In this article Theory of Complex Spectra II Giulio Racah defines $f(m_{1} m_{2} ; jm)$ by \begin{multline} \left(m_{1} m_{2} \mid j m\right) =(-1)^{j_{1}-m_{1}} f\left(m_{1} m_{2} ; j m\right)\left[\left(j_{1}+m_{1}\right) !\left(j_{2}+m_{2}\right) !(j+m) !\right]^{\frac{1}{2}} /\left[\left(j_{1}-m_{1}\right) !\left(j_{2}-m_{2}\right) !(j-m) !\right]^{\frac{1}{2}} \end{multline} where $\left(m_{1} m_{2} \mid j m\right)$ are the Clebsch-Gordan coefficients.Then, he shows that
\begin{multline} f\left(m_{1} m_{2} ; j m-1\right)=\\ \left(j_{2}+m_{2}+1\right)\left(j_{2}-m_{2}\right) f\left(m_{1} m_{2}+1 ; j m\right)-\\ \left(j_{1}+m_{1}+1\right)\left(j_{1}-m_{1}\right) f\left(m_{1}+1 m_{2} ; j m\right) \tag{1} \end{multline} Now he claims that \begin{multline} f\left(m_{1} m_{2} ; j m\right)=f\left(m_{1} m_{2} ; j j -u\right)=\\ A_{j} \sum_{t}(-1)^{t}\left(\begin{array}{l} u \\ t \end{array}\right) \frac{\left(j_{1}+m_{1}+t\right) !\left(j_{1}-m_{1}\right) !\left(j_{2}+m_{2}+u-t\right) !\left(j_{2}-m_{2}\right) !}{\left(j_{1}+m_{1}\right) !\left(j_{1}-m_{1}-t\right) !\left(j_{2}+m_{2}\right) !\left(j_{2}-m_{2}-u+t\right) !} \tag{2}\end{multline} The summation parameter takes on all integral values consistent with the factorial notation, the factorial of a negative number being meaningless.To demonstrate (2) he says that it suffices to verify that it satisfies (1). This what I tried.We have that \begin{multline} \left(j_{1}+m_{1}+1\right)\left(j_{1}-m_{1}\right) f\left(m_{1}+1 m_{2} ; j m\right)=\\ A_{j} \sum_{t}(-1)^{t}\left(\begin{array}{l} u \\ t \end{array}\right) \left(j_{1}-m_{1}-t\right)\left(j_{1}+m_{1}+1+t\right) \times \\ \frac{\left(j_{1}+m_{1}+t\right) !\left(j_{1}-m_{1}\right) !\left(j_{2}+m_{2}+u-t\right) !\left(j_{2}-m_{2}\right) !}{\left(j_{1}+m_{1}\right) !\left(j_{1}-m_{1}-t\right) !\left(j_{2}+m_{2}\right) !\left(j_{2}-m_{2}-u+t\right) !} \end{multline}

\begin{multline} \left(j_{2}+m_{2}+1\right)\left(j_{2}-m_{2}\right) f\left(m_{1} m_{2} +1 ; j m\right)=\\ A_{j} \sum_{t}(-1)^{t}\left(\begin{array}{l} u \\ t \end{array}\right) \left(j_{2}-m_{2}-u+t\right)\left(j_{2}+m_{2}+1+u -t \right) \times \\ \frac{\left(j_{1}+m_{1}+t\right) !\left(j_{1}-m_{1}\right) !\left(j_{2}+m_{2}+u-t\right) !\left(j_{2}-m_{2}\right) !}{\left(j_{1}+m_{1}\right) !\left(j_{1}-m_{1}-t\right) !\left(j_{2}+m_{2}\right) !\left(j_{2}-m_{2}-u+t\right) !} \end{multline}

