# Weyl Spinors Transformation, QFT1, Peskin, Chapter 3

Homework Statement:
The transformation law for Weyl spinors is as following (3.37); these transformation laws are connected by complex conjugation; using the identity (3.38)
Relevant Equations:
(3.37) and (3.38) peskin
\begin{align}
\psi_L \rightarrow (1-i \vec{\theta} . \frac{{\vec\sigma}}{2} - \vec\beta . \frac{\vec\sigma}{2}) \psi_L \\
\psi_R \rightarrow (1-i \vec{\theta} . \frac{{\vec\sigma}}{2} + \vec\beta . \frac{\vec\sigma}{2}) \psi_R
\end{align}
I really cannot evaluate these from boost and rotation generator which was introduced in (3.26) and (3.27) Peskin.
Although the main porblem is the identity introduced below:
$$\sigma^2 \vec\sigma^* = -\vec \sigma \sigma^2$$
My attempt:
$$\sigma^1 = \begin{pmatrix} 0 & 1 \\ 1 & 0 \\ \end{pmatrix}; \sigma^2 = \begin{pmatrix} 0 & -i \\ i & 0 \\ \end{pmatrix}; \sigma^3 = \begin{pmatrix} 1 & 0 \\ 0 & 1 \\ \end{pmatrix}$$
the h.c. of the Pauli Sigma matrices is as following:
$$(\sigma^1)^* = \begin{pmatrix} 0 & 1 \\ 1 & 0 \\ \end{pmatrix}; (\sigma^2)^* = \begin{pmatrix} 0 & i \\ -i & 0 \\ \end{pmatrix}; (\sigma^3)^* = \begin{pmatrix} 1 & 0 \\ 0 & 1 \\ \end{pmatrix}$$
so ##\sigma^2 = diag(1,1)##,right?
I think my problem is that I cannot write the identity in components form.

Last edited:

## Answers and Replies

vanhees71
Science Advisor
Gold Member
Of course ##\sigma^3=\mathrm{diag}(1,-1)##. I'd say the most simple way is just to prove the equation by doing the matrix multiplications for the three matrices :-).

• Pouramat
Of course ##\sigma^3=\mathrm{diag}(1,-1)##. I'd say the most simple way is just to prove the equation by doing the matrix multiplications for the three matrices :-).
Ohhh, Yes, it was a typo :)
the LHS:
$$\large \begin{pmatrix} 0 & 1 \\ 1 & 0 \\ \end{pmatrix} + \begin{pmatrix} 0 & i \\ -i & 0 \\ \end{pmatrix} + \begin{pmatrix} 1 & 0 \\ 0 & -1 \\ \end{pmatrix} = \begin{pmatrix} 1 & 1+i \\ 1-i & -1 \\ \end{pmatrix}$$
which should be equal to RHS:
$$-\large( \begin{pmatrix} 0 & 1 \\ 1 & 0 \\ \end{pmatrix} + \begin{pmatrix} 0 & -i \\ i & 0 \\ \end{pmatrix} + \begin{pmatrix} 1 & 0 \\ 0 & -1 \\ \end{pmatrix}) = \begin{pmatrix} -1 & 1+i \\ 1-i & 1 \\ \end{pmatrix}$$
but it is not!

vanhees71
Science Advisor
Gold Member
I have no idea what you are doing here. You are supposed to show that (I'll give one example)
$$\sigma^2 \sigma^{1*}=-\sigma^1 \sigma^2.$$
Take the right-hand side and use the anti-commutation relations
$$\{\sigma^j,\sigma^k\}=2 \delta^{jk}$$
to get
$$-\sigma^1 \sigma^2 =+ \sigma^2 \sigma^1.$$
Since ##\sigma^1## is real this hows that
$$-\sigma^1 \sigma^2=\sigma^2 \sigma^{1*}.$$
Now you can prove the other two identities!

