Heat up some sugar on the stove, and it sure seems to melt, but if it were just a simple phase change, then melted sugar would be perfectly clear, not brown. Melted sugar would taste simple and sweet, not bitter and complex. Melted sugar would cool into crystals, not glass. None of those things are true because heat doesn't cause sugar to change phases. Heat breaks it down into something new: caramel.
In fact, caramel is so unlike sucrose, C12H22O11, that its nature can't be expressed by a single chemical formula. Instead, it's a mixture of caramelan (C15H18O9), caramelane (C12H9O9), caramelen (C36H48O24), caramelene (C36H25O25), caramelin (C24H26O13), and over a thousand other compounds "whose names," one scholar lamented in 1894, "science seems to have invented in a fit of despair.
"Realizing that caramel contained more permutations of carbon, hydrogen, and oxygen than their cutesy naming scheme could accommodate, scientists gave up on trying to define it. They even gave up on nailing down a definitive melting point for sucrose. Unlike ice, or even coconut oil, sugar refused to liquefy at a single, consistent, scientifically reproducible temperature. The point of caramelization proved even trickier to isolate, in some experiments occurring at a mere 340°F and in others holding out past 360°F.
If you think that's confusing, you're not alone. Scientists have a better grasp of quantum mechanics than of caramel, which is still poorly understood. But in 2011, a team of researchers finally established that caramelization is a product of thermal decomposition, not melting (you can read the paper here). Not only that, they determined that it's technologically impossible to document the melting of sucrose independent of caramelization.