and \begin{multline} f\left(m_{1} m_{2}; j m-1\right)=\\ A_{j} \sum_{t}(-1)^{t}\left(\begin{array}{l} u \\ t \end{array}\right) \times \\ \frac{u+1}{u+1-t}\left(j_{2}+m_{2}+u+1-t\right)\left(j_{2}-m_{2}-u+t\right) \times \\ \frac{\left(j_{1}+m_{1}+t\right) !\left(j_{1}-m_{1}\right) !\left(j_{2}+m_{2}+u-t\right) !\left(j_{2}-m_{2}\right) !}{\left(j_{1}+m_{1}\right) !\left(j_{1}-m_{1}-t\right) !\left(j_{2}+m_{2}\right) !\left(j_{2}-m_{2}-u+t\right) !} \end{multline} and so we arrive at the following \begin{multline} \frac{(j_{1}+m_{1}+u+1)!(j_{1}-m_{1})!}{(j_{1}-m_{1}-u-1)!(j_{1}+m_{1})!}A_{j}(-1)^{u+1}+\\ A_{j} \sum_{t}(-1)^{t}\left(\begin{array}{l} u \\ t \end{array}\right) \frac{\left(j_{1}+m_{1}+t\right) !\left(j_{1}-m_{1}\right) !\left(j_{2}+m_{2}+u-t\right) !\left(j_{2}-m_{2}\right) !}{\left(j_{1}+m_{1}\right) !\left(j_{1}-m_{1}-t\right) !\left(j_{2}+m_{2}\right) !\left(j_{2}-m_{2}-u+t\right) !}\times \\ \Bigg[ \frac{u+1}{u+1-t}\left(j_{2}+m_{2}+u+1-t\right)\left(j_{2}-m_{2}-u+t\right)+\left(j_{1}-m_{1}-t\right)\left(j_{1}+m_{1}+1+t\right) \\-\left(j_{2}-m_{2}-u+t\right)\left(j_{2}+m_{2}+1+u -t \right) \Bigg] =0 /tag{3} \end{multline}

where $m=m_1+m_2$ , $u=j-m$ and $j=j_1+j_2$

Now $j_1-m_1=j_1-(m-m_2)=j_1-(j-u-m_2)=-j_2+m_2+u$ and so $-\left(j_{1}-m_{1}-t\right)=(j_2 -m_2-u +t)$

After some algebra last equation becomes \begin{multline} \frac{(j_{1}+m_{1}+u+1)!(j_{1}-m_{1})!}{(j_{1}-m_{1}-u-1)!(j_{1}+m_{1})!}A_{j}(-1)^{u+1}+ \\ A_{j} \sum_{t}(-1)^{t}\left(\begin{array}{l} u \\ t \end{array}\right) \frac{\left(j_{1}+m_{1}+t\right) !\left(j_{1}-m_{1}\right) !\left(j_{2}+m_{2}+u-t\right) !\left(j_{2}-m_{2}\right) !}{\left(j_{1}+m_{1}\right) !\left(j_{1}-m_{1}-t\right) !\left(j_{2}+m_{2}\right) !\left(j_{2}-m_{2}-u+t\right) !}\times \\ (j_2 -m_2-u +t)\Bigg[ \frac{u+1}{u+1-t}\left(j_{2}+m_{2}\right) -j-m-1 \Bigg] =0 \end{multline}

Can anyone tell why this last expression is zero?

  • $\begingroup$ Would this fit better on Math.SE? $\endgroup$
    – knzhou
    Mar 14, 2021 at 20:17
  • $\begingroup$ What is youe question? The algera derivation between the last two equation? $\endgroup$
    – ytlu
    Mar 15, 2021 at 9:26
  • $\begingroup$ My question is why the last expression is zero $\endgroup$ Mar 15, 2021 at 13:49
  • $\begingroup$ It is derived from the previous equation, and the previous one is zero as well. I still cannot get your meaning. $\endgroup$
    – ytlu
    Mar 15, 2021 at 13:56
  • $\begingroup$ I am trying to proof that the previous equation is zero $\endgroup$ Mar 15, 2021 at 17:30

1 Answer 1


Staring from the given definition of $ f(m_{1}, m_{2} ; j ,m)$: \begin{multline} \tag{1} f(m_{1}, m_{2} ; j ,m)=f(m_{1}, m_{2} ; j, j -u)=\\ A_{j} \sum_{t}(-1)^{t} \left( \begin{array}{l} u \\ t \end{array} \right) \frac{(j_{1}+m_{1}+t) ! (j_{1}-m_{1}) ! (j_{2}+m_{2}+u-t) ! (j_{2}-m_{2}) !}{(j_{1}+m_{1}) ! (j_{1}-m_{1}-t) ! (j_{2}+m_{2}) ! (j_{2}-m_{2}-u+t) !} \end{multline}

