In order to solve the problem of the pattern on the surface reference sample I had to move the pump probe by 2°. The resulting path was outside the periscope mirror, so I had to move the periscope. It also went outside the chopper, so I had to move the chopper as well. First I aligned the pump beam without the focusing lens and measured the transmitted peak from the pinhole at different Z
  Y                 Z
121.384100  55
120.684361  35
119.984621  15
angle 0.034987rad = 2.004deg

Then I put the focusing lens again and  after putting the lens maximize the AC signal on the surf ref sample. Then measured again the transmitted peak position:

  Y                 Z
121.923354 55
121.28          35
120.57          15
angle = 1.94deg

and moving the pinhole around the crossing point I maximized the transmitted power of both pump and probe at the following coordinates:
X 327.432
Y 121.255
Z   34.900

this is the new position of the crossing point

The first screenshot is the surface reference scan along Z.
The second screenshot is the surface reference scan along Y, that shows that the pattern amplitude is now of about 5%
The third screenshot is the bulk reference scan along Z.

I also made a map of the surface reference sample
Please note that the map Absorption[ppm] axis is not calibrated.