Many particle sensors are useless in foggy/hazy conditions, which ruins many citizen science projects in terms of data quality. Currently, the best solution is to calculate the dew point and then switching them off once you hit a specified limit.
Fog is condensed water droplets suspended in the air. Radiant heat , especially if it's optically focused, could remove them in the area where the PM sensor is looking.
How does this model deal with this?