We are aiming at proving the following equation: \begin{multline}\tag{2} f(m_{1}, m_{2} ; j, m-1)=\\ (j_{2}+m_{2}+1) (j_{2}-m_{2}) f(m_{1} ,m_{2}+1 ; j ,m)-\\ (j_{1}+m_{1}+1) (j_{1}-m_{1}) f(m_{1}+1, m_{2} ; j ,m) \end{multline}


First write down the left-hand-side of the equation (2), defining $u_1 \equiv j -m +1$.: \begin{multline}\tag{3} f(m_{1}, m_{2} ; j, m-1)=\\ A_{j} \sum_{t}(-1)^{t} \left( \begin{array}{l} u_1 \\ t \end{array} \right) \frac{(j_{1}+m_{1}+t) ! (j_{1}-m_{1}) ! (j_{2}+m_{2}+u_1-t) ! (j_{2}-m_{2}) !}{(j_{1}+m_{1}) ! (j_{1}-m_{1}-t) ! (j_{2}+m_{2}) ! (j_{2}-m_{2}-u_1+t) !}, \end{multline}

Then, examine the two terms in the right-ahnd-side, using $u_2 \equiv j -m = u_1 - 1$. The frist term in the RHS: \begin{multline}\tag{4} (j_{2}+m_{2}+1) (j_{2}-m_{2}) f(m_{1} ,m_{2}+1 ; j ,m) = (j_{2}+m_{2}+1) (j_{2}-m_{2}) \times\\ A_{j} \sum_{t}(-1)^{t} \left( \begin{array}{l} u_2 \\ t \end{array} \right) \frac{(j_{1}+m_{1}+t) ! (j_{1}-m_{1}) ! (j_{2}+m_{2}+1+u_2-t) ! (j_{2}-m_{2}-1) !}{(j_{1}+m_{1}) ! (j_{1}-m_{1}-t) ! (j_{2}+m_{2}+1) ! (j_{2}-m_{2}-1-u_2+t) !}\\ = A_{j} \sum_{t}(-1)^{t} \left( \begin{array}{l} u_2 \\ t \end{array} \right) \frac{(j_{1}+m_{1}+t) ! (j_{1}-m_{1}) ! (j_{2}+m_{2}+1+u_2-t) ! (j_{2}-m_{2}) !}{(j_{1}+m_{1}) ! (j_{1}-m_{1}-t) ! (j_{2}+m_{2}) ! (j_{2}-m_{2}-1-u_2+t) !} \end{multline} The two multipliers are combined into the factorial functions in Eq.(4).

The second term in the RHS: \begin{multline}\tag{5} (j_{1}+m_{1}+1) (j_{1}-m_{1}) f(m_{1}+1 ,m_{2} ; j ,m) = (j_1+m_1+1) (j_1-m_1) \times\\ A_{j} \sum_{t}(-1)^{t} \left( \begin{array}{l} u_2 \\ t \end{array} \right) \frac{(j_{1}+m_{1}+1+t) ! (j_{1}-m_{1}-1) ! (j_{2}+m_{2}+u_2-t) ! (j_{2}-m_{2}) !}{(j_{1}+m_{1}+1) ! (j_{1}-m_{1}-1-t) ! (j_{2}+m_{2}) ! (j_{2}-m_{2}-u_2+t) !}\\ = A_{j} \sum_{t}(-1)^{t} \left( \begin{array}{l} u_2 \\ t \end{array} \right) \frac{(j_{1}+m_{1}+1+t) ! (j_{1}-m_{1}) ! (j_{2}+m_{2}+u_2-t) ! (j_{2}-m_{2}) !}{(j_{1}+m_{1}) ! (j_{1}-m_{1}-1-t) ! (j_{2}+m_{2}) ! (j_{2}-m_{2}-u_2+t) !} \end{multline} The two multipliers are also combined into the factorial functions in Eq.(5).

Now we will proceed to show that Eq.(4)-Eq.(5) = Eq.(3). 

