unique solution means det =/= 0
for both no solutions and infinite solutions, det = 0
to clarify let us imagine two random lines:
ax + by = e ----> gradient = -a/b
cx + dy = f ----> gradient = -c/d
we know that if -a/b = -c/d, then the equations either have no solutions or infinite solutions, since they are either parallel to each other, or are exactly the same line. rearranging the equation...-ad = -bc, ad - bc = 0. [HINT HINT SIMILARITIES WITH DETERMINANT]
we also know that if -a/b =/= -c/d, then the equations have only one solution, since lines can only ever INTERSECT at one point. rearranging equation...-ad =/= -bc, ad - bc =/= 0. that is ad - bc is a non-zero number.
now look at the matrix version:
[a,b;c,d][x;y] = [e;f]
to find [x;y] we must take the multiply both sides by [a,b;c,d]^(-1) = 1/determinant [d,-b;-c,a]
but for the inverse matrix to exist, it is clear that the determinant must be a non-zero number. that is ad - bc =/= 0. [RELATE TO ABOVE]
if the determinant is a non-zero number, we end up solving the matrix, and there will only be one solution.
however if the determinant is 0, that is ad - bc = 0, then there is no inverse, hence we cannot multiply both sides by the inverse matrix and hence there will effectively be no way in using matrix methods, in which case we can only rely on intuitive methods to find out whether the answer is no or infinite solutions.