the probability of measuring spin-down and you will see that it doesn't work.
It does wrk: it gives it spin-down probability as ##sin^2(\frac{\theta}{2})##

DrClaude said:
The entire expression should be attached to measuring spin-up, even though a minus sign appears.
The entire expression evaluates to ##cos(\theta)## and indicates that ##cos(\theta)## is in fact a net result of spin-up and spin-down during any non-zero interval of time ##\delta t##, whereas at any single 'point' in time it is either spin-up xor spin-down.

DrClaude said:
The harm is in having an understanding that is actually incorrect.
To say that, while not measured along ##\vec r_2## the spin is in a superposition state of being spin-up and spin-down at each individual point in time is not only incorrrect, its meaningless.

The spin component in direction ##\vec{n}=(\sin \vartheta \cos \varphi,\sin \vartheta \sin \varphi,\cos \vartheta)## is represented in the usual basis of ##\hat{s}_3##-eigenstates by
$$\hat{s}_n=\vec{n} \cdot \vec{\sigma}/2=\frac{1}{2} \begin{pmatrix} \cos \vartheta & \exp(-\mathrm{i} \varphi) \sin \vartheta \\ \exp(\mathrm{i} \varphi) \sin \vartheta & -\cos \vartheta \end{pmatrix}.$$
The eigenvectors for ##\hat{s}_n## are
$$|\vec{n},1/2 \rangle=\begin{pmatrix}\exp(-\mathrm{i} \varphi/2) \cos \vartheta) \\ \exp(\mathrm{i} \varphi/2) \sin \vartheta) \end{pmatrix},$$
$$|\vec{n},1/2 \rangle=\begin{pmatrix} -\exp(-\mathrm{i} \varphi/2) \sin \vartheta \\ \exp(\mathrm{i} \varphi /2) \cos \vartheta \end{pmatrix}.$$
For a particle being prepared in ##|\vec{e}_z,1/2 \rangle## the probabilities to measure ##s_n=\pm 1/2## obviously is
$$P_{1/2}=|\langle \vec{n},1/2|\vec{e}_z,1/2 \rangle|^2=\cos^2(\vartheta/2), \quad P_{-1/2} = \sin^2(\vartheta/2).$$

kuruman and topsquark

Replies
11
Views
1K
Replies
2
Views
847
Replies
1
Views
812
Replies
6
Views
999
Replies
1
Views
824
Replies
1
Views
749
Replies
1
Views
989
Replies
7
Views
1K