Scrap yard alloy sorting: separating 304 from 316 with XRF

At the gate of a recovery centre, the price of a load is set before the weighbridge, at the moment someone decides what is in it. Two bins of stainless with the same weight can be worth very different sums. If the molybdenum-bearing steel is still mixed in with 304, the buyer tends to pay for everything as 304, or applies a precautionary discount. For a scrap dealer, sorting by alloy is the most direct lever on margin. It is also the first line of defence when a steel mill or foundry disputes a load. The tool that made it practical in the yard is the portable X-ray fluorescence analyzer: a few seconds of measurement per piece, the grade on the display, the spectrum saved as evidence. It works well on stainless, copper alloys, nickel alloys and titanium. On aluminium and carbon it has precise limits, and they are worth knowing before buying.

Why a mixed lot gets paid like its poorest part

Metal scrap is bought and sold by specification. Trade bodies publish detailed classifications, such as the Scrap Specifications Circular of the US Recycled Materials Association (ReMA). In Europe, Regulation (EU) No 333/2011 requires scrap that is to cease being waste to be graded to a customer specification, an industry specification or a standard. Each grade has its expected composition and its price. When different fractions share a bin, the buyer has no way to check them piece by piece and protects itself the simplest way: the lot goes under the lowest grade it contains, or is accepted at a discount.

The mechanism bites hardest where value depends on a few expensive elements. 316 differs from 304 through molybdenum, 2 to 3% in the AISI designation and 2.0 to 2.5% in the European equivalent 1.4401 of EN 10088-1:2023, along with slightly more nickel. Nickel alloys such as UNS N06625 or N07718 are more than half nickel by weight. Clean electrolytic copper fetches more than brasses and bronzes. In every one of these cases the valuable metal sits inside the piece and can't be seen from outside. Whoever fails to identify it sells it at the price of the commoner material.

Magnet, spark and colour: where traditional sorting stops

The yard has always had its own tools. The magnet separates ferrous from non-ferrous, and sends the ferritic and martensitic 400-series stainless along with the ferrous. The grinding-wheel spark test gives a sense of carbon content from the shape of the sparks. Colour and density point the way among copper, brass and aluminium. ASTM E1476-04 (reapproved 2022), a guide to metals identification and sorting that describes itself as intended for tutorial purposes, lists these qualitative methods alongside X-ray fluorescence and optical emission: eddy current, conductivity, thermoelectric testing, chemical spot tests and, as a special case, the spark test itself.

For a first sort they remain useful and cheap. They stop, though, where value begins. A 304 and a 316 look the same, respond to the magnet the same way (barely at all, except in cold-worked areas) and throw the same spark. A nickel-alloy bar that has ended up among the stainless can't be picked out by eye. A chemical spot test for molybdenum exists, but it takes time and a prepared surface spot on every piece, and it leaves no documentary trace. The portable analyzer covers exactly that step, and it does so after triage rather than in its place.

Separating 304 from 316 in the yard: what the instrument measures

Once the window sits on the piece, the analyzer irradiates the surface with a small X-ray tube and counts the fluorescence photons that each element re-emits at a characteristic energy. The software turns the spectrum into a composition and matches it against a grade library, returning the most likely name together with the measured contents. On stainless the technique is on home ground. Chromium and nickel set austenitic grades apart from ferritic ones such as 430, which carries almost no nickel, molybdenum separates 316 from 304, and titanium identifies 321, which relies on that element for stabilisation against intergranular corrosion after welding.

The same principle works on the other families that matter in a recovery centre. In copper alloys, zinc marks the brasses and tin the bronzes, while lead is measured with good sensitivity, which is useful because leaded and lead-free brasses can go to different outlets. Nickel and titanium alloys are recognised by the combination of their alloying elements. Next to the grade, the instrument always shows the spectrum, and the spectrum is what gets archived.

