High-Precision Characterization of III-Nitride Semiconductor Alloys with Secondary Ion Mass Spectrometry (SIMS)
- PDF / 389,103 Bytes
- 6 Pages / 414.72 x 648 pts Page_size
- 79 Downloads / 193 Views
Cs' ions with negative SIMS. The primary beam was rastered over a square region 100-125 microns on a side, and the positive or negative secondary ions were collected from a circular 363
Mat. Res. Soc. Symp. Proc. Vol. 395 01996 Materials Research Society
image area 30 microns in diameter, centered in the rastered region. The sputter rates were obtained by measuring the analytical crater depths with a Tencor P 10 stylus profilometer. Samples of GaN, AIN, InN, AIGaN, GaInN, AlInN, and AIGaInN (Figure 1) were obtained from various sources and the imnuritv levels in each were estimated usinQ SIMS: ion vield systematics permitted inference of concentrations to within a factor of five based on analogies with quantification in GaAs. The III-nitride samples (except for those of AIGaInN) and a silicon witness sample were then implanted with known doses of 'H, 12 C, 160, 24 Mg, 2 8Si, and 1 2Cd. The doses were adjusted to provide implant peaks well above the impurity concentrations, which in several cases were close to 1%. The purpose of the silicon witness sample was to provide an independent means of checking the nominal implant doses against archival data, since SIMS quantification is best understood for silicon (the only NIST standard reference material for SIMS is an implant of B into silicon). The doses in the Si witness sample were checked with SIMS. InN
*hard to grow Figure 1. The AIGaInN alloy system, showing the various compositions of the samples used in this work.
*
0 AIN
RESULTS AND DISCUSSION The experimental results provide an intriguing picture of how the ion formation mechanisms change with composition in the AlGaInN system. The relative sensitivity factors (RSFs) used in quantification typically are measured with respect to various atomic or molecular matrix ions such as Al', Ga', In', NAI-, NGa-, or NIn-. Since multiple measurements were made under a range of different instrumental conditions in many cases, relative standard deviations in the RSFs can be estimated. The statistics suggest that calibration with respect to the matrix ion NGa- is preferable to the use of Ga- in negative SIMS, since the Ga- intensity is more sensitive to the instrumental tuning conditions and produces a wider range of RSFs. Similar considerations were used to choose the best matrix ions under the other SIMS analysis conditions. Figure 2 shows how SIMS data can be calibrated with photoluminescence (PL), using a light-emitting diode structure of GaN/InGaN/GaN. 1.0 _10_22 GaN / InGaN / GaN blue LED MOCVD Laboratory
"courtesy of UCSB A blue LED • Figure 2. GaN/InGaN/GaN structure with high quantum efficiency, grown at the UCSB MOCVD " lab. With good crystal quality and a thin active layer, such structures may be precursors to reliable blue laser diodes. 0
I
Si dopant-•
. " 14% In(+Cs)
/_
1......__ U
0 _
_
_1___
DEPTH (micron)
364
..
_ -
U.t
10
16
Complications exist for such calibration--the nominal wavelength of the InGaN layer is not only related to the In content, but also to the dopant concentration. Hi
Data Loading...
