Midpoint and endpoint characterization factors for mineral resource dissipation: methods and application to 6000 data sets

Abstract

Purpose
The accessibility to most metals is crucial to modern societies. In order to move towards more sustainable use of metals, it is relevant to reduce losses along their anthropogenic cycle. To this end, quantifying dissipative flows of mineral resources and assessing their impacts in life cycle assessment (LCA) has been a challenge brought up by various stakeholders in the LCA community. We address this challenge with the extension of previously developed impact assessment methods and evaluating how these updated methods compare to widely used impact assessment methods for mineral resource use.

Methods
Building on previous works, we extend the coverage of the average dissipation rate (ADR) and lost potential service time (LPST) methods to 61 metals. Midpoint characterization factors are computed using dynamic material flow analysis results, and endpoint characterization factors, by applying the market price of metals as a proxy for their value. We apply these methods to metal resource flows from 6000 market data sets along with the abiotic depletion potential and ReCiPe 2016 methods to anticipate how the assessment of dissipation using the newly developed methods might compare to the latter two widely used ones.

Results and discussion
The updated midpoint methods enable distinguishing between 61 metals based on their global dissipation patterns once they have been extracted from the ground. The endpoint methods further allow differentiating between the value of metals based on their annual average market prices. Metals with a high price that dissipate quickly have the highest endpoint characterization factors. The application study shows that metals with the largest resource flows are expected to have the most impacts with the midpoint ADR and LPST methods, metals that are relatively more expensive have a greater relative contribution to the endpoint assessment.

Conclusion
The extended ADR and LPST methods provide new information on the global dissipation patterns of 61 metals and on the associated potentially lost value for humans. The methods are readily applicable to resource flows in current life cycle inventories. This new information may be complementary to that provided by other impact assessment methods addressing different impact pathways when used in LCA studies. Additional research is needed to improve the characterization of the value of metals for society and to extend the methods to more resources.

Full text

Vol:.(1234567890)
The International Journal of Life Cycle Assessment (2022) 27:1180–1198
https://doi.org/10.1007/s11367-022-02093-2
1 3
LCIA OFIMPACTS ONHUMAN HEALTH ANDECOSYSTEMS
Midpoint andendpoint characterization factors formineral resource
dissipation: methods andapplication to6000 data sets
AlexandreCharpentierPoncelet1 · PhilippeLoubet1· ChristophHelbig2,3· AntoineBeylot4· StéphanieMuller4·
JacquesVilleneuve4· BertrandLaratte5· AndreaThorenz2· AxelTuma2· GuidoSonnemann1
Received: 23 May 2022 / Accepted: 10 August 2022 / Published online: 8 September 2022
© The Author(s) 2022
Abstract
Purpose The accessibility to most metals is crucial to modern societies. In order to move towards more sustainable use of
metals, it is relevant to reduce losses along their anthropogenic cycle. To this end, quantifying dissipative flows of mineral
resources and assessing their impacts in life cycle assessment (LCA) has been a challenge brought up by various stakeholders
in the LCA community. We address this challenge with the extension of previously developed impact assessment methods
and evaluating how these updated methods compare to widely used impact assessment methods for mineral resource use.
Methods Building on previous works, we extend the coverage of the average dissipation rate (ADR) and lost potential
service time (LPST) methods to 61 metals. Midpoint characterization factors are computed using dynamic material flow
analysis results, and endpoint characterization factors, by applying the market price of metals as a proxy for their value.
We apply these methods to metal resource flows from 6000 market data sets along with the abiotic depletion potential and
ReCiPe 2016 methods to anticipate how the assessment of dissipation using the newly developed methods might compare
to the latter two widely used ones.
Results and discussion The updated midpoint methods enable distinguishing between 61 metals based on their global dissipa-
tion patterns once they have been extracted from the ground. The endpoint methods further allow differentiating between the
value of metals based on their annual average market prices. Metals with a high price that dissipate quickly have the highest
endpoint characterization factors. The application study shows that metals with the largest resource flows are expected to
have the most impacts with the midpoint ADR and LPST methods, metals that are relatively more expensive have a greater
relative contribution to the endpoint assessment.
Conclusion The extended ADR and LPST methods provide new information on the global dissipation patterns of 61 met-
als and on the associated potentially lost value for humans. The methods are readily applicable to resource flows in current
life cycle inventories. This new information may be complementary to that provided by other impact assessment methods
addressing different impact pathways when used in LCA studies. Additional research is needed to improve the characteriza-
tion of the value of metals for society and to extend the methods to more resources.
Keywords Dissipation· Losses· Metals· Mineral resources· Life cycle impact assessment· Circularity
Communicated by Matthias Finkbeiner
* Christoph Helbig
christoph.helbig@uni-bayreuth.de
Alexandre Charpentier Poncelet
alexandre.charpentier.poncelet@gmail.com
1 UMR 5255, Univ. Bordeaux, CNRS, INP, ISM,
33400Talence, Bordeaux, France
2 Resource Lab, University ofAugsburg, Universitaetsstr. 16,
86159Augsburg, Germany
3 Ecological Resource Technology, University ofBayreuth,
Universitaetsstr. 30, 95447Bayreuth, Germany
4 BRGM, 45060Orléans, France
5 Arts Et Métiers, University ofBordeaux, CNRS, INP, I2M,
33400Talence, Bordeaux, France
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

1181The International Journal of Life Cycle Assessment (2022) 27:1180–1198
1 3
1 Introduction
All metals in the periodic table can provide a socio-economic
benefit for modern societies (Graedel etal. 2013). Yet, min-
eral resources are non-renewable and accessible in constrained
quantities (Drielsma etal. 2016; Schulze etal. 2020), driving
the will to keep them in the economy for as long as possible
through, e.g., recycling or other circular economy strategies
(Blomsma and Tennant 2020; European Commission2020a;
Reuter etal. 2019; UNEP 2013). Dissipative flows of mineral
resources are flows that become inaccessible for future use
(Beylot etal. 2020b; Helbig etal. 2020). They can include
flows to the environment, to waste disposal facilities, and
flows to materials where their specific physicochemical char-
acteristics are no longer positively contributing to the material
characteristics (non-functional recycling) (Beylot etal. 2020b;
Charpentier Poncelet etal. 2021; Helbig etal. 2020).
In this context, it is proposed to develop methods to assess
the dissipation of mineral resources and its impacts in life
cycle assessment (LCA) (Berger etal. 2020; Beylot etal.
2020b, 2021; van Oers etal. 2020; Zampori and Sala 2017).
Charpentier Poncelet etal. (2019) proposed a framework
to consider the impacts of the dissipative flows of metals
on the area of protection (AoP) natural resources in LCA.
Based on that idea, we developed two dissipation-oriented
midpoint methods for life cycle impact assessment (LCIA)
applicable to extraction flows in the life cycle inventories
(LCI) and computed characterization factors (CF) for 18
metals (Charpentier Poncelet etal. 2021). These methods
relied on the results from the dynamic material flow analysis
(MFA) model of Helbig etal. (2020). The first method is the
lost potential service time (LPST), which quantifies the lost
opportunity to use metallic elements in the economy due to
dissipative flows over time horizons of 25, 100, or 500years.
The second one is the average dissipation rate (ADR), which
assesses the expected dissipation rates of metals from extrac-
tion until their complete dissipation.
The main objective of this article is to increase the
coverage of the LPST and ADR methods based on our
extension of the aforementioned dynamic MFA model to
61 metals (Charpentier Poncelet etal. 2022b). We also
explore the use of price-based endpoint CFs to represent
the loss of value associated with the dissipation of metals.
Complementarily, we investigate potential impact assess-
ment results using these new methods, and compare them
with widely used LCIA methods. To do so, we apply the
newly developed CFs to all non-empty market data sets
from the ecoinvent database version 3.7.1 (Moreno Ruiz
etal. 2020; Wernet etal. 2016) and compare the LCIA
results with those for the abiotic depletion potential (ADP)
ultimate reserves method (van Oers etal. 2019, 2002), the
surplus ore potential (SOP) method(Vieira etal. 2016)
and the surplus cost potential (SCP) method (Vieira etal.
2016). The latter two methods are included in ReCiPe
2016 (Huijbregts etal. 2017). Materials and methods are
detailed in “Sect.2,” results are presented and analyzed in
“Sect.3,” and conclusions are drawn in “Sect.4.”
2 Materials andmethods
2.1 Extending thecoverage oftheADR andLPST
methods
The MaTrace model initially developed by Nakamura etal.
(2014) allows quantifying losses of a cohort of extracted met-
als to the environment, other material flows (non-functional
recycling), and waste disposal facilities (including tailings
and slags) over time (Helbig etal. 2020). As in former studies
(Beylot etal. 2020a, 2021; Charpentier Poncelet etal. 2021;
Helbig etal. 2020), these losses are considered as dissipa-
tive flows in this study. With this model, the global anthro-
pogenic cycles of metals are studied one at a time, meaning
there are no direct links between the material flow models
of each metal. Contrastingly, Helbig etal. (2021) analyzed
seven major metals simultaneously with MaTrace-multi,
which adds another level of information and another level
of modeling complexity that would not be possible to handle
for 61 metals at this point. Therefore, despite the potential
limitations associated with the study of single cycles (e.g.,
mass-balance discrepancies between different metals used
in the same applications), we here focus on the individual
cycles of 61 metals as studied by Charpentier Poncelet etal.
(2022b). In the latter article, the losses of metals are evalu-
ated over time and their average lifetimes in the economy
are estimated based on the most recent data possible (most
typically between 2010 and 2020). The latter correspond to
the average duration over which metals remain in use in the
economy after extraction.
Midpoint CFs for 61 metals are derived from the results
of that article. The Python code and compiled datasets
underlying that study are accessible online (Helbig and
Charpentier Poncelet 2022). The overview of developments
proposed in this article based on our previous work is pre-
sented in Fig.1.
2.1.1 Midpoint characterization factors based ondynamic
MFA data
The midpoint CFs for the ADR and LPST methods are com-
puted using the approach described by Charpentier Poncelet
etal. (2021). The ADR method allows distinguishing
between the relative dissipation rates of metals after extrac-
tion in current conditions of consumption and recycling in
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

1182 The International Journal of Life Cycle Assessment (2022) 27:1180–1198
1 3
the economy. The LPST method measures the (relative) lost
opportunity to make use of metals over time once dissipated
based on these same conditions. The rationale for the LPST
method is summarized in the supporting information (SI).
The main equations for the ADR and LPST methods are
replicated below.
The ADR is calculated as the inverse of the average life-
time of metals in the economy, referred to as the total service
time (STTOT) in Eq. (1):
where the STTOT represents the total expected service time
of metal i in the economy after extraction and until its com-
plete dissipation (expressed in kg.yr/kg = yr). The ADR is
expressed in kg/kg.yr = yr−1.
The LPST is calculated as the difference between the opti-
mal service time, OST, defined as the total service time if
no dissipation occurred, and the expected service time, ST,
(1)
ADR
i
=1∕ST
TOT
i
.
given the expected dissipation pattern of metal i over a given
time horizon of 25, 100, or 500years, as shown in Eq. (2):
where the LPST, OST, and ST are expressed in kg.yr/kg (= yr).
The CFs of the ADR method are calculated as the ratio
between the ADR of metal i and that of iron (Fe), as shown
in Eq. (3):
where the
CFADR
for metal i is expressed in kg Fe-eq./kg.
There is no time horizon for the ADR method, as it inte-
grates the time function in its computation to provide a
yearly rate of dissipation (Charpentier Poncelet etal. 2021).
Similarly, those of the LPST method are calculated as
shown in Eq. (4):
(2)
LPST
i
,TH = OST
i
,TH − ST
i
,TH
(3)
CFADRi= ADR
i
∕ADRFe
(4)
CF
LPSTi,
TH = LPST
i,TH
∕LPST
Fe,TH
.
Loss rates&
lifetimeof 61
metals
CFLPST CFLPVAoP
Natural
resources
Endpoint/damage
Midpoint
Life cycleinventory
Previous work
(references presentedincaption)
FurtherLCIAmethod
developments (thisarticle)
Extension of
midpoint CFs to
61 metals
Dynamic MFA
resultsfor 61
metals
Method forcomputing
midpoint CFs&
computation of CFs for
18 metals
CFPVLR
Price-based
endpoint CFs
for61metals
CFADR
Resource
extraction from
ground
Dissipation overtime
afterextraction
Inaccessibility foruse
in technosphere
Lost potentialtomakeuse of
valueof inaccessible metals
Extraction
of metals
CF: characterizationfactor
ADR: averagedissipation rate
LPST: lostpotential servicetime
Legend
LPV: lost potentialvalue
PVLR:potential valuelossrate
Fig. 1 Overview of impact pathway and further development of the
ADR and LPST methods based on previous work (adapted from
Charpentier Poncelet et al. 2021). References to previous works:
method for computing midpoint CFs (Charpentier Poncelet et al.
2021); dynamic MFA results for 18 metals (Helbig etal. 2020); and
extended dynamic MFA results (Charpentier Poncelet etal. 2022b)
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

