The main cause of fronting , splitting and also tailing of peaks is the presence of sodium phosphate salt in the mobile phase because it may precepitate inside the column so you must wash the column with distilled water after certain number of injections inorder to get rid of any precepitated salt in the column .
Such a generalization is nonsense. It's true that phosphate will precipitate if you go to high % acetonitrile, but in this case it will rather precipitate before the column, in the mixer or the capillaries. If there would be a general problem with phosphate, why would it be a standard buffer for HPLC? As long as you're not going too high with the ACN content, there won't be any problems.
And btw washing hydrophobic C18 columns with pure water may lead to phase dewetting...
Concerning the original question, I'd supect that with phosphate buffer pH 7.0 at 45°C you might be well over the pH tolerance of that column. These conditions are quite harsh for a standard silica based C18. Dissolution of the base silica can occur, leading to a void at the column head which causes fronting. I'd work at a lower pH or use some sort of high-pH-stable column.
BTW with a modern column you usually don't need any TEA in the mobile phase. That was used decades ago to supress peak tailing...