Publikationsrepositorium - Helmholtz-Zentrum Dresden-Rossendorf

Dilute nitride semiconductors (DNS), such as GaAsN, with a nitrogen content y of a few percent or even less, have recently attracted considerable interest due to the giant bowing effect. That, in turn, offers the possibility to tailor the band structure of new devices, like LEDs, lasers, solar cells, and infrared photodetectors by varying the nitrogen content [1]. Determining proper values of the effective mass (EM) of DNS is a topic of interest because of the inconsistency of previous results (e.g. [2, 3]). To clarify the conflict we study a series of GaAsN samples (y = 0%-1%) by cyclotron resonance (CR) spectroscopy, Fourier spectroscopy and photoluminescence spectroscopy in magnetic fields in order to deduce the EM via the CR frequency, plasma frequency and the dielectric shift, respectively. So far, we are able to show that the discrepancies of former publications are most likely caused by the particular choice of the experimental technique. Probably the most direct and reliable method is the CR spectroscopy, which has rarely been used due to the low electron mobility in GaAsN. The CR does not signicantly change with different N contents and thus the EM. Our Magneto-PL spectroscopy results exhibit a completely dierent behavior. We use the same method as e.g. [2] and see a huge increase of the EM with the N content which is even bigger than in previous publications. On the other hand, the slope of ΔE, which is the key parameter in the calculation of the EM in this method, strongly depends on the region of interest. In our opinion, this method is not accurate enough for the EM determination.

[1] A. Erol, Dilute III-V Nitride Semiconductors and Material Systems, Springer-Verlag Berlin Heidelberg (2008).
[2] F. Masia et. al., Appl. Phys. Lett. 82, 4474 (2003).
[3] Y. J. Wang et. al., Appl. Phys. Lett. 82, 4453 (2003).