Hier finden Sie wissenschaftliche Publikationen aus den Fraunhofer-Instituten.

Impact of temperature increments on tunneling barrier height and effective electron mass for plasma nitrided thin SiO2 layer on a large wafer area

: Aygun, G.; Roeder, G.; Erlbacher, T.; Wolf, M.; Schellenberger, M.; Pfitzner, L.


Journal of applied physics 108 (2010), No.7, Art. 073304, 9 pp.
ISSN: 0021-8979
ISSN: 1089-7550
Journal Article
Fraunhofer IISB ()

Thermally grown SiO2 layers were treated by a plasma nitridation process realized in a vertical furnace. The combination of a pulsed-low frequency plasma and a microwave remote plasma with N-2/NH3/He feed gas mixture was used to nitride the thermally grown SiO2 gate dielectrics of MIS structures. Temperature dependency of effective masses and the barrier heights for electrons in pure thermally grown SiO2 as well as plasma nitrided SiO2 in high electric field by means of Fowler-Nordheim regime was determined. It is frequently seen from the literature that either effective electron mass or barrier height (generally effective electron mass) is assumed to be a constant and, as a result, the second parameter is calculated under the chosen assumption. However, in contrast to general attitude of previous studies, this work does not make any such assumptions for the calculation of neither of these two important parameters of an oxide at temperature ranges from 23 to 110 degrees C for SiO2, and 23 to 130 degrees C for nitrided oxide. It is also shown here that both parameters are affected from the temperature changes; respectively, the barrier height decreases while the effective mass increases as a result of elevated temperature in both pure SiO2 and plasma nitrided SiO2. Therefore, one parameter could be miscalculated if the other parameter, i.e., effective mass of electron, was assumed to be a constant with respect to variable physical conditions like changing temperature. Additionally, the barrier heights were calculated just by taking constant effective masses for both types of oxides to be able to compare our results to common literature values.