Intermediate Precision of Single-Molecule Break-Junction Conductance: Measurement, Uncertainty Budget, and Design Rules
Abstract
A break-junction experiment records thousands of single-molecule junctions in one session and reports a feature of their conductance histogram as the molecule’s conductance, with an uncertainty from resampling the traces. That is a withinsession sampling figure; how far the value moves between sessions is not measured or budgeted. We measure it in ISO 5725 and GUM form on archived STM-BJ recordings of DNA:RNA hybrid duplexes with repeat sessions at identical settings. On seven single-duplex groups, each two sessions from different recording rounds or dates, the intermediate-precision standard deviation of a session’s median is a factor of 1.46 (95% interval 1.28 to 2.15). Read in recording order, the median also wanders within a session with a correlation length near a hundred traces, so resampling understates the within-session uncertainty about threefold; counted consistently, the between-session term holds 74% of a single-session variance budget, and on records of this length it cannot be separated from slow drift. Five candidate remedies, tested under pre-registration, do not remove it. The expanded uncertainty of a single-session value, a lower bound that excludes calibration terms, is a factor of 2.4; spreading a fixed number of traces over more sessions reduces it, and a design rule gives the sessions a comparison needs. Public deposits from a second laboratory show a between-occasion component and the same within-session dependence. The figures are specific to the system; the method is not.
Related
- Beyond Go/No-Go Decisions: A Regional Selection Framework for Uncertainty-Aware Molecule Screening
- Solid-Phase Standard Enthalpies of Formation from SMILES with a Calibrated Per-Molecule Uncertainty: Application to a 13 606-Compound CHONFCl Inventory
- Interfacial thermal conductance between a polyethylene glycol polymer chain and water: A molecular dynamics study
- Charge-Induced Conductance Modulation and Capacitive Properties of Two-Layer–like 2γ-In2Se3: Insights from Density Functional Theory
- Probing Interfacial Thermal Conductance Under Extremely Large Temperature Differences via Ultrafast X-Ray Diffraction
BibTeX
@misc{minhas2026intermediate,
title = {{Intermediate Precision of Single-Molecule Break-Junction Conductance: Measurement, Uncertainty Budget, and Design Rules}},
author = {Viraaj Minhas},
howpublished = {ChemRxiv},
year = {2026},
month = oct,
doi = {10.26434/chemrxiv.15010359/v1},
url = {https://chemrxiv.org/doi/full/10.26434/chemrxiv.15010359/v1},
note = {CC BY},
}
RIS
TY - UNPB
TI - Intermediate Precision of Single-Molecule Break-Junction Conductance: Measurement, Uncertainty Budget, and Design Rules
AU - Viraaj Minhas
PY - 2026
DA - 2026/10/13
PB - ChemRxiv
DO - 10.26434/chemrxiv.15010359/v1
UR - https://chemrxiv.org/doi/full/10.26434/chemrxiv.15010359/v1
AB - A break-junction experiment records thousands of single-molecule junctions in one session and reports a feature of their conductance histogram as the molecule’s conductance, with an uncertainty from resampling the traces. That is a withinsession sampling figure; how far the value moves between sessions is not measured or budgeted. We measure it in ISO 5725 and GUM form on archived STM-BJ recordings of DNA:RNA hybrid duplexes with repeat sessions at identical settings. On seven single-duplex groups, each two sessions from different recording rounds or dates, the intermediate-precision standard deviation of a session’s median is a factor of 1.46 (95% interval 1.28 to 2.15). Read in recording order, the median also wanders within a session with a correlation length near a hundred traces, so resampling understates the within-session uncertainty about threefold; counted consistently, the between-session term holds 74% of a single-session variance budget, and on records of this length it cannot be separated from slow drift. Five candidate remedies, tested under pre-registration, do not remove it. The expanded uncertainty of a single-session value, a lower bound that excludes calibration terms, is a factor of 2.4; spreading a fixed number of traces over more sessions reduces it, and a design rule gives the sessions a comparison needs. Public deposits from a second laboratory show a between-occasion component and the same within-session dependence. The figures are specific to the system; the method is not.
ER -