© SCF, 2021

1 Introduction

MFA is a young discipline, born in the 1980s, independently, in different laboratories across the world, and it solidified into a structured academic discipline in the 1990s and early 2000s. It appeared roughly at the same time as LCA, to which it is sometimes compared, something that we will do here also. Both methods share part of their methodology: they are based on mass balances applied in a carefully defined system and thus constitute an application of physics, thermodynamics, chemical or process engineering but also economics or, even more simply, accounting. As a matter of fact, one acceptation for MFA is Material Flow Accounting, while the other is Material Flow Analysis, two sister but separate disciplines.

This article was written in support of a presentation to a meeting of the SFI program in Trondheim in 2019 [1] in a session dedicated to MFA, where the author was asked to act as a discussant. Hence the oral style of the article and the focus on his own personal experience and expertise. This is an assumed impressionistic article, given as the author’s opinion and not a review paper, except in as far as the author’s personal publications are concerned.

The article will start from the author’s personal recollection of how, at Usinor¹, the team to which he belonged was confronted with specific practical problems and thus invented something that would be called MFA today, as it was initially unaware of what was being invented elsewhere. Then MFA and LCA will be compared, to point out similarities and differences in the way they picture and model the world, especially from the standpoint of environmental issues and how they are integrated in a business worldview about that topic. Comments will be given regarding the Society and Materials conferences, which were created as a consequence of the team’s exploration of both worlds and their frustration at their limitations. Then the paper will focus on the connection of MFA with various user communities, especially as far as environmental issues are concerned. In the conclusions, a list of strengths and weaknesses of MFA will be proposed, from this utilitarian, i.e. business and industry-oriented standpoint.

2 Usinor’s attempt at inventing MFA

In the late 1980s, Usinor was confronted with the need to change its raw material base from iron ore to scrap². The need was local in Lorraine and it was an answer to the iron and steel crisis (end of the 1970s, early 1980s) based on the late reckoning³ that growth in steel production had stopped for good and that productivity in steelmaking (hours/ton) was set to jump significantly. The way to achieve this was to replace blast furnaces and oxygen converters by Electric Arc Furnaces (EAF), switching raw materials from iron ore to scrap and thus cutting energy consumption by 2/3 and the workforce correspondingly. Usinor’s Steel Mill in Gandrange was set to be rebuilt in this manner, in a practical example of a Schumpeter destructive transformation process. French Steel was also restructuring into a single company at the time, to privatize the activity, after a long period of protection under state custody⁴ [2].

Top management needed a storyline to justify this major transformation and the question to be addressed was whether enough scrap would be available, then and in the future, to justify a thoroughly new EAF steel mill to replace the Rombas integrated mill. There was thus a need to estimate the scrap resource in terms of scrap annual generation and of its connection with the recycling rate of steel.

Usinor at the time was looking at Japan for models of change. Pioneering work had been carried out there to estimate both scrap generation and the amount of scrap potentially available for recycling (cf. Fig. 1) [3].

Applying the Japanese model to the French situation seemed to indicate that only a small part of the scrap potential (30%) was collected then and therefore that a new EAF steelshop could easily tap into this resource [4].

To flesh out these intuitions, Usinor launched a special project called the Cycle of Iron (CdF⁵), to bring research on board and develop original knowledge on the scrap process route to accompany the transformation of Gandrange⁶ (cf. Fig. 2) [5].

One of the components of the Cycle of Iron Project was devoted to the study of the scrap resource (“explore the scrap deposit”). It was headed by Usinor’s Chief Economist, Alain Moreau, whose task was to work out a dynamic model of scrap generation, based on the time-series of steel production. The model was a flow model, assuming that obsolete scrap (consumer scrap) would be generated at the end of life of goods, with a spread in their lifetime and provision for spare parts to help maintain the goods during their use time; scrap not collected at end of life was “lost”, i.e. dissipated in the environment or lost in landfills, either regulated or wild ones. The study was never published in details, except recently [6,7], but the final outcome showed the evolution of the collecting rate of scrap from 1950 until 1990 as given in Figure 3 [5]. In the late 1960s, the collecting rate was larger than 1: this showed that scrap was indeed coming from a reservoir which released material late compared to a model of end-of-life and immediate release. This suggested that a stock and flow model needed to be built rather than a simple flow model (cf. Fig. 4).

Moreover, the study showed that the collecting rates were much higher than the 30% intuited from the initial study [4]. There was enough scrap to feed the new EAF, as the many years of operation of the steelshop demonstrated in a practical way, although not as much as first imagined.