1183The International Journal of Life Cycle Assessment (2022) 27:1180–1198
1 3
where the
CFLPST
for metal i is expressed in kg Fe-eq./kg.
2.1.2 Endpoint characterization factors based onmetal
prices
As the final step of the impact pathway, we evaluate the
potential socio-economic impacts due to the dissipation of
different mineral resources (cf. Fig.1). The Joint Research
Centre (JRC) of the European Commission suggested that
the average prices of resources over a given period can be
used as a proxy to reflect the complex utility that resources
have for humans and are practical to do so since the data is
easily available (Beylot etal. 2020a). The underlying assump-
tion of using the prices of metals as an indication of their
socio-economic value is that prices reflect at least to some
extent the value of metals for society, albeit not perfectly
(see discussion in, e.g., Beylot etal. 2020a; Ecorys 2012;
Henckens etal. 2016; Huppertz etal. 2019; and Watson and
Eggert 2021). The most expensive metals are generally used
in specialized applications, because their high price does not
justify their use in low value-added applications in which less
specific or cheaper materials can be used.
For this study, we consider that recent price statistics are
most likely to be representative of the value of resources
answering the current demand for different applications. We
further assume that this value is maintained over time as
long as metals remain in the economy. This assumption is
likely to be more realistic for the short than the long term
for most metals, given that it is impossible to predict the
long-term trends in demand for given applications, nor
the development of new applications. Whenever possible,
the 10-year average prices from 2006 to 2015 are consid-
ered. This period is chosen because most price statistics
are obtained from US Geological Survey (USGS) statistics
(Kelly and Matos 2014) for which 2015 is the most recent
year for which data are available for most metals. Details
for specific metals are presented below. Price averages and
references are provided in the SI.
The price of barium is derived from statistics for barites
because barium is almost exclusively consumed in compound
forms (Johnson etal. 2017). We considered a barium content of
58.9% in barites (BaSO4) based on its stoichiometric content.
Other data are gathered to fill data gaps and cover the plati-
num group metals and rare earth elements that are not covered
separately in USGS statistics. All of the prices are adjusted to
$US1998 to match the reference unit of the USGS price data.
Prices for individual platinum group metals (PGMs) except
osmium are compiled from Johnson Matthey (2021). The price
of niobium is calculated from data provided by Metalary (2021)
considering a niobium content of 69.9% in niobium oxides
(Nb2O5). Prices for ten of the rare earth elements (REEs) are
compiled from data underlying a report of the French geological
survey (BRGM) (Bru etal. 2015, p. 151–158). These are yttrium
(Y), lanthanum (La), cerium (Ce), praseodymium (Pr), neodym-
ium (Nd), samarium (Sm), europium (Eu), gadolinium (Gd),
dysprosium (Dy), and terbium (Tb).
Different periods are considered to estimate the price of
other metals due to a lack of data. Given that iron ores are
directly refined into steel, the price of iron is derived from that
of steel. The 10-year period for the latter ranges from 2001 to
2010 because no data are available for subsequent years (Kelly
and Matos 2014). Osmium (one of the PGMs), for which no
price data is available, is estimated to remain at a constant price
of $US400 per ounce (Labbé and Dupuy 2014). The price
of scandium is estimated from the price of scandium oxides
between 2012 and 2018, which ranged between $US4600 and
$US5400 per kilogram (European Commission2020b). For
the latter, we consider an average yearly price of $US5000
per kg from 2012 to 2018 and a scandium content of 65.2%
in scandium oxides (Sc2O3). The price of thulium (Tm) is cal-
culated based on the price of its oxides in the third semester
of 2015 (Bru etal. 2015) and considering a stoichiometric
content of 87.4%. Similarly, the prices of the remaining REEs,
namely erbium (Er), ytterbium (Yb), holmium (Ho), and lute-
tium (Lu), are calculated from the average yearly prices of
their oxides between 2009 and 2014 as reported by Stormcrow
(2014) and considering their respective stoichiometric contents
in oxides (approximately 87.5%). It should be noted that the
prices of REEs underwent an important price peak in 2011
due to Chinese bans on exports in the early 2010s (Bru etal.
2015), which has a notable effect on the reported prices for
these metals. Such potential limitations associated with using
the market price data are discussed in “Sect.3.1.2.”
Endpoint CFs are computed by multiplying the midpoint
indicators with the average price of metals, allowing to com-
pare the relative potential value lost due to the dissipation of
different metals over time. This approach differs from the
JRC approach proposing to directly characterize dissipative
flows identified in the LCI with price-based CFs (Beylot etal.
2020a). Yet, metals retain value for humans for as long as they
are in use and quantifying the problem of inaccessibility neces-
sitates a time dimension (Dewulf etal. 2021) that is taken into
account with the proposed ADR and LPST methods.
The ADR and LPST values are used for the calculation
of endpoint CFs in order to keep the units in mass and mass.
years (rather than their respective normalized midpoint CFs
expressed in iron equivalents). The endpoint CFs for the ADR
method represent a potential value loss rate (PVLR) due to
average yearly dissipation rates of metals and are calculated
as shown in Eq. (5):
where
CFPVLRi
is measured in $US1998/kg.yr.
(5)
CFPVLRi= ADR
i
× pricei
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

1184 The International Journal of Life Cycle Assessment (2022) 27:1180–1198
1 3
Multiplying LPST with the price of metals indicates the
lost potential value (LPV) due to the inaccessibility of met-
als over time. It is assumed that the potential value of metals
remains the same over time, thus no discounting is applied.
Endpoint CFs for the LPV are calculated as shown in Eq.
(6):
where
CFLPVi
is measured in $US1998/kg.
The total PVLR (TPVLR) and the total LPV (TLPV) are
calculated analogously to their corresponding midpoint cat-
egory totals as defined by Charpentier Poncelet etal. (2021),
as shown in Eqs. (7) and (8):
where the
TPVLR
is measured in $US1998/yr and the
TLPV
is measured in $US1998.
2.1.3 Uncertainty
A Monte Carlo simulation of 1000 iterations allowed
computing 95% confidence intervals for the underlying
dynamic MFA results (Charpentier Poncelet etal. 2022b).
The uncertainty is also taken into account to compute the
midpoint CFs derived from these results. It should be noted
that the computed uncertainty reflects the uncertainty due
to variation of input variables, and not the model’s ability
to correctly reflect the real-life global system. Price uncer-
tainties are not accounted for given that a potentially large
and unquantifiable uncertainty can be associated with the
assumption of the representativeness of the annual average
market price information for the relative value of metals for
humans over time.
2.2 Application ofcharacterization factors to6000
life cycle inventory data sets
This objective of this application study is to investigate
the general trends that could arise from using the ADR
and LPST methods to evaluate the potential impacts due to
the dissipation of mineral resources, and to compare these
impact assessment results with those of widely used LCIA
methods characterizing flows of metal resources in order to
determine how they might differ. The assessment is realized
by applying LCIA methods to multiple data sets rather than
compare their CFs together. In order to do so, CFs from the
selected LCIA methods are applied to the metal resource
flows for all of the 5999 non-empty market LCI data sets
(6)
CFLPVi= LPST
i
× pricei
(7)
TPVLR =
∑
n
i=1
mi× CF
PVLRi
(8)
TLPV = ∑n
i=1
mi× CF
LPVi
of the ecoinvent database version 3.7.1 (Moreno Ruiz etal.
2020; Wernet etal. 2016), using allocation at point of sub-
stitution (APOS). Market data sets represent the average
consumption mixes for a given region and product (Wernet
etal. 2016).
We consider flows of metal resources included in the
ecoinvent 3.7.1 database that are covered by the ADR and
LPST methods. These include 45 flows, categorized as “met-
als, in ground” (e.g., aluminum, in ground) as identified in
TableS4 of the SI. In order to support the analysis of the
contribution of different sections of economic activity to
the inventory totals, the LCI data sets are subdivided by the
International Standard Industrial Classification of All Eco-
nomic Activities (ISIC) classification (United Nations2008).
Sections of economic activity regroup multiple sectors (e.g.,
section A includes the agriculture, forestry and fishing sec-
tors). This facilitates the study’s intelligibility and indicates
general trends expected from using the developed ADR and
LPST methods. The number of data sets included in each
economic section and the total mass of extracted metal flows
for each section are detailed in section S4 of the SI.
This approach is inspired by the study of Rørbech etal.
(2014). However, unlike the latter study, we only charac-
terize the impacts from metal resource flows and not other
mineral forms of metallic elements nor energy minerals. It
should be kept in mind that this study assesses the potential
impacts of data sets whose functional units are not necessar-
ily comparable (cf. “Sect.3.2.3”).
2.2.1 Selected LCIA methods
We compare the midpoint and endpoint CFs for the LPST
and ADR methods developed in this article with the latest
ADP ultimate reserves method for elements based on the
cumulative production in 2015 (van Oers etal. 2019, 2002)
and the midpoint and endpoint CFs for the “mineral resource
scarcity” category in the ReCiPe 2016 method (Huijbregts
etal. 2017). The latter method includes 18 impact pathways
covering three areas of protection. Its midpoint CFs for the
mineral resource scarcity category originate from the SOP
method (Vieira etal. 2017), and its endpoint CFs, from the
SCP method (Vieira etal. 2016).
The ADP ultimate reserves, SOP and SCP methods are
briefly described in sectionS3 of the SI. Midpoint CFs are
available for all methods, while no endpoint CFs are pro-
posed in the ADP method. All of the CFs are normalized to
kg Fe-eq./kg to facilitate the comparison of impact scores
between metals and methods. The normalization is done by
dividing all of the CFs by that of iron for the corresponding
method. It should be noted that endpoint CFs for metals
included in ReCiPe 2016 are equivalent to midpoint ones
when normalized to iron equivalents because the former are
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

