Method for shallow impurity characterization in ultrapure silicon using photoluminescence

Abstract

The advantages of very low excitation density photoluminescence for the quantitative determination of shallow impurity concentrations in ultrapure Si are demonstrated using the weakly-absorbed 1047 nm output of a Nd:YLF laser. Compared to the usual high excitation method, which uses an argon ion laser, the weak bound exciton luminescence is significantly enhanced relative to that of the free exciton, and competing luminescence from multiexcitonic species, which complicates the accurate determination of the bound exciton intensities, is eliminated. Also, the presence of the laser Raman line in the same spectral region as the luminescence allows for the accurate determination of relative luminescence quantum efficiencies and can reveal the presence of deep traps. Finally, we report calibrations for boron, phosphorus, and arsenic in the concentration range 1011 to 1015 cm−3.

Determination of the phosphorus content in diamond using cathodoluminescence spectroscopy

“…As excitons are bound to neutral donors, ͓P͔ is basically the neutral phosphorus concentration. Practically, it is equal to the substitutional concentration due to an efficient neutralization of charged impurities by the injected carriers, observed even under weak injection conditions in PL 3 . As a result, once calibrated, the ratio r measured at low temperature can give access to the concentration of substitutional dopants.…”

“…As a result, once calibrated, the ratio r measured at low temperature can give access to the concentration of substitutional dopants. To further illustrate this point, it has to be mentioned that PL has been successfully used to determine simultaneously B ͑acceptor͒ and P ͑donor͒ contents in highly compensated silicon crystals, 3 as well as compensation ratios in GaAs. 6 In the case of diamond, we also report here the determination of the boron substitutional content, thanks to the calibration of Kawarada et al, 18 in an n-type phosphorus-doped diamond layer.…”

“…1 In the limit of low temperatures, the ratio between the bound-exciton ͑BE͒ and free-exciton ͑FE͒ emissions observed in luminescence experiments is proportional to the concentration of impurities in a wide range of concentrations and this, almost independently of the excitation densities. [2][3][4] These unique properties have made possible the use of luminescence spectroscopies as sensitive tools for dopant quantification in semiconductors. The relationships between the photoluminescence ͑PL͒ spectra and the dopant concentrations were established for Si, 2,3,5 SiC, 4 GaAs, 6 etc., and are commonly used to qualify the purity of semiconductor materials.…”

Luminescence Spectroscopy of Bound Excitons in Diamond

“…The sensitivity is almost four orders of magnitude higher in silicon, which probably reflects the four order longer Auger lifetimes compared to diamond. Practically, this limits the use of bound exciton spectroscopy in silicon to ultra‐high purity crystals (<10 15 cm −3 ), while the typical measure range in diamond (10 13 –10 18 cm −3 ) is better suited to doping levels used in electronic devices.…”

Dislocation related band-edge photoluminescence in boron-implanted silicon

“…An additional advantage of using the 1064 nm excitation is that the zonecentre, fundamental optical phonon for Si can also be observed overlapping with the band-edge luminescence. The intensity of this (phononic) Raman band is not affected by the concentration of traps or defect centres which usually reduce the number of free excitons through their capture and nonradiative recombination [8]. As these centres may significantly affect the PL intensity, the Raman band may be used as a scattering quantum counter to independently assess changes in the emission yields.…”