Obirch

Optical beam induced resistance change (OBIRCH) is a laser-based failure analysis technique that scans a focused laser beam across a powered, functioning chip while continuously monitoring current — pinpointing exactly where a resistive defect like a void, crack, or thinned metal line hides, by watching for an abnormal change in current as the laser's heat sweeps across it. It's a non-destructive technique performed on a live, operating device, which is exactly what lets it localize a defect precisely enough to guide subsequent physical cross-sectioning, rather than destroying the evidence while searching for it.

Local laser heating is the physical mechanism OBIRCH relies on entirely — the laser spot doesn't image anything directly, it simply heats whatever small spot of metal it's currently illuminating, and that heating is what reveals a defect's presence. As the laser sweeps across the device, it slightly raises the temperature of the metal line directly beneath it at that moment, and because metal resistance depends on temperature, that local heating produces a small, measurable change in current through the line — exactly the signal OBIRCH is built to detect and map across the entire scanned area.

Normal metal shows a small, predictable current change as the laser passes over it, because ordinary metal has a well-understood positive temperature coefficient of resistance — resistance rises slightly and predictably as temperature rises. This expected, modest current response is the baseline OBIRCH measurement establishes across all the healthy metal it scans, and it's specifically the departure from this predictable baseline behavior that signals something is actually wrong at a given location, not the baseline response itself.

A defect — a void, a crack, or an abnormally thin section of metal — produces an anomalously large or even inverted current change precisely because the defect concentrates the same laser heating into a much smaller effective cross-section of metal. A void or thin spot has less metal volume to absorb and dissipate the laser's heat, meaning the same heating input produces a disproportionately larger local temperature rise, and correspondingly disproportionate resistance change, at that specific point compared to the surrounding healthy metal. This outsized or anomalous response is exactly the signal OBIRCH is designed to flag, pinpointing the defect's precise location directly from where in the scan the anomalous current change actually occurred.

OBIRCH's defining operational advantage — performing this entire analysis on a powered, functioning device, non-destructively — is what lets it localize a defect precisely enough to guide subsequent, necessarily destructive physical analysis like cross-sectioning. Because OBIRCH doesn't damage or alter the device during the scan itself, it can map out exactly where a suspected defect sits before anyone commits to the destructive step of actually cutting into the chip to directly examine that location. This sequencing — non-destructive localization first, destructive confirmation second, at exactly the location OBIRCH identified — is standard failure-analysis practice specifically because destructive analysis performed blindly, without this kind of precise localization first, wastes enormous effort searching randomly through a chip that could contain a defect essentially anywhere.

OBIRCH and EMMI (emission microscopy) are complementary techniques precisely because they detect fundamentally different categories of defect, which is exactly why failure analysis labs routinely use both rather than treating either as sufficient alone. EMMI detects light emission from active, electrically-shorting defects — a defect that's actually conducting current abnormally and emitting light as a byproduct. OBIRCH instead detects purely resistive, passive defects — voids, cracks, thinning — that don't emit any light signal at all but still represent a genuine electrical problem. A failure analysis investigation that only used one technique would systematically miss whichever category of defect that technique simply isn't designed to detect, which is exactly why both remain standard, complementary tools in any serious interconnect failure-analysis workflow.