Resources, Conservation and Recycling 83 (2014) 63–76

Contents lists available at ScienceDirect

Resources, Conservation and Recycling

journal homepage: www.elsevier.com/locate/resconrec

Full length Article

A future perspective on lithium-ion battery waste flows

from electric vehicles
Kirti Richa, Callie W. Babbitt ∗ , Gabrielle Gaustad, Xue Wang
Golisano Institute for Sustainability, Rochester Institute of Technology, 111 Lomb Memorial Drive, Rochester, NY 14623, United States

a r t i c l e i n f o a b s t r a c t

Article history: As a proactive step towards understanding future waste management challenges, this paper presents a
Received 5 September 2013 future oriented material flow analysis (MFA) used to estimate the volume of lithium-ion battery (LIB)
Received in revised form wastes to be potentially generated in the United States due to electric vehicle (EV) deployment in the near
12 November 2013
and long term future. Because future adoption of LIB and EV technology is uncertain, a set of scenarios was
Accepted 16 November 2013
developed to bound the parameters most influential to the MFA model and to forecast “low,” “baseline,”
and “high” projections of future end-of-life battery outflows from years 2015 to 2040. These models were
Keywords:
implemented using technology forecasts, technical literature, and bench-scale data characterizing battery
Lithium-ion batteries
Electric vehicles
material composition. Considering the range from the most conservative to most extreme estimates, a
Material flow analysis cumulative outflow between 0.33 million metric tons and 4 million metric tons of lithium-ion cells could
Waste management be generated between 2015 and 2040. Of this waste stream, only 42% of the expected materials (by
weight) is currently recycled in the U.S., including metals such as aluminum, cobalt, copper, nickel, and
steel. Another 10% of the projected EV battery waste stream (by weight) includes two high value materials
that are currently not recycled at a significant rate: lithium and manganese. The remaining fraction of this
waste stream will include materials with low recycling potential, for which safe disposal routes must be
identified. Results also indicate that because of the potential “lifespan mismatch” between battery packs
and the vehicles in which they are used, batteries with high reuse potential may also be entering the
waste stream. As such, a robust end-of-life battery management system must include an increase in
reuse avenues, expanded recycling capacity, and ultimate disposal routes that minimize risk to human
and environmental health.
© 2013 Elsevier B.V. All rights reserved.

1. Introduction U.S. Energy Information Administration (EIA, 2012), J.D. Power and
Associates (Humphrey et al., 2010), Credit Suisse (Jobin et al., 2009),
Lithium-ion batteries (LIBs) have emerged as a promising energy International Energy Agency [IEA] (2011), Deutsche Bank (Watabe
storage solution for electric vehicles (EVs) and renewable energy and Mori, 2011), Deloitte Consulting (Giffi et al., 2010), Lazard Cap-
systems, but their potential environmental tradeoffs are not well ital Markets (Shrestha et al., 2010) and Morgan Stanley (Steinmetz
characterized. Although recent work has focused on supply side and Shankar, 2008). The range of deployment scenarios by these
issues, such as lithium availability, key uncertainties surround agencies vary significantly across parameters (economic growth, oil
the emergence and management of these batteries in the waste price, proposed Corporate Average Fuel Economy [CAFE] standards,
stream and the ability of domestic recycling infrastructure to battery technology etc.), and indicate anywhere between 0.45 mil-
recover scarce and valuable materials from a highly variable mix lion and 4 million EVs sold in the United States in 2020 (Fig. 1(a))
of discarded batteries. A proactive approach is required to pre- and international sales ranging between 5.2 million and 19.8 mil-
vent unanticipated environmental impacts of end-of-life (EOL) lion in the same time frame (Fig. 1(b)). Powering these vehicles will
battery generation associated with forecast growth in electric vehi- clearly require a large scale deployment of energy storage systems
cle deployment. (Gaines and Nelson, 2010; Gruber et al., 2011; Kushnir and Sandén,
Several agencies have predicted widespread diffusion of 2012).
electric-drive vehicles in the future, both in the U.S. and at a global This rapid growth in LIB demand comes with its own sustaina-
level. Forecasts of future EV sales (Fig. 1) have been produced by the bility tradeoffs: as replacements for nickel metal hydrides (NiMH),
LIBs reduce demand for rare earth metals but increase consump-
tion of lithium, cobalt, manganese, and nickel (Alonso et al., 2012;
∗ Corresponding author. Tel.: +1 585 475 6277. Gruber et al., 2011). Several studies have investigated the implica-
E-mail address: [email protected] (C.W. Babbitt). tions of EV penetration on material demand, particularly lithium

0921-3449/$ – see front matter © 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.resconrec.2013.11.008
64 K. Richa et al. / Resources, Conservation and Recycling 83 (2014) 63–76

a 4,000
4,000

3,600

3,200
US EV sales (1,000 units)

2,800

2,400
2,000
2,000
1,781

1,600

1,146 1,196
1,200

800 736
449 465
400

0
EIA Low Deloitte EIA Reference EIA High Lazard Capital J.D. Power Morgan Deutsche Bank
(PHEV+BEV) (PHEV+BEV) Stanley
(PHEV+HEV)

b 20,000
19,800
Global EV sales (1,000 units)

17,500

15,000 14,403

12,500

10,000

7,500 6,900
5,195
5,000

2,500

0
J.D. Power IEA Credit Suisse Deutsche Bank

Fig. 1. (a) EV sales forecast-2020 (U.S.). (b) EV sales forecast-2020 (Global). Literature references for each sales forecast provided in the main text.