Aluminium is less straightforward. Series alloyed with copper (2xxx), manganese (3xxx) and zinc (7xxx) separate well. Those that depend on magnesium and silicon, the 5xxx and 6xxx series, push the technique to its limit, because in air the fluorescence of light elements is absorbed before it reaches the detector. Instruments with a silicon drift detector (SDD) measure magnesium and silicon with longer counting times, while LIBS, which analyses the plasma produced by a laser pulse, starts with an advantage on light elements. The review by Gaustad, Olivetti and Kirchain in Resources, Conservation and Recycling (2012) lists silicon and magnesium, together with iron, copper and zinc, among the impurities that build up in recycled aluminium streams and limit their reuse.

What fluorescence doesn't see

The first limit concerns carbon. A portable XRF analyzer measuring in air doesn't detect elements lighter than magnesium, and carbon is one of them. On its own, then, it can't tell 304L from 304: under EN 10088-1, grade 1.4307 allows at most 0.030% carbon and 1.4301 up to 0.07%, while the rest of the composition overlaps. The same goes for classifying carbon steels and cast irons by carbon content. Where that distinction changes the price or the conformity, the answer is spark optical emission spectrometry (OES), mobile or in the lab, or a portable LIBS analyzer with argon purge, in the models that state carbon measurement. The full comparison is in our article on LIBS and XRF for carbon.

The second limit is depth. The fluorescence that reaches the detector comes from the first few micrometres of metal. Our own calculation from the attenuation coefficients tabulated by NIST puts it, in steel, at a few micrometres for nickel and a few tens of micrometres for molybdenum. Paint, rust, oil, galvanising or chrome plating are therefore measured instead of the base metal, and the countermeasure is a pass of a grinder or disc over a couple of square centimetres, or a reading on a fresh cut. It takes little, but it has to be done every time.

The third concerns geometry. Turnings, thin wire and small pieces don't cover the measuring window, and the result is affected by air and by whatever sits underneath. For these materials, sample cups are used or a representative sample is compacted. All three conditions belong in the written yard procedure.

One bin of demolition stainless, from triage to the mill

An example, built on stated assumptions. A recovery centre receives from an industrial demolition a bin of about 2 tonnes of mixed stainless: pipes, flanges, sheet and a few valve parts, an estimated 400 pieces in all. Magnet triage first removes the ferritic grades and any carbon-steel pieces that slipped into the load. In this scenario some 300 austenitic pieces remain, and they go to the analyzer after a quick pass of the disc on the measuring spot.

Then there is time. Each reading takes a few seconds of acquisition on a clean surface. With handling and cleaning it comes to 20 to 30 seconds per piece, our estimate for one operator working on manageable pieces. Three hundred pieces therefore mean a good two hours of work. The expected result is three or four fractions: 304 and 304L together, because XRF can't separate them, then 316 and 316L, 321, and any nickel-alloy pieces, which go into a separate container. If the pieces whose classification hinges on carbon carry weight in the value, a sample of them goes to spark OES. The 316 lot and the nickel-alloy lot, the most valuable ones, also get a lab confirmation before sale, for example by wavelength-dispersive X-ray fluorescence, the bench technique covered by the guide ASTM E1621-21.

What the operation earns depends on the day's prices and on the share of 316 in the bin, and any generic figure here would be invented. The sum is done with your own price lists: weight of the recovered fraction times the price gap between the two grades, minus the labour hours and the cost of any lab confirmation. If the 316 fraction is small, or the two grades are quoted close together, the two hours don't pay back, and selling the lot as mixed is the better call.

Who disputes a load, and what evidence answers them

The party that disputes a load is usually the steel mill or foundry that has to melt it, and the most serious reason is residual elements the melt can't remove. In steel, copper and tin stay in the bath after refining and, above certain thresholds, cause cracking during hot working. Copper is the best-studied case. A University of Cambridge study published in Environmental Science & Technology (Daehn, Cabrera Serrenho and Allwood, 2017) notes that there is currently no commercial process to remove copper from steel scrap, and estimates that around 2050 the copper in scrap is likely to exceed what the products in demand can tolerate. For anyone selling ferrous scrap, electric motors, windings and cables left in the load end up costing a downgrade.

