Can one still use principle component analysis if there is multicollinearity?

Answer 1

Yes. In fact, Principal Component Analysis is by design one of the ways in which multicollinearity can be addressed. That is, a situation in which a number of explanatory variables are highly correlated, typical regression models will deliver inaccurate estimates. This article might be helpful:

https://www.whitman.edu/Documents/Academics/Mathematics/2017/Perez.pdf

Hope this helps.