(Gaines and Nelson, 2010; Gruber et al., 2011; Grosjean et al., Clearly, a better understanding of the ultimate management and
2012; Kushnir and Sandén, 2012; Yaksic and Tilton, 2009). Though fate of batteries in the waste stream is required, but such an anal-
concerns over lithium scarcity in the long-term have been less- ysis is complicated by key uncertainties, including the expected
ened by reassuring results from such studies, there may still be timing and volumes of batteries reaching their end of life; the
future challenges for the U.S. to access world lithium resources. A quality, concentrations, and variability of specific materials con-
large portion of lithium deposits are found in only a few countries tained in spent batteries; and the capacity for recycling systems to
of the world, with the U.S. accounting for only 0.3% of current recover scarce and valuable materials from a highly variable bat-
lithium reserves (USGS, 2012) and about 3.7% of the world lithium tery waste stream. While the lag in deploying EV technologies may
reserve base (USGS, 2009). Trade embargoes or political instability suggest that battery waste will not be a priority for several years,
in the future may impact the U.S. EV and LIB industries, as many “lessons learned” from our current sub-optimal management of
lithium-supplying countries are already politically volatile. Fur- electronic waste show the perils of introducing complex products
thermore, cobalt, manganese and nickel, which are major inputs without proactive development of a waste management system.
to the lithium-ion battery industry, are not significantly mined In the case of electronic waste, low EOL value, difficulty recovering
in the U.S., requiring primary dependence on imported supplies valuable materials and insufficient domestic infrastructure has lead
(USGS, 2012). Hence, EV battery recycling may contribute to a sta- to exploitation of developing countries and loss of valuable mate-
ble domestic supply chain for these critical materials. Currently, rial resources (Babbitt et al., 2011; Williams et al., 2008; Widmer
most EOL batteries from consumer electronics contain high levels et al., 2005; Wang and Gaustad, 2012). Since many of these fac-
of cobalt, a metal whose high economic value catalyzes current tors are similar to LIBs, avoiding negative environmental, economic,
LIB recycling systems, but the trajectory of battery technology and social outcomes at EOL requires a more proactive approach in
could result in introduction of different material and value streams, planning for this new waste stream.
which may change the economic and policy implications of battery As a step towards addressing EOL LIB management, this paper
recycling (Wang et al., 2013,2014). applies a scenario-driven material flow analysis (MFA) to project
 K. Richa et al. / Resources, Conservation and Recycling 83 (2014) 63–76 65

the potential volume and timing of lithium-ion batteries entering 2,000

the waste stream as a result of their forecasted deployment in elec- 1,875
1,750
tric vehicles. Towards this objective, the number of EV LIB units
1,625

US EV sales (1,000 units)

entering the waste stream as well as the mass of battery cells in 1,500
that stream is estimated on an annual basis between years 2015 1,375
and 2040 for three different scenarios. To estimate recycling poten- 1,250
tial and waste management needs of EOL EV LIBs in the future, this 1,125
paper also aims to characterize the materials that would be present 1,000
875
in the EV battery waste stream on the basis of their recyclability
750
and their commodity value under different technology trajectories 625
of battery chemistry and form factor. Furthermore, this MFA model 500
also seeks to characterize the potential for diverting EV batteries 375
from this waste stream into reuse applications depending on the 250
remaining battery life. 125
0
MFA is a well-established method for investigating the material,

2013

2023

2033
2009

2011

2015

2017

2019

2021

2025

2027

2029

2031

2035
energy and environmental implications of commodity products
(Oguchi et al., 2008; Yoshida et al., 2009; Chang et al., 2009; Year
Steubing et al., 2010). While some MFA studies have addressed
High Oil Price Scenario Reference Case Low Oil Price Scenario
issues related to LIBs, the existing literature focuses on analyzing
the stock and flows of laptops and cell phones batteries (Chang Fig. 2. Energy Information Administration (2012) EV sales forecasts.
et al., 2009), tracking flows of cobalt (Harper et al., 2012), and
assessing supply and demand for lithium due to EV technology
(e.g., Gaines and Nelson, 2010; Grosjean et al., 2012; Kushnir
2.1. Model formulation
and Sandén, 2012). To our knowledge, no study has yet applied
MFA to fully model future outflows of batteries from EV sys-
The EV battery MFA model was implemented in three sequential
tems.
steps, each described in more detail in the following sec-
Because such an analysis is complicated by significant uncer-
tions:
tainty about technology adoption and performance, this MFA
is also informed by approaches used in previous studies to
develop scenario-based MFA for materials ranging from steel (1) Material flow analysis to estimate the waste flows of entire EV
(Park et al., 2011; Pauliuk et al., 2011; Michaelis and Jackson, battery packs.
2000) to electronic waste (Steubing et al., 2010; Kang and (2) Estimation of individual lithium-ion cells in the EV battery pack
Schoenung, 2006; Streicher-Porte et al., 2005). From a method- waste stream.
ological standpoint, this paper also highlights the uncertainties (3) Assessment of specific materials comprising each cell within
associated with conducting a scenario-based MFA of EV LIBs, the EV battery waste stream.
as a means of establishing future research priorities that must
be resolved as additional data and system parameters become Furthermore, based on the material and mass composition of the
available. Key variables addressed here include EV adoption EV battery waste stream, the economic value of the waste stream
dynamics, battery lifespan and constituent materials in lithium-ion was estimated on an annual basis.
cells.
2.1.1. Material flow analysis to estimate the flows of waste EV
battery packs
2. Methods The first part of the model calculated the number of lithium-
ion EV battery packs entering the U.S. waste stream on an annual
A future oriented top-down material flow analysis (MFA) was basis from years 2015 to 2040. This time period was chosen based
conducted to estimate the volume of lithium-ion batteries pro- on available data from the Department of Energy on both near-
jected to enter the waste stream in the near and long term future, and long-term EV deployment projections. The annual inflow of
after use in electric vehicles. MFA is a systematic assessment of EV batteries was estimated from EV sales forecasts, and the annual
the flows and stocks of materials within a defined temporal and outflow of waste batteries was determined based on the battery
spatial system (Brunner and Rechberger, 2004) that can be used lifespans once they entered vehicle use, as well as the lifespan of
to track the flow of a specific substance or of products within a the EV itself.
system. In the top-down MFA methodology, the product inflows
are determined from specific ‘final goods’ categories entering the 2.1.1.1. EV sales forecast. U.S. level EV sales forecasts were obtained
system and the outflows are determined from discards, based on from the Light Duty Vehicle (LDV) Sales Projections through the
product lifespan, with the material stocks being inferred from these year 2035 provided by the U.S. Energy Information Administra-
inflows and outflows (Graedel and Allenby, 2010). Here, the annual tion [EIA] (EIA, 2012). Three types of electric-drive vehicles were
inflow of EV batteries was estimated from projected EV sales, and considered in the EIA LDV sales forecasts: hybrid- [HEVs], plug-
the annual outflow of batteries was calculated based on battery in hybrid- [PHEVs] (10 miles and 40 miles ranges) and all-electric
and vehicle lifespans. Given the significant uncertainty about future or battery electric vehicles [BEVs]. For the baseline scenario, the
EV adoption rates and battery technologies, bounding scenarios EIA Annual Energy Outlook “reference case” LDV sales projections
were developed to forecast “low,” “baseline,” and “high” projec- were used. The reference case used in EIA projections is a baseline
tions of future waste battery outflows and their attendant material scenario assuming business-as-usual with current laws and regu-
implications. Key differences among these scenarios stemmed from lations being the same across the timeline of the projections (EIA,
variability in EV sales projections, battery lifespan distribution and 2012). The low and high scenarios reflect EIA forecasts that consider
parameters governing number of cells per battery pack, which will low and high oil prices, respectively. These forecasts are shown in
be discussed in the subsequent sections. Fig. 2.
66 K. Richa et al. / Resources, Conservation and Recycling 83 (2014) 63–76

