# Anti-commutator version of Zassenhaus formula

The Zassenhaus formula goes like $$e^{t(X+Y)}= e^{tX}~ e^{tY} ~e^{-\frac{t^2}{2} [X,Y]} ~ e^{\frac{t^3}{6}(2[Y,[X,Y]]+ [X,[X,Y]] )} ~ e^{\frac{-t^4}{24}([[[X,Y],X],X] + 3[[[X,Y],X],Y] + 3[[[X,Y],Y],Y]) } \cdots,$$ where $X$ and $Y$ are operators may not commute.

Do people know and derive the anti-commutator version of Zassenhaus formula that expresses in terms of anti-commutator $\{X,Y\}$ in a very compact form?

I haven't found it on the literature (yet).

(1) Let us consider Grassmann-parity of $X$ and $Y$ are both even, so that $X$ and $Y$ both contain even number of fermionic operators $f/f^\dagger$, where $$\{f_i,f_j^\dagger\}=\delta_{ij}$$

(2) What if $X$ or $Y$ contain an odd number of fermionic operators?

• Such an expansion is (in general) impossible. – AccidentalFourierTransform Nov 15 '17 at 18:55
• There cannot be any such general formula unless you specify some special properties of $X$ and $Y$. The formula for a commutator exists because it is associated with adjoint action of a group. – Prahar Nov 15 '17 at 19:07
• What's the Grassmann-parity of $X$ and $Y$? – Qmechanic Nov 15 '17 at 19:07
• Let us consider Grassmann-parity of X and Y are both even. – annie heart Nov 15 '17 at 19:20
• $\{f_i,f_j^\dagger\}=\delta_{ij}$ is then not relevant for $\{X,Y\}$. – Qmechanic Nov 15 '17 at 19:26