Workflow for S1 imagery processing

I intend to derive some polarimetric parameters in PolSAR, based on SLC products processed in SNAP as follows: Apply orbit file > S1 TOPS Split > Radiometric calibration > S1 TOPS deburst > Polarimetric Speckle filter > Terrain Correction > Export C2 matrix to PolSAR

It works relatively well from some individual scenes, however I would like to know how can I deal with data from two [or more] images from the same area (contiguous not overlapped) shall I mosaic the image first and the move to PolSAR or obtain the products in PolSAR and the merge them externally? When I tried to mosaic the images in SNAP I get an error telling that imaginary and real data cannot be handled.