I'm working with LMT85 temperature sensor and I want it to be waterproof, once it will be working submersed by water. I'm thinking on use hot glue on connections and then use a heat shrink tube.
Is it a good idea use hot glue?
I tend to coat the entire assembly in hot melt glue, and put a sleeve of heat shrink over it. On gentle heating, the sleeve shrinks, extruding the glue out of the ends, minimising the bulk of glue around the assembly, and sealing round the wires. This way, the bulk of the waterproofing is provided by the heatshrink tube, with the glue sealing the ends. Squeezing the ends down further reduces the cross sectional area of glue exposed to the fluid.
Note that stranded PVC sleeved wire is not waterproof between the insulation and the wire. The insulation should come all the way out to the dry area.
If the assembly needs only two connections, then I often use co-ax so that there's only one seal being made.
It would be better to find a more conductive epoxy, because the hot glue is probably more thermally insulative which will slow down your time constant. I'm not quite sure if hot glue is really waterproof over very long times.
I have used thermally conductive laquers in the past to coat thermistors and waterproof them. Now since discontinued, but you get they have other similar product lines.
I used exactly this approach, and make sure the sleeve is nice and long. I have not used for a submersed sensor. It is only if long life (commercial use) matters to you that you need to worry.
This suggests that your glue sticks might be EVA, and I would be a bit dubious of it cf. PVA glue and water, self-hydrolysis of EVA shoe materials. Put a sliver of it into a sealed jar of water on the window sill, and see what it looks like in 6 months.
I have also dipped in electronic flowable silicone (dow 3140) then when that is set, hot melt sleeved it as above for protection.
Plastics are not generally waterproof, oxygen proof, or anything proof when looked at closely, they tend to be resistant to one thing but not others (which is why juice cartons have multiple layers of different materials)
The thing not to use is epoxy, which is very porous to H2O. The manufacturers of IC encapsulants claim urethanes (PUR) are very good for water resistance.
You did not say how deep you plan to submerge your sensor. Note that water pressure increases about 0.43 psi per foot. Thus even at only ten feet, the water pressure will be over 4 psi. I do not recommend heat shrink tubing. That may work for reasonable water proofing in air. It will undoubtedly leak underwater. I do recommend a two part polyurethane. Polyurethane has been extensively used for years for encapsulating sonar hydrophones for use at depths of hundreds of feet. You do have to be careful to avoid too many air bubbles although that is more critical for acoustic applications (since air is a poor sound conductor) than for a temperature sensor.