45% batteries in the baseline and high scenarios respectively. In spite of

Percentage of EV LIBs reaching end-of-life

these variations, the distributions selected result in a majority of

40%
EV LIBs used in EVs modeled as having a lifespan in the range of
35% 8–10 years, consistent with warranty terms and recent literature.
Low Similar to EV batteries, the lifespan of electric vehicles too would
30% scenario
follow a distribution which may be even wider than that for bat-
25% teries, depending on early vehicle failure or car crashes as well
Baseline
20% scenario as extended life through multiple resales. However, to keep this
initial MFA model tractable, the EV lifespan has been fixed. In gen-
15% eral, traditional vehicle lifespan assumptions vary across studies in
High
10% scenario the range of 10–16 years (Huang et al., 2011; Greene and DeCicco,
2000; Lemp and Kockelman, 2008; Greene et al., 2005; Kumar and
5%
Sutherland, 2008). Only limited information is available on elec-
0% tric vehicle lifespan, but this parameter is modeled as 10 years in a
4 years 6 years 8 years 10 years 12 years 14 years
recent study by Gruber et al. (2011). While we recognize the uncer-
EV LIB lifespan tainty associated with lifespan and the need for future work in this
area, this MFA model assumed a moderate, fixed EV lifespan of 10
Fig. 3. Truncated lifespan distribution of EV batteries for three scenarios.
years as a starting point for analysis, with sensitivity analysis on
a 16 year EV lifespan shown in section S7.3 of the supplementary
2.1.1.2. Battery lifespan. The lifespan or service life of a lithium-ion information (SI).
battery can be expressed either in terms of its cycle life or its cal- The lifespan distribution shown in Fig. 3, contrasted against
endar life. Cycle life is defined as the number of charge-discharge the vehicle life, raises an important point: there will likely be a
cycles the battery can undergo before failing to meet specific “mismatch” between vehicle and battery lifespans. Some batter-
performance criteria. Calendar life on the other hand is defined as ies entering use in a given year would likely reach the end of their
the length of time a battery can be stored with minimal discharges life before the vehicles in which they are used. These vehicles then
before capacity diminishes. In general, a battery is considered need new batteries to continue operation in subsequent years. On
to have reached its end of life in EV application when it reaches the other hand, if a vehicle were to reach the end of its life before its
about 80% of its original capacity (Williams and Lipman, 2010). batteries, it is assumed that the battery would not be refitted into
EV battery lifespan is highly uncertain and dependent on many a new vehicle (although it may be reused in other applications)
factors which are still poorly understood. Marano et al. (2009) (Williams and Lipman, 2010; Cready et al., 2003). Thus, batteries
indicates that lithium-ion batteries usually have a calendar life entering the waste stream at any given time can be loosely classified
of 10 years, subject to favorable operating conditions that avoid into two types:
overcharging, aggressive driving leading to rapid discharge and Type 1 EOL EV batteries are those that have reached their end-
more frequent charging, and operation at high temperatures. of-lives in EV application due to capacity fade, either before or
Most previous studies have assumed a fixed EV battery lifespan of coinciding with the vehicles’ end of life. In general, an EV bat-
either 8 or10 years (Gruber et al., 2011; Yaksic and Tilton, 2009; tery has 70–80% of its original capacity intact once it reaches the
Harper et al., 2012), which is consistent with the length of many end of its utility for EV applications (Neubauer and Pesaran, 2011).
vehicle manufacturers’ warranty terms. However, some literature Though insufficient for automotive use, there is some potential that
indicates lower lifespan of about 5 years for EV LIBs (Anderman, these batteries can be reused in off-grid and grid-based stationary
2007). As per Dinger et al. (2010), EV battery lifespan could be energy storage applications instead of entering the waste stream
anywhere between 5 and 10 years, while Nemry et al. (2009) (Neubauer and Pesaran, 2011; Williams and Lipman, 2010; Cready
assume a lifespan of 10–15 years. Significant research efforts are et al., 2003).Type 2 EOL EV batteries are those found in vehicles that
aimed toward achieving lifespans of up to 15 years for EV batteries reach their end-of-lives before their batteries, which is likely the
(Kalhammer et al., 2009; Chalk and Miller, 2006). case in early vehicle failure or crash or if a vehicle has a battery
Applying a lifespan distribution to determine the EV-LIB out- replacement later in its useful life. This set of non-EOL EV batteries
flows would address the fact that the lifespan of a battery would could technically still meet the criteria for reuse in EVs, but actual
depend on its usage and charging patterns, which vary from user reuse in this manner is unlikely, given concern about reliability and
to user. Assuming that electric vehicles are charged 1.5 times per technical compatibility of “pre-aged” batteries (Cready et al., 2003;
week, the U.S. Department of Energy (DOE, 2010) predicted that the Burke, 2009). These batteries may have high potential for other
calendar life of a typical EV battery would increase from 4 years in reuse markets, like those described above. The distinction between
2009 to 14 years in 2015 owing to the ongoing innovation in this these two battery types is intended to indicate the potential for
field. Hence, the different scenarios in the model have considered diverting batteries from the waste stream into reuse applications.
battery lifespans ranging from 4 to14 years. Rather than a single
point estimate, a lifespan distribution (Fig. 3) was applied to model 2.1.1.3. Lithium-ion battery use in hybrid electric vehicles. Currently
a more realistic scenario, taking into consideration early battery most HEVs on the market use nickel metal hydride [NiMH] batter-
failures as well as batteries surviving for more than 10 years. Since ies, rather than lithium-ion, and NiMH batteries would continue to
a lifespan distribution of EV LIBs is not yet established, this tech- be a feasible option for HEV for several years (Frost and Sullivan,
nology being in its early stages of adoption, a truncated normal 2009). However, it is predicted that lithium-ion batteries’ share of
distribution of EV LIB lifespan has been used in the three scenarios the HEV market would grow and eventually surpass NiMH usage
(with a mean lifespan of 8 to 10 years). The variation in assumed between 2018 and 2025 (Jobin et al., 2009; Madani, 2009; Fu, 2009).
battery lifespan distribution among the three scenarios not only Estimates from a Credit Suisse report prepared by Jobin et al. (2009)
indicate the uncertainty in the lifespan of EV LIBs but also highlight were applied to the scenarios used here, as their study provided
that the volume of EV battery waste stream would be dependent both conservative and optimistic estimates for HEV lithium-ion
on battery lifespan to a certain extent. For instance, in the low sce- battery adoption. The high and baseline scenarios started from
nario, 70% of EV LIBs have been assumed to have lifespan exceeding the Credit Suisse bottom-up estimates, which were optimistic
8 years, whereas this percentage is 50% and 35% respectively for towards rapid LIB adoption in HEV (Jobin et al., 2009), leading to an
 K. Richa et al. / Resources, Conservation and Recycling 83 (2014) 63–76 67

