OK8 Global Third-Party Aluminium Dross Recyclers: Befesa, Altek & Commercial Models世界主流铝渣回收公司的技术特点与商业模式对比

Technical Features and Commercial Models of Leading Global Dross Recycling Enterprises | 世界主流铝渣回收公司的技术特点与商业模式对比

Industry Benchmarks · Published 2026-10-31 · Sources: audited annual reports, SEC filings and government award records where they exist; company-published descriptions are marked as such

Aluminium dross leaves a smelter as a residue and returns to it as a revenue line, but the route between those two points is frequently not owned by the smelter. In Europe, the Gulf, North America and increasingly India, a tier of specialist enterprises handles it — some as contractors operating inside the producer’s plant, some as centralised plants receiving material from many producers, some selling the equipment and licensing the process rather than processing anything at all.

This page sets out what distinguishes those enterprises technically, and how they are paid. It covers two corrections to statements that circulate widely, and it marks clearly where a figure comes from an audited filing and where it comes from the company that benefits from it.

1. The technical axis that matters: salt versus salt-free

Most comparisons of dross recyclers are organised by company. That is the wrong axis, because the same company may run both. The division that determines everything downstream is whether the route uses salt flux.

The distinction is not a quality ranking. A salt-free route avoids producing a hazardous residue; it does not automatically recover more metal, and the two claims are routinely conflated in secondary literature. Where a salt-free route is chosen, the practical question becomes what happens to the oxide fraction it produces in place of salt cake — a question with a commercial answer, not a technical one, because it depends on whether a customer for that oxide exists within transport distance.

1.1 Where each enterprise sits on that axis

This is a list of enterprises whose activity in dross we could trace to a named document. It is not a ranking, and absence from it is not a judgement. Several market-research summaries circulate longer lists including equipment manufacturers and producers with captive operations; we did not verify those entries individually and have not reproduced them.

2. Correction: who owns ALTEK

The claim that ALTEK is “an Epiroc company” is wrong, and the error matters because the correct owner is the same group whose audited accounts contain a USD 27.2 million charge for salt cake disposal.

ALTEK was acquired in May 2018 by Harsco Corporation for £45 million (approximately USD 60 million) on a debt-free, cash-free basis, plus contingent consideration tied to future performance. Harsco’s Form 10-K states the purchase directly and places ALTEK’s results inside the Harsco Environmental segment. ALTEK’s own website states that it is “part of the Enviri family,” and its published sales contact remains an @harsco.com address.

Epiroc has no connection to it. Epiroc is a Swedish mining-equipment manufacturer; its 2024 and 2025 annual reports list its acquisitions — Stanley Infrastructure, Weco, Yieldpoint, ASI Mining and others — and none is an aluminium dross business.

Practical consequence for a reader: any page naming the owner of ALTEK needs a date attached. Three valid answers exist depending on the year — Harsco, Enviri, New Enviri — and “Harsco Metals” is a fourth name referring to the same business under earlier branding.

3. Six commercial models

The trade press usually draws a single line — on-site service versus centralised plant. That line is real but incomplete. Six distinct models can be identified from named documents, and an enterprise may run more than one.

3.1 Why the choice is a transport question before it is a technology question

Salt slag and secondary dross are reactive. Aluminium nitride hydrolyses to ammonia; metallic aluminium hydrolyses to hydrogen; the residue is classified as hazardous in China (HW48, code 321-026-48 for secondary aluminium and fabricators), as an absolute entry in the UK/EWC lists, and as UN 3170 for transport. Every kilometre the material travels is a permitted movement of a hazardous residue.

• On-site processing does not remove the hazard — it removes the journey. The residue is treated where it is generated, so the hazardous-goods movement is avoided rather than managed.
• Centralised processing accepts the journey and buys scale: one 170 kt plant can carry equipment and environmental permits that no individual casthouse could justify.

This is visible in ALTEK’s own pitch for AluSalt. The stated saving is not mainly a higher recovery — it is the avoidance of shipping: the technology is described as removing the need to move “many hundreds of thousands of tonnes” of salt slag “across Europe to a small number of specialist recycling facilities.” The value being sold is the elimination of a hazardous-goods logistics chain.

Befesa makes the same point from the opposite geometry, naming proximity as the commercial mechanism: “close proximity to our clients is key in Befesa’s business model as it ensures reliable services and reduces transportation cost.”

