The 3.1 certificate arrives with the bars: heat number, chemical analysis, mechanical tests and the inspection manager's signature. Three weeks later a flange leaves the shop turned, drilled and welded to a pipe stub, and nothing of that number is left on the finished part, because the first lathe pass took it away with the chips. For the customer's auditor the document only counts if someone can show that this flange came from that bar. That is where material traceability, which a piece of paper solves, turns into metal part traceability, which needs a method instead. It takes marking that survives machining, the heat number carried across every cut, physical alloy verification where the risk justifies it, and a record that ties the measurement to the lot.
Tracing the material and tracing the part: two duties the standards keep apart
The difference lies in the unit being followed. The steel mill traces the heat, a chemical unit that can weigh tens of metric tons and become hundreds of bars, plates or forgings covered by the same certificate. Whoever processes the material has to follow the object instead, which is born from a cut and changes shape at every step. ISO 9001:2015 holds the two together in clause 8.5.2, which requires the organization to use suitable means to identify outputs. When traceability is a requirement, the organization shall also control their unique identification and retain the documented information needed to enable it. The new edition, ISO 9001:2026, came out on September 16, 2026. According to certification bodies, organizations certified to the 2015 version have until September 30, 2029 to complete the transition. Either way, the requirement is conditional. The customer, the contract or the law switches it on, and whoever switches it on also decides how fine the mesh must be, down to the lot or to the single part.
For pressure equipment the law always switches it on. Directive 2014/68/EU (PED) was transposed in Italy by Legislative Decree 93/2000, as amended by Legislative Decree 26/2016. In Annex I, section 3.1.5, it requires suitable procedures to identify the material of the components that contribute to pressure resistance "from receipt, through production, up to the final test of the manufactured pressure equipment". Section 4.3 adds the documentary side, because for all materials the equipment manufacturer must obtain the material manufacturer's documentation affirming compliance with a specification. For the main pressure-bearing parts of equipment in categories II, III and IV that documentation must be a certificate of specific product control.
Structural steelwork has a similar reference in EN 1090-2:2018+A1:2024, whose clause 5.2 ties traceability to the execution class. For EXC3 and EXC4, constituent products shall be traceable at all stages from receipt to handover after incorporation in the works. From EXC2 upwards, whenever different grades or qualities are in circulation together, each item shall carry a mark identifying its grade. Fabricators welding under EN ISO 3834-2:2021 find identification and traceability in clause 17, between calibration of measuring equipment (clause 16) and quality records (clause 18). The standard thus places it inside the welding process, next to the activities that secure its result.
What an EN 10204 certificate guarantees, and what to check before accepting it
EN 10204:2004 (UNI EN 10204 in Italy) defines four types of inspection document and splits them into two families. Types 2.1 and 2.2 rest on non-specific inspection, meaning tests the manufacturer runs on its own production but not necessarily on the delivered lot. Types 3.1 and 3.2 rest instead on specific inspection, carried out on the test unit set by the product standard or the order. The document type is chosen when the order is written, because at receiving it's too late. When the order says nothing, the product standard applies, and that is often less than the end customer expects.
| Document | Based on | Validated by | What it proves at receiving |
|---|---|---|---|
| 2.1 Declaration of compliance with the order | No test results | Manufacturer | Only the manufacturer's commitment |
| 2.2 Test report | Non-specific inspection | Manufacturer | Typical production values, not those of the delivered lot |
| 3.1 Inspection certificate 3.1 | Specific inspection on the lot or heat | Manufacturer's authorized inspection representative, independent of production | That this test unit was tested, with these results |
| 3.2 Inspection certificate 3.2 | Specific inspection on the lot or heat | Manufacturer's representative together with the purchaser's representative or an inspector designated by official regulations | The same as a 3.1, with a second independent signature |
For structural and pressure materials the 3.1 is the sensible starting level. The 3.2 is requested where the customer or a notified body wants to witness the tests, as for some oil & gas and shipbuilding components. For the main pressure-bearing parts of categories II, III and IV, the specific product control required by section 4.3 of the PED translates in practice into a 3.1 or a 3.2. None of the four documents, however, says anything about the part the warehouse is holding. All of them certify a test unit, and the physical link between that unit and the part is the receiver's job. At receiving, the useful document check on a 3.1 comes down to a few items worth writing into the acceptance procedure:
- the heat or lot number on the certificate matches the one marked on the material, package by package;
- the designation and product standard match the order, edition included (for example EN 10025-2 S355J2+N, EN 10216-2 P265GH, ASTM A182 F316L);
- chemical analysis and mechanical tests fall within the product standard's limits, not just those stated by the supplier;
- delivery condition and heat treatment are as required (normalized, quenched and tempered, solution annealed);
- the document carries the validation of whoever is entitled to give it and, for a 3.2, the second signature can be traced to the purchaser or the inspection body;
- if the material comes through an intermediary, the document is the manufacturer's original or an unaltered copy, with a reference linking it to the delivery.
The last item is the one most often skipped. EN 10204 expects the intermediary, meaning the trader or distributor who supplies the product without changing its properties, to pass on the manufacturer's document as it is, in original or as a copy, with an identification that keeps paper and material together. A certificate retyped on the distributor's letterhead, with the values copied over, has lost the validation of whoever ran the tests. EN 10168:2004 helps when reading certificates from different mills. It lists the information an inspection document for steel products may contain and gives every section a code number, so the heat number, designation and results sit in the same field even when the form changes.
Marking the part: the next operation decides the method
A mark lasts as long as the harshest operation it meets. Hard stamping deforms the metal, so it survives blasting, painting and heat treatment, but it introduces a notch. On pressure or fatigue-loaded components, low-stress stamps with rounded tips are used, in agreed positions away from welds and highly stressed areas. Dot peen marking writes with a series of dots, is programmable and can reproduce text and two-dimensional codes. Laser marking either engraves by removing material or, on stainless steel, marks by surface annealing that changes the color of the outer layer. On stainless surfaces in contact with process fluids, the corrosion resistance of the marked area must be checked, because the local thermal cycle can reduce it.
| Method | Effect on the part | Survives | Lost through |
|---|---|---|---|
| Hard stamping (including low-stress) | Deformation, notch | Blasting, painting, heat treatment | Machining of the marked area |
| Dot peen | Shallow dots | Handling, thin paint coats | Aggressive blasting, thick coatings |
| Laser engraving | Surface material removal | Washing, light blasting | Machining of the marked area |
| Laser annealing (stainless) | Color change without removal | Handling, washing | Pickling, blasting |
| Electrochemical marking | Minimal etching | Handling, washing | Blasting |
| Paint, color code, label or tag with QR code | None | Storage and transport | Cutting, washing, blasting, heat treatment |
The rule that follows is to mark with the method that survives the next harshest step, and to re-mark after every operation that can erase it. Where the finished surface can't take a mark, the part is marked in the machining allowance or on a tag tied to it, and the transfer is recorded. Data Matrix, the two-dimensional symbology defined by ISO/IEC 16022:2024, suits direct part marking for two reasons. The Reed-Solomon error correction of ECC 200 keeps it readable even when partly damaged, and in a few square millimeters the code carries a unique identifier that the ERP links to heat, certificate and order. A QR code does the same job on labels, where surface condition isn't an issue.
The heat number survives cutting only if someone copies it first
Every operation is a point where the original marking can disappear, and cutting comes first. A six-meter bar with the heat number stamped at one end becomes twenty pieces, and nineteen of them are anonymous unless the number is transferred to each one before cutting. Afterwards it's too late, because pieces from different heats end up in the same bin. The practice that holds is to transfer the marking to the remaining length and to every piece, with the same method or with an internal code.
The internal code is needed when machining inevitably removes the original mark, and it is a short identifier, stamped on the part and entered in a register that links it uniquely to heat, certificate and order. That register has to exist before the part changes shape, otherwise the code is just a label. Welding then brings in a second material that usually nobody marks, the filler metal. Wire, rods and fluxes have their own lot and certificate, and a joint is only traceable if the weld record also carries that lot, together with the welding procedure specification (WPS) and the welder. Heat treatment closes the loop from the other side. A part quenched and tempered or stress relieved in the shop has properties the heat certificate no longer describes, and its identity card becomes the heat treatment report, tied to the part by the same code.
Foundries start the chain from the inside. The casting is born with its heat or melt number, often cast in relief from the mold or stamped after cleaning, and the spectrometric analysis run on the heat sample becomes the basis of the certificate the foundry itself issues. In every one of these cases the practical carrier is the job traveler, on paper or digital, which follows the lot from one step to the next and on which each department records what it did and with which material. When the traveler gets separated from the part, traceability breaks, even with every certificate on file.
Who decides how many parts get PMI, and why sampling misleads on mixed lots
The extent of verification is decided by the asset owner or, along the supply chain, by the purchaser who writes it into the order. That is the approach of API RP 578, the American Petroleum Institute recommended practice now in its 4th edition, published in February 2023 as Material Verification Program for New and Existing Assets. Its status is that of a recommendation, and it becomes binding when a contract or plant specification invokes it, which is common in oil & gas and petrochemical specifications. The program it describes rests on PMI (positive material identification), the check that a component's nominal composition matches the specified material. The owner sets the extent based on risk, meaning service criticality, the consequences of an alloy mix-up and supply-chain reliability. The way of counting, meaning what a lot is, what percentage and which parts, is set by whoever writes the plan. None of the standards cited here fixes a percentage that fits everyone.
Sampling misleads on a mixed lot. Take 200 bars with two wrong-alloy pieces among them, 1%, and a 10% random check. The probability that none of the 20 tested pieces is one of the two wrong ones is (180/200) × (179/199), about 0.81. Four times out of five the lot passes acceptance with the mix-up inside, and doubling the sample to 40 pieces still leaves around a 64% chance of missing them. The calculation holds for random sampling without replacement, and if the wrong pieces all sit at the bottom of the bundle it's worse.
Do your lots pass through service centers or stockists?
Send us the alloys in circulation, incoming volumes and the points where marking gets lost. We will tell you where 100% PMI is needed and with which instrument, with a test on your own material at no obligation and a tailored quote within 24 hours.
Email us Message us on WhatsAppThat calculation explains why potentially mixed lots, those that have passed through a service center, a warehouse holding offcuts or a construction site, are checked 100%. Sampling stays defensible on homogeneous deliveries arriving from the mill with intact marking. On the methods side, ASTM E1476-04(2022) is the general guide to metals identification, grade verification and sorting, with the principles, fields of use and, where available, the expected precision of each technique. ASTM E1916-24 deals specifically with separating mixed lots in the plant using methods within reach of plant staff. It also states two limits worth knowing, because some alloy combinations can't be reliably separated with those means and sorting doesn't replace the analysis that verifies compliance with a specification.
Material analysis with XRF, LIBS or a laboratory: the question decides the instrument
PMI analysis covers several chemical techniques for material analysis, and each one answers a single question well. The handheld XRF analyzer, the most common metal analyzer at receiving, uses X-ray fluorescence to report chromium, nickel, molybdenum, copper, niobium, titanium and manganese contents in a few seconds, with no preparation on a clean part and without leaving a mark. Its limit is the light elements. Silicon and aluminum are read with reduced sensitivity and carbon stays out of reach, for physical reasons explained in the comparison of LIBS and XRF for carbon measurement. XRF separates a 304 from a 316 or a P11 from a P22 with ease, but it can't separate a 316 from a 316L.