Fig. 4. Conceptual basis of estimating future EV Li-ion battery outflows.

assumption that all HEVs use lithium-ion batteries by year 2015 Si,t = sales of new EVs of type i that use LIBs in year t and
and 2025, respectively in these two scenarios. The Credit Suisse Wi,t = non-EOL EVs of type i requiring a replacement LIB in year t.
top-down estimate for HEV lithium-ion battery adoption was used The above relationship was distinguished by “i” vehicle types:
for the low scenario, with an assumption that 100 percent of HEVs BEV, PHEV10, PHEV40, and HEV (the percentage of HEVs that use
would not rely on lithium-ion batteries until the year 2032. Details lithium-ion batteries). The number of batteries entering new EVs
about HEV lithium-ion battery adoption is provided in the SI. (Si,t ) was determined by the sales forecast for that year, as described
in a previous section. The non-EOL EVs requiring a replacement
2.1.1.4. Estimation of EV battery pack outflows. Considering the battery (Wi,t ) was based on the scenario-specific cases of first-use
sales and lifespan assumptions stated above, Fig. 4 illustrates the batteries with a shorter lifespan than the vehicles in which they
conceptual basis of estimating EV battery outflows by this model. were used. In cases of extreme “lifespan mismatch,” vehicles with
The number of new LIBs entering EV use in any year t would very long lifespans paired with batteries with very short lifespans
depend on EV sales in year t, as well as the number of non-EOL EVs may require two battery replacements. Hence, for a given year, t,
which would require a replacement battery in that year (Fig. 4). 
Wi,t = Pl × (Si,(t−l) + Wi,(t−l) ) (2)
Here, it was assumed that all non-EOL EVs would use a replacement
battery, while we do recognize that realistically, all vehicles may i l

not be put back into use due to high replacement battery cost or l = EV battery lifespan, s.t. l < EV lifespan, Pl = percentage of EV LIBs
damages due to automotive accidents. The lifespan distribution was sold in any given year to have a useful life of l years in EV application,
based on “Pl ”, the percentage of batteries sold in any given year to Si,(t−l) = sales of new EVs of type i in year (t − l), and Wi,(t−l) = non-EOL
have a useful life of l years in EV application, which varied based EVs of type i requiring a replacement battery in year (t − l).
on the scenario (Fig. 3). Kt , the total number of lithium-ion battery Thus, the number of EV LIB packs entering the waste stream (B)
packs entering use in EVs in year t was determined as follows: in a given year t after an l year lifespan is expressed as:
  
Kt = Si,t + Wi,t (1) Bt = Pl × (Ki,(t−l) ) (3)
i i l
68 K. Richa et al. / Resources, Conservation and Recycling 83 (2014) 63–76

Table 1
Vehicle consumption rate, battery efficiency, percent available energy and EV battery energy storage for the three scenarios.