1185The International Journal of Life Cycle Assessment (2022) 27:1180–1198
1 3
calculated from the latter using the same conversion factor
calculated for copper (Berger etal. 2020).
The ADP ultimate reserves method is currently recom-
mended for use in the product environmental footprint (PEF)
(Zampori and Pant 2019). The ADP ultimate reserves and
SOP methods are recommended by the life cycle initiative to
answer different questions linked with mineral resource use
(Berger etal. 2020). The former is recommended to answer
the question “How can I quantify the relative contribution
of a product system to the depletion of mineral resources?.”
The latter is interim recommended to answer the question
“How can I quantify the relative consequences of the con-
tribution of a product system to changing mineral resource
quality?.” The midpoint ADR method could address the
question “How can I quantify the relative contribution of a
product system to the dissipation of mineral resources?” and
the LPST method, “How can I quantify the relative contri-
bution of a product system to the inaccessibility of mineral
resources due to dissipation?.” Their endpoint versions could
answer the question “How can I quantify the relative contri-
bution of a product system to the potential mineral resource
value lost due to dissipation?.”
While we here focus on two widely used LCIA meth-
ods addressing mineral resource use, we refer readers to
the study of Rørbech etal. (2014) and the critical review
of the life cycle initiative’s task force on mineral resources
(Berger etal. 2020; Sonderegger etal. 2020). Rørbech etal.
(2014) applied the CFs from eleven LCIA methods to min-
eral resource flows (including energy minerals, minerals,
and metals) for all market data sets included in the ecoin-
vent version 3.0 database. The life cycle initiative’s taskforce
on mineral resources critically reviewed 27 LCIA methods
assessing the impacts of mineral resource use (Sonderegger
etal. 2020).
2.2.2 Coverage ofmineral resource flows byselected LCIA
methods
The ReCiPe 2016 method includes CFs for 75 resource flows
of mineral resources, 26 of which are mineral compounds
or ores. Parts of the latter 26 flows are no longer included in
the ecoinvent database because many mineral compounds
were converted to pure flows of metal content since ver-
sion 3.6 (Moreno Ruiz etal. 2019). The latest update for the
ADP ultimate reserves method for elements (van Oers etal.
2019) covers 76 elements. These include the 61 metallic ele-
ments that are also covered by the ADR and LPST methods,
in addition to calcium, cesium, potassium, sodium, and a
few non-metals (e.g., halogens). TableS3 in the SI presents
the 65 metallic elements included in the latest ADP method
for elements (van Oers etal. 2019) and indicates whether
they are included in the ReCiPe 2016 and the ADR/LPST
methods or not. TableS4 shows all of the CFs considered
for the selected methods considered in the application study,
including their value.
3 Results anddiscussion
3.1 Extended ADR andLPST methods
3.1.1 Midpoint andendpoint characterization factors
Table1 presents the ADR and STTOT for 61 metals as com-
puted by Charpentier Poncelet etal. (2022b) and their cor-
responding midpoint and endpoint CFs calculated for the
ADR and LPST methods. The 95% confidence intervals for
CFs are provided in TablesS1 and S2of the SI.
Higher CFADR and CFLPST indicate that a metal has a higher
average dissipation rate and thus a shorter lifetime in the econ-
omy (based on data for the recent past; see Charpentier Poncelet
etal. 2022b). Given the coverage of 61 metals, we may not pro-
vide extensive details for each of them. The underlying dynamic
MFA model and results as well as its supplementary materials
should be consulted for additional information on losses for each
metal (Charpentier Poncelet etal. 2022b). We here discuss a few
examples. It can be seen from the latter work that iron is second
best preserved in the economy after gold. It has an expected
lifetime of 154years thanks to relatively long-lived applica-
tions (e.g., infrastructure and mechanical equipment), a small
percentage of dissipation in use, and a combined yield of about
80% for the collection and recycling processes (Charpentier
Poncelet etal. 2022b). Its midpoint CFADR is 1.0kg Fe-eq./kg.
The corresponding CFADR of chromium (Cr), which is relatively
well conserved in the economy (e.g., as chromium and stainless
steels used in long-lived applications such as infrastructure and
transport) although less than iron, is 5.5kg Fe-eq./kg. In com-
parison, gallium is quickly dissipated at the production phase
(> 99%) for technical and economic reasons (Helbig etal. 2020;
Løvik etal. 2016, 2015), resulting in relatively high midpoint
CFADR, CFLPST25, and CFLPST100, with 1383kg Fe-eq./kg, 6.0kg
Fe-eq./kg, and 3.2kg Fe-eq./kg, respectively. At the same time,
the CFADR of indium, which is also relatively rapidly lost after
extraction due to overall low process yields and a lack of end-of-
life recycling, is 105kg Fe-eq./kg. The corresponding midpoint
CFs for, e.g., the LPST100 method, are 2.3kg Fe-eq./kg for
chromium and 3.1kg Fe-eq./kg for indium.
Below, we discuss general trends per categories of met-
als established by the UNEP (2011): ferrous, non-ferrous,
precious, and specialty metals. Figures2 and3 show mid-
point and endpoint CFs for the ADR and LPST100 methods,
respectively. Figs.S2 and S3 provided in the SI depict CFs
for the LPST25 and LPST500 methods.
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

1186 The International Journal of Life Cycle Assessment (2022) 27:1180–1198
1 3
Table 1 Total expected service times (STTOT) corresponding to average lifetimes in the economy calculated by Charpentier Poncelet etal. (2022b), average dissipation rate (ADR), computed as
loss rates by Charpentier Poncelet etal. (2022b), lost potential service time (LPST) with time horizons of 25, 100, and 500years, and the associated midpoint and endpoint characterization fac-
tors for the LPST and ADR methods
STTOT
(cf. ref. in
caption)
ADR
(cf. ref. in
caption)
LPST25 LPST100 LPST500 Midpoint CFs Endpoint CFs
CFADR CFLPST25 CFLPST100 CFLPST500 CFPVLR CFLPV25 CFLPV100 CFLPV500
Metal i kg.yr/kg kg/kg.yr kg.yr/kg kg.yr/kg kg.yr/kg kg Fe-eq./kg kg Fe-eq./kg kg Fe-eq./kg kg Fe-eq./kg $US1998/
kg.yr
$US1998/kg $US1998/kg $US1998/kg
03_Li 7.1E + 0 1.4E-1 1.9E + 1 9.3E + 1 4.9E + 2 2.2E + 1 4.5E + 0 3.0E + 0 1.4E + 0 4.2E-1 5.6E + 1 2.8E + 2 1.5E + 3
04_Be 2.6E + 1 3.8E-2 7.3E + 0 7.4E + 1 4.7E + 2 5.9E + 0 1.8E + 0 2.4E + 0 1.3E + 0 1.2E + 1 2.2E + 3 2.2E + 4 1.4E + 5
05_B 2.5E + 1 4.1E-2 1.1E + 1 7.5E + 1 4.8E + 2 6.3E + 0 2.6E + 0 2.4E + 0 1.3E + 0 2.4E-2 6.2E + 0 4.4E + 1 2.7E + 2
12_Mg 7.5E + 0 1.3E-1 2.1E + 1 9.2E + 1 4.9E + 2 2.0E + 1 5.0E + 0 2.9E + 0 1.4E + 0 4.8E-1 7.4E + 1 3.3E + 2 1.8E + 3
13_Al 7.6E + 1 1.3E-2 7.4E + 0 4.9E + 1 4.2E + 2 2.0E + 0 1.8E + 0 1.6E + 0 1.2E + 0 2.3E-2 1.3E + 1 8.4E + 1 7.3E + 2
14_Si 1.0E + 1 1.0E-1 1.6E + 1 9.0E + 1 4.9E + 2 1.5E + 1 3.9E + 0 2.9E + 0 1.4E + 0 1.7E-1 2.6E + 1 1.5E + 2 8.1E + 2
21_Sc 1.1E-2 8.7E + 1 2.5E + 1 1.0E + 2 5.0E + 2 1.3E + 4 6.0E + 0 3.2E + 0 1.4E + 0 4.6E + 5 1.3E + 5 5.3E + 5 2.6E + 6
22_Ti 1.4E + 1 7.4E-2 1.2E + 1 8.6E + 1 4.9E + 2 1.1E + 1 3.0E + 0 2.8E + 0 1.4E + 0 6.7E-1 1.1E + 2 7.8E + 2 4.4E + 3
23_V 8.0E + 0 1.3E-1 2.0E + 1 9.2E + 1 4.9E + 2 1.9E + 1 4.8E + 0 2.9E + 0 1.4E + 0 2.5E + 0 4.0E + 2 1.9E + 3 9.9E + 3
24_Cr 2.8E + 1 3.6E-2 1.1E + 1 7.3E + 1 4.7E + 2 5.5E + 0 2.8E + 0 2.3E + 0 1.3E + 0 6.2E-2 2.0E + 1 1.3E + 2 8.3E + 2
25_Mn 1.8E + 1 5.4E-2 1.6E + 1 8.2E + 1 4.8E + 2 8.3E + 0 3.8E + 0 2.6E + 0 1.4E + 0 5.7E-2 1.6E + 1 8.6E + 1 5.1E + 2
26_Fe 1.5E + 2 6.5E-3 4.1E + 0 3.1E + 1 3.5E + 2 1.0E + 0 1.0E + 0 1.0E + 0 1.0E + 0 3.7E-3 2.4E + 0 1.8E + 1 2.0E + 2
27_Co 4.1E + 0 2.5E-1 2.1E + 1 9.6E + 1 5.0E + 2 3.8E + 1 5.1E + 0 3.1E + 0 1.4E + 0 7.0E + 0 6.0E + 2 2.7E + 3 1.4E + 4
28_Ni 5.9E + 1 1.7E-2 8.5E + 0 5.7E + 1 4.4E + 2 2.6E + 0 2.0E + 0 1.8E + 0 1.2E + 0 2.6E-1 1.3E + 2 8.7E + 2 6.7E + 3
29_Cu 4.5E + 1 2.2E-2 6.5E + 0 5.9E + 1 4.5E + 2 3.4E + 0 1.6E + 0 1.9E + 0 1.3E + 0 1.2E-1 3.4E + 1 3.1E + 2 2.4E + 3
30_Zn 2.5E + 1 4.0E-2 1.1E + 1 7.6E + 1 4.8E + 2 6.2E + 0 2.7E + 0 2.4E + 0 1.3E + 0 7.1E-2 2.0E + 1 1.3E + 2 8.4E + 2
31_Ga 1.1E-1 9.0E + 0 2.5E + 1 1.0E + 2 5.0E + 2 1.4E + 3 6.0E + 0 3.2E + 0 1.4E + 0 3.3E + 3 9.2E + 3 3.7E + 4 1.8E + 5
32_Ge 4.7E-2 2.1E + 1 2.5E + 1 1.0E + 2 5.0E + 2 3.3E + 3 6.0E + 0 3.2E + 0 1.4E + 0 2.2E + 4 2.5E + 4 1.0E + 5 5.1E + 5
33_As 1.9E + 0 5.3E-1 2.4E + 1 9.8E + 1 5.0E + 2 8.2E + 1 5.7E + 0 3.1E + 0 1.4E + 0 2.5E-1 1.1E + 1 4.7E + 1 2.4E + 2
34_Se 5.3E-1 1.9E + 0 2.5E + 1 9.9E + 1 5.0E + 2 2.9E + 2 5.9E + 0 3.2E + 0 1.4E + 0 1.1E + 2 1.4E + 3 5.7E + 3 2.9E + 4
38_Sr 5.4E + 0 1.9E-1 2.0E + 1 9.5E + 1 4.9E + 2 2.9E + 1 4.9E + 0 3.0E + 0 1.4E + 0 1.3E-1 1.4E + 1 6.5E + 1 3.4E + 2
39_Y 5.7E + 0 1.7E-1 2.0E + 1 9.4E + 1 4.9E + 2 2.7E + 1 4.9E + 0 3.0E + 0 1.4E + 0 8.7E + 0 1.0E + 3 4.7E + 3 2.5E + 4
40_Zr 1.5E + 1 6.5E-2 1.3E + 1 8.5E + 1 4.8E + 2 1.0E + 1 3.2E + 0 2.7E + 0 1.4E + 0 5.9E-2 1.2E + 1 7.6E + 1 4.4E + 2
41_Nb 2.4E + 1 4.1E-2 1.1E + 1 7.6E + 1 4.8E + 2 6.4E + 0 2.7E + 0 2.4E + 0 1.3E + 0 1.7E + 0 4.5E + 2 3.1E + 3 2.0E + 4
42_Mo 2.4E + 1 4.1E-2 1.3E + 1 7.7E + 1 4.8E + 2 6.3E + 0 3.3E + 0 2.4E + 0 1.3E + 0 1.1E + 0 3.8E + 2 2.1E + 3 1.3E + 4
44_Ru 1.3E + 1 7.5E-2 1.5E + 1 8.7E + 1 4.9E + 2 1.2E + 1 3.5E + 0 2.8E + 0 1.4E + 0 2.4E + 2 4.7E + 4 2.8E + 5 1.6E + 6
45_Rh 2.4E + 1 4.2E-2 1.2E + 1 7.7E + 1 4.8E + 2 6.5E + 0 2.8E + 0 2.5E + 0 1.3E + 0 2.0E + 3 5.5E + 5 3.6E + 6 2.2E + 7
46_Pd 2.9E + 1 3.5E-2 1.0E + 1 7.3E + 1 4.7E + 2 5.4E + 0 2.5E + 0 2.3E + 0 1.3E + 0 3.7E + 2 1.1E + 5 7.8E + 5 5.1E + 6
47_Ag 4.4E + 1 2.3E-2 7.6E + 0 6.2E + 1 4.6E + 2 3.5E + 0 1.8E + 0 2.0E + 0 1.3E + 0 1.1E + 1 3.6E + 3 2.9E + 4 2.1E + 5
48_Cd 8.5E + 0 1.2E-1 1.7E + 1 9.2E + 1 4.9E + 2 1.8E + 1 4.2E + 0 2.9E + 0 1.4E + 0 3.0E-1 4.4E + 1 2.3E + 2 1.2E + 3
49_In 1.5E + 0 6.8E-1 2.4E + 1 9.9E + 1 5.0E + 2 1.1E + 2 5.7E + 0 3.1E + 0 1.4E + 0 3.0E + 2 1.0E + 4 4.3E + 4 2.2E + 5
50_Sn 1.2E + 1 8.4E-2 1.5E + 1 8.8E + 1 4.9E + 2 1.3E + 1 3.5E + 0 2.8E + 0 1.4E + 0 1.4E + 0 2.4E + 2 1.5E + 3 8.2E + 3
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

