Corrected bug in last_pj calculation. Using the last active particle in the cell as the upper bound now works.
