Quantification of matrix and impurity elements in AlxGa1-xN compounds by Secondary Ion Mass Spectrometry

FBH research: 25.04.2016

Fig. 1: Determination of Al mole fraction in AlGaN/AlN/sapphire films using optical transmission measurements in combination with HRXRD analysis. The insets show the HRXRD reciprocal space map and the transmission spectra for the Al0.48Ga0.52N sample.

Fig. 2: Matrix ion intensity ratio AlCs+/GaCs+ vs. mole fraction ratios in AlxGa1-xN.

Sputter rates S (in µm³·s-1·nA-1), matrix ion intensity ratios and absolute sensitivity factors (in 1014 at·s/cm³·counts) of H, C, O, Si in AlxGa1-xN alloys.

The AlGaN material system is of increasing interest for UV-based optoelectronic devices such as light emitting diodes (LEDs), laser diodes and photodetectors. The electronic properties of AlxGa1-xN compounds are strongly influenced by the mole fraction x of AlN, as well as by the content of residual impurities and dopant elements like H, C, O, Si, and Mg. Thus, quantification of matrix and impurity elements is of utmost importance for any epitaxial growth development. Secondary Ion Mass Spectrometry (SIMS) is the method of choice to identify and quantify very low elemental concentrations in the range of ppm as well as the content of main lattice constituents.

In a collaboration between FBH and Berlin-based RTG Mikroanalyse GmbH a comprehensive SIMS calibration procedure for the quantification of matrix and impurity elements of epitaxially grown AlxGa1-xN layers over the full compositional range has been conducted [1]. For that, high-quality AlxGa1-xN layers (0 ≤ x ≤ 1) were heteroepitaxially grown on c-plane sapphire and implanted with H, C, O and Si. A comparative analysis of AlxGa1-xN composition using different characterization techniques was done to ensure a reliable matrix quantification of the SIMS calibration samples. A measurement example (room-temperature transmission measurement) is shown in Fig. 1. on the determination of AlGaN composition from optical absorption has to take into account the strain relaxation determined by HRXRD. As expected for the growth on AlN/sapphire templates the degree of relaxation increases from R=0.61 for the high Al-containing sample (Al0.86Ga0.14N) to R=0.81 for the low Al-containing material (Al0.29Ga0.71N). The compositional analysis using various techniques yielded consistent Al contents x with an accuracy of ±1%.

For the quantitative characterization of impurities by SIMS, calibration curves were generated using a 14.5 keV Cs+ primary beam. Measured sputter rates decrease with a nearly linear slope as a function of Al content in the range of 0 ≤ x < 0.48. At higher Al concentrations the sputter rates show only a weak dependence on AlN mole fraction. Matrix ion intensity ratios of AlCs+/GaCs+ change linearly with direct proportionality as a function of x/(1-x) (Fig.2). It has been demonstrated that the intensity ratio AlCs+/GaCs+ is appropriate to quantify the AlN and GaN content in the AlxGa1-xN material system with an accuracy of 2 %. The sensitivity factors for analyzing the impurity element concentrations of H, C, O, and Si were determined. Their dependence on the AlN mole fraction is reasonable considering the dependence of the sputter rate vs. Al content. Table I summarizes the sputter rates, matrix ion intensity ratios and the absolute sensitivity factors of H, C, O and Si for the different AlGaN layers. The concentrations of investigated impurities and dopants (down to a scale of ppm and lower) are detectable in the whole  AlxGa1-xN composition range with an accuracy of better 20%  above noise level.

This work provides the basis for an improved quantitative evaluation of the AlxGa1-xN material system which takes on greater significance in modern electronic and optoelectronic devices.
[1] Funding program 03WKBT06A/B ("Berlin WideBaSe") by the federal ministry of education and research.

Publication

P. Jörchel, P. Helm, F. Brunner, A. Thies, O. Krüger, M. Weyers, Journal of Vacuum Science & Technology B, Nanotechnology and Microelectronics: Materials, Processing, Measurement, and Phenomena, vol. 34, no. 3, p. 03H128, May 2016.