This post focuses on the question of whether the chemicals in heat pumps are safe. The chemicals do appear to be safe for home occupants, even if they leak into indoor air.
Heat pumps contain refrigerant fluid which circulates between their outdoor and indoor units. The refrigerants are chemicals in the hydrofluorocarbon family (HFCs). Concern was expressed to me recently that leaks of refrigerant HFCs might be toxic to home occupants because they belong to the “PFAS” family of chemicals. They also belong to the even broader family of “halogenated hydrocarbons.” Both of these chemical families have well-earned reputations as potentially harmful. However, inferences based on family membership alone can be inaccurate for chemicals just as they can for humans.
Available toxicological evidence appears to assuage concerns about home exposures to HFCs through heat pump leaks. The HFCs in common heat pumps, although members of two dangerous families, are much safer than their relatives.
HFCs were developed in the 80s and 90s as replacements for chlorofluorocarbons (CFCs). After scientists raised the alarm that CFCs were destroying the ozone layer, an intense international collaboration of regulators and chemical industry scientists began to search for replacements. Unlike their harmful relatives, the HFCs were all heavily vetted for safety before they went on the market as refrigerants. (The impact of HFCs on global warming is a different conversation.)
This post collects resources on the toxicology of HFC heat pump refrigerants. The outline is as follows:
- Defining chemical families
- Development and vetting of HFCs
- Worst-case home exposure levels to HFCs
- Conclusion
Defining chemical families
PFAS
The PFAS family of chemicals consists of hydrocarbons in which least one carbon atom is fully fluorinated — either a CF2 group in the middle of the structure or a CF3 group as branch or terminus. Major agencies and working groups have narrowed this very broad definition in different ways.
Roughly a dozen PFAS chemicals — mostly polyfluorinated alkyl acids (“PFAAs”) — were widely used around the globe before it surfaced that they are durable in the environment and very harmful to health even at small concentrations. Research is now rapidly expanding on the harms caused by these chemicals. In at least one instance, after the harms of PFAAs were known, a major chemical company shifted from PFAAs to another dangerous PFAS. To prevent additional “regrettable substitutions,” regulators are now requiring reporting on the very broad universe of all PFAS chemicals. While the harms of the core dozen PFAAs are increasingly well documented, there is much less knowledge about most of the thousands of other PFAS chemicals.
Halogenated hydrocarbons (HHCs)
Halogenated hydrocarbons are hydrocarbons with one or more of the hydrogen atoms replaced by a halogen — fluorine, chlorine, bromine, or iodine. In the broadest use of the term, other atoms or groups may also be attached. The carbon atoms may be arranged in multiple ways (chains, rings, branching structures) and they may be joined by single or double bonds. Broadly defined in this way, HHCs include a very wide range of chemicals, among them highly toxic pesticides. In fact, PFAS fit within the broadest definition of HHCs.
The HHC term may be used more narrowly as a synonym for alkyl halide — a hydrocarbon that is saturated (has no double bonds between carbon atoms), is not a ring, and in which some or all of the hydrogen atoms have been replaced by halogens (and no other atoms or functional groups of atoms are attached).
Alkyl halides are described generally in the Encyclopedia of Toxicology as “very toxic when inhaled in high concentrations or extended periods of time.” The Toxic Use Reduction Institute, in recommending reporting requirements for short-chain alkyl halides stated:
Based on the [Scientific Advisory Board’s] review, central nervous system (CNS) effects are found consistently across the chemicals in [the short-chain alkyl] category. Additional hazards noted for some of these chemicals include target organ toxicity; reproductive and developmental toxicity; carcinogenicity; and respiratory effects.
Toxics Use Reduction Institute Policy Analysis, February 2018, C1-C4 Halogenated Hydrocarbons/Halocarbons Not Otherwise Listed, page 2.
Some alkyl halides are naturally occurring.
