Number of 4-fermion, 2-derivative Lorentz invariants

[Simon_Tyler]

The notation below, is consistent with Wess and Bagger's https://www.amazon.com/dp/0691025304/?tag=pfamazon01-20.

Given a Majorana spinor field in 4D, written in 2-component notation as
[tex] \Psi(x) = \begin{pmatrix} \psi(x) \\\\ \bar\psi(x) \end{pmatrix} ,
\quad (\psi_\alpha)^* = \bar\psi_{\dot\alpha} \ ,
[/tex]
how many linearly independent Lorentz invariants can be formed using 2n spinors and n space-time derivatives, tied together with the sigma/Pauli matrices and various metrics?

And, more importantly for my application, how many are there modulo total derivatives?

Is there a general (eg representation theory) approach to this type of problem?

Notes:

1. I am mainly (at the moment) concerned with the n=2 case.
2. Due to anticommutativity, there are no such terms with n>4.
3. This can obviously be rewritten using 4-component Majorana spinors and Dirac matrices.

--------------------

For example with n=1 the only invariant is the standard kinetic term
[tex] \psi \sigma^a \partial_a \bar\psi
= \psi^\alpha \sigma^a_{\alpha\dot\alpha} \partial_a \bar\psi^{\dot\alpha}
= - (\partial_a\psi) \sigma^a \bar\psi + \text{total derivative} \ .
[/tex]

For [tex] n=2 [/tex], I believe (and want to prove) that there are only 6 invariants up to total derivatives.
Defining the matrix
[tex] v_a{}^b = i \psi\sigma^b\partial_a\bar\psi [/tex]
and its complex conjugate
[tex] \bar v_a{}^b = -i (\partial_a\psi)\sigma^b\bar\psi [/tex],
I chose the basis(?)
[tex] (\partial^a\psi^2)(\partial_a\bar\psi^2) \ ,\; tr(v)tr(\bar v) \ , \;
tr(v)^2\ ,\; tr(v^2)\ ,\; tr(\bar v)^2 \,\; tr(\bar v^2)\ .
[/tex]
Other terms being related by (for example)
[tex] tr(v\bar v) = tr(v)tr(\bar v)
+\tfrac12\Big(tr(v^2)-tr(v)^2+tr(\bar v^2)-tr(\bar v)^2\Big)
+ \text{total derivative}
[/tex]

 

[Simon_Tyler]

(bump)

Does anyone have any ideas how to systematically approach such a problem?