Analysis of the causes of variance in resistance measurements on metal-molecule-metal junctions formed by conducting-probe atomic force microscopy

Alkanethiol tunnel junctions were studied using conducting-probe atomic force microscopy to determine causes of variability in measured resistance behavior. Measurements were made on Au/decanethiol/Au monolayer junctions, and effects of substrate roughness, tip chemistry, presence of solvent, extensive tip usage, applied load, and tip radius were examined. Resistance measurements yielded log-normal distributions under a variety of conditions, indicating that the origin of the variance is likely to be either changes in tunneling length or electronic overlap. Spreads in resistance values for a given tip were much less when flat, template-stripped Au substrates were used rather than rough, evaporated Au substrates. Chemical modification of tips with ethanethiol (C-2) or butanethiol (C-4) and performing measurements under cyclohexane were also found to reduce variance by a factor of about 2-4. Experiments performed with unmodified tips showed an increase in junction resistance over the course of hundreds of consecutive measurements, whereas junctions made with modified tips or under cyclohexane did not. Attempts to ascribe variance between tips to varying tip radii failed; however, decreases in resistance with increasing applied load on the tip contact were observed and could be interpreted in terms of conventional contact mechanics models.