Scenario Vehicle consumption rate (Wh/mile) Battery efficiency Battery available energy (%) EV battery energy storage (kWh)

BEV PHEV HEV BEV PHEV10 PHEV40 HEV

Low 250 95% 90% 80% 30% 29 3.3 13 3.5

Baseline 300 90% 85% 75% 25% 39 4.4 18 5.3
High 350 85% 80% 70% 20% 51 5.9 24 8.2

Ki,(t−l) = total number of LIB packs entering use in EVs of type i in the energy storage of individual cells (Ecellj ), dependent on the cell
year (t − l). cathode chemistry j and was estimated as follows:
 
Epacki (Wh)
Di,j = (5)
2.1.2. Estimation of individual lithium-ion cells in the EV battery Ecellj (Wh)
pack waste stream
The approach described thus far focused on total battery packs,
which each may contain a varied number and type of cells, 2.1.2.2. Battery energy storage (Epacki ). The battery pack energy
depending on technical specifications such as EV type and cath- storage (Epacki ) depends on the EV type and its associated elec-
ode chemistry. Next, we estimated the number of lithium-ion cells tric range, as well as other parameters like vehicle consumption
in this EV battery waste stream, for a given year t as rate and percent battery efficiency and available energy:
 Epacki =
(Ri × C)
(6)
Nt = Bt × (PEi,t × PCj × Di,j ) (4) ( × Ai )
i j
Ri = electric range of EV type i (miles), C = electric vehicle con-
sumption rate (Wh/miles),  = percent efficiency of EV LIB, and
Bt = number of LIB packs in EV battery waste stream in year t,
Ai = percent available energy of the total EV LIB energy for a given
i = EV type (BEV, PHEV10, PHEV40, HEV), j = LIB cathode chemistry,
EV type i
PEi,t = percentage of waste LIB packs belonging to EV type i in year
While the vehicle electric range remains constant with each sce-
t, PCj = percentage of LIBs of battery chemistry j in EV battery waste
nario, the other three parameters in Eq. (6) will vary over the three
stream, and Di,j = number of cells per LIB pack, specific to EV type
scenarios:
and cathode chemistry.
Parameters Bt and PEi,t change with time as well as with the
(1) Vehicle electric range (Ri ): The electric ranges for the three EV
scenario under consideration as they are functions of annual EV
types were determined based on EIA (2012) and Gaines and
sales within a given scenario. On the other hand, PCj and Di,j were
Nelson (2010) and were kept fixed across all scenarios. The BEVs
assumed constant with time, though Di,j does vary across the sce-
were assumed to have 100 miles electric range, the HEVs were
narios as shown in Table S3.2.
assumed to have 4 miles electric range. In case of PHEVs, both
The number of battery packs in a given year t (Bt ) was obtained
10 and 40 miles electric ranges were considered.
from the EV battery MFA results discussed in the previous sec-
(2) Vehicle consumption rate (C): The consumption rate of an elec-
tion. PEi,t , percentage of waste batteries belonging to a given EV
tric vehicle can be defined as the electrical energy consumed
type in year t was based on the relative prevalence of each type
per mile of travel. Table S5.1 in the supplementary information
of EV sold, and thus entering the waste stream. Four prevalent
lists the energy consumption rates of electric vehicle models
lithium-ion cathode chemistries (i.e. j) were considered, namely,
in the recent years (according to EPA tests), and assump-
lithium cobalt oxide (LiCoO2 ), lithium manganese oxide (LiMn2 O4 ),
tions documented in the literature. Based on these values,
lithium iron phosphate (LiFePO4 ) and lithium nickel cobalt man-
the consumption rate of EVs was assumed to be 250 Wh/mile,
ganese (NCM) oxide, all having 18650 form factor cells (cylindrical
300 Wh/mile and 350 Wh/mile for the low, baseline and high
cells with 18 mm diameter and 65 mm length). The selection of this
scenarios, respectively.
form factor was based on data availability, with the recognition
(3) Battery efficiency (): The overall energy stored by the bat-
that results may change with alternative form factors, like the pris-
tery available for electric vehicle application depends on the
matic cells, expected to be used in most EVs. Sensitivity analysis
energy efficiency of the battery, so there is an inverse rela-
was conducted on this assumption as described in following sec-
tionship between efficiency and number of cells. The battery
tions. While the current LIB waste stream is almost entirely made
efficiency determines the amount of energy taken out dur-
up of consumer electronic batteries, which typically contain 100%
ing discharge after it was initially charged. The most common
LiCoO2 cathode chemistry (Wang et al., 2014), the distribution of
energy efficiency value for lithium-ion batteries reported in
cathode chemistries assumed in this paper (i.e. PCj ) for all three
literature is 90% (Van den Bossche et al., 2006; Gondelach,
scenarios is 10% LiCoO2 , 30% LiMn2 O4 , 30% LiFePO4 , and 30% NCM.
2010; Shiau et al., 2009; Karden et al., 2007; Tanaka et al.,
This distribution was selected to reflect that all three latter cath-
2001; Matheys et al., 2008). For the Tesla Roadster BEV, the
ode chemistries are likely candidates to replace the existing lithium
efficiency of the charge-discharge cycle of lithium-ion bat-
cobalt oxide based batteries for EV application, aside from limited
teries was reported to be approximately 86% (Eberhard and
application as in the case of Tesla vehicles. The number of cells per
Tarpenning, 2006). Campanari et al. (2009) have assumed a
battery pack for a given EV type using a given battery chemistry
92% efficiency of lithium-ion batteries used in electric vehi-
(Di,j ) varied with the scenario under consideration as described in
cles. According to Rydh and Sandén (2005), the efficiency of
the following section.
lithium-ion batteries can lie anywhere between 85% and 95%.
Hence, the EV battery efficiency was assumed to be 95%, 90%,
2.1.2.1. Determination of number of cells per battery pack (Di,j ). The and 85%, respectively for the low, baseline, and high scenarios
number of cells per LIB pack for a given EV type and a given battery (see Table 1).
chemistry (Di,j ) was estimated from the energy storage capacity (4) Available energy of EV battery (Ai ): The available energy of an
of the EV battery pack (Epacki ), dependent on the EV type i and EV battery is typically less than the total energy stored because
 K. Richa et al. / Resources, Conservation and Recycling 83 (2014) 63–76 69

