I recall reading interviews of some top mathematicians (probably at the end of the Princeton guide to Mathematics), who stated that they couldn't work on their

I recall reading interviews of some top mathematicians (probably at the end of the Princeton guide to Mathematics), who stated that they couldn't work on their problems deeply more than a few hours a day. The rest of the day would then be spent attending seminars, reading papers or typing. So yes, it is perfectly normal to be "mentally exhausted" after uninterrupted hours on a mentally challenging task, albeit you could probably spend 9 hours straight on some routine programming without problem, except perhaps boredom.