I thought the problem was primarily galvanic activity between copper and aluminum, not thermal growth. A bit of searching says this is so in environments with any wetting or very high humidity. The increased resistivity of aluminum's oxide layer is also part of it.

Basically: they just don't play well together.