3,000 8 billion cells

2,800

Number of EOL EV battery packs (1,000 units)

2,600
2,400
2,200 4.5 billion cells

2,000 3.3 billion cells

1,800
1,600
1,400
1,200 1.3 billion cells
880
1,000 million cells
800
600 30-500 million LIB cells

400
200 300 million cells

0
2020

2027

2036
2015
2016
2017
2018
2019

2021
2022
2023
2024
2025
2026

2028
2029
2030
2031
2032
2033
2034
2035

2037
2038
2039
2040
Year
Low Baseline High

Fig. 5. End-of-life Li-ion batteries generated from EV applications annually between 2015 and 2040.

the depth of discharge is restricted to preserve battery life and cells per battery pack for EV type i and battery chemistry type j
for safety purposes (Axsen et al., 2008). According to Srinivasan (Di,j ), the amount of any material y present in the EV battery waste
(2008), the available energy of a HEV battery is 20-30% of its stream for a given year t was estimated as,
total energy, while for a BEV or PHEV battery it could be as 
high as 70–80%. As per the Argonne National Laboratory [ANL] MOy,t = Bt × (PEi,t × PCj × Di,j × my,j ) (7)
(2012) BatPaC model, the energy utilized by a HEV battery is 25% i j
of the total energy, while it is 70–75% and 85–90% for a PHEV
my,j = mass of a given material y (aluminum, copper, lithium etc.) in
and BEV battery respectively. Based on these ranges, assump-
a lithium-ion cell of cathode chemistry j.
tions for the available energy percentage of the total battery
The variable my,j was obtained from the bill of materials of
energy for each of the vehicle type and for the three scenarios
lithium-ion cells of the four cathode chemistries from the disas-
are as shown in Table 1, which also includes assumptions for
sembly of 18650 lithium-ion cells (Wang et al., 2013) and remained
the vehicle consumption rate, battery efficiency and available
constant across the scenarios as well as with time (Table S4.1, SI).
energy for the three scenarios, and the calculation of EV battery
The other variables in this part of the model have been discussed
energy storage based on these factors.
in previous sections.

2.1.2.3. Cell energy storage (Ecellj ). The energy storage of 18650 2.2. Economic value of materials in EV battery waste stream
cells for the four battery chemistries considered was obtained as
the product of cell capacity and the nominal or average cell voltage The annual value of materials present in the EV battery waste
as described in section S3 in the SI. The cell capacity (mAh) was esti- stream was estimated using global spot prices (London Metal
mated as the product of the cathode mass and the specific capacity Exchange, 2012; Shanghai Metals Market, 2012) and USGS (2012)
(mAh/g) of lithium-ion cells for each of the four cell chemistries commodity values of LIB materials (Table S9.1, SI). This estima-
considered in the model. The specific capacity of the lithium-ion tion only included currently recycled materials (aluminum, cobalt,
cells was obtained from Dahn and Erlich (2011). The cathode mass copper, nickel, steel and iron) as well as high value materials not
of each of the cell types was estimated from their respective bill currently recycled in the U.S. but with high potential for recovery
of materials. The cell energy storage of each of the lithium-ion cell in the future (lithium and manganese) to calculate the “maximum
types was assumed to be constant with time as well as across the theoretical commodity value” of the EV battery waste stream. The
three scenarios. Using the approach described above, the final input future-oriented characterization of lithium and manganese as high
to the MFA model pertaining to number of cells per LIB pack was value materials is based on several factors, including current LIB
determined (summarized in Table S3.2 in SI, which distinguishes recycling efforts aimed at developing recovery processes for these
across scenarios, vehicle types, and cathode chemistries). materials (Paulino et al., 2008; Dunn et al., 2012; Zou et al., 2013;
Yang et al., 2013), limited lithium and manganese resources in the
2.1.3. Assessment of specific materials comprising each cell U.S. and the resultant dependence on import of these metals (USGS,
within the EV battery waste stream 2012), and the potential price rise of these metals with growing
In the final stage of modeling, the specific materials contained demand for EV LIBs. Manganese comprises about 20–25% of a typ-
in the battery cells were taken into account. Based on Bt , the total ical lithium-ion cell (Wang et al., 2013; ANL, 2012) making these
number of waste LIB packs in year t, the percentage of waste LIBs cells a viable source for recovery of manganese. Though lithium
belonging to EV type i in year t (PEi,t ), the percentage of battery constitutes only 1–2% of the total cell mass of typical LIBs (Wang
chemistry j in EV battery waste stream (PCj ), and the number of et al., 2013; ANL, 2012), considering an EV battery pack comprising
70 K. Richa et al. / Resources, Conservation and Recycling 83 (2014) 63–76

Fig. 6. Mass of Li-ion cells in EV battery waste stream. (In the pie-chart, “carbon” includes carbon black and graphite. “Others” include plastics, binders, electrolytes and other
non-metals like phosphorus etc.)

of thousands of cells, the amount of lithium available for recovery The battery waste flows were differentiated based on the “Type
would not be negligible. Although lithium carbonate is currently 1” and “Type 2” classifications of remaining life as described earlier
a lower cost input to LIB production (USGS, 2012), the forecast (Table 2, also summarized in Fig. 7 for the baseline). Characteriza-
increase in lithium demand by 2020 (Jobin et al., 2009) and poten- tion of EOL batteries into these categories provides some indication
tial lag in supply (Kushnir and Sandén, 2012) may trigger lithium of the volume of batteries with the highest potential for suitable
price rise in future. In fact, lithium spot prices of about $62/kg have reuse applications. For instance, type 2 EOL batteries still hypo-
been listed in the Shanghai Metals Market (2012). thetically have remaining EOL life, making them better suited
Recycling efficiencies of materials and the collection rate of for applications requiring high capacity. In each of the scenarios,
spent EV LIBs were not considered in estimating the commodity Type 2 batteries represent a sizeable fraction, and despite current
value of EV battery waste stream. Other materials in this waste hesitance surrounding reuse in vehicles, the number of batteries
stream that are unlikely to be recycled (graphite, electrolyte, plas- expected suggests that “re-matching” Type 2 batteries with older
tics, etc.) were excluded from this valuation. The baseline scenario vehicles or some other form of cascading use should be studied
MFA results were used as basis for these economic estimations. further.

