I have a system of $4$ spin $\frac{1}{2}$ particles, whose Hamiltonian looks like this:

$$ H = \alpha (s_1 \cdot s_2 + s_1 \cdot s_4 +s_2 \cdot s_3 +s_3 \cdot s_4 )$$

In order to find its eigenvalues and respective degeneracies I tried the typical method:

$$ \hat{\overrightarrow{S}} = s_1+ s_2+s_3+s_4 $$

$$ \hat{S}^2 = (s_1+ s_2+s_3+s_4) \cdot (s_1+ s_2+s_3+s_4)$$

which leads me to:

$$ \frac{1}{2} \left( \hat{S}^2 - \sum_{i=1}^4 |s_i|^2\right) = s_1 \cdot s_2 + s_1 \cdot s_4 +s_2 \cdot s_3 +s_3 \cdot s_4 + s_1 \cdot s_3 +s_2 \cdot s_4 $$

There are two extra terms $s_1 \cdot s_3$ and $s_2 \cdot s_4$ that will still appear in the Hamiltonian. How do I handle these in order to get the possible eigenvalues and degeneracies of this particular system?


1 Answer 1


You are composing four doublets, hence you are inspecting a $2^4\times 2^4= 16\times 16$ matrix. This reduces to the familiar spin 2, spin 1, and spin 0 blocks, specifically a quintet, three triplets, and two singlets, 16 states in all, $$ 1/2\otimes 1/2\otimes 1/2\otimes 1/2= 2\oplus (3) ~~1\oplus (2) ~~0. $$

But your hamiltonian $\alpha(𝑠_1+𝑠_3)⋅(𝑠_2+𝑠_4)$ is special: it does not care how spins 1 & 3 compose together symmetrically or anti symmetrically, in a triplet or a singlet, and likewise for 2 & 4. These two options are degenerate, respectively, in both cases, so you are really composing just $$ \vec S_a = {\vec s_1+\vec s_3} ~~~\hbox{with} \\ \vec S_b = {\vec s_2+\vec s_4}, $$ where the spins of each $S_a$ and $S_b$ are 0 and 1; so, four combinations in all for $\vec S= \vec S_a+\vec S_b$, $$ H= {\alpha\over 2} (S^2-S_a^2-S_b^2). $$

Recalling that the Casimirs for spin 0; 1; and 2 are 0; 2; and 6, respectively: you get the singlet-singlet combination (singlet) to have 0 energy eigenvalue; the two singlet-triplet combinations (triplets) to have 0 energy eigenvalue, as well.

The triplet-triplet combination reduces to the quintet, with eigenvalue α; a triplet with eigenvalue ; and a singlet with eigenvalue -2α.

Not all products of the reduction with a common spin are treated identically by the (a)symmetric hamiltonian.

  • 1
    $\begingroup$ Shouldn´t the possible values for $S_a$ and $S_b$ be $0$ or $1$? Using the theory of addition of angular momentum, that's what it should be, right? $\endgroup$ Dec 7, 2021 at 15:13
  • $\begingroup$ Also, for the eigenvalues I used: $H = \frac{\alpha}{2} (S^2 - S_a^2- S_b^2) = \frac{\alpha}{2} (S(S+1) - S_a(S_a +1)- S_b(S_b+1))$ and tried the different cases: $S=2, S_a=1 , S_b=1$; $S= 1 , S_a=1 , S_b=0$, etc... and got $3\alpha$, $2 \alpha$, $\alpha$ and $0$ for the eigenvalues. $\endgroup$ Dec 7, 2021 at 15:17
  • $\begingroup$ Pardon my typo; indeed, the intermediate components spins should be 0 or 1, but my typing was thinking "triplet" for the latter. The last paragraph is detailing $S_a=1=S_b$ for $S=2,1,0$, hence eigenvalues α, -α, -2α, as it states. Recall, quintet is spin 2, triplet is spin 1, and singlet is spin 0. $\endgroup$ Dec 7, 2021 at 15:25
  • $\begingroup$ Sorry for being annoying but, in the triplet-singlet combination (which I presume is for $S_a=1$, $S_b=0$ and $S_b=1$, $S_a=0$ ), when plugging in the values $S=0,1$ I found that the eigenvalue associated for $S=0$ is $H=- \alpha$ and for $S=1$ is $H=0$. Am I right? $\endgroup$ Dec 7, 2021 at 17:20
  • 1
    $\begingroup$ Nevermind, I got it now. Thank you! $\endgroup$ Dec 7, 2021 at 19:24

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.