Chlorofluorocarbons (CFCs)
Chlorofluorocarbons (CFCs) are alkyl halides in which all the hydrogen atoms have been replaced with a mix of fluorine and chlorine atoms. When they were developed in the 1930s, they were an important innovation because refrigerants used previously were toxic and/or explosive. At a convention in 1930, the inventor of CFCs inhaled a sample and blew out a candle — demonstrating faith in both their non-toxicity and their non-flammability. This article provides a useful overview of the history of refrigerants, including the half century dominance of CFCs (and HCFCs in which one or more hydrocarbon atoms remains).
CFCs were nonflammable, nonexplosive, noncorrosive, very low in toxicity, and odorless, and . . . their vapor pressures and heats of vaporization made them suitable for refrigeration applications. . . .
CFCs and HCFCs rapidly replaced other refrigerants in all but applications where companies accepted the increased risk of flammable and toxic refrigerant releases or in applications where the existing technologies were more energy efficient. For example, ammonia continued to be used in cold storage, ice making, ice rinks, and absorption refrigerators using gas flame as an energy source.
Stratospheric ozone, global warming, and the principle of unintended consequences—An ongoing science and policy success story (Andersen, et al. 2013). See also this EPA page describing history of refrigerants.
Of course, the tragedy of the CFCs was that they turned out to be harmful to the ozone layer and were ultimately phased out under the Montreal Protocol.
HydrofluoroCarbons (HFCs)
As discussed further below, HFCs were developed to replace CFCs. HFCs are also a subfamily of the alkyl halides — they are saturated hydrocarbon chains with some of the hydrogen atoms replaced by fluorine. So, they are all of the chemical form CnHmFk where m + k = 2n + 2. They are referred to by a special numbering system in which HFC-XYZ (sometimes R-XYZ) means the HFC in which X = n-1, Y = m+1 and Z = k. So, for example, R-125 is pentafluoroethane, C2H1F5.
R-125, Pentafluoroethane, C2HF5
Development and vetting of HFCs
Response to the Montreal Protocol
By the 1980s, safe refrigeration was an essential feature of modern life. Chemical companies mobilized to develop safe refrigerants to replace CFCs as the Montreal Protocol was adopted. This article describes the toxicity review mobilization as follows:
Following the adoption of the Montreal Protocol, four of the world’s largest producers of CFCs, AlliedSignal, Atochem, DuPont and ICI, formed a consortium to develop preliminary programs for the environmental and toxicological evaluation of alternatives to the then current chlorofluorocarbons. . . . In November 1987, just months after the signing of the Montreal Protocol, the major global producers of fluorocarbons met in New York and formed an international consortium called the Program of the Development of Alternative Fluorocarbon Toxicity Testing (PAFT). . . .
The PAFT consortium was composed of three committees, the Management Committee, responsible for financing the program and communication of the results to the business areas in the member companies, the Toxicology Committee who were responsible for the development of the toxicology data that would allow for a determination of the safety of these products when used in their intended applications and a Product Standards Committee who were responsible for setting product purity specifications that could be achieved by all producers and would meet the safety standards set by the Toxicology Committee . . .
Successful candidates had to have low toxicity as they were replacing the CFCs that were of low toxicity. Thus it was expected that they would have relatively high Occupational Exposure Levels. . . .
The development of environmentally acceptable fluorocarbons (Rusch, 2018)
Toxicological Vetting
Rusch (2018) describes the stage by stage assessments of the toxicity of candidate HFCs by the toxicologists collaborating through the Program of the Development of Alternative Fluorocarbon Toxicity Testing.
The first phase included the short term, genotoxicity and screening studies designed to rapidly identify substances not suitable for further testing. They are listed below:
- Acute inhalation toxicity (rat)
- Cardiac sensitization (dog)
- Ames assay
- Chromosome aberration in human lymphocytes
- In vivo mouse micronucleus assay
- 2-week repeat exposure inhalation toxicity screen (rat)
The second phase involved more complex, time consuming and expensive studies. It included:
- 4-week repeat exposure inhalation toxicity (rat)
- Teratology (rat and rabbit)
- Preliminary pharmacokinetics (blood levels, uptake and elimination)
- Environmental toxicology: Typical studies included toxicity to algae, invertebrate aquatic species (Daphnia magna) and fish (e.g. zebra fish and trout).