4. Befesa: the audited numbers

Befesa S.A. is a Luxembourg-incorporated company listed in Frankfurt and included in the MDAX. Unlike most participants in this field, it publishes segment-level operating and financial data in an audited annual report, which makes it the one case here where the numbers can be checked rather than estimated.

Q1 2026 (latest published quarter): salt slags 101.9 kt at 86.7% utilisation, revenue €29.8m, EBITDA €7.7m, margin 25.7%. Secondary aluminium 39.0 kt at 76.1% utilisation, revenue €85.3m, EBITDA €1.8m, margin 2.1%, EBIT −€0.2m.

4.1 A capacity figure that appears twice, differently

Befesa’s own website states 450 kt of annually installed capacity to recycle salt slags and SPL. Its 2024 annual report and 2026 quarterly report both state 470 kt. The plant-level list reconciles the difference: Valladolid is listed at 150 kt in the 2022 annual report and on the website, and at 170 kt in the 2024 annual report (Lünen 170 + Hannover 130 + Valladolid 170 = 470).

This page uses 470 kt, the figure used in the audited filings. We note the other figure rather than silently picking one — a reader comparing a company’s website against its own annual report will find a 20 kt gap.

4.2 Geography, and a stated intention rather than a fact

Befesa describes its plants as distributed across Europe, Asia and the US, and identifies Europe, South Korea, Turkey and the US as markets where strict regulation is already in place. It names China as a market where “ongoing increase in regulation creates opportunity… to enter.” That is an intention stated by the company, not an operation we verified. It is recorded as a statement of strategy, not as a footprint.

5. What actually comes out of a salt slag plant

Befesa describes its salt slag route as a five-stage process: collection and transport of the hazardous waste; grinding and particle-size classification; reaction of the dangerous components and dissolution of the salts; separation, washing and production of inert products; and crystallisation of the salts.

The description commonly used in secondary sources — “complete recovery of NaCl/KCl and aluminium oxides” — overstates what the primary source says. Three specifics:

• The salt is recovered as a mixture, not separated. Befesa’s own process description states that crystallisation yields “a salt… a mixture of NaCl and KCl.” Separation into pure NaCl and pure KCl is not described.
• The oxide fraction is a product, not a recovered metal. It leaves as “Paval,” used in ceramics, refractories and rock wool. That is a legitimate outlet, and it is not the same thing as recovered aluminium.
• The crystallisation step consumes energy. The documented route evaporates the brine and condenses the vapours, with natural gas boilers named as the steam source. A plant whose input is a hazardous waste and whose output includes reusable salt is still running an evaporator.

Befesa reports 16 registered patents and €7.12 million of R&D expenditure over five years across its aluminium business, with a dedicated testing facility in Valladolid operating since 2012 and pilot plants for salt slag treatment, brine crystallisation and oxide refining. Its Erandio plant holds ASI Performance Standard V3 (2022) certification.

6. The finding worth reading: service margin versus metal margin

The hazardous-waste service earns roughly ten to twelve times the margin of the metal production it feeds — and in 2025 the metal side lost money at the operating line while the service side did not. The direction of travel reinforces it: salt slag utilisation has held near 90% (90.8% in 2024, 89.8% in 2025, 86.7% in Q1 2026), while secondary aluminium utilisation fell from 83.6% to 75.1% to 76.1% over the same points, with the company attributing the decline to compressed metal margins and weak European automotive demand.

What this means for a producer evaluating a contractor: a third-party recycler is not primarily a metal buyer. Its profit is in the treatment fee and the compliance service, not in the aluminium it hands back. The residue is the contractor’s raw material and the producer’s liability, and the party capturing the margin is the one willing to hold the permit.

7. ALTEK, “zero waste,” and the salt cake bill

ALTEK’s public positioning is explicit. Its managing director is quoted describing the company as “pushing the aluminium industry to a zero-waste scenario”; its global sales director states that with AluSalt “there is no need to ship waste or utilize a landfill.” Harsco’s own release calls AluSalt “a zero-waste solution.”

The same corporate group’s audited accounts record something else:

The Q1 2026 filing describes the company as “exploring other options to resolve this matter in a cost-effective manner, including engaging certain entities to remove the remaining salt cakes from Bahrain, subject to a suitability analysis.”

Read those two blocks together. A business marketing a technology that removes the need for landfill has, on its own books, a USD 27 million provision to dispose of accumulated salt cake in Bahrain — the site of its Gulf joint venture — unresolved across at least five reporting periods.

We are not alleging that the technology does not work. The filing does not say where the salt cake came from, which process produced it, or why it remains. Several readings are consistent with the evidence: legacy material predating the plant, a plant that did not reach design throughput, a contractual dispute, or a residue the technology does not address. None of them can be distinguished from the public filings.

