Surface Topography Challenge > The Original Investigation

  • The Surface Topography Challenge began in 2022. In the original investigation, 153 people worked together to create a single comprehensive description of a surface. By combining measurements from many researchers, instruments, and techniques, we learned some fundamental lessons about our methods for describing surface topography.

    • The original Problem Statement was issued in 2022

    • 2 types of surfaces were studied in detail

    • 153 participants in 20 countries

    • 63 companies, national labs, and universities

    • 12 measurement techniques were used

    • 2088 measurements collected and made freely available

    • The final collaborative paper was published in 2025

  • Topography influences how surfaces stick, slide, wear, conduct, seal, leak, break, and more. The purpose of the Surface Topography Challenge is to develop better understanding and agreement on how to measure, report, and analyze surface topography.

    This is accomplished through three goals:

    1. Compare the advantages and disadvantages of different surface-topography measurement techniques.

    2. Generate the single most comprehensive description of a surface yet created.

    3. Aid the development of next-generation surface descriptors.

    Read the original problem statement

  • For the original investigation, we created two different surfaces, each of which was separated into hundreds of individual samples:

    • The “Smoother Surface” was made from a coating of chromium nitride (CrN) deposited on prime-grade polished silicon wafers.

    • The “Rougher Surface” was made from the same CrN coating deposited on the unpolished “backside” of silicon wafers.

    These samples were then shipped free of charge to any group that volunteered to measure them and share the raw topography measurements via contact.engineering, a free web app.‍

  • Once the samples were manufactured, we announced the Challenge in July of 2022 and set a deadline of approximately a year later.

    We were thrilled with the enthusiastic response. Ultimately, 153 scientists and engineers from 64 research groups and companies volunteered to characterize the two sample surfaces, using a wide variety of techniques.

    The results of the 2088 measurements constitute the most comprehensive surface description ever compiled.

    In all, 12 measurement techniques were used, from three categories:

    • Microscopy techniques such as white light interferometry, confocal microscopy, and focus variation microscopy

    • Contact techniques including stylus profilometry, nano indenture/scanning probe microscopy, and atomic force microscopy

    • Cross-section techniques, including scanning electron microscopy and transmitted electron microscopy

  • Participants uploaded their data to contact.engineering. This freely available, open-source topography-analysis platform ensures that all data were analyzed identically, removing user- and software-specific variations.

    Participants also submitted a concise description of their methods and results, including their understanding of the instrument's lateral resolution.

    To reduce each measurement to a single number, we computed the root-mean-square (RMS) height for each measurement.

    The process revealed a somewhat remarkable fact: the raw data for both surfaces showed RMS height variations from below 100 pm to 10 μm—spanning six orders of magnitude! The challenge then was to understand why.

  • We worked to combine these many measurements into a single descriptor of each surface. To do so, we:

    • used scale-dependent analysis,

    • performed outlier removal and detection, and

    • used a majority-rule approach.

    In the end, we established the complete topography of both surfaces over all length scales, with broad-base consensus across techniques and researchers.

Lessons learned

  • Each measurement technique has artifacts. Some are easy to see, others can fool even expert users. When multiple different measurement techniques are compared, these artifacts often become obvious.

    BEST PRACTICE 1: For the highest confidence in surface topography, combine data from multiple measurement techniques.

  • Even in a single measurement, some data is reliable and some is unreliable. All techniques have artifacts that introduce certain types of error. For example, electrical noise or vibration can introduce artificial signals. Or zooming in too far can introduce a kind of “blur”. By presenting the data as a function of length scale, we were able to reduce or eliminate these artifacts.

    Then, instead of dismissing whole measurements just because some portions were artifacted, we could separate out the reliable and unreliable portions of the data.

    BEST PRACTICE 2: Compute and report scale‑dependent parameters to account for length scales.

  • The complete surface measurements generated in the original investigation demonstrate that the topography of a surface cannot meaningfully be captured by any single number. To compute any parameter, even a scale-dependent parameter, we need to make context-dependent and sometimes empirical choices. While we can document the choices made in our calculations, a more robust solution is to instead always save and report the raw topography data rather than simply a handful of parameter values. Doing so allows subsequent context-dependent analyses to be performed with differing methods.

    BEST PRACTICE 3: Save, analyze, and report raw topography measurements, not just computed parameters