1187The International Journal of Life Cycle Assessment (2022) 27:1180–1198
1 3
Table 1 (continued)
STTOT
(cf. ref. in
caption)
ADR
(cf. ref. in
caption)
LPST25 LPST100 LPST500 Midpoint CFs Endpoint CFs
CFADR CFLPST25 CFLPST100 CFLPST500 CFPVLR CFLPV25 CFLPV100 CFLPV500
Metal i kg.yr/kg kg/kg.yr kg.yr/kg kg.yr/kg kg.yr/kg kg Fe-eq./kg kg Fe-eq./kg kg Fe-eq./kg kg Fe-eq./kg $US1998/
kg.yr
$US1998/kg $US1998/kg $US1998/kg
51_Sb 1.2E + 1 8.3E-2 1.5E + 1 8.8E + 1 4.9E + 2 1.3E + 1 3.6E + 0 2.8E + 0 1.4E + 0 5.1E-1 9.2E + 1 5.4E + 2 3.0E + 3
52_Te 6.3E-1 1.6E + 0 2.4E + 1 9.9E + 1 5.0E + 2 2.4E + 2 5.9E + 0 3.2E + 0 1.4E + 0 1.8E + 2 2.8E + 3 1.1E + 4 5.8E + 4
56_Ba 2.6E + 0 3.9E-1 2.2E + 1 9.7E + 1 5.0E + 2 6.0E + 1 5.4E + 0 3.1E + 0 1.4E + 0 4.8E-2 2.8E + 0 1.2E + 1 6.2E + 1
57_La 4.1E + 0 2.4E-1 2.1E + 1 9.6E + 1 5.0E + 2 3.8E + 1 5.1E + 0 3.1E + 0 1.4E + 0 4.5E + 0 3.9E + 2 1.7E + 3 9.0E + 3
58_Ce 5.6E + 0 1.8E-1 2.0E + 1 9.4E + 1 4.9E + 2 2.7E + 1 4.8E + 0 3.0E + 0 1.4E + 0 3.4E + 0 3.8E + 2 1.8E + 3 9.3E + 3
59_Pr 6.5E + 0 1.5E-1 1.9E + 1 9.4E + 1 4.9E + 2 2.4E + 1 4.6E + 0 3.0E + 0 1.4E + 0 1.1E + 1 1.3E + 3 6.5E + 3 3.4E + 4
60_Nd 7.8E + 0 1.3E-1 1.7E + 1 9.2E + 1 4.9E + 2 2.0E + 1 4.2E + 0 2.9E + 0 1.4E + 0 8.1E + 0 1.1E + 3 5.9E + 3 3.1E + 4
62_Sm 1.1E + 1 9.2E-2 1.5E + 1 8.9E + 1 4.9E + 2 1.4E + 1 3.5E + 0 2.8E + 0 1.4E + 0 3.0E + 0 4.8E + 2 2.9E + 3 1.6E + 4
63_Eu 2.6E + 0 3.9E-1 2.2E + 1 9.7E + 1 5.0E + 2 6.0E + 1 5.4E + 0 3.1E + 0 1.4E + 0 4.1E + 2 2.4E + 4 1.0E + 5 5.2E + 5
64_Gd 5.9E + 0 1.7E-1 1.9E + 1 9.4E + 1 4.9E + 2 2.6E + 1 4.6E + 0 3.0E + 0 1.4E + 0 1.1E + 1 1.2E + 3 6.0E + 3 3.1E + 4
65_Tb 5.1E + 0 2.0E-1 2.0E + 1 9.5E + 1 4.9E + 2 3.0E + 1 4.8E + 0 3.0E + 0 1.4E + 0 1.8E + 2 1.8E + 4 8.7E + 4 4.6E + 5
66_Dy 6.8E + 0 1.5E-1 1.8E + 1 9.3E + 1 4.9E + 2 2.3E + 1 4.4E + 0 3.0E + 0 1.4E + 0 6.4E + 1 8.0E + 3 4.1E + 4 2.1E + 5
67_Ho 1.4E + 1 7.4E-2 1.3E + 1 8.6E + 1 4.9E + 2 1.1E + 1 3.2E + 0 2.8E + 0 1.4E + 0 7.1E + 0 1.2E + 3 8.3E + 3 4.7E + 4
68_Er 1.4E + 1 7.0E-2 1.4E + 1 8.6E + 1 4.9E + 2 1.1E + 1 3.3E + 0 2.7E + 0 1.4E + 0 7.1E + 0 1.4E + 3 8.7E + 3 4.9E + 4
69_Tm 6.7E + 0 1.5E-1 1.9E + 1 9.3E + 1 4.9E + 2 2.3E + 1 4.5E + 0 3.0E + 0 1.4E + 0 3.2E + 1 4.0E + 3 2.0E + 4 1.1E + 5
70_Yb 6.6E + 0 1.5E-1 1.9E + 1 9.3E + 1 4.9E + 2 2.3E + 1 4.5E + 0 3.0E + 0 1.4E + 0 8.3E + 0 1.0E + 3 5.1E + 3 2.7E + 4
71_Lu 5.8E + 0 1.7E-1 2.0E + 1 9.4E + 1 4.9E + 2 2.6E + 1 4.8E + 0 3.0E + 0 1.4E + 0 1.4E + 2 1.6E + 4 7.6E + 4 4.0E + 5
72_Hf 7.2E-2 1.4E + 1 2.5E + 1 1.0E + 2 5.0E + 2 2.1E + 3 6.0E + 0 3.2E + 0 1.4E + 0 4.5E + 3 8.1E + 3 3.2E + 4 1.6E + 5
73_Ta 9.2E + 0 1.1E-1 1.7E + 1 9.1E + 1 4.9E + 2 1.7E + 1 4.0E + 0 2.9E + 0 1.4E + 0 1.6E + 1 2.4E + 3 1.3E + 4 7.1E + 4
74_W 5.7E + 0 1.7E-1 1.9E + 1 9.4E + 1 4.9E + 2 2.7E + 1 4.7E + 0 3.0E + 0 1.4E + 0 5.0E + 0 5.6E + 2 2.7E + 3 1.4E + 4
75_Re 1.0E + 1 9.8E-2 1.7E + 1 9.0E + 1 4.9E + 2 1.5E + 1 4.0E + 0 2.9E + 0 1.4E + 0 3.2E + 2 5.4E + 4 2.9E + 5 1.6E + 6
76_Os 3.6E + 0 2.8E-1 2.1E + 1 9.6E + 1 5.0E + 2 4.3E + 1 5.2E + 0 3.1E + 0 1.4E + 0 5.4E + 3 4.2E + 5 1.9E + 6 9.8E + 6
77_Ir 9.5E + 0 1.0E-1 1.7E + 1 9.0E + 1 4.9E + 2 1.6E + 1 4.1E + 0 2.9E + 0 1.4E + 0 1.3E + 3 2.1E + 5 1.1E + 6 6.0E + 6
78_Pt 4.1E + 1 2.4E-2 1.1E + 1 6.6E + 1 4.6E + 2 3.8E + 0 2.6E + 0 2.1E + 0 1.3E + 0 6.5E + 2 2.8E + 5 1.8E + 6 1.2E + 7
79_Au 1.9E + 2 5.2E-3 7.6E + 0 3.9E + 1 3.3E + 2 8.0E-1 1.8E + 0 1.2E + 0 9.4E-1 1.4E + 2 2.0E + 5 1.0E + 6 9.0E + 6
80_Hg 4.6E + 0 2.2E-1 2.1E + 1 9.5E + 1 5.0E + 2 3.3E + 1 5.0E + 0 3.0E + 0 1.4E + 0 5.8E + 0 5.5E + 2 2.5E + 3 1.3E + 4
81_Tl 7.0E + 0 1.4E-1 1.8E + 1 9.3E + 1 4.9E + 2 2.2E + 1 4.4E + 0 3.0E + 0 1.4E + 0 6.3E + 2 7.9E + 4 4.1E + 5 2.2E + 6
82_Pb 2.7E + 1 3.7E-2 9.7E + 0 7.4E + 1 4.7E + 2 5.6E + 0 2.3E + 0 2.3E + 0 1.3E + 0 6.3E-2 1.7E + 1 1.3E + 2 8.2E + 2
83_Bi 8.0E + 0 1.2E-1 1.7E + 1 9.2E + 1 4.9E + 2 1.9E + 1 4.1E + 0 2.9E + 0 1.4E + 0 2.0E + 0 2.7E + 2 1.5E + 3 7.8E + 3
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

1188 The International Journal of Life Cycle Assessment (2022) 27:1180–1198
1 3
The highest CFs for the midpoint ADR method are almost
entirely specialty metals because they are typically dissi-
pated the fastest. Endpoint CFs are dramatically different
for precious metals, whose price indexes are consistently
the highest. The latter are among the highest ranked end-
point CFs. Still, a few rapidly dissipating specialty metals
with a high annual average market price, i.e., scandium (Sc),
germanium (Ge), hafnium (Hf), and gallium (Ga), remain
among the highest CFs, i.e., the first, second, fourth, and
fifth CFs, respectively. Conversely, endpoint CFs for ferrous
and non-ferrous metals remain in the bottom half of CFs for
both the midpoint and the endpoint.
As for the ADR method, the highest CFs for the midpoint
LPST100 method are almost entirely those of specialty met-
als, and endpoint CFs are much higher for precious metals.
The most rapidly dissipating metals with the largest endpoint
CFs in the ADR method feature less prominently in the rank-
ing of CFs. The six highest CFs are precious metals, i.e.,
rhodium (Rh), osmium (Os), platinum (Pt), iridium (Ir),
gold (Au), and palladium (Pd). It shows that the price has
a greater influence on the ranking of endpoint CFs for the
LPST method than for the ADR method, because the mid-
point CFLPST are less differentiated than the midpoint CFADR.
3.1.2 Limitations fortheendpoint characterization factors
The limitations of the midpoint ADR and LPST methods
were highlighted in previous work (Charpentier Poncelet
etal. 2021). We here identify limitations linked with the
use of price statistics for the computation of endpoint CFs.
Firstly, the prices of metals include production costs influ-
enced by several factors such as the price of energy (Watson
10 310 210 1100101102103104105106107108
Fe
Al
B
Ba
Mn
Zr
Cr
Pb
Zn
Cu
Sr
Si
As
Ni
Cd
Li
Mg
Sb
Ti
Mo
Sn
Nb
Bi
V
Sm
Ce
La
W
Hg
Co
Er
Ho
Nd
Yb
Y
Ag
Pr
Gd
Be
Ta
Tm
Dy
Se
Lu
Au
Tb
Te
Ru
In
Re
Pd
Eu
Tl
Pt
Ir
Rh
Ga
Hf
Os
Ge
Sc
Midpoint CFs -ADR (kg Fe-eq./kg) Endpoint CFs -ADR ($US1998/kg.yr)
10
1
10
0
10
1
10
2
10
3
10
4
10
5
10
6
Au
Fe
Al
Ni
Cu
Ag
Pt
Pd
Cr
Pb
Be
Zn
B
Mo
Nb
Rh
Mn
Zr
Er
Ti
Ho
Ru
Sb
Sn
Sm
Re
Si
Ir
Ta
Cd
Bi
V
Nd
Mg
Li
Tl
Dy
Tm
Yb
Pr
Gd
Lu
W
Y
Ce
Sr
Tb
Hg
La
Co
Os
Ba
Eu
As
In
Te
Se
Ga
Hf
Ge
Sc
Ferrous metals Non-ferrous metals Specialty metalsPrecious metals
Fig. 2 Midpoint and endpoint characterization factors for the ADR
method. CFs are shown in ascending order and log scale to facilitate
comparison between methods. Black lines indicate the 95% confi-
dence intervals. Average values for all CFs are provided in Table1;
values for the 95% confidence intervals are provided in TablesS1 and
S2 of the SI
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

