Premium
Quantification of dopant species using atom probe tomography for semiconductor application
Author(s) -
Yeoh Wai Kong,
Hung ShihWei,
Chen ShihChe,
Lin YiHsiang,
Lee Jang Jung
Publication year - 2020
Publication title -
surface and interface analysis
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.52
H-Index - 90
eISSN - 1096-9918
pISSN - 0142-2421
DOI - 10.1002/sia.6706
Subject(s) - atom probe , dopant , doping , analytical chemistry (journal) , semiconductor , materials science , atom (system on chip) , transistor , field effect transistor , chemistry , optoelectronics , nanotechnology , physics , transmission electron microscopy , computer science , embedded system , chromatography , voltage , quantum mechanics
Doping of semiconductors serve various purposes in metal‐oxide‐semiconductor (CMOS) technology, eg, increase carrier concentration and modify electric field distribution. With the scaling down of device and the introduction of three‐dimensional fin field‐effect transistors (FinFET), precise and reliable dopant quantification of concentration at the nano‐scale is critical. Laser‐assisted atom probe tomography (APT) provides a unique approach to characterize and quantify the dopant in three dimensions at sub‐nanometer resolution. Nevertheless, quantification accuracy of APT is strongly influenced by the experimental conditions. Although B quantification has been widely studied, the correlation of B signal loss to B concentration is not yet established. In addition, no phosphorous quantification study has been reported. In this work, we found that, due to B multi‐hit effect in APT, high B dose sample has larger B loading compared with low B dose sample. For standard calibration with minimized impact from multi‐hit effect, we recommend B dose in the range of 1e14 atoms/cm 2 . Despite the fact that B loading is dose dependent, APT quantification of B achieves precision within 2% to 6% relative standard deviation (RSD), which demonstrates that APT has good accuracy. On the other hand, P quantification suffers from mass interference of 31 P + and 31 P 2 2+ at 31 Da resulting in a large loading between APT and secondary ion mass spectrometry (SIMS). Nevertheless, we recommend that 31 Da to be labeled as 31 P + for smaller P variation for the APT analysis.