Detection methods and their limits

What documentary, physical, and analytical methods for evaluating suspect electronic components can establish, what each cannot establish, and why no single result proves authenticity.

A suspect electronic component that reaches incoming inspection can be evaluated through several distinct forms of evidence: documentary review, external inspection, non-destructive internal and material analysis, electrical characterization, and, where justified, destructive forensic examination. Each method answers a bounded question. None produces an unrestricted conclusion about authenticity. This page sets out, for each, what the method can establish and — with equal weight — what it cannot, because treating a single test result as a conclusion is the most common error in component verification, and it is the error this page is written to prevent.

The reason detection has to be understood through its limits, rather than through its capabilities alone, is visible in the incident record. Among suspect-counterfeit parts reported to ERAI in 2025 that underwent electrical testing, 24 percent produced a passing result and 12 percent failed. ERAI notes that the passing figure is probably understated, because reporting organizations tend to submit failures and omit successful results, and it treats electrical results alone as insufficient to separate a nonconforming part from a suspect-counterfeit one. A part that passes a test has not been shown to be genuine. It has been shown to pass that test.

What detection can and cannot establish

Detection methods change the evidence position of a lot: they may resolve a defined uncertainty, identify a nonconformance, or expose a new contradiction. Their evidentiary effect is not symmetrical. A method can convert a suspicion into a verified finding of a specific nonconformance — for example, package construction inconsistent with the claimed device, a documented marking alteration, or an accessible material composition inconsistent with the applicable acceptance criteria — when the method, sample, chain of custody, and acceptance criteria support that conclusion.

Establishing the opposite is harder. A strong authenticity claim normally depends on traceable provenance supported by mutually consistent documentary and physical evidence; no single bench result can reconstruct provenance that was never recorded. Resolving enough relevant uncertainties may support an accept, restricted-use, further-verification, quarantine, or reject disposition for the intended application, but a passing result remains bounded by the question asked, the sample tested, and the coverage applied.

The methods are set out here in roughly increasing depth, corresponding to the L0 to L3 verification scale. Verification depth is not the same as standalone reliability. A deeper level adds questions, methods, or evidentiary coverage; it does not make evidence from lower levels unnecessary and does not eliminate every remaining limitation.

The methods, and what each one misses

L0 — Documentary Evidence Review. The first layer examines the record around the part rather than the part itself: certificates of conformance, test reports, invoices, packing lists, date and lot codes, and any traceability documentation. It catches internal contradictions — a date code inconsistent with verified manufacturer records, lifecycle information, or known package characteristics available for that part, a lot code that does not match the physical marking, documentation absent where it should exist. What it misses is competent forgery. Paperwork can be internally consistent and still describe goods other than the ones in the box, and a document review cannot see the box.

L1 — External Inspection. Microscopy of the package surface, and, where justified by the approved test plan, marking-permanency or resistance-to-solvents testing, surface the most common physical alterations: blacktopping and resurfacing, remarking, incorrect logos, fonts, or lead finishes, sanded texture, and ghost markings visible beneath a reprint. This is also where an inexperienced inspector is most often misled, because a clean, confident marking reads as reassurance, when a well-executed remark is precisely a clean, confident marking. What this level misses is the part that is physically correct but wrong in origin — a reclaimed device containing an originally genuine die but presented with an unsupported condition, history, grade, or provenance, or an imitative device whose external appearance is consistent with the claimed package.

L2 — Non-destructive Internal and Material Analysis. This level examines internal construction or material composition without intentionally destroying the device. X-ray radiography can reveal die dimensions, die placement, bond-wire patterns, lead-frame geometry, voids, and package construction inconsistent with the expected device; it does not ordinarily establish exact die identity or electrical performance. X-ray fluorescence answers a different question — it provides elemental-composition evidence for accessible finishes, coatings, and materials — and it does not image the internal construction of the package. Acoustic microscopy may be used where relevant to identify delamination or other internal package conditions. Because these methods are generally non-destructive, they may support broader screening than destructive examination, but the actual coverage remains a sampling and test-plan decision, and a structurally consistent image still establishes nothing about provenance, prior use, or conformity under operating conditions.

L3 — Advanced Electrical, Destructive, and Forensic Verification. This level addresses questions left unresolved by the layers above. Depending on the risk hypothesis and the approved test plan, it may include electrical parametric or functional testing, decapsulation, die-mark and die-layout examination, scanning electron microscopy, cross-sectioning, or other destructive analysis. It can identify die-mark or layout inconsistencies, construction inconsistent with the claimed device, parametric drift, functional failure under specified conditions, and indicators consistent with prior use, handling damage, ageing, or degradation that earlier levels could not resolve — subject to the methods selected, the parameters tested, the sample size, the acceptance criteria, and how representative the sample is. Two limits are structural. Electrical testing is often non-destructive, and it may produce a passing result for a device that remains suspect or nonconforming on other evidentiary grounds. Destructive examination consumes the tested sample, and so, without an adequate sampling basis and evidence of lot homogeneity, it cannot establish the condition of every unit in the lot.

