This thesis describes the potential of relaxor ferroelectrics for solid state cooling based on the
electrocaloric effect. The core of this investigation is to identify the reliable methods to
correctly evaluate the electrocaloric effect and develop materials with the properties required for
commercial electrocaloric cooling.
A thorough review of the state-of-the-art electrocaloric research reveals that too many research
groups still rely on the indirect evaluation of the electrocaloric effect (ECE) from polarization
measurements and highlights the need for direct ECE measurements. A direct electrocaloric effect
measurement set-up based on a modified-differential scanning calorimeter, allowing the acquisition
of both ther- mal (ECE, heat capacity) and electrical (P-E loops, leakage current) information
simultaneously, has successfully been constructed and benchmarked.
Direct ECE measurements have been performed on normal ferroelectrics, such as barium titanate, but
also well-known relaxor ferroelectrics, such as the PMN- PT system, for fundamental understanding
of the electrocaloric effect. These results highlight the importance of the polar direction of the
electrocaloric mate- rials with regard to the direction of applied electric field. A region with
negative ECE, which could be exploited to increase the efficiency of electrocaloric cooling cycles,
has been identified for <001 >-oriented PMN-30PT by both direct and indirect measurements. This
negative ECE is observed in the vicinity of the low temperature field-induced structural phase
transition, which forms intermediate lower-symmetry monoclinic phases. The occurrence of this
phenomenon requires the combination of several parameters related to the direction of application
of the electric field. The results on PMN-PT also show how the chemical disorder in ferroelectric
relaxors provides important entropy changes over the ferroelectric to paraelectric transition which
enables an extended cooling regime as the ECE maximum can be extended over several tens of degrees.
Direct ECE measurements have therefore been performed on novel relaxor fer- roelectrics, including
perovskite, Aurivillius phase and tungsten bronze structures, with a focus on lead-free, for
environmental purposes, highly disordered materi- als. For most of these systems, the direct ECE
measurement presented here are the first ever reported. The presence of a dual electrocaloric peak, sometimes far
above the ferroelectric to paraelectric transition, is confirmed in all the studied relaxor
ferroelectrics. This peak was attributed to the extra contribution to the field-induced entropy
change by the polar nanodomains. The presence of this ex- tra ECE peak confirms the great potential
of relaxor ferroelectrics for solid-state electrocaloric cooling over a range of temperature broad
enough for commercial applications.
Comparisons between direct and indirect measurements are performed on nu- merous systems throughout
this thesis, in order to identify the domain of validity of the indirect method still overly used
in the literature. It is shown that the indi- rect ECE method, although it gives satisfactory
results for normal ferroelectrics, is unreliable for strong relaxor ferroelectrics above the
ferroelectric to paraelectric phase transition, where the dual peak is observed by direct
measurements. These limitations are attributed to the inability of the indirect method to account
for the field-induced entropy contribution of the polar nanodomains to the electrocaloric
effect.