Richard McLaughlin - All You Ever Wanted to Know About Silicon Drift Detectors

Abstract
A revolution has occurred over the past 10 years in microanalysis with the advent of the silicon drift detector (SDD). Although the SDD has been around since the early 1980s, only recently has the technology migrated to the SEM and TEM. Understandably, great excitement has been generated mainly due to the high acquisition rates and excellent peak resolution possible with this type of detector. The drastic reduction in mapping time is an obvious benefit and has been the main reason for the early adoption of the technology. The earliest SDDs did have their limitations. They exhibited poor low energy performance, poor peak stability and due to the small sensor size (10 mm²), required beam currents in excess of 10 nA to take advantage of their high throughput. Such limitations prevented a broader migration away from the Si(Li) detector. However, recent improvements in sensor and digital processor designs have now removed the main obstacles preventing wider acceptance.

Sensors as large as 100 mm² are now found in detectors. This allows the analyst to achieve count rates > 30,000 cps under normal SEM imaging conditions (< 1 nA, 20 kV) and >200,000 cps at a quite reasonable 5 nA. The improved low energy performance allows detection down to Si L-peak (~0.09 keV) and due to the excellent peak stability and high count rates, good quantitative analysis can be performed in seconds.

It is evident that the performance of the SDD will allow the analyst to become more efficient in collecting data. But silicon drift technology also permits the analyst to acquire better data. The better signal/noise in high count maps can reveal subtle chemical differences and high count spectra have lower effective detection limits (~0.05 wt%). Larger sensors will also make low kV (< 5 kV) mapping a more practical application.

The SDD is capable of generating large amounts of data. Large datasets, in particular extremely large data cubes, present a challenge and an opportunity for a modern EDS system. The challenge is in handling and processing gigabytes of data in a smooth and easy manner. The opportunity is in devising new ways to process (e.g. fitted least square maps) and utilize (e.g. principle component analysis) the data cube to aid the analyst.

Biography
Dr. Richard McLaughlin is a Senior Applications Specialist for Oxford Instruments’ NanoAnalysis division, based in Concord, MA.

A Canadian by birth and a geologist by training, he has been a field geologist exploring for base metals and a geophysical surveyor searching for gold. He has several degrees in the Earth Sciences and has used various analytical techniques during his studies including TIMS, ICP-MS, ICP-AES, EDS, and powder X-ray diffraction. He obtained his Ph.D. degree in Geology at McMaster University studying water/rock interactions, an M.Sc. degree at Lakehead University studying rare-metal mineralization, and a B.Sc. at Queen’s University.

Richard has channeled his diverse analytical background to provide expertise for Oxford’s line of energy dispersive spectrometers (EDS), wavelength dispersive spectrometers (WDS) and Electron Backscattered Diffraction (EBSD) systems. He has been with Oxford Instruments for over 13 years.