The rungs between a beaker and a plant

Research organisations produce results at a scale nobody can build from, and the ladder between that result and an operating circuit is both formally defined and reliably underestimated.

Reviewed by Peter Uppal

A small waist-high skid-mounted test rig on wheels standing on the floor directly beside a full-scale process vessel and its pipework. — Illustrative artwork: pilot and full-scale process equipment of the kind used in scaling up mineral processing. Not a facility, equipment or material connected to this project. · Illustration · Osmond Hub

The short version

Between a promising laboratory result and a working plant sit several defined stages, each of which exists because the previous one cannot answer the questions the next one raises. The European Commission's scale runs from TRL 1 – basic principles observed(opens in a new tab) to TRL 9 – actual system proven in operational environment(opens in a new tab). In minerals the intermediate rung has a name and a purpose: pilot plants 'are used to reduce the risk of scaling laboratory results to full scale production'(opens in a new tab). The gap is not a formality, and the historical record on crossing it is sobering.

A ladder somebody wrote down

The reason a research finding is not a technology is easier to see once the intervening stages are named. The Commission's definitions are terse and worth having in full at the middle rungs, where the interesting failures happen: TRL 4 – technology validated in lab(opens in a new tab); TRL 5 – technology validated in relevant environment (industrially relevant environment in the case of key enabling technologies)(opens in a new tab); TRL 6 – technology demonstrated in relevant environment (industrially relevant environment in the case of key enabling technologies)(opens in a new tab); TRL 7 – system prototype demonstration in operational environment(opens in a new tab).

The scale originates with NASA, whose version says the same thing in more physical language — component/subsystem validation in laboratory environment(opens in a new tab) at level four, and, at nine, actual system 'mission proven' through successful mission operations(opens in a new tab). The distance between 'validated in lab' and 'validated in a relevant environment' is one line of text and, frequently, several years.

Minerals has its own version of the same ladder

A process-development account written for the minerals industry describes the sequence concretely: batch laboratory tests to evaluate the major variables, then locked cycle testwork, then — where warranted — a pilot campaign. The source gives three reasons for that last step, not one: pilot plant tests 'are usually required to demonstrate a high risk process, to increase the confidence level on a marginal project or for the production of concentrate or other downstream products for other testwork'(opens in a new tab). Usually, and for any of those three: the hedge is the source's, and it matters, because piloting is expensive and is not done reflexively.

What a pilot buys is continuity. A bench test is a sequence of discrete experiments; a plant is a circulating system in which the tailings of one step become the feed of another and the reagents accumulate. Scale-up literature is explicit that these are issues difficult to investigate during bench scale tests(opens in a new tab), and that the mechanical and operating difficulties which arise are part of what is being tested.

The historical record, quoted carefully

The most quantified thing available on how often this goes wrong is old and comes from a different industry. A 1981 study of pioneer process plants — a sample its publisher describes as 44 plants built by 34 oil, chemical, and mineral companies — found that over 50 percent of the plants in the sample 'failed to achieve their production goals in the second six months after start-up'(opens in a new tab).

That figure is not about mineral processing, and it is not transferred to it here. It is cited because nothing equivalent for this industry was obtained, and because the direction it points is the one every practitioner account also points in. The honest summary is that a laboratory result is a weak predictor of plant performance, that this is well known, and that the size of the effect in minerals specifically is not something this article can source.

A bench test asks whether the chemistry works. A pilot asks whether it keeps working when its own output comes back around the loop.

Why some of this equipment is public

Piloting is where the funding question becomes structural. The facilities are expensive, intermittently used, and useless to own if you only need one campaign — which is precisely the profile of infrastructure that ends up publicly held.

A national geological survey describes its own as a unique service package combining mineral research laboratories with an industrial-scale pilot plant, a testing platform for the long-term behaviour of extractive waste and a tailing area(opens in a new tab), offering users a unique opportunity to test new procedures and processes(opens in a new tab). The European Commission generalises the category, defining technology infrastructures as facilities, equipment, capabilities and resources required to develop, test, upscale and validate technology — including test beds, pilot lines, pilot plants and demonstration facilities(opens in a new tab), whose stated purpose is to let firms derisk their R&D&I investments before market introduction(opens in a new tab).

That is the answer to the question of what a public research organisation is for at this stage: it owns the rung nobody can justify owning alone.

How the work is meant to travel

The routes out of a research organisation are also formally defined. The Commission distinguishes dissemination — the public disclosure of the results not only by scientific publications but via any pertinent medium(opens in a new tab) — from exploitation, defined as the use of results in developing, creating and marketing or improving a product or process, or in creating and providing a service in standardisation activities or shaping a policy(opens in a new tab). Publication and use are different obligations, and a finding can satisfy the first for years without touching the second.

Which is a useful frame for reading any early metallurgical result, including this publisher's own. Osmond's preliminary testwork, carried out at a commercial laboratory, produced a monazite concentrate of 19.4% TREO excluding yttrium at approximately 76% recovery and a zircon concentrate of 50.2% ZrO₂ at approximately 70% recovery(opens in a new tab) — work the company describes as preliminary, performed on a fraction of a bulk composite, on crushed bulk ore that was not optimised. On the ladder above, that is a low rung, honestly labelled. The distance from there to a plant is the subject of this article, and it is not short.

Exploration results and mineralogical estimates only. Orión has no JORC-compliant Mineral Resource or Reserve; maiden MRE and Scoping Study pending, targeted Q3 CY26.

Related