LIBS, laser-induced breakdown spectroscopy, vaporizes a microscopic spot of the surface with a pulse and analyzes the light emitted by the plasma. With an argon-purged measurement chamber it also reads carbon, and it is the portable instrument that separates L grades and verifies carbon steels. In exchange it needs local grinding of the surface and leaves a micro-mark. Spark optical emission spectrometry (OES) remains the quantitative reference for carbon and low-alloy steels, codified by ASTM E415-21 for the simultaneous determination of 21 elements. For stainless and alloy steels, in the laboratory, wavelength dispersive XRF under ASTM E572-21 is also used. These methods require sample preparation, and their results, when issued by a laboratory accredited to ISO/IEC 17025:2017, carry the weight needed in a dispute.
| Question to answer | Suitable instrument | Limit to know |
|---|---|---|
| Is it the ordered alloy family and grade? (stainless, Cr-Mo, nickel, copper, titanium alloys) | Handheld XRF | No carbon; light elements at reduced sensitivity |
| Is it 316L or 316? Does the carbon content meet the weldability requirement? | Handheld LIBS or mobile OES | Local grinding, micro-mark on the part |
| Is the certificate wrong? You need a result to put to the supplier | Accredited laboratory (OES, combustion) | Sample removal, turnaround time |
The verdict depends on the material mix. For a fabrication or machine shop working mostly with stainless and low-alloy steels, XRF covers most incoming checks and LIBS earns its place when the L grade or carbon is an order requirement. The criteria for choosing configuration and detector are collected in the guide on how to choose a portable XRF analyzer. When the shop receives almost only unalloyed structural steels, an in-house instrument rarely pays off, because serious document control and an external laboratory for doubtful cases cost less. Suppliers to oil & gas, or builders of pressure equipment with different alloys in circulation, on the other hand, find in-house PMI the only way to check 100% without stopping production. The article on PMI on piping and welds in oil & gas plants covers that case in detail. Foundry work shifts the logic again, because verification starts from the charge and the heat sample, as the article on alloy verification in the foundry explains.
From reading to lot: what to record so PMI holds up in an audit
For an auditor, a measurement that leaves no record is a measurement that never happened. The minimum record, for each part or group of parts, contains the part identifier, heat number and certificate number, the expected alloy and the measured one with its main contents. The result, the instrument with its serial number, the operator and the date complete the line. Upstream, two reliability checks are needed, namely verification of the instrument on a certified reference sample at the start and end of each shift, recorded as well, and operator qualification under the internal procedure. The chain the auditor must be able to walk through, in one line, is part ID → heat no. → 3.1 certificate → PMI result (alloy, Cr/Ni/Mo) → operator and instrument → release.
Current analyzers store spectrum and result together with the identifier typed in or read from a barcode, and export to the quality management system. That step turns a reading into evidence, more than the instrument's precision does. After measurement the verified part is marked with a sticker, a paint dot or the code itself, so the warehouse can tell at a glance what has been tested and what hasn't.
Where the chain breaks: the mix-ups that actually reach the shop floor
Almost every material mix-up starts at the same point, a hand-over where the part loses its identity and nobody gives it back. It happens at the service center that cuts bars from different heats and attaches a single certificate to the bundle. It also happens in the warehouse where offcuts go back on the rack unmarked next to a grade that looks identical, such as a 304 next to a 316 or an S275 next to an S355. It happens with a certificate that is right on designation but refers to another heat, and with one retyped by the intermediary. In welding it happens with filler metal drawn from the wrong lot, in maintenance with spare parts that enter the plant outside the acceptance flow. The worst case, the falsified certificate, can give itself away precisely on traceability: a heat number that doesn't follow the mill's format, or identical analysis values on different heats.
The remedies follow the same cause. Material is re-identified at every hand-over, look-alike grades are kept physically apart, anything arriving without a legible number goes into quarantine and, where marking is a requirement, unmarked material is treated as non-conforming.
Setting up incoming inspection in a mid-sized manufacturer: who does what
A traceability solution for metal parts is designed as a process, and the analyzer occupies only one link of it. With a few dozen people on staff, incoming inspection works when every department knows which link it holds. A split that stands up in audits is this:
- Purchasing: writes into the order the product standard with its edition, the EN 10204 document type, heat number marking on every part or package, whether copies from intermediaries are acceptable and any PMI to be performed by the supplier.
- Receiving and warehouse: compares delivery note, certificate and marking, and sends anything that doesn't match to a marked quarantine area, not onto the rack.
- Quality: checks the certificate against the checklist, performs or orders PMI according to the plan (100% or sampling), records the result and releases the material.
