How to Identify Counterfeit IC Chips: A Buyer's Guide

Counterfeit electronic components are not a victimless crime of "cheap substitutes that mostly work." They are a direct threat to product reliability, brand reputation and — in medical, automotive or industrial safety applications — human life. The Semiconductor Industry Association estimates that counterfeit parts cost the global electronics industry $75 billion annually, and the incidence rate has climbed sharply since the 2020-2023 supply-chain crisis when desperate buyers turned to unvetted brokers.

The most common forms of fraud fall into five categories: remarked genuine parts (a lower-spec chip is sanded and relabeled as a higher-spec one), recycled pulls (used chips harvested from e-waste boards, cleaned and sold as new), clone dies (a functionally similar but electrically inferior chip is packaged with a major brand's markings), ghost lots (fake date codes and lot numbers that cannot be traced back to any real production run), and empty reels (real packaging with fake or missing contents inside).

For a B2B buyer, the practical question is not "does fraud exist" — it does, everywhere — but "what specific steps do I take on every incoming lot to catch it before it reaches SMT." This guide lays out those steps in the order we run them at our Shenzhen quality lab, from zero-cost visual checks through destructive physical analysis.

Visual inspection under a 20-40× stereo microscope remains the fastest and cheapest counterfeit screen. The trained eye catches six out of ten fake lots before any electrical test runs. Focus on these seven markers in order:

Marking quality. Genuine ICs from TI, ST, Analog Devices, Infineon, Microchip, NXP, onsemi and Nexperia use laser etching — the characters are physically engraved into the silicon package surface, sharp-edged, and cannot be scratched off with a fingernail or dissolved by acetone. Counterfeit parts often use silk-screened ink or pad-printed markings that feel slightly raised, blur at the edges under magnification, and wipe off with isopropyl alcohol on a cotton swab. If a swab test removes any ink, the part is fake.

Font and spacing consistency. Every major manufacturer uses a proprietary marking font with tight tolerances. Compare suspect parts against a known-genuine reference under the same magnification: character height, stroke width, spacing between characters, and alignment should match exactly. Cloned dies frequently use a generic sans-serif font that is visually "close" but measurably different.

Date code plausibility. The date code (usually 4 characters: 2-digit year + 2-digit week) should be consistent with the manufacturer's production history. A part marked with a future week, a week when the fab was known to be shut down, or a date code that contradicts the reel label is a red flag. TI, ADI and Microchip all publish date-code checkers on their websites.

Pin finish. Genuine leads have uniform matte-tin plating with a slightly rough texture. Recycled pulls show shiny, re-plated leads (the original tin oxidizes during use, so counterfeiters re-plate them — the result is brighter and smoother than factory finish). Look for solder residue, scratches or discoloration near the lead frame — signs the part was previously soldered onto a board.

Package surface. Injection-molded packages have a consistent matte texture with a small mold-dot or notch on pin 1. Sanded-and-remarked parts often show a slightly glossier surface (from sanding) or swirl marks from the abrasive. Under 40× magnification, the mold compound should look uniform, not polished.

Weight. A reel of 2,000 SOP-8 parts from TI weighs about 180 grams. A reel of 2,000 QFN-28 CP2102 parts from Silicon Labs weighs about 95 grams. Counterfeit reels frequently weigh 10-20% less because the dies inside are smaller or the lead frames are thinner. Use a 0.01 g precision scale and compare against known-genuine weights.

Reel and packaging. Original manufacturer reels have specific label fonts, barcodes and lot-code formats. Fake reels often have blurry labels, mismatched lot codes between the reel and the bag, or use generic packaging that the manufacturer never ships in. If the reel label says "TI" but the bag inside is a plain anti-static bag with no markings, suspect a repack operation.

Visual inspection misses remarked genuine parts (real package, real markings, but a lower-spec die inside) and clone dies that look perfect externally. Electrical testing catches these. The approach depends on the part type:

Analog parts (op-amps, comparators, references, regulators). Measure the parameters that are hardest to fake cheaply: input offset voltage, gain-bandwidth product, quiescent current, and temperature drift. A genuine LM358B should show input offset under 1 mV at 25 °C; a remarked standard-grade LM358 will show 3-7 mV. A genuine TL431A regulates at 2.495 V ±1%; a clone drifts 2-3% across temperature. Build a small test fixture with a precision DMM (Keysight 34461A or equivalent), a thermal chamber or hot plate, and a curve tracer for I-V sweeps.

Digital parts (MCUs, memory, logic). Read the device ID registers. Genuine STM32F103C8T6 reports Flash size 0x0040 (64 KB) at address 0x1FFFF7E0; a remarked C8T6 (actually a 32 KB die) reports 0x0020. ATmega328P signature bytes are 0x1E 0x95 0x0F; clones report different values. For EEPROMs (AT24C02), write to the full address range and read back — a remarked 24C01 (128 bytes) will fail above 0x7F. For MOSFETs, use a curve tracer to measure Rds(on) at specified Vgs — clones consistently show 2-5× higher resistance.

Interface chips (USB-UART bridges, RS-232 transceivers). Run a real-world stress test at the highest supported baud rate. A genuine CP2102 handles 1 Mbps with zero errors over an hour; remarked CH340 dies drop bytes above 921,600 bps. For MAX3232, measure the charge-pump output voltage — genuine parts produce ±5.5 V; fakes often produce only ±3 V which fails RS-232 spec compliance.

Power MOSFETs and discretes. A curve tracer is non-negotiable. Measure breakdown voltage (BVdss), gate threshold (Vgs(th)), on-resistance (Rds(on)) at multiple Vgs levels, and gate charge (Qg). Clones routinely fail at rated voltage or show 3-10× higher Rds(on) than the datasheet value. For IRF540N, genuine parts switch 33 A at 44 mΩ; fakes overheat at 10 A.

Batch sampling strategy. For lots under 500 pcs, test 100% (cost-effective). For lots 500-5,000 pcs, test a random sample of 30-50 pcs. For lots above 5,000 pcs, use MIL-STD-1916 or ANSI/ASQ Z1.4 sampling with tightened inspection levels for new or unknown suppliers. Any single failure escalates the entire lot to 100% inspection.

When visual and electrical tests give ambiguous results — or when the lot value justifies definitive verification — X-ray imaging and chemical decapsulation provide irrefutable evidence.

X-ray inspection reveals die size, lead-frame geometry, bond-wire count and routing, and internal voids or delamination. Every genuine part from a given manufacturer has a specific die size and bond-wire pattern that is consistent across production lots. A remarked LM358B will show a smaller die than a genuine LM358B because the underlying silicon is a standard-grade LM358. A recycled pull will show solder residue on the lead frame inside the package or reflow-induced voids in the mold compound.

Industrial X-ray systems (Nordson Dage, YXLON, Bruker) cost $80k-300k new, but contract labs in Shenzhen, Hong Kong and Singapore offer per-sample X-ray at $30-80 per part with 24-hour turnaround. For a $50,000 lot of STM32F103CBT6, spending $500 on X-ray of 10 samples is trivially cheap insurance.

Decapsulation (chemically removing the mold compound with hot fuming nitric acid or sulfuric acid) exposes the bare die, allowing you to read the manufacturer's die markings under a metallurgical microscope. Every major fab laser-etches a unique die ID, mask revision and sometimes a wafer lot code onto the silicon itself. These markings cannot be faked without access to the original mask set — if the die says "CKS32F103" instead of "STM32F103," the part is a Chinese clone regardless of what the package says.

Decap services run $100-250 per part with 3-5 day turnaround. Reserve this for high-value lots (FPGAs, high-end MCUs, automotive-grade parts) or when a supplier relationship is in question.