What can be said is narrower and still useful: a “zero waste” claim and a USD 27 million disposal liability sat in the same group at the same time. For a producer evaluating a contractor, the question worth asking is not whether zero waste is achievable in principle, but which specific residues the contract places on which party’s books — and what happens to material the process does not convert.

8. Salt-free entrants: what is claimed and what is documented

Three enterprises market salt-free routes. Their published figures follow a pattern this site has documented before, and it is worth stating plainly because it determines how the numbers can be used.

8.1 TAHA International

TAHA describes a two-stage, salt-free, energy-conserving process operated from Bahrain, stating: over 90% recovery of the available aluminium, 97% available-aluminium recovery efficiency, 81% reduction of greenhouse gas emissions versus TRF, 50 kt of dross processed globally, 20+ years in service, and membership of the Aluminium Stewardship Initiative. It names three downstream products: steel slag conditioner, agricultural fertiliser, and aluminium trihydrate for flame retardants.

One comparative claim is attributed to a named third party — Dr. Oliver Zamal, CEO of KMF Industrial Technologies GmbH — describing a trial at a Gulf rolling mill in which dross was split into two equal portions, one processed by a TRF-type centrifuge and one by TAHA’s system, with TAHA reported as superior. That is a testimonial published by the beneficiary, not a measured trial report. No split-sample analysis, no metal content of the feed, and no recovery basis are given.

The denominator question applies directly. TAHA states “recovery of the available aluminium” — that is recovery of the metal contained in the dross, not of the furnace charge mass. The two are different quantities and the difference is large, as documented on this site’s technology comparison page. A figure stated against contained metal cannot be compared with a figure stated against charge mass. 97% is not contradicted here; it is placed.

8.2 PyroGenesis — DROSRITE

DROSRITE is documented on this site’s technology comparison page, where four published recovery figures for the same process — 98%, about 90%, 74%, and 18–40% — are shown to be reconcilable once the stated basis of each is identified. That page should be read before any DROSRITE figure is quoted. The commercial model is equipment sale preceded by paid demonstrations; the company has also described expansion into zinc dross recovery.

8.3 Runaya

Runaya describes “green aluminium recovery” from dross at an Odisha facility under a zero-waste, zero-discharge framing, and is reported as collaborating with TAHA. The published description is a positioning statement. No throughput, recovery rate or residue analysis was traced. It is listed here because the enterprise is named in multiple independent summaries of the field, not because its figures could be verified.

9. North America: a public-record case

One project in this field can be checked against a government award record rather than a company statement, which makes it useful as a calibration point for what a salt slag recycling facility is expected to achieve.

Project ZAAR (Zero Waste Advanced Aluminium Recycling), led by Real Alloy Recycling LLC, was selected by the US Department of Energy’s Office of Clean Energy Demonstrations under the Industrial Demonstrations Program. The award record (award ID DECD0000094, awarded 11 November 2024) shows:

Two things in that record deserve attention. First, 70% is a diversion target, not a measured recovery rate — it describes what the facility is meant to do, and the facility was not stated as operating at the time of the record. Second, the list of target products includes ammonium sulfate and reusable hydrogen, which are the products of the nitrogen and hydrogen management problem documented on this site’s hazardous-waste page. A salt slag facility in this model is, in part, a gas plant.

10. Captive operators: the model with no third party

Several producers run the circuit themselves, and their published positions are useful because they are not sales material.

• Raffmetal (Italy) has operated an in-house salt slag recovery plant since 1989, recovering salt, granulate and aluminium back into its own cycle and selling aluminium oxides as the product “RAPAL.” It states that since 1989 it has not sent melting residues to landfill. A continuous operation of that duration is the single most useful datum here, because it speaks to durability rather than to design.
• Norsk Hydro is described as using screw-extruder compaction to minimise dross formation before recovery is considered — an upstream lever — combined with hot dross processing and hydrometallurgical wet leaching to recover salt fluxes.
• Novelis is reported as following a salt-free melting strategy specifically to avoid generating hazardous salt slag.
• Stena Aluminium combines proprietary rapid cooling with mechanical and pyrometallurgical upcycling in-house.
• AMAG Casting (Ranshofen) remelts white and black dross under salt flux in on-site rotary and tilting drum furnaces.

A joint venture sits between captive and third-party: Alba–Daiki Sustainable Solutions (ADSS), announced at Gateway Gulf on 3 November 2024, with Alba holding 70% and Daiki Aluminium 30%, to establish an aluminium dross processing facility in Bahrain. Both partners targeted commencement of operations by September 2026. The structure is worth noting: a producer taking the majority stake in a dedicated processing facility, rather than contracting the work out.

11. The integrated package: REKS and the salt loop

REKS GmbH & Co. KG, a joint venture of K+S and REMEX, offers the secondary aluminium industry a package under the name REKAL: supply of smelting salts, recycling of the salt slag, extraction of aluminium granulate, and disposal of aluminium filter dust — one counterparty for the whole loop. The company states that it is the only company to offer this complete package.

The distinctive element is the disposal end. K+S operates underground disposal at Herfa-Neurode and Zielitz, where material is isolated in disused sections of potash mines, and underground recovery — using the structural properties of less-contaminated waste to secure mine areas at risk of subsidence — at Hattorf, Wintershall, Unterbreizbach, Bernburg and Zielitz.

Why this matters for the comparison: the salt loop does not close by itself. Whatever cannot be reused as flux has to go somewhere, and in this model the somewhere is a permitted underground repository. An enterprise that supplies the salt, takes back the salt slag and can dispose of what remains is selling something a pure technology vendor cannot: an exit. That is the same function Befesa performs for its customers, delivered through a different asset base.

12. Claims that cannot be audited

12.1 “More than 45 percent aluminium”

The published formulation is: “recovering more than 45 percent aluminium, depending on existing recovery techniques, than has traditionally been possible.” It cannot be evaluated as written, for two reasons. It is stated as an improvement, so 45% is a delta, not a recovery rate — and the baseline it is a delta from is not named. And the qualifier means the figure moves with the customer’s starting point, which is exactly the variable a buyer needs fixed. A number whose value depends on an unstated baseline is not a specification.

12.2 Installation counts on four different denominators

ALTEK’s installed base has been published as “over 550 units globally” (2017), “nearly 60 countries with over 400 installations” (2018), “over 500 global customers” (2017), and “active equipment in over 600 locations around the world” (2020). Units, installations, locations and customers are four different denominators, and the growth they appear to show is partly a change of what is being counted.

12.3 The size of the salt slag market

Harsco’s 2018 acquisition release sizes salt slag processing as “a nearly USD 2 billion potential market.” That is an acquirer’s estimate published to support a transaction. Separately, market-research summaries circulate values for a “salt slag recycling service market” with figures broken down by service type and application. We do not reproduce those. They are proprietary estimates whose methodology cannot be examined, and a figure that cannot be re-derived is not usable for comparison here.

13. What this page could not establish

• The per-tonne fee structure. Trade descriptions of build-own-operate and tolling contracts commonly state a processing charge per tonne handled. No published contract rate was traced. The models are described here qualitatively; the pricing is not quantified because the source does not exist in public filings.
• Recovery rates for any named contractor’s plant. No operator publishes a recovery figure with its denominator in an audited document. Where a recovery number appears in this field without stating “per tonne of what,” it is not comparable across sites.
• Which process produced the Bahrain salt cake. Not stated in the filings reviewed.
• Whether the May 2024 CD&R sale completed. A completion announcement was not traced. Treat the ownership question as open until a filing states it.
• Comparable segment data for ALTEK alone. Its results are reported inside the Harsco Environmental segment and not broken out.
• Verification of TAHA’s 97% and 81% figures, and of any throughput or recovery figure for Runaya. Both rest on company publication.
• Whether the Alba–Daiki facility commenced operations by its stated September 2026 target.

14. On sources, and what this page does not do

Three grades of source appear here, and they are distinguished throughout. Audited and public-record material — Befesa’s segment data, Enviri’s provisions, the DOE award record, the Alba–Daiki announcement — can be checked against the document named. Company-published descriptions of what each enterprise offers are identified as such: they are claims, not measurements, and section 12 explains why several cannot be checked. Unverified entrants are listed where an enterprise is named across multiple independent summaries of the field, with the absence of figures stated.

This page names companies and carries no links to them. Naming a listed company and citing its audited accounts is not a recommendation; the site recommends, ranks and endorses no supplier, and links to none. Where a reader needs to select a contractor, that decision requires site-specific data — residue composition, national classification, transport distance, permit status — that no published comparison can supply.

Three items in this page will move: the ownership of Harsco Environmental, the Bahrain salt cake reserve, and the commissioning status of the Alba–Daiki facility. All are dated here so that a reader can tell how current the statement is.