I tried a sample of sanded Lexan & was disappointed with result. Distance for 2.5" spread was about 8" with a non uniform field. Might do some more sanding. Tried a parabolic concentrator,
5mm exit, 13 mm entrance, 10 mm long with an acceptance angle of about 45 deg. @ US $2.00.
A cone made from Al tape, 5 X 15 X 20 mm also about 45 deg. worked. A large cone , .25" X 2" X 3.5" covered the 2" dia. at 0 deg. but only about 20 deg. acceptance angle. Light source 3mW red laser non modulated in semi dark room.

Lexan is not a good choice for optical applications of any kind. Lucite or plexiglass are far better.

 

Here is what I have suggested. Namely, use the acrylic as a light guide/pipe. There is no space between the diffusing surface and the acrylic body. The sanded surface acts to get the light inside the acrylic at other than perpendicular.


I just did a quick experiment with an acrylic scraper about 1/4" thick x 1" x 3" and a laser level. If the laser was aimed at a polished surface, it went straight through with only slight scattering. If it hit a sanded area, it looked like the whole inside lit up. Ideally, you might want to embed the detector in the acrylic with a clear plastic to match the acrylic's index of refraction, but that might not be necessary.

Here is the IR transmission spectrum from typical acrylic:



In the near IR, it is quite transparent.

@btebo makes a good point. Are you sure the IR light that is being detected is from an IR laser and how narrow a beam is it?

 

I think it is a 5 mW. red laser, post #3.

 

Red or near IR makes no difference. The laser I did the experiment with was actually red (Leica level).

 

I know your PCB design is already done, but I wonder how well placing the sensor on the edge of a sheet of acrylic would work. My thinking is if the outside face of the sheet is sanded to scatter the light, it is then reflected to the edge where the sensor is located. You might get away with using only 4 sensors, one on each edge of the sheet.

Kinda like those artistic designs were they put LEDs on the edge shining into the acrylic and then sand/etch areas they want to glow:

 

I was skeptical but crude test prove otherwise. Two Ledtech LT 9593-0125 Si photo transistors are spaced 2.5 ", each with cone reflector 5mm X 15 mm X 20 mm overlaid with red filter. Sanded Lexan 5 " in front . Light is unmodulated narrow beam 5 mW. red laser @ 2'. Transistors monitored with ohm meters. Beam passing between sensors is detected from second sensor before loosing signal from first. Angle of detection about 45 deg.

 

I was skeptical but crude test prove otherwise. Two Ledtech LT 9593-0125 Si photo transistors are spaced 2.5 ", each with cone reflector 5mm X 15 mm X 20 mm overlaid with red filter. Sanded Lexan 5 " in front . Light is unmodulated narrow beam 5 mW. red laser @ 2'. Transistors monitored with ohm meters. Beam passing between sensors is detected from second sensor before loosing signal from first. Angle of detection about 45 deg.

It will work even bbetter if you use acrylic plastic.

 

It will work even bbetter if you use acrylic plastic.

I said polystyrene earlier, because I couldn’t think of “acrylic plastic”. I second @MisterBill2 !

 

I said polystyrene earlier, because I couldn’t think of “acrylic plastic”. I second @MisterBill2 !

I was commenting about LEXAN. OK, I am not certain about the properties of polystyrenem probably some place in between,