The third phase included:
- 13-week repeat inhalation toxicity (rat)
- Metabolism
The fourth and final phase included:
The development of environmentally acceptable fluorocarbons (Rusch, 2018)
- Chronic toxicity/carcinogenicity inhalation study (rat)
- Generation inhalation reproduction study (rat)
Rusch (2018) details the findings from this research program individually for each major commercial HFC. The paper then summarizes the findings as to the major commercial HFCs as follows:
As a class, these products have very low water solubility, . . . They all have high vapor pressures and most are gasses at 25°C. As a consequence, these materials will not accumulate in water and will partition rapidly into air. . . . all HFCs and HFOs used in commercial applications had LC50s above 100,000 ppm (10%). None of these commercial substances were shown to be highly mutagenic, developmental toxins or carcinogenic. In conclusion based on available evidence, working together on toxicology and environmental issues, the Global Fluorocarbon Industry was able to develop new, safe refrigeration products and have them into commercial markets in less than 10?years.
The development of environmentally acceptable fluorocarbons (Rusch, 2018). See also An overview of environmental hazards and exposure risk of hydrofluorocarbons (HFCs) (Tsai 2005) and Reproductive and developmental toxicity of hydrofluorocarbons used as refrigerants (Ema, et. al, 2009).
While the author of the above quoted conclusion does have industry affiliations, the body of studies generated by this systematic program of collaborative investigation also appears to have satisfied regulators as to the relative safety of the HFCs which are currently used in heat pumps. During both the Clinton years and the GW Bush years, the EPA recognized the toxicology teams that vetted HFCs as environmental heroes. OECD found in 2004 and 2005 that the most common heat pump chemicals had “a low hazard profile” for human health (finding as to HFC-125 and as to HFC-32).
The 1990s consortium studies also appear to have satisfied the broader scientific community. My searches of several chemical and toxicology data collections mostly surfaced reports on the consortium studies and did not surface new toxicology findings within the past twenty years. My searches focused on studies of HFC-125 and HFC-32, the two ingredients of the most common heat pump refrigerant, R-410A. HFC-125 is being phased out due to global warming concerns, but HFC-32 is likely to remain in use for some years.
Common heat pump HFCs — latest year of toxicology studies found by searches
Collection Searched for Toxicology Findings | HFC-125 | HFC-32 |
---|---|---|
EPA Comptox hazard findings list | 1992 | 1996 |
EPA Comptox Pubmed abstract sifter | 2000 | 1996 |
EPA Human health assessment (last update date) | 2002 | n/a |
European Chemical Agency Dossier (Studies cited in toxicological data tab) | 2001 | 1994 |
OECD Existing Chemicals Database — Final Assessment Date | <2005* | <2004* |
PubChem Hazardous Substances Databank | <2005** | 1996*** |
ASHRAE Standard 34, Addendum F, Safety Classification of Refrigerants | <2005**** | <2005**** |
** Citing the 2005 OECD results and other secondary sources, apparently referring to the pre-2000 studies.
*** Also cites a review article dated 2009 which was not accessed.
****Addendum to 2019 Standard cites secondary summary sources dated before 2005, mostly a commissioned literature review dated 2000.
Worst-case home exposure levels to HFCs
The available evidence seems to indicate that neither acute nor chronic exposure to HFC refrigerants in homes reaches levels that pose material risk to health.
The refrigerant content of heat pumps is generally well under 10 pounds. The volume of air in a 1250 square foot apartment with 8 foot ceilings (10000 cubic feet) weighs approximately 750 pounds under typical atmospheric conditions. So, the refrigerant concentration in a catastrophic sudden leak confined in the apartment would be on the order of 1.3%, likely much less. In the animal lab studies for R-125 and R-32, animals survived exposures to concentrations up to 40 times higher for hours without enduring effects. A sudden catastrophic indoor leak is hard to imagine since heat pumps are filled at the outdoor head. In any event, the concentration would drop rapidly due to air circulation.