Why the methods have to be layered

Because each method has a blind spot that another can address, the methods are complementary rather than interchangeable. ERAI’s 2025 discussion makes the point in operational terms: electrical results alone may be insufficient to distinguish a nonconforming part from a suspect-counterfeit one, and the report emphasizes the value of combining electrical, visual, marking, and other relevant methods, particularly for high-risk applications involving parts without traceability to the original manufacturer.

Evidence from one level does not substitute for the questions addressed at another. The L0–L3 scale is cumulative as an evidentiary framework, while the exact methods, sequence, sample coverage, and stopping rules are determined by the approved test plan and the evidence gaps being addressed. Completion of a level therefore means that its defined evidentiary questions have been addressed to the extent required by the plan; it does not mean that every possible method associated with that level was used or that questions assigned to a deeper level were resolved.

Sampling, and the problem the destructive methods introduce

When a test consumes the part, the tested unit is no longer available for use, and the finding describes the sample rather than every unit in the delivery. The inference is valid only within the bounds of the sampling plan and the evidence supporting lot homogeneity. A lot assembled from several sources, or one whose custody has gaps, undermines the assumption that the sample represents the whole. This is one reason lot-level verification and chain-of-custody are treated together, and why the depth of verification a lot warrants is decided from its evidence situation rather than applied uniformly to every delivery — the subject of lot-level verification.

How CILM uses detection evidence

CILM does not perform laboratory testing. It determines which evidentiary questions remain open, routes the lot to an appropriate verification depth, and records qualified results as evidence.

A verification result is incorporated as a Lab Evidence Event only after its provenance and scope have been assessed. The event record identifies the tested sample, the chain of custody, the laboratory, the methods, the coverage, the acceptance criteria, the findings, the timestamp, the responsible reviewer, and a file-integrity record. A report that lacks these elements may still inform a human decision, but it does not strengthen evidence confidence in the way a fully qualified event does.

The Lab Evidence Event supports a bounded, governed update to the evidence-supported factor values and to the confidence score, described under measuring supply-chain risk. The update is bounded by design: verification reduces residual risk but does not, on its own, outweigh weak provenance or documentation, and the exact governance parameters are held in restricted technical documentation rather than published here. The prior assessment and the routing decision are retained separately, so an auditor can reconstruct what was known before testing and what changed afterward.

The Digital Component Passport is then updated with the verification event, its evidence, its coverage, its findings, its limitations, and the resulting disposition. It does not state that the lot is authentic. It records the evidence position reached and the residual uncertainty that remains.

The bounded result

Every detection method has evidentiary boundaries that better equipment, competent execution, and broader coverage can reduce but cannot eliminate entirely. A result without identified anomalies at the deepest level establishes only that the tested sample met the acceptance criteria for the specific methods, parameters, conditions, and coverage applied. It does not, by itself, establish that every unit in the lot is genuine and traceable to its manufacturer.

When the methods are relevant, competently performed, and fully documented, the residual uncertainty may be materially reduced, but it is not zero and the methodology does not describe it as zero. Parts with an obsolete or end-of-life status accounted for 60.02 percent of the parts reported to ERAI in 2025, compared with 36.15 percent for active parts. It is when authorized supply becomes constrained, fragmented, or exhausted that buyers are increasingly pushed toward channels where provenance evidence is weaker.

Detection often carries the greatest decision weight where provenance evidence is weakest, even though testing alone cannot reconstruct provenance that was never recorded. CILM therefore treats each result as one qualified input to a documented disposition rather than as the disposition itself.

[DIAGRAM: a four-rung vertical ladder labelled L0 to L3 from bottom to top. Each rung has two columns: “can establish” on the left and “cannot establish” on the right. L0 Documentary Evidence Review — can establish: inconsistent date/lot codes, missing traceability; cannot establish: competent forgery, correct paperwork on the wrong goods. L1 External Inspection — can establish: blacktopping, remarking, incorrect markings and finishes; cannot establish: a reclaimed device with an originally genuine die in a correct package, imitative devices matching the claimed package. L2 Non-destructive internal and material analysis — X-ray radiography: die and package-construction anomalies; XRF elemental analysis: composition anomalies in accessible materials; cannot establish: exact die identity, electrical behaviour, provenance. L3 Advanced electrical, destructive, and forensic — can establish: layout or construction inconsistencies, parametric drift, functional failure; cannot establish: lot-wide authenticity from a sample, the origin of an electrically-passing part. Caption below: depth increases upward; each level addresses additional questions, but no level removes every residual limitation. Author-developed.]