- Production: transfers the marking before every cut, updates the traveler, records filler lots and heat treatments.
- Shipping: together with quality, carries out final release, with confirmation PMI on critical parts and a dossier linking every component to its certificates.
The verification plan also needs an exit rule that works both ways. A supplier checked 100% for an agreed number of consecutive deliveries without deviations can move to sampling. The first discrepancy sends it back to 100% and opens a non-conformity against the supplier. Quality sets the threshold, and it is best written down before starting, while no problem is shaping it yet.
What to send us for an assessment
To work out where physical verification is needed and with which instrument, a few precise details are enough:
- alloys and grades in circulation, with the product standards called out in the orders;
- monthly incoming volumes and number of suppliers, flagging those that go through intermediaries;
- material forms: bars, tubes, plate, forgings, castings;
- points where the original marking gets lost: cutting, machining, welding, heat treatment;
- customer requirements: EN 10204 certificate type, PED, EN 1090-2, API RP 578, verification plans;
- how checks are recorded today: paper, quality management system or not at all;
- material mix-ups or non-conformities that have already happened.
You can send them by email to info@pitech-solution.com or on WhatsApp, with a sample certificate and a photo of the marking.
Part traceability is designed starting from the point in the chain where the original marking disappears, and that is where you decide how to transfer it, who checks it and where it's recorded. If different alloys circulate at that point, physical verification is needed, and the choice between XRF, LIBS and laboratory depends on the question it has to answer. To size it on real numbers, meaning alloys in circulation, incoming volumes, hand-over points and customer requirements, PITECH proposes the most consistent analysis configuration and a test on your own material. The starting point is the page on XRF and PMI material analysis, with the pages on handheld XRF analyzers for alloy identification and LIBS analyzers for carbon and L grades.
Frequently asked questions about metal part traceability and PMI
What is metal part traceability and how does it differ from material traceability?
It is the ability to link every finished part to the heat it came from, to its certificates and to the operations it went through. Material traceability stops at the heat and its EN 10204 certificate. Part traceability has to survive cutting, machining, welding and heat treatment, so it needs marking, heat number transfer at every step and a record that ties the part code to the documents.
What should be checked on an EN 10204 3.1 certificate at receiving?
First, that the heat number on the certificate matches the one marked on the material. Then that the designation and product standard match the order, and that the chemical analysis and mechanical tests fall within the product standard's limits. Delivery condition and heat treatment must be as required, and the document must be validated by an authorized representative independent of production. If the material comes through an intermediary, the certificate must be the manufacturer's original or an unaltered copy.
What is the difference between a 3.1 and a 3.2 certificate?
The difference is who validates it. Both report results of specific inspection on the delivered lot. A 3.1 is validated by the manufacturer's authorized inspection representative, independent of the manufacturing department, while a 3.2 also carries the signature of the purchaser's representative or an inspector designated by official regulations. A 3.2 is requested when the customer or a notified body wants to witness the tests.
What is PMI analysis and which metal analyzer do you need?
PMI (positive material identification) analysis is the non-destructive check that a component's composition matches the specified alloy. A handheld XRF analyzer identifies stainless steels, low-alloy steels, nickel, copper and titanium alloys in seconds, but it can't measure carbon. Handheld LIBS also measures carbon and separates L grades such as 316L, while laboratory spark OES remains the reference when you need a quantitative result to put to the supplier.
Should PMI be done on every part or by sampling?
It depends on risk and on who writes the verification plan. API RP 578 leaves the extent to the asset owner, based on service criticality, the consequences of a mix-up and supply-chain reliability. On potentially mixed lots sampling offers little protection. With 2 wrong parts in 200 and a 10% random sample, the probability of missing both is about 81%, so those lots are checked 100%.
How do you mark a metal part so traceability isn't lost?
With the method that survives the harshest operation that follows. Hard stamping survives blasting and painting but creates a notch. Dot peen and laser can also write Data Matrix codes to ISO/IEC 16022:2024, while paint and labels only work for storage and transport. The heat number must be transferred to every piece before cutting and, where machining removes it, replaced by a recorded internal code.