The cheapest counterfeit screen is to never buy from a channel where counterfeits enter in the first place. Apply these four verification layers before you place an order:

Authorized distribution check. Every major manufacturer publishes a list of authorized distributors on their website. TI has Arrow, Avnet, Digi-Key, Mouser, Future, and a handful of regional partners. ST has Arrow, Avnet, Future, Rutronik, and WPG. If a broker claims to sell "TI authorized stock" but is not on TI's published list, they are either lying or they are selling gray-market parts that may or may not be genuine. Gray market is not automatically counterfeit, but it carries no manufacturer warranty and no traceability.

Certificate of Conformance (C-of-C). A genuine C-of-C from the manufacturer includes the lot code, date code, quantity, test results summary, and an authorized signature or stamp. Request the C-of-C before shipment and verify the lot code matches the reel labels you receive. Fake C-of-Cs are common — cross-check the format against known-genuine examples from the manufacturer's website or from previous orders.

Traceable lot codes. Ask the supplier for the full chain: which distributor they bought from, which lot codes, which date codes, which country of origin. A legitimate authorized distributor can trace any reel back to the fab within hours. A broker who cannot answer these questions, or who answers vaguely ("we got it from a TI agent in Shenzhen"), is almost certainly selling gray-market or worse.

Price plausibility. If a part is allocated at the authorized distributor with a 26-week lead time and a broker offers 10,000 pcs "in stock" at 30% below authorized price, ask yourself: where did this stock come from? The three possibilities are (a) a cancelled order that was legitimately resold, (b) gray-market surplus from an over-order, or (c) counterfeit. Options (a) and (b) exist but are rare; option (c) is the default. A price that is too good to be true is a counterfeit signal, not a bargain.

Every shipment from our Shenzhen warehouse passes through a four-stage gate before it leaves the building:

Stage 1 — Incoming inspection. All 100% of lots receive visual inspection under 20× magnification, checking marking quality, pin finish, date code plausibility, and reel/label consistency against our reference database of known-genuine examples.

Stage 2 — Electrical sampling. Lots under 500 pcs: 100% electrical test on the parameters most likely to expose fraud (offset voltage, Flash size, Rds(on), baud rate accuracy). Lots 500-5,000: 30-pc random sample. Lots above 5,000: MIL-STD-1916 tightened sampling.

Stage 3 — X-ray on high-value lots. Any lot with unit value above $5 or total value above $5,000 receives X-ray of 5-10 samples to verify die size and bond-wire pattern against the manufacturer's reference.

Stage 4 — Documentation package. Every shipment includes a C-of-C, lot traceability sheet, and photographs of the actual reels being shipped (not stock photos). You can verify the lot codes against the manufacturer's database before the parts arrive.

If any stage fails, the entire lot is quarantined and returned to the supplier. We do not ship "probably genuine" parts.

If you suspect a shipment contains counterfeit parts, act quickly to preserve evidence and your legal position:

Stop using the lot immediately. Do not put suspect parts on your SMT line. Quarantine the entire lot — reel, bag, labels, packaging — in a secure location. Once parts are soldered onto boards, verification becomes exponentially harder and your return rights may be compromised.

Document everything. Photograph the reels, labels, individual parts under magnification, and any test results you have. Record the supplier name, PO number, invoice number, lot codes, date codes, and quantity. Save all correspondence with the supplier.

Run independent verification. Send 5-10 samples to an independent test lab (Keystone Compliance, White Horse Labs, Purity Solutions, or our partner lab in Shenzhen) for visual, electrical, X-ray and — if warranted — decap analysis. Cost is typically $500-1,500 and provides a defensible technical report.

Notify the supplier in writing. Send a formal non-conformance report with the lab findings. Request full refund or replacement from authorized channel within 10 business days. Keep a copy for your records.

Escalate if necessary. If the supplier refuses to cooperate, file a complaint with the manufacturer (TI, ST, ADI all have counterfeit reporting portals), with your local chamber of commerce or trade association, and — for significant losses — with law enforcement. In the US, the FBI and Homeland Security Investigations actively prosecute counterfeit component cases. In the EU, national IP enforcement agencies handle these.