I don't read anything indicating "not working" in your previous post. You reported unstable operation, which can be expected under the said conditions. It may be the case, that the LD has been already damaged by overheating or exceeded optical power rating, hopefully not.
I think, you misunderstood several points in the datasheet.
- 87 mA is specified as
typical operating current with 20 mW optical output. It's not guaranteed to keep the diode power below 22 mW maximum rating.
- As you can easily calculate, the electrical input will be about 180 mW. The maximum case temperature is specified with 40°C. This means, it's impossible to achieve specified operation without a heatsink.
- - - Updated - - -
An additional point about LD behaviour in open loop. Not shown in your short form datasheet, but the laser threshold is also strongly temperature dependent. This leads to a reduction of optical output, or even stop of laser operation with increasing temperature, see the characteristic of a red 30 mW LD from Mitsubishi.
In so far, the observed behaviour, is pretty understandable.