• Pouramat
I have no idea what you are doing here. You are supposed to show that (I'll give one example)
$$\sigma^2 \sigma^{1*}=-\sigma^1 \sigma^2.$$
Take the right-hand side and use the anti-commutation relations
$$\{\sigma^j,\sigma^k\}=2 \delta^{jk}$$
to get
$$-\sigma^1 \sigma^2 =+ \sigma^2 \sigma^1.$$
Since ##\sigma^1## is real this hows that
$$-\sigma^1 \sigma^2=\sigma^2 \sigma^{1*}.$$
Now you can prove the other two identities!
thank you @vanhees71. Yes you are right.

dear @vanhees71
And 1 more question, do you have any idea to explicitly show that ##\sigma^2 \psi^*_L## transforms like a right-handed spinor? using the identity.

vanhees71
Science Advisor
Gold Member
Take the conjugate complex of the transformation law for ##\psi_L## and multiply with ##\sigma^2##. Then use the (anti-)commutation relation for the complex conjugated Pauli matrices just proven.

• Pouramat
Take the conjugate complex of the transformation law for ##\psi_L## and multiply with ##\sigma^2##. Then use the (anti-)commutation relation for the complex conjugated Pauli matrices just proven.
I have a question to that. Peskin asks us to show that ##\sigma^2\psi_L## transforms like a right-handed spinor. So why can we just consider the transformation of ##\psi_L## alone to then multiply by ##\sigma^2##, and not the transformation of whatever the object ##\sigma^2\psi_L## is?

Edit: I think that my question is why we don't need to take the ##\sigma^2## in front into account when doing the transformation. After all, ##\sigma^2\psi_L\ne\psi_L##, so it should be taken into account when transforming, right?

Last edited:
vanhees71
Science Advisor
Gold Member
You now the transformation representing Lorentz transformations. The infinitesimal version is already sufficient, i.e., the generators for left- and right-handed Weyl spinors. Then it's just Pauli-matrix gymnastics to prove that ##\sigma^2 \psi_L## transforms like ##\psi_R##. Mathematically it's of course not the special pauli matrix ##\sigma^2## which is behind this but the antisymmetric product of two spinors given by ##\epsilon_{ij}## (with ##\epsilon_{12}=-\epsilon_{21}=1##) which is relevant here.

That it's an antisymmetric product instead of a symmetric hints at the fact that one should rather argue with Grassmann-number valued fields for fermions, which is what comes out also using the path-integral formulation.

You now the transformation representing Lorentz transformations. The infinitesimal version is already sufficient, i.e., the generators for left- and right-handed Weyl spinors. Then it's just Pauli-matrix gymnastics to prove that ##\sigma^2 \psi_L## transforms like ##\psi_R##. Mathematically it's of course not the special pauli matrix ##\sigma^2## which is behind this but the antisymmetric product of two spinors given by ##\epsilon_{ij}## (with ##\epsilon_{12}=-\epsilon_{21}=1##) which is relevant here.

That it's an antisymmetric product instead of a symmetric hints at the fact that one should rather argue with Grassmann-number valued fields for fermions, which is what comes out also using the path-integral formulation.
I’m sorry, but I fail to see how this answers my question. What doesn’t make sense for me is the following:
We are asked to show that the object ##\sigma^2\psi_L^*## (in my previous post I forgot the complex conjugate, sorry) transforms as a right-handed spinor. For me, this means that we need to consider $$\sigma^2\psi_L^*\to(\sigma^2\psi_L^*)’.$$However, you are saying that we should consider $$\psi_L^*\to(\psi_L’)^*,$$to then just multiply by ##\sigma^2##. I can see that this gives the correct result, but I fail to see why we are allowed to do it like this, considering that, in general, these are two different quantities. I hope that this clarifies my confusion.

vanhees71
Science Advisor
Gold Member
But then you solved the problem. I don't understand, what's still unclear.

But then you solved the problem. I don't understand, what's still unclear.
I wouldn't consider it solved, since I don't know why I am allowed to do the transformation like this. The goal for me is to not blindly solve it and move on, but to understand why it works like this. What I want to know is why I can consider the transformation ##\psi_L^*\to(\psi_L')^*##, even though one is asked to consider the transformation ##\sigma^2\psi_L^*\to(\sigma^2\psi_L)'##, since we are asked to show that ##\sigma^2\psi_L^*## transforms as a right-handed spinor. I fail to see what confuses you in my question...

vanhees71
Science Advisor
Gold Member
Why? If you know how ##\psi_L## transforms, you can just take the conjugate complex of the equation to know how ##\psi_L^*## transforms. Then you multiply by ##\sigma^2## to find out our ##\sigma_2 \psi_L^*## transforms and compare this with the transformation of ##\psi_R##.