It’s not about the mass

Electrocaloric cooling devices traditionally comprise sub-millimetre-thick ceramic working bodies surrounded by relatively massive apparatus. Now, cooling devices that are each based on a thin polymer layer have been stacked to yield a composite lightweight device that pumps heat across a wide temperature span.

There is currently growing interest in cooling devices based on caloric materials, in which phase transitions, and thus thermal changes, are driven near room temperature by magnetic fields, electric fields or mechanical fields1. Electric fields are easy to administer as one simply applies a voltage, but the resulting electrocaloric (EC) effects cannot be driven hard without breakdown in monolithic sub-millimetre-thick ceramic working bodies, which has limited the temperature span of traditional prototype EC coolers2,3,4 to around 13 K.

Writing now in Nature Energy5, Qibing Pei and colleagues at the University of California show that a stack of much lighter EC devices, each based on a 0.1-mm-thick polymer bilayer, can pump heat across a temperature span as wide as 8.7 K. This performance catches the eye because it compares rather well with recently unveiled prototypes6,7 based on much thicker ceramic multilayer capacitors8. These state-of-the-art working bodies are roughly as thick as the monolithic ceramics of old2,3,4, but they comprise many thin layers that can be driven hard.

The temperature span between the hot and cold ends of an EC cooling device need not be limited by the voltage-driven temperature change that can be achieved under adiabatic conditions in the EC working body on which it is based. For example, one can create temperature gradients along loops (Fig. 1a) or columns (Fig. 1b) of flowing fluids that exchange heat with EC working bodies, at the cost of requiring a pump. Alternatively, one can develop temperature gradients along extended caloric working bodies, either using a flowing fluid as above9, or by translating one line of EC working bodies that make and break contact with a second such line at the cost of requiring an actuator (Fig. 1c). EC prototypes have therefore traditionally required a substantial mass of non-caloric components, some of which compromises performance by virtue of its thermal mass. Moreover, device complexity has precluded serial stacking for an enhanced temperature span.

Fig. 1: Miniaturization of electrocaloric prototypes.

ad, Simplified schematics, not to scale, show devices that pump heat from cold ends (blue) to hot ends (red) by cyclically driving EC working bodies (yellow) with a voltage while cyclically translating either a fluid (a,b) or the EC working bodies themselves (c,d). The working bodies comprise sub-millimetre-thick plates of PST (PbSc0.5Ta0.5O3)2,3,4 (a), sub-millimetre-thick MLCs of PST6 (b), the same MLCs of PST7 (c) and a much thinner 0.1-mm-thick polymer bilayer of P(VDF-TrFE-CFE), that is, poly(vinylidene fluoride-trifluoroethylene-chlorofluoroethylene)11 (d). The device in d is simpler, smaller and lighter than the others, and can be stacked in order to pump heat across 8.7 K (ref. 5).

Conceptually and technologically, it is perhaps simpler to repeatedly translate a single EC working body between a hot end where it dumps heat, and a cold end where it absorbs heat10. In their 2017 report11, Pei and colleagues employed this strategy in an elegant way by electrostatically actuating the flexible EC bilayer itself (Fig. 1d). However, the device temperature span was limited to 2.8 K, and it remained unclear whether EC devices based on thin polymers can pump enough heat to compete with EC devices based on thicker ceramics.

By stacking four of the lightweight polymer devices, Pei and colleagues have now achieved a temperature span of 8.7 K at zero cooling power, or a cooling power of 906 mW at zero temperature span5. Interestingly, these figures are similar to the values of 13 K and 1.22 W for one of the prototypes based on relatively massive ceramic multilayer capacitors (Fig. 1b). This equivalence would appear to challenge the conventional wisdom that it is desirable to employ the largest possible mass of EC material, albeit dispersed to ensure that no region lies far from the surface given the low thermal conductivity of EC materials.

The ability to stack four devices was facilitated by the simplicity and compactness that arose from electrostatically actuating EC bilayers. By thus avoiding the pumps and actuators that have been employed elsewhere (Fig. 1a–c), it was possible to reduce the mass of these and other non-caloric components. Moreover, two positive outcomes arose from the fact that each bilayer device in the stack was cycled in antiphase with respect to a neighbour. First, the instantaneous flow of heat was relatively smooth because one bilayer could absorb heat from its cold end while its counterpart dumped heat at its hot end, mimicking the antiphase operation seen elsewhere10 (Fig. 1a). Second, energy efficiency was improved because each discharging bilayer could help charge its neighbour, as originally demonstrated in a prototype based on commercially available ceramic multilayer capacitors10.

In the future, EC devices based on flexible bilayers could find some niche application because they are small and lightweight. Although it may be challenging to improve cooling power because the active thermal mass is small, there is cause for optimism given the aforementioned similarity of performance with respect to one of the prototypes based on relatively massive ceramic multilayer capacitors (Fig. 1b). Therefore, it remains to be seen whether the push for applications will be best served by using low-mass polymer bilayers with a mechanical flexibility that permits total mass to be low, or thicker ceramic multilayer capacitors that require higher total mass. However, as long as we remain stuck with the circumstantial correlation between active and total mass, the low mass solution should be relatively attractive for any application.


  1. 1.

    Moya, X. & Mathur, N. D. Science 370, 797–803 (2020).

    Article  Google Scholar 

  2. 2.

    Sinyavsky, Y. V. et al. Ferroelectrics 90, 213–217 (1989).

    Article  Google Scholar 

  3. 3.

    Sinyavskii, Y. V. & Brodyansky, V. M. Ferroelectrics 131, 321–325 (1992).

    Article  Google Scholar 

  4. 4.

    Sinyavskii, Y. V. Chem. Petrol. Eng. 31, 295–306 (1995).

    Article  Google Scholar 

  5. 5.

    Meng, Y. et al. Nat. Energy https://doi.org/10.1038/s41560-020-00715-3 (2020).

  6. 6.

    Torelló, A. et al. Science 370, 125–129 (2020).

    Article  Google Scholar 

  7. 7.

    Wang, Y. et al. Science 370, 129–133 (2020).

    Article  Google Scholar 

  8. 8.

    Nair, B. et al. Nature 575, 468–472 (2019).

    Article  Google Scholar 

  9. 9.

    Pecharsky, V. K. & Gschneidner, K. A. Jr J. Magn. Magn. Mater. 200, 44–56 (1999).

    Article  Google Scholar 

  10. 10.

    Defay, E. et al. Nat. Commun. 9, 1827 (2018).

    Article  Google Scholar 

  11. 11.

    Ma, R. et al. Science 357, 1130–1134 (2017).

    Article  Google Scholar 

Download references

Author information



Corresponding authors

Correspondence to Xavier Moya or Neil D. Mathur.

Ethics declarations

Competing interests

X.M. is director of research at Barocal Ltd.

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Moya, X., Mathur, N.D. It’s not about the mass. Nat Energy 5, 941–942 (2020). https://doi.org/10.1038/s41560-020-00741-1

Download citation


Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing