| _ ______
First, let's prove that det(B) = det(B). This is trivially true for a 1x1
matrix. Using induction, we assume it's true for any nxn matrix. Then
_ __ __ ___ __ ___ __
det(B) = B det(S ) - B det(S ) + ... � B det(S ), where the S are the
11 1 12 2 1n n i
appropriate submatrices (from the usual definition of det). Then
_ __ _______ __ _______ __ _______
det(B) = B det(S ) - B det(S ) + ... � B det(S )
11 1 12 2 1n n
____________________________________________
= (B det(S ) - B det(S ) + ... � B det(S ))
11 1 12 2 1n n
because the product of complex conjugates is the complex conjugate of
the product, and the sum of complex conjugates is the complex conjugate
of the sum. We see that this last equation is simply
_ ______
det(B) = det(B)
Which, by the inductive hypothesis, is true for all square matrices.
QED.
(This result could've been reached more succinctly -- it's true of any
polynomial having real coefficients).
_
Suppose now that B is unitary. That is, B B = I. We have that
_ _ _ ______
det(B B) = det(B) det(B) = det(I) = 1. But det(B) = det(B).
_
Let det(B) = a + ib. We have det(B) det(B) = (a + ib)(a - ib) = a� + b� = 1.
That is, |det(B)| = 1.
|