Dependence of Physical Properties of Piezoelectric Aluminum-Scandium Nitride on Scandium Concentration

Abstract

In this work the physical properties of the piezoelectric aluminum-scandium nitride (ASN) solid solution as a function of scandium concentration were studied using the density functional theory and experimental methods. The phase transition from the wurtzite phase to the rock salt phase at a Sc concentration of 43% was shown. The barriers of transformation from the wurtzite phase to the rock salt phase for various Sc concentrations were obtained. The behavior of the ASN piezoelectric constant d33 calculated by the piezoelectric constants e33, e31, and e15 shows a sharp increase with increasing Sc concentration compared to aluminum nitride AlN. The relationship between the increase in the piezoelectric response of ASN and the softening of the lattice, accompanied by a decrease in the main elastic constants C11, C33, C44 and C66, as well as a decrease in the c/a ratio with increasing Sc concentration, is shown. ASN films with a predominance of the crystal orientation (00·2) were obtained experimentally by magnetron sputtering. The structural properties of the films were studied by X-ray diffraction analysis. A comparison of the experimentally obtained dependence of the c/a ratio on the Sc concentration with the theoretical values showed a good correspondence. Studies of the physical properties of ASN thin films were performed using microwave multi-overtone composite resonators on diamond substrates with a longitudinal bulk acoustic wave (BAW) as the operating mode in the range of 0.5 – 20 GHz. The frequency dependences of the Q-factor of BAW-resonators with different ASN films were obtained, and the frequency dependences of the square of the modulus of the form factor as |m|2 were calculated. The dependences of the elastic constant С33 and the piezoelectric constant e33 for the ASN films with different Sc concentrations were calculated. The calculated and measured values of these constants are agreed within the experimental error.