The carbon part

If you are switching to a heat pump for the carbon reasons, the math should pencil out. We will walk through it.

The grid mix question

A heat pump runs on electricity. The carbon footprint of that electricity depends on where the electricity comes from. In a place where the grid is mostly coal, the carbon per kilowatt-hour is high. In a place where the grid is mostly hydro, nuclear, wind, and solar, it is low. The American grid is a mix that varies state by state, and the mix is shifting toward lower-carbon sources every year.

Even with the dirtiest current grid mix in the continental United States, a heat pump produces less carbon per unit of useful heat than a combustion furnace, because the multiplier effect (the COP, which we cover in the trick) does enough work to overcome the grid emissions. As grids decarbonize, the carbon advantage of the heat pump grows on its own without you touching the install.

The way to think about it is that a heat pump install is a forward-looking decision. The carbon today is good. The carbon in ten years is much better, because the grid is much cleaner. A combustion install is a backward-looking decision. The carbon today is what it is, and it does not improve, because there is no version of natural gas that gets cleaner over time.

The refrigerant question, handled honestly

Heat pumps contain refrigerant. The refrigerant is the working fluid that carries the heat around the loop. If it leaks, the refrigerant itself has a greenhouse impact, often a high one relative to its mass. The honest math on a heat pump install includes the refrigerant.

The two refrigerants you will most often see in current residential heat pumps are R-410A and R-32. R-410A is the older standard, with a global warming potential (GWP) of about 2,088 times that of carbon dioxide on a per-kilogram basis. R-32 is the newer standard, used by most current cold-climate models, with a GWP of about 675. Both are legal. R-32 is roughly a third the impact, on a leak, of R-410A. The industry is shifting toward R-32, and beyond it toward lower-GWP refrigerants like R-290 (propane) for certain applications.

A typical residential heat pump contains two to five kilograms of refrigerant. If the install is done well, the refrigerant stays in the system for the life of the equipment, gets recovered at end of life, and the per-year leak rate is low. If the install is done poorly, the refrigerant slowly leaks, the system loses performance, and the carbon hidden in the leak is real.

This is one more reason the contractor matters. A clean braze, a proper vacuum, a leak-tested system. These are basic install practices and the difference between them and a sloppy install is the difference between zero refrigerant leak over fifteen years and meaningful leak over the same period.

The bottom-line read

For a typical American household, switching from a gas furnace and a separate air conditioner to a heat pump avoids on the order of one to three tonnes of CO2 per year. For a household switching from oil or propane, the avoided emissions are larger, often three to five tonnes per year. The numbers vary with house size, climate, grid, and install quality.

Over fifteen years of operation, the savings compound into the range of tens of tonnes. The refrigerant impact, on a well-installed system, is a small offset against that. The overall direction of the math is not subtle.

If you want a deeper read, the National Renewable Energy Laboratory and Rewiring America both publish work on this, in plain language and with their sources visible. The site maintains a current set of references in the primer.