Electrical – Beer lambert law ratio

First of all, the relevant regulation requires lasers to emit less than 5mW of radiation. This is not the same as having 5mW electrical power:

21 CFR 1040.11(b) and 1040.11(c), limit surveying, leveling, and alignment, and demonstration laser products to Class IIIa. This means that pointers are limited to 5 milliwatts output power in the visible wavelength range from 400 to 710 nanometers.

LD efficiency is about 30% for the best lasers, the one you have bought probably has about 10%. So you can safely multiply the power you apply by at least 3.

Start by measuring the voltage drop \$V_D\$ on the LD with the resistor you already have, and also measure the current (voltage drop on the resistor multiplied by its resistance: \$I_D=\frac{V_R}R=\frac{9-V_D}R\$). Your target is to have \$P=V_D*I_D=0.005*3=0.015\$. Gradually lower the resistance until you get the right power.

Second, instead of rising the threshold to turn your detector ON, you may consider lowering the threshold for it to go off. Cover your detector with a piece of glass / transparent plastic with a color close to that of your laser beam. It should let most of the laser light through, while rejecting a significant portion of ambient light.

Finally, if you need a bigger detection surface, consider installing a lens of that size, and place the detector in focus behind the lens. In that case, a laser beam hitting anywhere on the lens surface along its optical axe will end up in the detector:

It's a particular classification of nonlinear behavior that has important applications.

In the same way as you can consider, say, a BJT as linear over a limited range, you can consider something like a diode as square law over a limited range. That simplification allows you to analyze functions such as RF detectors analytically.

See, for example, this Agilent paper "Square Law and Linear Detection".