Figure 4 shows the model that was imagined at the time. It is not fully self-coherent (scrap stock on the left means incrementation of the scrap stock by the annual collection, while scrap stock on the right means the stock incremented over the long time) as it mixes static and dynamic elements. Moreover, it connects with a concept of the anthropogenic mine, an expression not yet used at the time. It also does not show the stock of steel in use, although the transformation from steel consumption to scrap potential, via a lifetime function, toggles between flux and stock, steel coming in and scrap going out − moreover, the transformation also changes the name of iron from steel to scrap. The CdF project stopped before this model was properly worked out.

This was the time when T. Graedel at the Yale University launched his STAFF project and the author was invited to come to the annual review meetings for the whole duration of the project − the author was one of the very few industry participants at the time. He became aware of MFA then and of its conceptual development and he surmised that the new methodology would eventually bring USINOR the answers that were needed.

At the end of this experience, there was no clear answer from MFA in terms of recycling, recycling rate and anthropogenic mines, however: when the question was raised, the answer was that the concept of recycling rate was not clear, as, indeed, many definitions could be proposed! The question was indeed felt by the MFA community as looking backwards to the past, while the field was more intent on developing foresight visions and looking at the future!

ArcelorMittal stopped doing research in that area, although it kept close to the academic work and to the expanding number of research teams active in MFA. Moreover, the demand was for the industrial researchers to get involved in LCA, a field that the team had also helped bring to life in the 1980s as it seemed to hold more immediate returns for an industry like steel⁸.

In particular, LCA was addressing many environmental issues, contrary to MFA: the approach was related to consumer or investment goods incorporating steel, thus offering the opportunity of fleshing out the industry’s green dimension organized around the products it puts on the market. Whether, with hindsight, this promise was fulfilled was another matter that ought to be explored in a different paper: the experience of the author and the evidence of the SAM conferences seem to suggest that it was only partially delivered and that LCA has been continuing on its momentum [15].

Fig. 1 Steel stock in Japan since the Meiji era (in 103 tons).

Fig. 2 Structure of the Iron Cycle project (1995–2002).

Fig. 3 Collecting rate of scrap (observed) as a function of historical time, expressed as a fraction of scrap potential actually collected as scrap7.

Fig. 4 Schematic stock and flow model developed as part of the CdF project.

3 LCA vs. MFA

The similarities and differences of LCA and MFA are fairly simple to describe from a conceptual and methodological standpoint, although the way they were and are still perceived by the economic sectors ought to probably be analyzed more carefully, which will be tentatively done here.

Conceptually, MFA and LCA look at time and space (geographic and economic space) from very different standpoints (cf. Fig. 5) – so that they can be compared to the dichotomy of history/geography (Tab. 1 [8]).

Speaking first of temporality [9], LCA is focused on time, the time of the life cycle (LC) of the functional unit, of which it is telling the story. On the other hand, MFA looks at what happens hic (et nunc) during a particular year. This is what literary criticism calls, respectively, diachronic and synchronic approaches [8].

Note that LC time is different from calendar time in as far as it does not actually flow in the “LCA universe” except for the functional unit and indeed everything else, in the background in particular but also in the foreground, is (usually) not time-dependent. MFA, on the other hand, is also interested in time, as its dynamic version deals with a pile-up of annual static MFAs and various years connect together by an exchange of material to represent the generation of end-of-life waste.

Paradoxically, a fully dynamic version of LCA still remains in limbo today [10,11]: LCA is foremost a STEM approach, which does not explore the conceptual subtleties that SSH would provide, for example, the multidimensional nature of time.

In terms of space, LCA deals with a system, usually a small one⁹, which the approach isolates and follows through time. One might use the analogy of fluid dynamics and describe LCA as a Lagrangian coordinate approach: this is indeed the theoretical framework in which the mass and energy balances that form the base of LCA may be conducted to match physical principles. This functional unit is described with a large set of input and output variables that describe its relationship with the environment, in terms of energy and raw materials footprints (upstream part of environmental issues) but also of emissions and waste generation (downstream part of environmental issues): LCA is time-consuming to implement because of its complexity, which is why its conclusions are often organized in terms of middle- or end-point indicators; its final interpretation may be difficult, as weighing for example Climate Change (GHG emissions) vs. public health (toxicology) is a complex issue. Last, LCA is a micro-vision of the world, a concept common in economy.

MFA, on the other hand, usually looks at a particular substance (e.g. phosphorus or lithium), or element (phosphorus, lithium or iron) or material (steel) and follows it through a variety of trajectories inside the anthroposphere, or, less often, inside anthroposphere, biosphere and geosphere. There are variants that look at energy (EFA).. and it would be nice to have one looking at exergy (E_xFA?). MFA is a macro-vision of the world.

