Nanoindentation Hardness Thin Film
# Nanoindentation Hardness and Modulus
Wafer bow measurement extracts a film's average stress from how much it bends the wafer, but stress alone doesn't tell a reliability engineer whether that film is hard and stiff or soft and compliant — and CPI simulation, via stress migration modeling, and underfill-modulus selection all need real mechanical-property numbers, not just a stress value. Nanoindentation supplies those numbers directly, by pressing a tiny, precisely shaped diamond tip into the film and reading hardness and elastic modulus straight out of the resulting load-versus-displacement curve.
The indenter is small enough to probe a single thin film without the substrate underneath corrupting the measurement. A Berkovich indenter — a three-sided pyramidal diamond tip, the standard geometry for this technique — presses into the film under a continuously monitored, precisely controlled load, typically reaching contact depths of only tens to a few hundred nanometers. That shallow depth matters enormously: press too deep relative to the film's own thickness and the substrate underneath starts contributing to the measured response, corrupting the result with a mechanical property that isn't really the film's own. Keeping indentation depth to roughly a tenth of film thickness or less is the standard rule of thumb precisely to keep the substrate out of the measurement.
The load-displacement curve's unloading slope is where the elastic modulus actually comes from, not the loading curve. As the indenter presses in, the load-vs-depth curve captures a mix of elastic and plastic (permanent) deformation together, but as the indenter retracts, the material's elastic recovery dominates the unloading response — the plastic deformation stays put, while the elastic part springs back. Analyzing the initial slope of that unloading curve, using the Oliver-Pharr method that's become the field's standard analysis framework, separates out the film's elastic modulus from its hardness, which comes instead from the peak load divided by the actual contact area at maximum indentation. One test, one load-unload cycle, and both numbers come out of the same curve.
These numbers feed directly into exactly the stress and reliability models this session's other entries rely on. Via stress migration's vacancy-flux modeling and CPI finite-element simulation both need a film's elastic modulus as an input, not just its measured stress — stress alone tells you the driving force, but modulus tells you how the material actually responds to and redistributes that stress mechanically. Low-k dielectrics specifically get nanoindentation-characterized as a standard qualification step precisely because their engineered porosity drops modulus and hardness substantially relative to dense oxide, and knowing exactly how much softer a given low-k formulation is compared to its predecessor is what lets a packaging team predict whether that film will survive CPI stress or crack under it, rather than finding out only after a failed qualification run. Equipment from Hysitron (now part of Bruker) and KLA's nanomechanical testing lines are the production-characterization standard for this measurement across the industry.
Wafer bow measurement answers "how much stress does this film carry?" Nanoindentation answers a different, complementary question — "how stiff and hard is this film, mechanically?" — and a reliability engineer needs both numbers together before a stress value alone can actually predict whether a structure survives.