In Eq.(5), we change the summation dummy index $t+1 = t'$. \begin{multline}\tag{6} A_{j} \sum_{t'=1}^{t'=u_2+1} (-1)^{t'-1} \left( \begin{array}{l} u_2 \\ t'-1 \end{array} \right) \frac{(j_{1}+m_{1}+t') ! (j_{1}-m_{1}) ! (j_{2}+m_{2}+u_2-t'+1) ! (j_{2}-m_{2}) !}{(j_{1}+m_{1}) ! (j_{1}-m_{1}-t') ! (j_{2}+m_{2}) ! (j_{2}-m_{2}-u_2+t'-1) !} \\ = A_{j} \sum_{t'=0}^{t'=u_1} (-1)^{t'-1} \left( \begin{array}{l} u_2 \\ t'-1 \end{array} \right) \frac{(j_{1}+m_{1}+t') ! (j_{1}-m_{1}) ! (j_{2}+m_{2}+u_1-t') ! (j_{2}-m_{2}) !}{(j_{1}+m_{1}) ! (j_{1}-m_{1}-t') ! (j_{2}+m_{2}) ! (j_{2}-m_{2}-u_1+t') !} \end{multline} In the last expression, we use $u_2 = j-m = u_1 - 1$, and extend the summation lower limit from $1$ to $0$, since the term $t'=0$ vanishes having $(-1)!$ in the denominator of the combinatory function.

Also in Eq.(4), we replace $u_2 +1 = u_1$, and extended the summation upper limit from $u_2$ to $u_1 = u_2 +1$, the extra term has a $-1 !$ in the denominator of the combinatory function, thus vanishes. \begin{multline}\tag{7} A_{j} \sum_{t=0}^{t=u_1} (-1)^{t} \left( \begin{array}{l} u_2 \\ t \end{array} \right) \frac{(j_{1}+m_{1}+t) ! (j_{1}-m_{1}) ! (j_{2}+m_{2}+u_1-t) ! (j_{2}-m_{2}) !}{(j_{1}+m_{1}) ! (j_{1}-m_{1}-t) ! (j_{2}+m_{2}) ! (j_{2}-m_{2} -u_1+t) !} \end{multline}

Finally, the Eq.(4) - Eq.(5) now becomes Eq.(7) - Eq.(6):

\begin{multline}\tag{8} \text{Eq.(7) } - \text{ Eq.(6) } = A_{j} \sum_{t=0}^{t=u_1} (-1)^{t} \left\{ \left( \begin{array}{l} u_2 \\ t \end{array} \right) - (-1) \left( \begin{array}{l} u_2 \\ t-1 \end{array} \right) \right\} \times \\ \frac{(j_{1}+m_{1}+t) ! (j_{1}-m_{1}) ! (j_{2}+m_{2}+u_1-t) ! (j_{2}-m_{2}) !}{(j_{1}+m_{1}) ! (j_{1}-m_{1}-t) ! (j_{2}+m_{2}) ! (j_{2}-m_{2} -u_1+t) !} \end{multline}

Evaluate the combinatroy functions in the curry braket: \begin{multline} \tag{9} \left( \begin{array}{l} u_2 \\ t \end{array} \right) + \left( \begin{array}{l} u_2 \\ t-1 \end{array} \right) = \frac{u_2 !}{t! (u_2-t)!} + \frac{u_2 !}{(t-1)! (u_2 - t + 1)!} \\ = \frac{u_2 !}{t! (u_2-t+1)!} \left( u_2 - t + 1 + t\right) = \frac{(u_2+1) !}{t! (u_2-t+1)!} = \frac{(u_1) !}{t! (u_1-t)!} = \left( \begin{array}{l} u_1 \\ t \end{array} \right) \end{multline}

Using the result of Eq.(9), Eq.(8) becomes: \begin{multline}\tag{10} A_{j} \sum_{t=0}^{t=u_1} (-1)^{t} \left( \begin{array}{l} u_1 \\ t \end{array} \right) \frac{(j_{1}+m_{1}+t) ! (j_{1}-m_{1}) ! (j_{2}+m_{2}+u_1-t) ! (j_{2}-m_{2}) !}{(j_{1}+m_{1}) ! (j_{1}-m_{1}-t) ! (j_{2}+m_{2}) ! (j_{2}-m_{2} -u_1+t) !} \end{multline}

Eq. (10) is exactly the same as Eq.(3), the left hand side of Eq.(2). Therfore Eq.(4) - Eq.(5) = Eq.(3), conclude the proof of the equality in Eq.(2).

  • 1
    $\begingroup$ Thank you for your response. $\endgroup$ Mar 17, 2021 at 6:33

Your Answer

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

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