1189The International Journal of Life Cycle Assessment (2022) 27:1180–1198
1 3
and Eggert 2021) that may cause a diversion from the
assumed relationship between the price and the actual value
of metals for humans over time. We decided to maintain
this information in the price statistics because there is not
much precise information available on the share of different
production costs on the price of most metals (Huppertz etal.
2019), and because we assumed that the efforts put into pro-
duction also partly reflect the utility of metals for humans. It
should be noted that market prices tend to be more volatile
than the value of metals over time that we intend to quantify
using these data. Secondly, the large indirect investments
and production costs associated with by-product metals may
lead to both low production yields and market inefficien-
cies, potentially keeping their prices up (Watson and Eggert
2021). Consequently, some by-product metals like gallium
and scandium have relatively high midpoint CFs (production
losses are here accounted for as dissipation, leading to highly
dissipative profiles for these metals) and even greater end-
point CFs (due to their relatively high prices). Thirdly,
unquantified uncertainty may arise from using price data
from different sources, for metals of possibly different quali-
ties or purities, and in a few instances, over different time
series. Fourthly, some metals that are often used in a com-
pound form, such as the magnesium content of magnesia or
the lanthanum content of lanthanum oxides, may not have
the same potential value as their refined metal form. Nev-
ertheless, most metals are almost exclusively used as pure
metals or alloys (Graedel etal. 2022; UNEP 2011). Fifthly,
the price of a few metals are determined from the price of
compounds by considering the stoichiometric content of
the metals in the compounds, which may not be consist-
ent with how other price statistics are calculated. Finally,
Midpoint CFs - LPST100 (kg Fe-eq./kg) Endpoint CFs - LPST100 ($US1998/kg)
Ferrous metals Non-ferrous metals Specialty metalsPrecious metals
10 1100101
Fe
Au
Al
Ni
Cu
Ag
Pt
Pd
Cr
Pb
Be
B
Zn
Nb
Mo
Rh
Mn
Zr
Er
Ti
Ho
Ru
Sb
Sn
Sm
Re
Si
Ir
Ta
Cd
Bi
V
Nd
Mg
Li
Tl
Dy
Tm
Yb
Pr
Gd
Lu
W
Y
Ce
Sr
Tb
Hg
La
Co
Os
Ba
Eu
As
In
Te
Se
Ga
Hf
Ge
Sc
101102103104105106107
Ba
Fe
B
As
Sr
Zr
Al
Mn
Pb
Cr
Zn
Si
Cd
Li
Cu
Mg
Sb
Ti
Ni
Bi
Sn
La
Ce
V
Mo
Hg
W
Co
Sm
Nb
Y
Yb
Se
Nd
Gd
Pr
Ho
Er
Te
Ta
Tm
Be
Ag
Hf
Ga
Dy
In
Lu
Tb
Ge
Eu
Ru
Re
Tl
Sc
Pd
Au
Ir
Pt
Os
Rh
Fig. 3 Midpoint and endpoint characterization factors for the
LPST100 method. CFs are shown in ascending order and log scale
to facilitate comparison between methods. Black lines indicate the
95% confidence intervals. Average values for all CFs are provided
in Table1; values for the 95% confidence intervals are provided in
TablesS1 and S2 of the SI
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

1190 The International Journal of Life Cycle Assessment (2022) 27:1180–1198
1 3
the volatility of market prices may have a significant effect
on the average price of metals. For instance, the price of
different REEs increased by a factor of 3 to 27 during the
price peak of 2011 (Bru etal. 2015). For other metals, price
variations typically ranged between ± 50 and 150% of the
average price over the 2006–2015 period. Thus, using the
average of yearly prices over a decade reduces the short-term
variations associated with volatility and may provide more
stable indications of the value of metals. Additional research
is needed to address these challenging limitations.
3.2 Application study
3.2.1 Midpoint impact assessment persection ofeconomic
activity
Figure4 shows the relative contribution of metals to inven-
tory results split in eight sections of economic activity and
relative midpoint impacts for selected LCIA methods. Met-
als contributing to over 10% of the total impacts for at least
one method are shown individually on the figure for all
columns of the corresponding economic section; others are
grouped altogether.
Iron largely dominates resource extraction in the inven-
tories among all sections, with 76% (section C: manufac-
turing) to 98% (section B: mining and quarrying) of their
respective inventory shares. Despite its CF being one of the
two smallest for the ADR and LPST methods, iron consist-
ently comes out as one of the main contributors for the ADR
method, and is increasingly important for the LPST25, 100
and 500 methods. It should be noted that, given the overall
great recyclability of iron (or steel) in most of its applica-
tions (Pauliuk etal. 2017), its relative impacts would likely
be much lower for LCI that account for the recycling in the
end-of-life modeling beyond what is already accounted for
in LCI databases (the same could be expected for other well-
recycled metals). In this case, the extraction of metal would
be allocated between different product systems (or between
different life cycles) and therefore would be lower than a
product system using only primary metal. As for other met-
als, iron’s relative share of impacts is increasingly similar
to its share of the inventory flows when considering longer
time horizons for the assessment using the LPST method,
because the midpoint CFLPST25 are more differentiated than
the CFLPST100 and the CFLPST500 (i.e., they spread from 1.0
to 6.0, 1.0 to 3.2, and 1.0 to 1.4kg Fe-eq./kg, respectively).
Aside from iron, the only metals showing up among the
highest relative contributions to the midpoint ADR and
LPST impact assessments are other widely extracted metals
typically representing around 1 to 5% of inventory shares for
different economic sections. For instance, barium is more
often than not part of the main contributors for the total
dissipation impacts as assessed with the ADR and LPST
methods. Indeed, it is mostly used for gas and oil well drill-
ing under its mineral form of barites, explaining both its
extensive use in multiple economic sections and its highly
dissipative profile that is best distinguished with the mid-
point ADR method. Zinc also importantly contributes to the
impact scores for the ADR and LPST methods for section
C (manufacturing) and the other sections. As an indication,
barium and zinc represent 0.7% and 4.5% of the inven-
tory flows by weight for section C, while they respectively
account for 18% and 12% of the total impacts for the mid-
point ADR method. Nickel is widely used in section F (con-
struction) with 3.1% of the inventory total, and comes up as
the third highest contribution to the ADR and LPST assess-
ments for that section with approximately 3 to 5% of their
total impacts. Similarly, aluminum flows represent 4.7% of
the inventory totals for section C (just over zinc), but its
midpoint CFADR and CFLPST100 are 67% and 35% lower than
the corresponding CFs for zinc, explaining its lower share
of the impacts for that section (i.e., 4.0% for ADR and 5.8%
for LPST100). For the same reason as aluminum and nickel,
chromium and manganese do not contribute over 10% of the
impacts for the midpoint ADR and LPST methods for any
section of economic activity despite being widely extracted.
Similarly to the ADR and LPST methods, widely
extracted copper, iron, and nickel recurrently show up as
important contributors to the impact assessment with the
SOP method included in ReCiPe2016. For that method, iron
contributes over 10% of the total impacts for all economic
sections except section C (manufacturing), i.e., from 36% of
the total impacts for section H (transportation and storage)
to 87% for section B (mining and quarrying). Meanwhile,
nickel generally contributes to a larger share of the impacts
for the SOP method than for the ADR and LPST methods,
with, e.g., 45% of the relative impacts of section F (con-
struction) and 11% of those of section H (transportation and
storage). Finally, copper represents 12% of the total impacts
for the SOP method for section A (agriculture, forestry and
fishing), 14% for section D (electricity, gas, steam and air
conditioning supply), 10% for section E (water supply; sew-
erage, waste management and remediation activities), and
9.0% for section F (construction).
In contrast to the other methods, the relative impacts of
iron are negligible for ADP ultimate reserves because its
CF is among the lowest for that method, while those of the
scarcest metals are several orders of magnitude higher (van
Oers etal. 2019). For example, the CF of gold in the latter
method is 2 billion times higher than iron (cf. TableS4).
Copper, which has a lower crustal concentration (28ppm)
than other widely extracted metals (e.g., 63ppm for bar-
ium, and 774ppm for manganese: van Oers etal. 2019), is
the only widely extracted metal with over 10% of the total
impacts for the ADP method for at least one section of eco-
nomic activity. It contributes most to the impact assessment
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

1191The International Journal of Life Cycle Assessment (2022) 27:1180–1198
1 3
for sections A (18% of the impact total), B (15%), D (16%),
E (12%), and F (16%), while its shares of inventory totals
range from 0.94 to 1.5% across these economic sections.
Aside from copper, only scarce metals with very low
crustal concentrations recurrently come up as important
contributors for the ADP ultimate reserves method. These
Section B: Mining and quarryingSection A: Agriculture, forestry and fishing
Iron
Copper
Gold
Barium
Zinc
Te llurium
Platinum
Palladium
Nickel
Silver
Magnesium
Section C: Manufacturing Electricity, gas, steam & air
conditioning supply
Water supply; sewerage, waste
management & remediation activities Section F: Construction
Section H: Tr ansportation and storage Other sections
Inventory
ADP 2015
ReCi
Pe 2016
LPST25
LPST100
LPST500
ADR
Metals
Others
0%
20%
40%
60%
80%
100%
0%
20%
40%
60%
80%
100%
0%
20%
40%
60%
80%
100%
0%
20%
40%
60%
80%
100%
0%
20%
40%
60%
80%
100%
0%
20%
40%
60%
80%
100%
0%
20%
40%
60%
80%
100%
0%
20%
40%
60%
80%
100%
Inventory
ADP 2015
ReCiPe 2016
LPST25
LPST100
LPST500
ADR
Section D:
Section E:
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

1192 The International Journal of Life Cycle Assessment (2022) 27:1180–1198
1 3
are gold, palladium, platinum, silver, and tellurium. Pre-
cious metals are also revealed to be important contributors
to the ReCiPe 2016’s impact assessments, but only for sec-
tions with the highest shares of these metals in the inventory
totals, i.e., gold, palladium, and platinum in sections C and
H, and silver in other sections. Indeed, although the differ-
ence is less marked than for ADP, the CFs for precious met-
als are also among the highest in the SOP method because
they require a lot of additional ores to produce. For instance,
the CFs of gold and platinum are 5 and 6 orders of magni-
tude greater than iron’s and rank thirtieth and thirty-third
highest out of 33 CFs of the ReCiPe 2016 method consid-
ered in this study, respectively (cf. TableS4). In contrast to
the ADP and SOP methods, precious metals do not appear
among highest contributors to total impact scores for the
midpoint ADR and LPST methods because their midpoint
CFs consistently rank among the smallest, as observable in
Figs.2 and3.
3.2.2 Midpoint versusendpoint impact assessment
We here investigate differences between the midpoint and
the endpoint impacts assessed with the selected LCIA meth-
ods, as depicted in Fig.5. Especially, we observe the effect
of considering the price of metals in the endpoint assess-
ment. For brevity, we here focus on general observations,
and the comparison is elaborated with quantitative examples
in section S5 of the SI. ReCiPe 2016’s relative impacts are
the same for the midpoint and endpoint assessment for the
reason exposed in “Sect.2.2.1.” Contrastingly, there is an
important shift in the relative shares of impacts between
midpoint and endpoint for the ADR and LPST methods,
because the price information allows further differentia-
tion between metals. The impacts due to expensive metals
increase dramatically, while those of cheaper ones diminish.
Because of the way the CFADR are computed, the effect is
most dramatic for metals that dissipate relatively quickly and
that have relatively high annual average market prices (such
as gallium) putting forward potentially substantial differ-
ences between the endpoint impact assessment results for the
ADR or the LPST methods. Since the prices have a stronger
effect on the endpoint CFs for the LPST method, the increase
is more dramatic for the most expensive metals (e.g., gold)
in the LPST method than the ADR method. Cheaper met-
als such as barium and iron see their shares of the impacts
diminish importantly in the endpoint assessment.
This convergence of results for the endpoint LPST100
and ReCiPe 2016 methods can be explained by the simi-
larities between the surplus amount of ores assumed to be
required to produce scarcer metals in the future, its assumed
cost, and the higher economic value of these same metals.
Indeed, production costs represent approximately 50–75%
of the market price of metals (Huppertz etal. 2019). The
price strongly influences the computation of endpoint CFs
for the LPST methods, especially over longer time horizons
(cf. Figure3 for the LPST100 method, and Figs.S2 and
S3 for the LPST25 and LPST500 methods, respectively).
As shown in Fig. S5 of the SI, these endpoint methods are
similarly sensitive to the relative shares of inventory totals
for iron, copper, gold, and nickel. The dissipation of scarcer
metals is costlier for society because they require relatively
more ore to produce (as accounted for in the midpoint ReC-
iPe 2016 method, i.e., SOP), and the endpoint ReCiPe 2016
method, i.e., SCP, assumes they are also more costly to pro-
duce. The cost of production is also reflected in the price
information underlying the endpoint CFs for the ADR and
LPST methods.
3.2.3 Comparison oftheimpacts ofdata sets
acrossmethods
Figure6 shows the total impact scores for the selected LCIA
methods applied to the forty-five studied metals across all
market processes from the ecoinvent 3.1.7 database. Both
X- and Y-axes are in log scale. Please note that the scale of
the X-axes may vary between graphs. R2 values represent the
correlation of the log–log regression of impact scores per
data set. Lower R2 values indicate that the compared meth-
ods present different impact hotspots among the covered
metals. Three additional scatter plots are presented in the SI.
The impact scores range over 18 orders of magnitude
across the data sets. The relatively high correlations between
impact scores can partially be explained by the differences
in the relative size of functional units (i.e., the variations
in their total resource extraction flows). Indeed, the sum of
metal resource flows spreads over nine orders of magnitude
across data sets. Normalizing or rescaling functional units
could reduce such variations, going beyond what could be
achieved in this article. For instance, the results of Rørbech
etal. (2014) show that correlations between impact scores
per data set could be expected to be much smaller after res-
caling functional units to a more comparable scale.
The impact scores for the midpoint ADR and LPST100
methods, as shown in Fig.6c, are rather well correlated
(R2 = 0.991) given that they both rely on the service time
Fig. 4 Contribution of metals to inventory totals and impact scores
for four midpoint LCIA methods. Graphs present results per section
of economic activity established in the ISIC. ADP 2015: ADP ulti-
mate reserves method for elements based on the cumulative produc-
tion in 2015 (van Oers etal. 2019); ReCiPe 2016: results for the SOP
method (Vieira etal. 2017) included in the latter. Metals contribut-
ing over 10% of the total impacts for at least one impact method are
shown individually for all columns of the corresponding section of
economic activity; others are grouped altogether. The inventory col-
umn presents the relative mass of resource flows in the compiled LCI
data sets
◂
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