Here XRF plays a different role from alloy sorting. On a load of shredded ferrous scrap it doesn't measure the average copper content, which would take sampling and lab analysis. It helps find and remove the pieces that carry it, and confirms that a lot declared as stainless or nickel alloy really is one. At intake it checks that the material matches the supplier's declaration; at dispatch it documents that the lot sold has the expected composition. How many readings, on how many pieces and under which sampling rule is not fixed by any scrap standard. The buyer's specification sets it, or the parties agree it, and it pays to put it in writing before the first dispute.

A dispute closes sooner if the lot file holds at least the following:

• lot identifier, weight, date and reference to the specification or sales grade;
• number of pieces measured and the agreed sampling rule;
• grade and element contents for every reading, with the spectrum saved and the instrument serial number;
• the reading of a reference sample at the start of the shift, showing that the instrument was measuring correctly that day;
• the radiometric monitoring certificate for the load;
• for end-of-waste material, the statement of conformity required by the applicable regulation.

Radiation protection and radiometric monitoring: two separate duties

In Italy, a recovery centre meets Legislative Decree No 101 of 31 July 2020, which transposes Directive 2013/59/Euratom on basic safety standards against ionising radiation, from two different sides. The first concerns the load. Article 72 requires radiometric monitoring from anyone who imports, collects, stores or melts metal scrap, to intercept orphan sources and abnormal radioactivity, and assigns it to second- or third-degree radiation protection experts. This duty exists regardless of any XRF analyzer, and Regulation 333/2011 also calls for it when it requires qualified staff to monitor the radioactivity of each consignment.

The second side concerns the instrument. An XRF analyzer contains an X-ray tube, and using it is a practice involving ionising radiation. As a rule it requires the practice notification under Article 46 of the same decree, sent to the listed bodies at least thirty days before start-up, together with an assessment by a radiation protection expert, operator training and written rules of use. Whether a given model qualifies for an exemption is for the expert to decide, based on the device's technical data. It is worth planning before purchase, because the notification lead time adds to the delivery time.

End-of-waste: what Regulations 333/2011 and 715/2013 actually require

Council Regulation (EU) No 333/2011 sets the criteria under which iron, steel and aluminium scrap cease to be waste; Commission Regulation (EU) No 715/2013 does the same for copper scrap. Neither prescribes XRF. They require scrap to be graded to a customer specification, an industry specification or a standard for direct use in steelworks or foundries, and foreign materials to stay within a weight limit: 2% for iron and steel, 5% for aluminium or alternatively a metal yield of at least 90%, 2% for copper. They also require a quality management system and a statement of conformity for each consignment, and Regulation 333/2011 adds radioactivity monitoring of every consignment.

For iron and steel scrap, the regulation counts non-ferrous metals among foreign materials (alloying elements in the ferrous substrate excluded), alongside earth, glass, plastics, wood and processing residues. The analyzer fits this framework as a means of proof: it is one of the ways the quality system shows that a lot matches the declared specification grade and that non-ferrous pieces have been removed. Checking the weight of non-metallic foreign materials remains a job for visual inspection and weighing, which XRF doesn't replace.

The test for putting an analyzer in your yard

The instrument pays for itself where mixed lots of stainless, copper alloys, nickel alloys or titanium come through every week, and where the price gap between grades is wide enough to cover measuring time. A yard that handles almost only shredded ferrous scrap, sold under grades whose composition and impurities are checked by lab sampling, would gain little from it. Likewise, if the classification that matters to your customers depends on carbon, the choice should shift towards optical emission or LIBS. The practical test comes down to one look at the bin. If it holds a metal that is worth more than the rest and can't be seen, measuring it pays; if all the material is worth the same, sorting adds no margin.

For the technical side of alloy libraries, measuring times and the operating flow, see the in-depth page on XRF analysis of metal scrap, while the criteria for choosing the instrument are in the guide on how to choose a portable XRF analyzer. If you are weighing up an instrument for your yard, the next step is a test on your own material: PITECH compares materials, volumes and working conditions, and prepares a tailored quote within 24 hours.