In the second medium,
2θ1−Δθ−Δα=90o
where,
Δθ=θ1−θ2
Δα=α1−α2
which implies 2θ1>90o and so, θ1>45o
So,
θ1+θ2−α1+α2=90o
both material properties ε and n are involved. Alternatively,
2θ1−90o=Δθ+Δα
the change in orientation needed, is from both a change in θ and α.
When we set,
α1=θ1 --- (1)
then,
θ2+α2=90o --- (2)
the ray splits into two perpendicular directions with orthogonal polarization. Given a light source with a spread of values in θ, there can be a number of θ for which a split of the refracted rays orthogonal to each other can occur. Expressions (1) and (2) are the same as those we obtained in the post "Still Looking For Brewster" dated 16 Aug 2015. From that post,
tan(α1)tan(α2)=ε2ε1
sin2(α1)cos(α1)cos(α2)=ε2ε1n2n1
but α1+α2≠90o as oppose to the often quoted α1+α2=90o in the derivation for Brewster's angle. Instead,
sin(α2)=n1n2sin(α1)
cos(α2)=√1−(n1n2)2sin2(α1)
So,
tan(α1).n1n2sin(α1)√1−(n1n2)2sin2(α1)=ε2ε1
tan2(α1)sin2(α1)=(ε2ε1n2n1)2{1−(n1n2)2sin2(α1)}
tan2(α1)sin2(α1)+(ε2ε1)2sin2(α1)=(ε2ε1n2n1)2
sin4(α1)+(ε2ε1)2sin2(α1)(1−sin2(α1))=(ε2ε1n2n1)2(1−sin2(α1))
sin4(α1)+(ε2ε1)2sin2(α1)−(ε2ε1)2sin4(α1)=(ε2ε1n2n1)2−(ε2ε1n2n1)2sin2(α1)
sin4(α1){1−(ε2ε1)2}+(ε2ε1)2{1+(n2n1)2}sin2(α1)=(ε2ε1n2n1)2
sin4(α1){(ε1ε2)2−1}+{1+(n2n1)2}sin2(α1)=(n2n1)2
sin4(α1){(ε1ε2)2−1}+{1+(n2n1)2}sin2(α1)=(n2n1)2