[Zurück]


Zeitschriftenartikel:

A. Ababneh, M Alsumady, H. Seidel, T. Manzaneque, J. Hernando-Garcia, J.L. Sànchez-Rojas, A. Bittner, U. Schmid:
"c-axis orientation and piezoelectric coefficients of AlN thin films sputter-deposited on titanium bottom electrodes";
Applied Surface Science, 259 (2012), S. 59 - 65.



Kurzfassung deutsch:
Aluminum nitride (AlN) reactively sputter deposited from an aluminum target is an interesting compound material due to its CMOS compatible fabrication process and its piezoelectric properties. To obtain high piezoelectric coefficients it is a necessary pre-request to synthesize films with c-axis orientation. Besides the influence of sputter conditions on the microstructure of AlN thin films the condition of the substrate
surface is another important factor of utmost importance. In this study, the influence of 350 nm thick titanium metallization DC sputter-deposited on SiO2/Si substrates at varying back pressure levels bp,Ti in the range of 2×10−3 to 14×10−3 mbar on the c-axis orientation and the piezoelectric coefficients of 600 nm thick AlN thin films is investigated. Besides the plasma power for Ti deposition (Pp,Ti = 100 W) the parameters for AlN synthetization are fixed to Pp = 1000 W and bp,AlN = 4×10−3 mbar in 100% N2 atmosphere. Basically, the surface roughness of the Ti bottom layer is the dominating factor resulting either in a high degree of c-axis orientation (i.e. at low bp,Ti values) or in an amorphous AlN microstructure
(i.e. at high bp,Ti values). Under low pressure conditions, a smooth and dense surface characteristics is achieved due to a higher kinetic energy associated with the adatoms what is especially important at nominally unheated substrate conditions. The piezoelectric coefficient d33 decreases from 2.55 to 1.7 pm−1 when increasing the titanium sputter pressure from 2×10−3 to 14×10−3 mbar. When decreasing the Ti film thickness to 60 nm and hence, reducing the root mean square roughness by a factor of about 2, the intensity associated with the AlN (0 0 2) peak is increased by a factor of about 1.7 demonstrating the direct impact. Furthermore, the highest values for d33 and d31 (i.e. 3.15 pm V−1 and
−1.28 pm V−1) are determined.

Kurzfassung englisch:
Aluminum nitride (AlN) reactively sputter deposited from an aluminum target is an interesting compound material due to its CMOS compatible fabrication process and its piezoelectric properties. To obtain high piezoelectric coefficients it is a necessary pre-request to synthesize films with c-axis orientation. Besides the influence of sputter conditions on the microstructure of AlN thin films the condition of the substrate
surface is another important factor of utmost importance. In this study, the influence of 350 nm thick titanium metallization DC sputter-deposited on SiO2/Si substrates at varying back pressure levels bp,Ti in the range of 2×10−3 to 14×10−3 mbar on the c-axis orientation and the piezoelectric coefficients of 600 nm thick AlN thin films is investigated. Besides the plasma power for Ti deposition (Pp,Ti = 100 W) the parameters for AlN synthetization are fixed to Pp = 1000 W and bp,AlN = 4×10−3 mbar in 100% N2 atmosphere. Basically, the surface roughness of the Ti bottom layer is the dominating factor resulting either in a high degree of c-axis orientation (i.e. at low bp,Ti values) or in an amorphous AlN microstructure
(i.e. at high bp,Ti values). Under low pressure conditions, a smooth and dense surface characteristics is achieved due to a higher kinetic energy associated with the adatoms what is especially important at nominally unheated substrate conditions. The piezoelectric coefficient d33 decreases from 2.55 to 1.7 pm−1 when increasing the titanium sputter pressure from 2×10−3 to 14×10−3 mbar. When decreasing the Ti film thickness to 60 nm and hence, reducing the root mean square roughness by a factor of about 2, the intensity associated with the AlN (0 0 2) peak is increased by a factor of about 1.7 demonstrating the direct impact. Furthermore, the highest values for d33 and d31 (i.e. 3.15 pm V−1 and
−1.28 pm V−1) are determined.

Erstellt aus der Publikationsdatenbank der Technischen Universität Wien.