1193The International Journal of Life Cycle Assessment (2022) 27:1180–1198
1 3
(albeit with a different conceptual approach; cf. Equa-
tions1 and2) and on the same underlying data. Thus, their
CFs rank almost the same across metals. The comparison
between impact scores for the midpoint ADR method and
those of the ADP ultimate reserves method have the lowest
correlation (Fig.6d; R2 = 0.927), followed by those of mid-
point LPST100 versus ADP methods (Fig.6a; R2 = 0.937).
Data sets responsible for most impacts for the ADP 2015
are expected to be quite different from those for the ADR
and LPST methods. The former are influenced almost
exclusively by data sets with larger resource flows of the
scarcest elements in the crust (e.g., precious metals and
tellurium), whereas the latter are influenced not only by
the dissipation patterns of different metals but also by the
mass of their resource flows (cf. “Sect.3.2.1”). The impact
scores for the latter midpoint methods thus result mostly
from iron flows because they are the largest resource flows,
and from other widely extracted metals with higher CFs
than others (e.g., barium and zinc).
The impact scores of the LPST100 midpoint method are
better correlated with ReCiPe 2016 (Fig.6b; R2 = 0.985)
than those of ADP 2015 (Fig.6a; R2 = 0.937). Logically,
ReCiPe 2016’s impact scores have a lower correlation
with those of ADP 2015 (R2 = 0.967; cf. Fig.S6 of the
SI) than with the other two methods. This can partially
be explained by the fact that the CFs of the ReCiPe 2016
method present a stronger convergence with those of the
LPST100 method than the ADP2015 method (cf. TableS4
in the SI). Another reason for the stronger convergence
between ReCiPe 2016 and the LPST100 results is that
their CFs are much less differentiated than those of ADP
2015. Therefore, their impact scores are more deter-
mined by the relative size of metal flows in the LCI in
comparison to the ADP 2015 method. Contrastingly, the
pronounced differentiation between CFs for the ADP 2015
method (cf. TableS4 in the SI) leads to acute hotspots for
a few very scarce metals, despite their very small resource
flows in the inventory.
The correlation between endpoint ADR and LPST100
results is similar to their midpoint assessments, as shown
by comparing Fig.6c (R2 = 0.991) and Fig.6e (R2 = 0.991).
Even though the price of metals have a greater influence
on the CFLPST100 than the CFADR, their endpoint CFs have
similar rankings. The few metals with much larger CFs
in the ADR method are metals with the highest dissipa-
tion rates and high prices. Relatively small resource flows
of such metals in the inventories translate into relatively
large shares of the total impacts, the most obvious exam-
ple being gallium (as observable in Fig.5). It should be
noted that other similar metals with high CFs do not have
resource flows in the ecoinvent 3.7.1 database (e.g., ger-
manium and scandium). Thus, data sets are expected to
present noticeably greater impact scores with the endpoint
ADR method than the endpoint LPST100 method due to
higher-than-average resource flows of highly dissipative,
relatively expensive metals (e.g., gallium and germanium;
cf. Figure2). For instance, gallium contributes an aver-
age of 0.016% to the inventory totals for the 100 data
sets with the largest difference between the total impact
scores for the endpoint ADR and LPST100 methods. In
contrast, gallium’s average share of inventory totals is of
no endpoint CFs
To tal midpoint impacts
0%
20%
40%
60%
80%
100%
Inventory
ADP 2015
ReCiPe 2106
LPST25
LPST100
LPST500
ADR
To tal endpoint impacts
Inventory
ADP 2015
ReCi
Pe 2106
LPST25
LPST100
LPST500
ADR
0%
20%
40%
60%
80%
100%
Te llurium
Platinum
Palladium
Nickel
Manganese
Copper
Chromium
Aluminium
Gallium
Selenium
Gold
Barium
Magnesium
Metals
Others
Iron
Fig. 5 Inventory contributions and total midpoint and endpoint
impact scores for selected LCIA methods applied to 5999 market
data sets from the ecoinvent 3.7.1 database. ADP 2015: ADP ultimate
reserves method for elements based on the cumulative production in
2015 (van Oers etal. 2019); ReCiPe 2016 (midpoint): results for the
SOP method (Vieira etal. 2017); ReCiPe 2016 (endpoint): results for
the SCP method (Vieira etal. 2016). Metals representing the top five
contributions to the total inventory flows and for each LCIA method
are shown in their respective columns; others are grouped altogether
Fig. 6 Comparison of impact assessment for six selected pairs of
LCIA methods, covering 5,999 market data sets organized by section
of economic activity defined in the ISIC. a Midpoint LPST100 vs.
ADP 2015; b midpoint LPST100 vs. ReCiPe 2016; c midpoint ADR
vs. LPST100; d midpoint ADR vs. ADP 2015; e endpoint ADR vs.
LPST100; and f endpoint LPST100 vs. ReCiPe 2016. R2 values rep-
resent the correlation of the log–log regression of impact scores per
data set
◂
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

1194 The International Journal of Life Cycle Assessment (2022) 27:1180–1198
1 3
1E-05 1E-02 1E+01 1E+04 1E+07 1E+10 1E+13
ADP 2015 (kg Fe-eq.)
LPST 100 (midpoint, kg Fe-eq.)
1E-08
1E-05
1E-02
1E+01
1E+04
1E+07
1E+10
ADR (midpoint, kg Fe-eq.)
1E-08
1E-05
1E-02
1E+01
1E+04
1E+07
1E+10
1E-08 1E-05 1E-02 1E+01 1E+04 1E+07 1E+10
LPST100 (midpoint, kg Fe-eq.)
Section B: Mining and quarrying
SectionA:Agriculture, forestry and fishing
Section C: Manufacturing
Section D: Electricity, gas, steam & air conditionning supply
Section F: Construction
Section H: Transportation and storage
Other sections
Section E: Water supply; sewerage, waste management & remediation activities
LPST 100 (midpoint, kg Fe-eq.)
1E-08
1E-05
1E-02
1E+01
1E+04
1E+07
1E+10
ReCiPe 2016 (kg Fe-eq.)
1E-08 1E-05 1E-02 1E+01 1E+04 1E+07 1E+10
b
c
ADR (midpoint, kg Fe-eq.)
ADP 2015 (kg Fe-eq.)
1E-08
1E-05
1E-02
1E+01
1E+04
1E+07
1E+10
1E-05 1E-02 1E+01 1E+04 1E+07 1E+10 1E+13
d
a
R = 0.937
2R = 0.985
2
R = 0.927
2
R = 0.991
2
ADR (endpoint, kg Fe-eq.)
LPST100 (endpoint, kg Fe-eq.)
1E-08
1E-05
1E-02
1E+01
1E+04
1E+07
1E+10
1E-08 1E-05 1E-02 1E+01 1E+04 1E+07 1E+10
e
R = 0.991
2
1E-08
1E-05
1E-02
1E+01
1E+04
1E+07
1E+10
1E-08 1E-05 1E-02 1E+01 1E+04 1E+07 1E+10
ReCiPe 2016 (kg Fe-eq.)
LPST 100 (endpoint, kg Fe-eq.)
f
R = 0.995
2
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

1195The International Journal of Life Cycle Assessment (2022) 27:1180–1198
1 3
0.0054% across all of the data sets. Finally, the compari-
son between endpoint category totals for the LPST100
method and those of the ReCiPe 2016 method (Fig.6f;
R2 = 0.995) shows an even higher convergence than for
the respective midpoint assessments, as partly explained
by the relation between the SCP and the price information
used for the computation of endpoint CFs for the LPST
method exposed in “Sect.3.2.2.”
3.3 Evaluation ofADR andLPST methods
We evaluated the extended ADR and LPST methods against
five criteria adapted from the ILCD handbook (European
Commission2010). We hereby summarize our observations;
the detailed evaluation can be found in section S7of the SI.
Completeness of scope: the methods provide a significant
coverage of metal resources at the global scale, but do not
include mineral compound flows (e.g., sand and talc), nor
fossil fuels and their derived products like plastics. Relevance
for the assessment of mineral resource use on the AoP natural
resources: the methods rely on real-life statistics and provide
a global evaluation of dissipation patterns for the studied met-
als. The CFs apply to extraction flows, and hence do not allow
distinguishing between different product systems that may dis-
sipate metals differently. They address the safeguard subject
for mineral resource use established by the Life Cycle Initia-
tive, i.e., the “potential to make use of the value that mineral
resources can hold for humans in the technosphere” (Berger
etal. 2020). Scientific robustness and certainty: the methods
rely on data and methods that are published or to be published
as peer-reviewed articles. Uncertainty is addressed and where
practical accounted for Documentation, Transparency, and
Reproducibility: The underlying data and model are docu-
mented transparently and are available in open-access (Helbig
and Charpentier Poncelet 2022). The methods can be repli-
cated, and additional CFs can be recomputed or newly devel-
oped with the help of published articles (dynamic MFA meth-
ods: Charpentier Poncelet etal. 2022b and LCIA methods:
Charpentier Poncelet etal. 2021 and this article). Applicabil-
ity: The large-scale application study presented in “Sect.3.2”
demonstrated that the ADR and LPST methods are readily
usable with current LCI databases in LCA studies, albeit with
some limitations to keep in mind (cf. “Sect.3.1.2”).
3.4 Perspectives fortheADR andLPST methods
Some perspectives to improve the ADR and LPST methods
can be identified. Additional CFs could be developed for
other minerals such as construction aggregates and sand. It
would be possible for third-party users to generate CFs for
additional mineral resources by using the framework and
methods for dynamic MFA of Charpentier Poncelet etal.
(2022b), along with the methods presented by Charpentier
Poncelet etal. (2021) and complemented in this article. It
could also be possible to adapt the model to assess the dissi-
pation of plastics and that of biotic resources used in similar
sectors as mineral resources (e.g., wood).
Furthermore, additional research could aim to improve
the evaluation of the economic and use values (i.e., the value
of services provided by applications) of mineral resources,
as defined by Charpentier Poncelet etal. (2022a). A differ-
ent methodology could be developed to evaluate endpoint
damage in a more reliable way than by using annual aver-
age market prices. Other metrics could also be considered
such as the global economic importance of resources com-
puted for criticality studies (Graedel etal. 2012). Moreover,
efforts could be spent on developing CFs considering dif-
ferent assumptions on the future recovery of metals from,
e.g., waste disposal facilities (see e.g. Dewulf etal. 2021).
Such assumptions could build on different cultural perspec-
tives, as discussed by Charpentier Poncelet etal. (2022a).
Finally, CFs of the ADR and LPST methods could apply to
dissipative flows (or “resource inaccessibility” flows) rather
than extraction flows if these were identified in the LCI.
For instance, dissipative flows due to a given product sys-
tem could be estimated by subtracting functionally recycled
metal flows from extracted metal flows. We discuss such
perspectives in more details in section S8of the SI, and
investigate how different future recovery scenarios could be
implemented in such an assessment under different cultural
perspectives.
4 Conclusion
The extended ADR and LPST methods presented in this
article allow evaluating the impacts of dissipation of metals
in LCA. Because of the lack of information on dissipative
flows in the LCI, they so far apply to extraction flows, and
thus no distinction is made between processes or product
systems that allow for recycling. Potential workaround solu-
tions addressing this issue are presented in section S8 of
the SI, along with potential ways forward to account for the
future recovery of metals in dissipation-oriented approaches.
Further research is needed to compare LCIA results when
applying the CFs from the ADR and LPST methods to dis-
sipative flows in the inventory (e.g., using the process-based
JRC approach; see Beylot etal. 2021) instead of extraction
flows and evaluate how that would influence impact assess-
ment results.
The midpoint CFs for the ADR and LPST methods rely
on replicable results from open-source data and model
(Charpentier Poncelet etal. 2022b; Helbig and Charpentier
Poncelet 2022). Endpoint CFs introduce the use of price
statistics in the assessment, which is helpful to distinguish
between the value of different metals but presents some
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

