Abstract
This thesis primarily deals with accuracy obtainable when using IBA (Ion Beam Analysis) techniques to characterize materials. RBS (Rutherford Backscattering Spectrometry) is the main technique used, together with EBS (Elastic Backscattering Spectrometry), ERDA (Elastic Recoil Detection Analysis) and NRA (Nuclear Reaction Analysis). An exhaustive literature review on these analytical methods is made in connection with accuracy issues such as stopping powers and multiple scattering. The experimental set-ups and procedures are described, with emphasis laid on critical aspects of work where the highest accuracy is required. The instrumentation for dosimetry on ion implanters is first estabhshed at the 1% level for high-dose heavy implants in silicon. A new parameterisation of He stopping power in Si is used, and this latter material, via the surface yield, is used as a calibration standard. A precision (standard uncertainty) in the determination of implantation doses by RBS is conclusively demonstrated at 1.5%. The IBA DataFurnace code is validated for such accurate analysis, which can now be made routinely and rapidly. The certified Sb sample IRMM-302/BAM-L001, which has a certification of 0.6% traceable to the international standard of weight in Paris, is measured, and more importantly this measurement demonstrates the reliability of the stopping power parameterisation at 1.4%. Using conventional ERDA, the H dose of an amorphised Si wafer, implanted with 6-keV H+ ions, is found to be 57.8(1.0)x1015 at/cm2, which is a 1.8% standard uncertainty. The estimated combined uncertainty of this measurement is ~6%, and this mainly comes from the determination of the ERDA solid angle by using standard Kapton. The Kapton composition is carefully determined using RBS. The RBS solid angle is obtained using the amorphised silicon surface yield as a calibration standard as in the dosimetry analysis mentioned above. The ERDA H absolute dose obtained is compared with the results from other participants from all over the world in a Round Robin exercise, which includes measurements by using both He-ERDA and HI-ERDA (Heavy Ion-ERDA) together using various detectors. The results from each participant are given and compared. The overall absolute dose obtained of the implant is 57.0(1.2)x1015 H/cm2, and this represents an inter-lab reproducibility of 2.2% (standard uncertainty). Unstable surface hydrogen contamination was observed, and this surface peak was resolved by some of the methods. This implant can now be used as a standard for quantitative analysis of hydrogen. Low-fluorine content SiO2:F films are analysed by RBS for absolute fluorine concentration determination. Prior to the RBS analysis, the uniformity of the films and stability of F under beam irradiation is investigated. Because the RBS is not very sensitive to F and the F signal has a large matrix background, an internally consistent method of data handling, which enables the relative collected charge to be determined very precisely for the spectra from different samples, is developed. This method has as a parameter the F content, which is then extracted iteratively. A F concentration of 10 at% is determined with an estimated uncertainty of 10% (one percentage point, i.e. 10 +/- 1%). The O stopping powers are found to be the main factor governing the accuracy of the absolute determination of the F content. All the other uncertainties add up to only ~1%. The elemental composition of residual deposits from an ion implanter is thoroughly investigated using several complementary analytical methods, namely, RBS, BBS and NRA. Preliminary SEM/EDAX results are used as a guide. Depth profiles of such non-homogeneous, non-fiat and brittle samples are obtained, which give an indication of the concentration of each element present. From this complete IBA elemental study, some unprecedented light is brought on both the history of the implanter and the way in which these deposits are formed. Such an investigation is essential for a better understanding and the development/miniaturisation of semiconductors as it impressively pushes the boundaries of accuracy obtainable in IBA material characterisation.