A much more likely scenario is chronic leakage over a period of months during which air would be changed many times. Even a relatively low ventilation rate means 0.5 air changes per hour. Over a heating season of 120 days , this would mean dilution of a worst case 10 pound leak by 1440 air changes (.05 * 24 * 120). The resulting average concentration through the heating season would be on the order of a few parts per million — several orders of magnitude below concentrations for which no effects were observed in various animal studies of chronic exposure to R-125, R-32, and other now common HFCs.
Conclusion
Toxicology is a complex specialty, but to this layman reviewing available sources, there appears to be plenty of evidence that the HFCs in heat pumps pose no meaningful risk to occupants and no evidence to the contrary. It is reasonable for regulatory reporting purposes to define the PFAS chemical family broadly. This results in greater visibility into emerging commercial chemicals which could be harmful and also reflects legitimate concerns for the likely durability of PFAS. However, the mere fact that a chemical is among thousands meeting a broad definition of a PFAS sheds little light on the health effects of exposure.
The studies collected in this post dispel for me any concern about health effects on occupants of homes with heat pumps with HFC refrigerants. However, we do need to phase out HFCs for a different reason — they are potent global warming gases and regulators are setting stricter and stricter limits on their use.
I’m impressed (and convinced) by the erudition of this analysis.
Thank you!
I am happy with heat pump information and agree with safety side of it. With that said there is a flaw in heat pump discussions. Right now incentives in Mass are geared towards home owners who occupy their units. There is little to no incentive for a land lord to make the switch to heat pumps especially as Mass Save is run via the utility asking for the person paying the bill rather than the property owner directly. With such a high percentage of residential units in Metro Boston being rented how should this be addressed?
Rental property is actually a huge focus of MassSave and the Energy Efficiency Advisory Council that oversees it.
It’s very challenging, because neither landlords nor tenants have strong incentives to invest in energy saving and decarbonization measures. The landlords don’t pay the heating bills and the tenants aren’t sure they’ll be around long enough to reap the benefits.
But the issue is front and center.
I hadn’t realized you were also a chemist. Great, comprehensive explanation that “chemicals” aren’t all bad.
Definitely not a chemist. Just a generalist doing his best to assess the state of the literature.
Thanks for your work on this subject. The wider effects on climate are still to be solved. It seems every “solution” has a downside.
This is a fantastic deep dive on a complicated subject! Thank you!
Thank you Will for this well documented, concise review.
It would be nice if you do a deep dive into vaccine safety – including the covid-19 vaccines. Even more broadly, I’d love to hear you reflect on how you think the state of Massachusetts did during the covid-19 3-year emergency saga.
I did give the vaccine safety issue a lot of time during the pandemic. Thank you for the hours we spent together on the issue, and also for the literature you sent me which I did review.
Thanks for the information Will, and for your sustained efforts to guide us.
As much as I would like to be fully reassured ( I have a full heat pump system), I noted that the researcher “does have industry affiliations”, and only concludes to the ” the relative safety of the HFCs which are currently used in heat pumps”.
Meanwhile Europe and other countries are already using the much improved R-454B refrigerant replacing R-410A which we are still using here.
I believe that the human health risks are the same and are non-issues for R-454B as for R-410A.
But yes, R-454B is a good substitution for R-410A to reduce global warming impact of leaks (as is straight R-32).
Heat pumps using R-410A will be banned for importation or manufacture in the United States after January 1, 2025 and will be banned for sale after January 1, 2026. Agreed, we are a little behind the Europeans on this.
See additional discussion here.
It’s unlikely technical innovations will forestall the runaway greenhouse effect. Yes, we have to take steps where we can to offset carbon and save money, but heat pumps, electric vehicles, solar panels on the roof are the Roman Catholic Indulgences of our time. We can’t buy forgiveness for our sins of over consuming. Out penny wise remedies like EVs, solar &c may cause is to make more pound foolish shopping choices. We need to have moral innovations to reduce the opiate of overconsumption that induces container ships to burn multiple olympic sized pool volumes of petrol because I clicked a button on Amazon.
This all assumes good faith too by industry to put what they say they are pitting in the system.
There are unintended consequences. Global reduction of ozone-depleting CFC 11 slowed by half as bad actor China produces the worlds home insulation panels.