The Planet Formation Imager (PFI) Project has formed a Technical Working Group (TWG) to explore possible facility architectures to meet the primary PFI science goal of imaging planet formation in situ in nearby star- forming regions. The goals of being sensitive to dust emission on solar system scales and resolving the Hill-sphere around forming giant planets can best be accomplished through sub-milliarcsecond imaging in the thermal infrared. Exploiting the 8-13 micron atmospheric window, a ground-based long-baseline interferometer with approximately 20 apertures including 10km baselines will have the necessary resolution to image structure down 0.1 milliarcseconds (0.014 AU) for T Tauri disks in Taurus. Even with large telescopes, this array will not have the sensitivity to directly track fringes in the mid-infrared for our prime targets and a fringe tracking system will be necessary in the near-infrared. While a heterodyne architecture using modern mid-IR laser comb technology remains a competitive option (especially for the intriguing 24 and 40{mu}m atmospheric windows), the prioritization of 3-5{mu}m observations of CO/H2O vibrotational levels by the PFI-Science Working Group (SWG) pushes the TWG to require vacuum pipe beam transport with potentially cooled optics. We present here a preliminary study of simulated L- and N-band PFI observations of a realistic 4-planet disk simulation, finding 21x2.5m PFI can easily detect the accreting protoplanets in both L and N-band but can see non-accreting planets only in L band. (abridged -- see PDF for full abstract)