1196 The International Journal of Life Cycle Assessment (2022) 27:1180–1198
1 3
limitations at this point (cf. “Sect.3.1.2”). The identified
limitations are expected to be most relevant for metals
with very small resource flows in the LCI or for which no
resource flows are currently reported in widespread data-
bases like ecoinvent. For instance, it was observed that by-
product metals like gallium are likely to be over-represented
in the endpoint results for the ADR method. Moreover, while
price statistics were here considered to be the best applicable
proxy to represent the value of metals for humans as defined
by the Life Cycle Initiative’s taskforce on mineral resources
(Berger etal. 2020; Sonderegger etal. 2020), other metrics
might be considered in the future to improve the way the
economic and use values of resources are taken into account.
Future research could aim to improve upon price statistics
or explore the use of other proxies to reflect the value of
metals for humans. Therefore, we recommend the use of
midpoint CFs to evaluate the physical dissipation rates of
metals (ADR) or the related inaccessibility to potential users
(LPST). The use of endpoint CFs provides additional infor-
mation on the potentially lost value of metals, but should be
done cautiously and keeping their limitations in mind.
The application study revealed that the midpoint impact
assessments are expected to diverge between selected LCIA
methods, as impact hotspots are expected to be due to differ-
ent metals. The divergence is expected to be smaller between
the ADR/LPST methods and ReCiPe 2016, because their
CFs spread over less orders of magnitude, thus being more
dependent on the size of resource flows than ADP 2015.
Moreover, the scarcest metals (e.g., precious metals) consist-
ently have higher CFs in the ADP 2015, midpoint and end-
point ReCiPe 2016 methods (SOP and SCP, respectively),
and endpoint ADR and LPST methods. However, their CFs
are so high in the ADP 2015 method that they consistently
dominate impact scores for that method, while they only
show up in top impact scores for data sets with relatively
high amounts of these scarce metals with the other methods.
Rather, widely used metals with relatively large resource
flows can generally be expected to come out as being respon-
sible for the most impacts when assessed with the midpoint
ADR, LPST, and midpoint/endpoint ReCiPe 2016 methods.
This remains generally true for the endpoint ADR and LPST
methods, although the metals that dissipate the fastest tend
to be more represented in the endpoint assessment using the
ADR method, and most expensive metals in the endpoint
assessment using the LPST method.
Finally, we would like to conclude this article with a short
discussion on assessing multiple aspects related to mineral
resource use in the AoP natural resources. Charpentier Poncelet
etal. (2022a) suggested that multiple LCIA methods should be
used to assess the impacts of mineral resource use on the AoP
natural resources exhaustively. Indeed, resources provide eco-
nomic values and use values to different users, and practitioners
with different cultural perspectives could attempt to characterize
the impacts of mineral resource use in different ways depending
on their beliefs. Therefore, the assessment of mineral resource
use on the AoP natural resources in the LCA or life cycle sus-
tainability assessment (LCSA) contexts could be realized by
using the ADR or LPST methods along with other LCIA meth-
ods with complementary pathways such as those covered by the
ADP 2015 and ReCiPe 2016 methods (cf. Charpentier Poncelet
etal. 2022a). Measuring the impacts of dissipation in terms of
lost potential to use the economic and use values (in monetary
units or in different units) could enable the comparison with the
impacts for other biotic and abiotic resources as well as values
obtained from ecosystems (cf. discussion in Charpentier Poncelet
etal. 2022a). Nevertheless, as discussed by the authors, signifi-
cant developments are needed to address impact pathways related
to natural resource use consistently. Moreover, important method
developments and harmonization between existing or new LCIA
methods would be needed for an exhaustive impact assessment
of different aspects of mineral resource use to become viable
(Charpentier Poncelet etal. 2022a).
Supplementary Information The online version contains supplemen-
tary material available at https:// doi. org/ 10. 1007/ s11367- 022- 02093-2.
Acknowledgements We thank F. Ardente for providing a compila-
tion of price statistics retrieved from the US Geological Survey data
available on the latter’s website. We are grateful to the two anonymous
reviewers whose comments greatly helped us improve the quality of
this manuscript.
Funding Open Access funding enabled and organized by Projekt
DEAL. This article was mostly developed and written during the PhD
thesis of A.C.P., which was co-financed by the French agency for the
ecological transition (ADEME) and the French geological survey
(BRGM).
Data availability Data and results depicted in the article are provided
in the Supporting information documents. The data and model used
to compute midpoint characterization factors may be accessed online
at https:// doi. org/ 10. 17605/ OSF. IO/ CWU3D (Helbig and Charpentier
Poncelet 2022). Price data and references are detailed in the article and
in the Supporting information. Data from the ecoinvent database are
subject to copyright and may not be divulged.
Declarations
Competing interests The authors declare no competing interests.
Open Access This article is licensed under a Creative Commons Attri-
bution 4.0 International License, which permits use, sharing, adapta-
tion, distribution and reproduction in any medium or format, as long
as you give appropriate credit to the original author(s) and the source,
provide a link to the Creative Commons licence, and indicate if changes
were made. The images or other third party material in this article are
included in the article's Creative Commons licence, unless indicated
otherwise in a credit line to the material. If material is not included in
the article's Creative Commons licence and your intended use is not
permitted by statutory regulation or exceeds the permitted use, you will
need to obtain permission directly from the copyright holder. To view a
copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

1197The International Journal of Life Cycle Assessment (2022) 27:1180–1198
1 3
References
Berger M, Sonderegger T, Alvarenga R, Bach V, Cimprich A, Dewulf
J, Frischknecht R, Guinée J, Helbig C, Huppertz T, Jolliet O,
Motoshita M, Northey S, Peña CA, Rugani B, Sahnoune A,
Schrijvers D, Schulze R, Sonnemann G, Valero A, Weidema BP,
Young SB (2020) Mineral resources in life cycle impact assess-
ment: part II – recommendations on application-dependent use of
existing methods and on future method development needs. Int J
Life Cycle Assess. https:// doi. org/ 10. 1007/ s11367- 020- 01737-5
Beylot A, Ardente F, Marques A, Mathieux F, Pant R, Sala S, Zampori
L (2020a) Abiotic and biotic resources impact categories in LCA :
development of new approaches. Publications Office of the Euro-
pean Union, Luxembourg. https:// doi. org/ 10. 2760/ 232839
Beylot A, Ardente F, Sala S, Zampori L (2021) Mineral resource dis-
sipation in life cycle inventories. Int J Life Cycle Assess. https://
doi. org/ 10. 1007/ s11367- 021- 01875-4
Beylot A, Ardente F, Sala S, Zampori L (2020b) Accounting for the
dissipation of abiotic resources in LCA: status, key challenges
and potential way forward. Resour Conserv Recycl 157:104748.
https:// doi. org/ 10. 1016/j. resco nrec. 2020. 104748
Blomsma F, Tennant M (2020) Circular economy: preserving materials or
products? Introducing the Resource States framework. Resour Conserv
Recycl 156:104698. https:// doi. org/ 10. 1016/j. resco nrec. 2020. 104698
Bru K, Christmann P, Labbé J-F, Lefebvre G (2015)Panorama 2014
du marché des Terres Rares. Orléans, France
Charpentier Poncelet A, Beylot A, Loubet P, Laratte B, Muller S, Villeneuve
J, Sonnemann G (2022a) Linkage of impact pathways to cultural per-
spectives to account for multiple aspects of mineral resource use in life
cycle assessment. Resour Conserv Recycl 176:105912. https:// doi. org/
10. 1016/j. resco nrec. 2021. 105912
Charpentier Poncelet A, Helbig C, Loubet P, Beylot A, Muller S, Villeneuve
J, Laratte B, Thorenz A, Tuma A, Sonnemann G (2022b) Losses and
lifetimes of metals in the economy. Nat Sustain. https:// doi. org/ 10.
1038/ s41893- 022- 00895-8
Charpentier Poncelet A, Helbig C, Loubet P, Beylot A, Muller S,
Villeneuve J, Laratte B, Thorenz A, Tuma A, Sonnemann G
(2021)Life cycle impact assessment methods for estimating the
impacts of dissipative flows of metals. J Ind Ecol jiec.13136.
https:// doi. org/ 10. 1111/ jiec. 13136
Charpentier Poncelet A, Loubet P, Laratte B, Muller S, Villeneuve J,
Sonnemann G (2019) A necessary step forward for proper non-
energetic abiotic resource use consideration in life cycle assess-
ment: the functional dissipation approach using dynamic material
flow analysis data. Resour Conserv Recycl 151:104449. https://
doi. org/ 10. 1016/j. resco nrec. 2019. 104449
Dewulf J, Hellweg S, Pfister S, León MFG, Sonderegger T, de Matos
CT, Blengini GA, Mathieux F (2021) Towards sustainable resource
management: identification and quantification of human actions that
compromise the accessibility of metal resources. Resour Conserv
Recycl 167:105403. https:// doi. org/ 10. 1016/j. resco nrec. 2021. 105403
Drielsma J, Russell-Vaccari AJ, Drnek T, Brady T, Weihed P, Mistry M,
Simbor LP (2016) Mineral resources in life cycle impact assessment—
defining the path forward. Int J Life Cycle Assess 21:85–105. https://
doi. org/ 10. 1007/ s11367- 015- 0991-7
Ecorys (2012)Mapping resource prices: the past and the future
European Commission (2020a) Communication from the Commission
to the European Parliament, the Council, the European Economic
and Social Committee and the Committee of the Regions. A new
Circular Economy Action Plan For a cleaner and more competi-
tive Europe. Brussels, Belgium.
European Commission (2020b). Study on the EU’s list of Critical Raw
Materials. Critical Raw Materials Factsheets. https:// doi. org/ 10.
2873/ 92480
European Commission (2010)Framework and requirements for life cycle
impact assessment models and indicators, international reference life
cycle data system (ILCD) handbook. https:// doi. org/ 10. 2788/ 38719
Graedel TE, Barr R, Chandler C, Chase T, Choi J, Christoffersen
L, Friedlander E, Henly C, Jun C, Nassar NT, Schechner D,
Warren S, Yang MY, Zhu C (2012) Methodology of metal
criticality determination. Environ Sci Technol 46:1063–1070.
https:// doi. org/ 10. 1021/ es203 534z
Graedel TE, Harper EM, Nassar NT, Reck BK (2013) On the materi-
als basis of modern society. Proc Natl Acad Sci 112:6295–6300.
https:// doi. org/ 10. 1073/ pnas. 13127 52110
Graedel TE, Reck BK, Miatto A (2022) Alloy information helps pri-
oritize material criticality lists. Nat Commun 13:1–8. https:// doi.
org/ 10. 1038/ s41467- 021- 27829-w
Helbig C, Charpentier Poncelet A (2022)ODYM MaTrace dissipation
- code and datasets. https:// doi. org/ 10. 17605/ OSF. IO/ CWU3D
Helbig C, Kondo Y, Nakamura S (2021)Simultaneously tracing the fate of
seven metals with MaTrace-multi (accepted manuscript). J Ind Ecol
Helbig C, Thorenz A, Tuma A (2020) Quantitative assessment of dis-
sipative losses of 18 metals. Resour Conserv Recycl. https:// doi.
org/ 10. 1016/j. resco nrec. 2019. 104537
Henckens MLCM, van Ierland EC, Driessen PPJ, Worrell E (2016)
Mineral resources: geological scarcity, market price trends, and
future generations. Resour Policy 49:102–111. https:// doi. org/ 10.
1016/j. resou rpol. 2016. 04. 012
Huijbregts MAJ, Steinmann ZJN, Elshout PMF, Stam G, Verones F,
Vieira M, Zijp M, Hollander A, van Zelm R (2017) ReCiPe2016:
a harmonised life cycle impact assessment method at midpoint
and endpoint level. Int J Life Cycle Assess 22:138–147. https://
doi. org/ 10. 1007/ s11367- 016- 1246-y
Huppertz T, Weidema B, Standaert S, De Caevel B, van Overbeke E
(2019) The social cost of sub-soil resource use. Resources 8:19.
https:// doi. org/ 10. 3390/ resou rces8 010019
Johnson CA, Piatak NM, Miller MM (2017)Barite (Barium), in: Criti-
cal mineral resources of the United States—economic and envi-
ronmental geology and prospects for future supply. U.S. Geologi-
cal Survey, Reston, Virginia.
Johnson Matthey (2021)Price charts [WWW Document].http:// www.
plati num. matth ey. com/ prices/ price- charts (Accessed 13 June 2021).
Kelly T, Matos GR (2014)Historical statistics for mineral and material
commodities in the United States (2016 version) [WWW Docu-
ment]. US Geol Surv Data Ser. 140.https:// www. usgs. gov/ cente rs/
nmic/ histo rical- stati stics- miner al- and- mater ial- commo dities- united-
states (Accessed 15 Aug 2021)
Labbé J-F, Dupuy J-J (2014)Panorama 2012 du marché des plati-
noïdes, Brgm/Rp-63169-Fr. Orléans, France.
Løvik AN, Restrepo E, Müller DB (2016) Byproduct metal availabil-
ity constrained by dynamics of carrier metal cycle: the gallium-
aluminum example. Environ Sci Technol 50:8453–8461. https://
doi. org/ 10. 1021/ acs. est. 6b023 96
Løvik AN, Restrepo E, Müller DB (2015) The global anthropogenic
gallium system: determinants of demand, supply and efficiency
improvements. Environ Sci Technol 49:5704–5712. https:// doi.
org/ 10. 1021/ acs. est. 5b003 20
Metalary (2021)Niobium price [WWW Document].https:// www. metal ary.
com/ niobi um- price/
Moreno Ruiz E, Valsasina L, FitzGerald D, Brunner F, Symeonidis A,
Bourgault G, Wernet G (2019)Documentation of changes imple-
mented in the ecoinvent database v3.6. Zürich, Switzerland
Moreno Ruiz E, Valsasina L, FitzGerald D, Symeonidis A, Turner D,
Müller J, Minas N, Bourgault G, Vadenbo C, Ioannidou D, Wernet
G (2020)Documentation of changes implemented in the ecoinvent
database v3.7 & v3.7.1. Zürich, Switzerland.
Nakamura S, Kondo Y, Kagawa S, Matsubae K, Nakajima K, Nagasaka
T (2014) MaTrace: tracing the fate of materials over time and across
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

