This question may also be asked in physics stackexchange, but I'll try my luck first here. I want to set up a water temperature sensor system. My temperature sensor will most probably be DHT22 due to its availability and its good performance with ESP8266.
The problem is DHT22 is not water resistant. Three solutions came to my mind:
Coating the DHT22 with hot silicone gun. Hot silicone shouldn't really damage an electronic devices, supposedly to operate perfectly between -10 to 85 degrees celcius. However, it is a lumpy solution at best and there is no guarantee water won't sneak in, in some time.
Making a probe out of metal and fit it to DHT22. II'll dip the probe into water and it will have a very large surface area like a large sphere, inside water. The probe and DHT22 will be in contact with a thermal gel. I can then coat the DHT22 with tape or hot silicone to avoid any damage due to splashing water. I can then calibrate the DHT22 reading with actual water temperature to get rid of any bias due to it being out of water. I can even coat the DHT22 with polystyrene to reduce bias from air temprature.
Making a metal box and fit the DHT22 in it and dip it into water. Now DHT22 will be in water. Again, it will be in contact with metal via a thermal gel. I can then screw the box, tape it with water resistant tape and then hot silicone over the tape. This time, it will again be lumpy but I'll have time to change the once a month or so and water resistant tape will offer extra resistance against small leaks.
I wonder which one, or some other solution, would be best for such a task.