I've been doing a bit of research into air-to-water heat pumps (using a conventional vapor compression cycle) as I'll be installing one soon (Bear with me—I pro

I've been doing a bit of research into air-to-water heat pumps (using a conventional vapor compression cycle) as I'll be installing one soon (Bear with me—I promise I'll get to thermoelectric devices by the end). They're currently able to replace a boiler for a heating system designed around relatively low-temperature (120°F/50°C) emitters, with lower temperatures leading to better performance (so a large radiant panel in a floor or ceiling is a good choice). One of the underutilized capabilities they have is to do radiant cooling. The tricky part is that any part of the system that drops below the dewpoint (in whatever space it's located in) will condense moisture out of the air, likely damaging itself or things around it over time. You can address this to some degree by either modulating the heat pump or mixing/recirculating water to keep the emitters a few degrees above the dewpoint, but that limits the amount of sensible heat that you can remove from the space and doesn't do anything for latent heat (i.e. humidity). One solution is to use an emitter (e.g. fan coil, either central or in a little wall unit like a mini-split) that has a drip tray that drains the condensate away to where it can't do damage, but the pipes leading to the emitters still need to be insulated and vapor-sealed. This isn't terribly expensive to add to a new system but could be quite expensive to retrofit into a system where the pipes are already buried into the walls. Another approach is to keep the fluid above the dewpoint and use a separate dehumidifier that condenses moisture out of the air and releases the resulting heat into the space. This isn't ideal because you're adding an additional heat source into the space that you're trying to cool. As I understand it, the strength of thermoelectric devices is that they're silent and relatively small, but their weakness is they're not terribly efficient, especially for cooling, and have a steeper drop-off of efficiency with increasing temperature difference between the cold and hot side. But for dehumidification, the temperature difference is likely small (from a few degrees above the dewpoint to a few degrees below it), and also the latent cooling load tends to be a fraction of the sensible cooling load. So the idea would be to build a device that is plumbed inline with the emitter in each room (or possibly just e.g. a bathroom or kitchen) that uses a thermoelectric device to cool a heatsink below the dewpoint (with a fan) and collects the condensate for disposal, but rejects the heat into the return pipe to the heat pump. The advantage over a conventional fan coil-based system is that the in-wall plumbing (and indeed all of the hydronic plumbing) remains above the dewpoint, so the indoor hydronic piping doesn't need insulating or vapor sealing. I could see this being especially useful in buildings converting from a condensing boiler to a heat pump and want to add cooling without tearing open all the walls.