LCA is usually understood as product-oriented and this resonates within businesses, which are focused on growing markets and market share. The environmental dimension is also important, in as far as LCA proposes a methodology to take the environment on board, where economics or traditional business practice would be at odds with integrating in their strategy what they consider as externalities. Many businesses have LCA teams which are looking for ways of using LCA beyond the marketing targets of the early adoption of LCA. Worldsteel has developed an LCI database for steel, mainly because it felt that “good data” was essential for the sector. The LCA community is huge, comprising academic and industrial teams as well as countless practitioners organized as consultants. Several high-visibility journals are published in connection with LCA and many general ones publish LCA studies. There are not many institutions bringing together the MFA community, like SETAC or EPLCA do for LCA [12], if one dismisses the ISIE as more general [13]. Moreover, LCA is standardized at ISO level, a rather unique case for a scientific discipline: didactic presentations of LCA usually refer to these ISO standards. LCA describes a life-cycle, therefore the value-chain that makes and uses the functional unit: it is thus fully embedded in the economic fabric. As mentioned before, LCA is more oriented towards the past than the future, but many researchers would disagree with this statement.

The MFA community belongs to the larger community of Industrial Ecology, which is why MFA is often described as the study of study of anthropospheric metabolism or socio-economic metabolism. The MFA community is smaller than the LCA community. Industry is less present there than academia. The same is true of policymakers, except in the field of Material Flow Accounting, which is slightly different from MFA_n (Analysis). There is no standardization of MFA_c, which allows a diversity of approaches but makes comparing different studies difficult. Contrary to LCA databases, MFA databases are few and organized at the level of individual publications [14]. The reference journal for MFA is the Journal of Industrial Ecology. The MFA literature does not deal comprehensively with the issues that attracted my applied research team to the field, i.e. estimates of recycling rates and of the features of the anthropogenic mine. MFAs have been widely developed for many rare and critical elements. Within the MFA field, a lot of attention is given to the upstream part of environmental issues. As mentioned before, MFA is more oriented to the future than to the past and therefore is an ideal tool for substantiating foresight studies, such as strategic explorations of broad issues like the future availability of critical materials, the future extent of the circular economy, strategies to fight or accommodate climate change, etc. On the other hand, the granularity of the environmental issues MFA tackles is coarser than LCA’s, at least for the time being.

Fig. 5 Temporalities in LCA and MFA: a comparison [8].

4 Society and materials conferences

In the early 2000s, we were convinced at ArcelorMittal that LCA and MFA should not develop independently in their own silos, which is why we created a series of conferences, which have been bringing together representatives of both communities, the Society and Materials (SAM), conferences [15,16]. The underlying assumption was that neither LCA nor MFA were sufficient to address the complexity of the non-economic and non-technical issues related to materials (and other commodities, like energy, electricity, water, etc. and functions like mobility/transport, shelter/construction/infrastructure, etc.): hence the introduction of the word “Society” in the conference title and of the concept of value of materials as its guiding light [17].

The conference was meant to create a forum for researchers to explore and experiment with new concepts in the nexus of materials and society and to give a chance to alternative methodologies, besides LCA and MFA, to emerge, what were called the new metrics. One clear conclusion now is that no other holistic methodology, capable of competing with LCA or MFA, emerged. This also means that both methods will endure and continue to adapt to new demands in the area.

Among the 471 papers presented during 12 SAM conferences, 15% were devoted to MFA.

Box 1 summarizes the features of MFA that they highlight.

The major innovation themes discussed in SAM meetings are given in Box 2.

A first conclusion point is the systematic exploration of the Mendeleev table achieved in terms of global flows (cf. Fig. 6): note that the coverage of the table was far from complete in 2011 (more work has been carried out since) and that dynamic MFAs were a minority. The second most important point is the use of stocks to derive foresight projections of future material use, mostly developed by Müller et al. The connection with the economics community through the input-output methodology (Leontiev matrices) is also significant, when the importance of establishing bridges between various scientific disciplines has been stressed in SAM meetings. Closer connections with MFA_c are also a strong trend, as MFA helps look at the environmental footprint of materials-related activities, such as the amount of waste generated by mining.

In conclusion, SAM conferences have been acting as methodological whistle blowers, pointing to new developments and new concepts before they became mainstream production in the high impact-factor literature. This has been true, in particular, for MFA progress.

Fig. 6 Elements for which global cycles were derived in 2011 [18].