1198 The International Journal of Life Cycle Assessment (2022) 27:1180–1198
1 3
products in open-loop recycling. Environ Sci Technol 48:7207–
7214. https:// doi. org/ 10. 1021/ es500 820h
Pauliuk S, Kondo Y, Nakamura S, Nakajima K (2017) Regional dis-
tribution and losses of end-of-life steel throughout multiple prod-
uct life cycles—insights from the global multiregional MaTrace
model. Resour Conserv Recycl 116:84–93. https:// doi. org/ 10.
1016/j. resco nrec. 2016. 09. 029
Reuter MA, van Schaik A, Gutzmer J, Bartie N, Abadías-Llamas A
(2019) Challenges of the circular economy: a material, metallurgi-
cal, and product design perspective. Annu Rev Mater Res 49:253–
274. https:// doi. org/ 10. 1146/ annur ev- matsci- 070218- 010057
Rørbech JT, Vadenbo C, Hellweg S, Astrup TF (2014) Impact assess-
ment of abiotic resources in LCA: quantitative comparison of
selected characterization models SUPP INFO. Environ Sci Tech-
nol 48:11072–11081. https:// doi. org/ 10. 1021/ es502 3976
Schulze R, Guinée J, van Oers L, Alvarenga R, Dewulf J, Drielsma J
(2020) Abiotic resource use in life cycle impact assessment—Part
II – Linking perspectives and modelling concepts. Resour Conserv
Recycl. https:// doi. org/ 10. 1016/j. resco nrec. 2019. 104595
Sonderegger T, Berger M, Alvarenga R, Bach V, Cimprich A, Dewulf
J, Frischknecht R, Guinée J, Helbig C, Huppertz T, Jolliet O,
Motoshita M, Northey S, Rugani B, Schrijvers D, Schulze R,
Sonnemann G, Valero A, Weidema BP, Young SB (2020) Mineral
resources in life cycle impact assessment—part I: a critical review
of existing methods. Int J Life Cycle Assess 25:784–797. https://
doi. org/ 10. 1007/ s11367- 020- 01736-6
Stormcrow (2014)Rare Earth Industry Report
UNEP (2013)Metal recycling: opportunities, limits, infrastructure, A
Report of the Working Group on the Global Metal Flows to the
International Resource Panel.
UNEP (2011)Recycling rates of metals: a status report, A Report of
the Working Group on the Global Metal Flows to the Interna-
tional Resource Panel. United Nations Environment Programme,
Nairobi, Kenya.
United Nations (2008)International Standard Industrial Classification
of All Economic Activities (ISIC), Rev. 4. United Nations Pub-
lication, New York.
van Oers L, de Koning A, Guinée JB, Huppes G (2002)Abiotic
resource depletion in LCA: improving characterisation factors for
abiotic resource depletion as recommended in the new Dutch LCA
Handbook, Road and Hydraulic Engineering Institute.
van Oers L, Guinée JB, Heijungs R (2019) Abiotic resource depletion
potentials (ADPs) for elements revisited—updating ultimate reserve
estimates and introducing time series for production data. Int J Life
Cycle Assess. https:// doi. org/ 10. 1007/ s11367- 019- 01683-x
van Oers L, Guinée JB, Heijungs R, Schulze R, Alvarenga RAF, Dewulf
J, Drielsma J, Sanjuan-Delmás D, Kampmann TC, Bark G, Uriarte
AG, Menger P, Lindblom M, Alcon L, Ramos MS, Torres JME
(2020) Top-down characterization of resource use in LCA: from
problem definition of resource use to operational characterization
factors for dissipation of elements to the environment. Int J Life
Cycle Assess. https:// doi. org/ 10. 1007/ s11367- 020- 01819-4
Vieira M, Ponsioen T, Goedkoop M, Huijbregts M (2016) Surplus
cost potential as a life cycle impact indicator for metal extraction.
Resources 5:2. https:// doi. org/ 10. 3390/ resou rces5 010002
Vieira MDM, Ponsioen TC, Goedkoop MJ, Huijbregts MAJ (2017)
Surplus ore potential as a scarcity indicator for resource extrac-
tion. J Ind Ecol 21:381–390. https:// doi. org/ 10. 1016/j. jcrc. 2017.
02. 030
Watson BJ, Eggert RG (2021) Understanding relative metal prices and
availability: combining physical and economic perspectives. J Ind
Ecol 25:890–899. https:// doi. org/ 10. 1111/ jiec. 13087
Wernet G, Bauer C, Steubing B, Reinhard J, Moreno-Ruiz E, Weidema
B (2016) The ecoinvent database version 3 (part I): overview and
methodology. Int J Life Cycle Assess 21:1218–1230. https:// doi.
org/ 10. 1007/ s11367- 016- 1087-8
Zampori L, Pant R (2019) Suggestions for updating the product envi-
ronmental footprint (PEF) method. Publications Office of the
European Union, Luxembourg. https:// doi. org/ 10. 2760/ 424613
Zampori L, Sala S (2017) Feasibility study to implement resource dis-
sipation in LCA. Luxembourg. https:// doi. org/ 10. 2760/ 869503
Publisher's Note Springer Nature remains neutral with regard to
jurisdictional claims in published maps and institutional affiliations.
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

1.
2.
3.
4.
5.
6.
Terms and Conditions
Springer Nature journal content, brought to you courtesy of Springer Nature Customer Service Center GmbH (“Springer Nature”).
Springer Nature supports a reasonable amount of sharing of research papers by authors, subscribers and authorised users (“Users”), for small-
scale personal, non-commercial use provided that all copyright, trade and service marks and other proprietary notices are maintained. By
accessing, sharing, receiving or otherwise using the Springer Nature journal content you agree to these terms of use (“Terms”). For these
purposes, Springer Nature considers academic use (by researchers and students) to be non-commercial.
These Terms are supplementary and will apply in addition to any applicable website terms and conditions, a relevant site licence or a personal
subscription. These Terms will prevail over any conflict or ambiguity with regards to the relevant terms, a site licence or a personal subscription
(to the extent of the conflict or ambiguity only). For Creative Commons-licensed articles, the terms of the Creative Commons license used will
apply.
We collect and use personal data to provide access to the Springer Nature journal content. We may also use these personal data internally within
ResearchGate and Springer Nature and as agreed share it, in an anonymised way, for purposes of tracking, analysis and reporting. We will not
otherwise disclose your personal data outside the ResearchGate or the Springer Nature group of companies unless we have your permission as
detailed in the Privacy Policy.
While Users may use the Springer Nature journal content for small scale, personal non-commercial use, it is important to note that Users may
not:
use such content for the purpose of providing other users with access on a regular or large scale basis or as a means to circumvent access
control;
use such content where to do so would be considered a criminal or statutory offence in any jurisdiction, or gives rise to civil liability, or is
otherwise unlawful;
falsely or misleadingly imply or suggest endorsement, approval , sponsorship, or association unless explicitly agreed to by Springer Nature in
writing;
use bots or other automated methods to access the content or redirect messages
override any security feature or exclusionary protocol; or
share the content in order to create substitute for Springer Nature products or services or a systematic database of Springer Nature journal
content.
In line with the restriction against commercial use, Springer Nature does not permit the creation of a product or service that creates revenue,
royalties, rent or income from our content or its inclusion as part of a paid for service or for other commercial gain. Springer Nature journal
content cannot be used for inter-library loans and librarians may not upload Springer Nature journal content on a large scale into their, or any
other, institutional repository.
These terms of use are reviewed regularly and may be amended at any time. Springer Nature is not obligated to publish any information or
content on this website and may remove it or features or functionality at our sole discretion, at any time with or without notice. Springer Nature
may revoke this licence to you at any time and remove access to any copies of the Springer Nature journal content which have been saved.
To the fullest extent permitted by law, Springer Nature makes no warranties, representations or guarantees to Users, either express or implied
with respect to the Springer nature journal content and all parties disclaim and waive any implied warranties or warranties imposed by law,
including merchantability or fitness for any particular purpose.
Please note that these rights do not automatically extend to content, data or other material published by Springer Nature that may be licensed
from third parties.
If you would like to use or distribute our Springer Nature journal content to a wider audience or on a regular basis or in any other manner not
expressly permitted by these Terms, please contact Springer Nature at
onlineservice@springernature.com