14.1 Rotation of Vectors 261
reflects that thePμν(m)are ‘eigen-tensors’ of the tensorHμν. On the other hand,Hμν
can be represented as a linear combination of the projection tensors, viz.
Hμν=
∑^1
m=− 1
imPμν(m). (14.12)
Some additional formulas involving second rank projection operators are:
hνhμaν =
(
2
3
Pμν(^0 )+
1
2
Pμν(^1 )+
1
2
Pμν(−^1 )
)
aν, (14.13)
hνhκhμbνκ =
(
3
5
Pμν(^0 )+
8
15
Pμν(^1 )+
8
15
Pμν(−^1 )
)
bνκhκ, (14.14)
whereaνandbμνare a vector and a second rank tensor.
14.1.3 Rotation Tensor for Vectors
With the help of the projection tensors, the rotation tensor for vectors can now be
expressed as
Rμν(φ)=
∑^1
m=− 1
Pμκ(m)(exp[φH])κν=
∑^1
m=− 1
exp[imφ]Pμν(m). (14.15)
Decomposition into real and imaginary parts leads to
Rμν(φ)=Pμν(^0 )+cosφ
(
Pμν(^1 )+Pμν(−^1 )
)
+sinφi
(
Pμν(^1 )−Pμν(−^1 )
)
. (14.16)
Notice that
Pμν(^0 )=hμhν≡Pμν‖, Pμν(^1 )+Pμν(−^1 )=δμν−hμhν≡Pμν⊥, (14.17)
correspond to the projection tensors onto the direction parallel and perpendicular to
h, denoted byP‖andP⊥, respectively. Furthermore, one has
i
(
Pμν(^1 )−Pμν(−^1 )
)
=Hμν. (14.18)
Thus the rotation tensor also reads
Rμν(φ)=hμhν+sinφεμλνhλ+cosφ(δμν−hμhν). (14.19)