5 MFA and various user communities

Who is active and who is interested in MFA?:

• first of all, academics working in the field. The initial investment in modeling tools, modeling “tours de main” and background databases is such that the teams specializing in MFA have a kind of economic rent, which constitutes a barrier to entry for newcomers in the field, especially since application software and databases are rare. This gives a flavor of activism to the field. And a tendency or a risk of “siloing”…

• industry, at the level of individual companies, is only weakly involved. Usinor/ArcelorMittal is an exception in the field of structural materials. Businesses focusing on rare, scarce or critical elements are keenly aware of MFA and have been using it to explore availability of resources and discuss the need for a stronger involvement in the circular economy (recovery and recycling), cf. for example UMICORE [19]. This a somewhat restrictive way of using MFA compared to the broad scope of concepts developed by academia and the involvement is usually one-shot, i.e. commissioning a study or hiring a PhD with that background but not developing an internal lab to keep the competence active and alive in-house;

• industry associations, particularly in the non-ferrous metals and aluminum sectors have been active in commissioning industry-wide MFA studies. Interestingly, the steel sector has kept out of the trend (EUROFER and worldsteel), except in Japan. Their goals were to create a narrative about the future, mainly for communication purposes vs. institutions like the EC, for example: long-term strategic plans are formulated elsewhere, inside companies. Other disciplines are in competition with MFA, especially foresight economics (P. Criqui, in France [20]) but also the large high-level consultants such as McKinsey, BCG, etc.;

• academic research is driven by the research teams themselves but also by the institutions that provide funding and develop program calls in which MFA could be fitted. Among international institutions the UN has been proactive (International Resource Panel [21] and Global Materials Flow Resource Panel) as well as the EC in its programs devoted to raw materials (EIT Raw Materials and EIP Raw Materials) plus the H2020 calls [22]), although MFA is not very explicitly mentioned, the focus being mostly on critical raw materials. The IEA, which has been switching attention from energy to raw materials resources, has not yet adopted MFA as a “pet” methodology − their culture being economics. Governments are also focusing more on MFA_c than on traditional MFA [23]. Note also the focus on waste, pioneered by the Vienna school of MFA (Professor Paul Brunner).

6 Strengths and weaknesses of MFA

This section summarizes the comments already expressed in the previous sections (Tab. 2). It is not a criticism of the field of MFA but rather a collection of suggestions for it to move forward and further.

7 Conclusions

Scientific communities based on methodologies that measure the environmental and social connections of economic activities (LCA, MFA) are well and alive.

MFA is more “confidential” than LCA, although the practical importance of LCA is probably overblown (LCA is oversold).

The economic world is not listening very well to MFA, because MFA pursues its own targets and does not provide some of the answers that industry would need (recycling rates), nor speak its language (products, price, economic value). However, this may also be because industry is often short-sighted when looking at the future or because the executives who do look at the future are not fully aware of MFA, nor are their high-level consultants.

I do not have any magic wand to remove this difficulty, apart from preaching, working, publishing, lecturing, on and on…

However, MFA is a good idea, it should prevail, eventually!

List of acronyms

ACV: Analyse de cycle de vie (French)

AFM: Analyse de flux de matière

BCG: Boston Consulting Group

CdF: Cycle du fer (French)

EAF: Electric Arc Furnace

EFA: Energy Flow Analysis

EFA: Element Flow Analysis

EIP: Raw materials The European Innovation Partnership (EIP) on raw materials

EIT: Raw Materials European Institute & Innovation and Technology on Raw Materials

EPLCA: European Platform on Life Cycle Assessment

EU: European Union

ExFA: Exergy Flow Analysis

GDP: Gross Domestic Product

GHG: Greenhouse gases

IEA: International Energy Agency

IoF: Intensity of Stock

IoF: Intensity of Flow

ISIE: International Society of Industrial ecology

ISO: International Standard Organization

LCA: Life-Cycle Analysis

LCC: Liffe-Cycle Costing

LiU: Life-in-Use

MFA: Materials Flow Analysis

MFAc: Materials Flow Accounting

MFC: Materials Flow Costing

RR: Recycling Rate

SAM: Society & Materials (conference)

SETAC: Society of Environmental Toxicology and Chemistry

SFA: Substance Flow Analysis

SFI: Sentre for Forskningsdrevet Innovasjon (Norwegian)

SOVAMAT: SOcial VAlue of MATerials

SSH: Social Sciences and Humanities

STEM: Science, Technology, Engineering & Mathematics

TMR: Total Materials Requirement

Acknowledgements

Thanks to Jean-Sébastien Thomas, who picked up the responsibility for Sustainability at AM Research after I retired, but who passed away since [24], and to Gaël Fick of IRT-M2P, who has been very active in developing an MFA program within IRT-M.

References

All Tables

All Figures

	Fig. 1 Steel stock in Japan since the Meiji era (in 103 tons).
In the text

	Fig. 2 Structure of the Iron Cycle project (1995–2002).
In the text

	Fig. 3 Collecting rate of scrap (observed) as a function of historical time, expressed as a fraction of scrap potential actually collected as scrap7.
In the text

	Fig. 4 Schematic stock and flow model developed as part of the CdF project.
In the text

	Fig. 5 Temporalities in LCA and MFA: a comparison [8].
In the text

	Fig. 6 Elements for which global cycles were derived in 2011 [18].
In the text