3. Results and discussion 3.2. Lithium-ion cells and attendant material flows in the EV
battery waste stream
3.1. Estimation of EOL battery packs from electric vehicles
The EOL EV batteries generated on an annual basis would con-
Based on the parameters defined for each of the three scenarios, tain hundreds or even thousands of cells, each consisting of differ-
the number of EV LIB packs potentially entering the waste stream ent metals, carbonaceous materials (carbon black and graphite) and
on an annual basis was estimated (Fig. 5). other miscellaneous materials such as organic carbonates, lithium
While the three scenarios projected similar increase in EV bat- salts, binder, plastics, etc. Considering the parameters specified for
tery waste flows in the U.S. during the first five years of the analysis, the baseline scenario, approximately 3.3 billion individual lithium-
results quickly diverge due to differences in input sales and battery ion cells may be entering the waste stream annually by 2040. By
lifespans. As per the baseline, approximately 1.9 million LIB packs that point, the cumulative outflows between years 2015 and 2040
(each consisting of many cells) could be entering the waste stream would be on the order of 30 billion cells requiring EOL management.
annually by year 2040. However, considering the range from the The annual waste flows could be as low as 0.88 billion cells (low sce-
most conservative to most extreme estimates, the waste stream nario) or as high as 8 billion cells (high scenario) per year by 2040.
could hypothetically fall anywhere between 0.83 and 2.87 million
LIB packs per year by 2040. The cumulative baseline outflow of
Table 2
LIB packs between 2015 and 2040 (21 million packs) was approx-
Percentage of Type 1 and Type 2 EOL EV batteries accrued in the waste stream
imately two and a half times greater than the total number of EV between 2015 and 2040.
battery packs calculated in the “low” scenario (8.7 million packs)
and about two times fewer than that of the “high” scenario (40 Scenario Percentage of Type Percentage of Type
1 EOL batteries 2 EOL batteries
million packs). Of these LIB packs, between 27 and 35% would be
coming from all-electric and plug-in hybrid electric vehicles, and Low 57% 43%
Baseline 63% 37%
the remaining 65–73% were estimated to be from hybrid electric
High 62% 38%
vehicles, reflecting the projected sales of each vehicle type.
 K. Richa et al. / Resources, Conservation and Recycling 83 (2014) 63–76 71

Fig. 7. EV battery material inflow and outflows-Baseline Scenario. Thickness of each bar corresponds to the relative mass of material in each category.

For the baseline scenario in year 2020, the LIB waste stream 2” batteries may have been discarded before their true EV end-
could contain approximately 3,400 metric tons of lithium-ion cells of-life. Previously shown materials analysis (Fig. 6) distinguishes
just from EV application, which is about 4 times the estimated col- different materials contained in the battery waste stream, but the
lection volume of waste LIBs from consumer electronics in 2012 ultimate fate of these materials depends on whether an infrastruc-
(Wang et al., 2014). In terms of the resultant material flows, the ture and market exist for their recycling back into productive use.
range of scenarios indicate a total EV battery waste stream between The potential of each material to be recycled was determined by
0.33 and 4 million metric tons, with a baseline estimate of 1.3 mil- assessing current recycling practices and secondary markets avail-
lion metric tons generated cumulatively between 2015 and 2040. able for these materials (USGS, 2012). Based on the potential to be
Fig. 6 summarizes annual outflows of battery materials on five- recycled, the materials expected in the EV battery waste stream
year increments over the long and short term future (Extensive were categorized as currently recycled, high value-not recycled,
annual estimates of EOL EV LIB material outflows are provided in and low value-not recycled materials.
SI, section S6.3). In this initial estimate, the material-specific com- Of the estimated battery outflows, low value materials, which
position of that waste stream does not vary, and is also summarized are currently not being recycled and are not expected to be in
in Fig. 6. the future, could constitute 48% of the EV battery waste stream
In comparison with the small body of recent literature on and include graphite, carbon black, lithium hexafluorophosphate
lithium demand for EVs, this MFA predicted relatively conserva- (LiPF6 ), organic carbonates (such as ethylene carbonate or dimethyl
tive outflows, even for the “high” scenario. For instance, Gaines and carbonate), binder (polyvinylidene fluoride) and mixed plastics
Nelson (2010) estimated a maximum waste flow of 20,000 metric (polypropylene, polyethylene). Apart from plastics, none of these
tons of lithium in 2040 from “optimistic” EV deployment, a predic- materials have a secondary market at present and it would not
tion about 10 times greater than our baseline scenario estimates for be economically viable to recover them from the waste stream.
that year. To put our estimates in a global context, this study fore- Moreover, as a mixed grade of plastics would be present in the
cast cumulative lithium outflows between 2020 and 2040 between battery waste stream, their recovery would not be likely due to
4.5 thousand to 55 thousand metric tons for the U.S. On a global high contamination. Because these materials are not suitable for
basis, Gruber et al. (2011) estimated 860 thousand metric tons recycling, infrastructure must be equipped to accommodate their
of potentially recoverable lithium from EV batteries in the same introduction to landfills or other disposal routes. As such, relevant
time frame (with 100% recycling participation and 100% lithium environmental and health impacts should be anticipated. The car-
recovery). bonaceous material present in the EV battery waste stream could
The disparity observed between the low and high scenario in raise concern in the future owing to their large quantity in the
this paper is indicative of the variability in estimates of EV sales, waste stream (Fig. 6) and knowledge of potential health impacts
the battery lifespan (and resultant need for replacement battery of particulate carbon (e.g., exposure to graphite dust can adversely
packs, particularly in the high scenario) as well as the parameters affect respiratory system and pulmonary function (NIOSH, 2007)).
determining the number of cells per EV battery pack. Even with The electrolyte used in LIBs can have toxicity concerns as well.
these uncertainties, we can begin to analyze results further, using For instance, the electrolyte salt LiPF6 is a hygroscopic substance
the baseline scenario as a focal point (to minimize the amount of and in presence of moist air or water forms hydrogen fluoride
data presented in the main text). The baseline scenario was further gas (Archuleta, 1995), which has severe environmental risks and
characterized on the basis of battery inputs, outputs, and mate- toxicity concerns (EPA, n.d.). Similarly, organic carbonates used as
rial characteristics. Fig. 7 summarizes these characteristics for the electrolyte solvents are mildly toxic, volatile and flammable com-
cumulative input of LIBs in electric vehicles between 2009 and pounds, producing toxic fumes on decomposition (Vimmerstedt
2034, and the net EOL battery outflows between 2015 and 2040. et al., 1995). Environmental impacts of EV battery waste could also
The majority (80%) of new batteries entering use would be be a concern due to the non-biodegradability of binder and other
paired with new EVs sold in the market, while the remaining 20% plastics in lithium-ion cells.
would be replacement batteries for existing in-use EVs (Fig. 7). Another 42% of the materials in the cumulative EV battery
About 63% of the batteries leaving use were “Type 1,” with no waste stream would include metals that are currently and expected
remaining life for EV applications; while the remaining 37% “Type to continue being recycled according to statistics from the USGS
72 K. Richa et al. / Resources, Conservation and Recycling 83 (2014) 63–76

