Statistical variation analysis of sub-5-nm-sized electron-beam-induced deposits

Abstract

We report on the statistical analysis of the variations in the size and position of sub-5nm tungsten-containing dots in regular arrays deposited by electron-beam-induced deposition. Full widths at half maximum of the dots are 4.2 and 2.0nm in average. It can be observed in the recorded annular dark-field images that there is a variation in intensity for these dots. We have analyzed these variations and it is found that the relative standard deviation for the mass per dot is 0.092 for the 4.2nm dots and 0.26 for the 2.0nm dots. Comparing this to a relative standard deviation in the estimated number of precursor molecules that are pinned down per dot of 0.041 for the 4.2nm dots and 0.11 for the 2.0nm dots, it appears that the dot-to-dot variation in mass for both dot sizes compares reasonably well with the values expected from Poisson statistics on the number of molecules per dot. It can be concluded that at these dimensions, the statistics on the number of pinned precursor molecules dominates the control of feature sizes.

Substrate temperature and electron fluence effects on metallic films created by electron beam induced deposition

“…% to 59.2 at. % when the substrate temperature was increased from 25 C to 280 C. Collectively, these results on W(CO) 6 and Pt(PF 3 ) 4 strongly suggest that thermal as well as electron stimulated reactions can also contribute to the overall decomposition process and thus the ultimate chemical composition in the deposit. However, the underlying reactions responsible for these temperature dependent variations in film composition have yet to be elucidated.…”

“…1(a) and also with recent results obtained by Mulders et al [ Fig. 1(b)], who studied the composition of EBID films created from MeCpPtMe 3 in a SEM as a function of substrate temperature in the range 25 C to 360 C. 15 Indeed, despite the different deposition conditions and analytical techniques (XPS in the present study, and EDS in the study by Mulders), a comparison of Figs. 1(a) and 1(b) shows that the chemical composition of the EBID materials is very similar in the two studies.…”

“…The choice of these three molecules was predicated on their widespread use by the EBID community as precursors, and our understanding of the purely electron stimulated decomposition process. 7,8,10,11,16,19,[21][22][23][24][25][26][27][28][29][30][31][32] In addition, our results can be compared and benchmarked against compositional data that already exist from previous EBID studies where deposition was accomplished under more typical conditions. Our experimental approach involved exposing nanometer thick films of parent precursor molecules to different electron fluences.…”

“…% at 80 C). In another example, Wang et al showed that for EBID films created from Pt(PF 3 ) 4 in an Auger electron spectrometer the (Pt/PtþP) atomic percent increased systematically as the substrate temperature increased from 25 C to 120 C. 14 More recently, Mulders et al performed a systematic temperature dependent study on a range of EBID precursors: W(CO) 6 , TEOS, MeCpPtMe 3 , Co(CO) 3 NO, Co 2 (CO) 8 , and Me 2 Au(acac). This study revealed that increases in metallic content can occur when the substrate temperature was increased, although the importance of this processing variable depended on the precursor.…”

Molecule-by-Molecule Writing Using a Focused Electron Beam

“…focused electron-beam-induced deposition ARTICLE amount of deposited material is not linearly proportional to the electron exposure time but is instead determined by the statistics on the number of dissociated molecules. 9 This makes it impossible to extend the existing strategy of merely decreasing the electron exposure time to reduce the deposit size. If we want to go beyond the current resolution limit, it is necessary to use a technique that compensates for the randomness of the process.…”