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.
Expert fallacy: I know a lot about X, so I'm also an expert on adjoining topics V, W, Y, and Z. Also on topic "ecks". And "x-rays". And "excerpts" and "excommunication". Ah, hell: I'm smarter than most, so I'm usually right. Especially about things I'm sure I right about.