You might have wondered why 4-vectors, scalar products and Dirac matrices all have a bar, like \bar{p}^\mu or \bar{p} \cdot \bar{q}. The bar is there to specify that they are 4-dimensional objects. Objects that live in D dimensions do not have a bar, cf.
[p, \[Mu]]
FVD% // FCI // StandardForm
p^{\mu }
(*Pair[LorentzIndex[\[Mu], D], Momentum[p, D]]*)
[\[Mu], \[Nu]]
MTD% // FCI // StandardForm
g^{\mu \nu }
(*Pair[LorentzIndex[\[Mu], D], LorentzIndex[\[Nu], D]]*)
This origin of this notation is a publication by Breitenlohner and Maison on the treatment of \gamma^5 in D dimensions in the t’Hooft-Veltman scheme. The main idea was that we can decompose indexed objects into 4- and D-4-dimensional pieces, e.g. p^\mu = \bar{p}^\mu + \hat{p}^\mu. Consequently, in FeynCalc we can also enter D-4-dimensional objects
[p, \[Mu]]
FVE% // FCI // StandardForm
\hat{p}^{\mu }
(*Pair[LorentzIndex[\[Mu], -4 + D], Momentum[p, -4 + D]]*)
[p, q]
MTE% // FCI // StandardForm
\hat{g}^{pq}
(*Pair[LorentzIndex[p, -4 + D], LorentzIndex[q, -4 + D]]*)
When we contract Lorentz tensors from different dimensions, the contractions are resolved according to the rules from the paper of Breitenlohner and Maison, e. g.
[p, \[Mu]] FV[q, \[Mu]]
FVD[%] Contract
p^{\mu } \overline{q}^{\mu }
\overline{p}\cdot \overline{q}
[p, \[Mu]] FVE[q, \[Mu]]
FV[%] Contract
\hat{q}^{\mu } \overline{p}^{\mu }
0
Sometimes we need to switch from one dimension to another, e.g. to convert a 4-dimensional object to a D-dimensional one or vice versa. This is done via
[p, \[Mu]]
FVD[%, 4] ChangeDimension
p^{\mu }
\overline{p}^{\mu }
The second argument of ChangeDimension
is the new dimension . The most common choices are 4, D or D-4
[p, \[Mu]]
FVD[%, D - 4] ChangeDimension
p^{\mu }
\hat{p}^{\mu }