(2012). This fraction includes aluminum (57,000 metric tons), 3.4. Uncertainty analysis
cobalt (52,000 metric tons), copper (75,000 metric tons), nickel
(32,000 metric tons), steel (295,000 metric tons) and iron (43,300 3.4.1. EV sales and battery lifespan
metric tons). These material masses are on a cumulative basis It was established by the range of scenario results that the
between 2015 and 2040. The remaining 10% of the EV battery volume of LIBs in the EV battery waste stream would be highly
waste stream would include two high value materials that are cur- dependent on EV sales and the actual battery lifespan. The sales
rently not recycled in the U.S., i.e., lithium (18,000 metric tons) and of electric vehicles will depend on a number of factors in the mid-
manganese (116,000 metric tons). Many of these metals (lithium, and long-term, such as oil prices, battery and vehicle cost, EV and
aluminum, nickel and cobalt) have high embodied energy when battery technology, government subsidies, policies and regulations
extracted from virgin resources (ecoinvent Centre, 2007). Hence, etc. (EIA, 2012). The lifespan of lithium-ion batteries in EV applica-
recycling of LIBs offers a dual benefit: avoided energy inputs for tion will depend on battery technology progress as well as usage
production of primary metals and potential economic revenue patterns at the consumer level. Long battery lifespans would have a
from material recovery, which is particularly high for cathode two-fold benefit, first reducing the need for a second (or even third)
chemistries like Li2 CO3 and NCM that contain 10–17% by weight replacement battery and thereby reducing the cost of ownership
of high-value cobalt. of electric vehicles, and second, raising the potential for post-EV
battery reuse, which can also defray costs across the battery life
cycle.
3.3. Economic value of materials in EV battery waste stream Uncertainty analysis was performed to tease apart the role of
sales and lifespan parameters on LIB waste flows, by holding one
Considering the baseline scenario with a mix of lithium-ion cell parameter constant and varying the other (SI section S7.1). When
chemistries, the total EV LIB waste stream may contain materi- the EV sales estimates are held constant at the baseline level, cumu-
als valued at approximately 3.8 billion USD on a cumulative basis lative (2015–2040) LIB outflows increased 16% or decreased 15%
between 2015 and 2040. This estimate is the maximum theoretical from the shortest lifespan (“high” scenario) to the longest (“low”
commodity value of the EV battery waste stream considering the scenario). On the other hand, when the baseline EV LIB lifespan
potential value for materials that currently have recycling infra- distribution was combined with the high and low scenario EV
structure in the U.S, and does not take into account material losses sales figures, cumulative (2015–2040) outflows of EOL packs could
that would occur due to recycling inefficiencies. The total possible increase by as much as 62% and decrease by 52%, respectively. It is
waste stream value would be increased by over 1.5 billion USD if Li evident that even though the battery lifespan distribution plays a
and Mn are also included. role in influencing the volume of EV battery waste stream, electric
The actual economic value of the EV battery waste stream would vehicle sales will be the governing factor influencing EV LIB waste
depend on the LIB collection rates, the recovery rates of the vari- flows in the future.
ous materials present in the stream, and the cost of recycling itself. When a longer EV lifespan of 16 years was tested for model
Considering recent recycling efficiencies (see SI, Table S10.1), com- sensitivity, the cumulative (2015–2040) outflows of these batteries
modity value of approximately 3 billion USD could be obtained into the waste stream changed by less than 2%, although the annual
between 2015 and 2040 by recovery of metals such as aluminum, waste stream volumes varied, as shown in SI section S7.3. Further,
copper, nickel, cobalt, iron and steel assuming that 100% of batter- the percentage of Type 1 and Type 2 EV batteries also changed with
ies in the waste stream can be collected for recycling. Wang et al. increasing lifespan of EVs. For instance, when a longer EV lifespan
(2014) analyzed the profitability of LIB recycling facilities for sev- of 16 years was assumed for the baseline scenario, the percent-
eral possible future co-mingled LIB waste streams based on the age of Type 2 EV batteries estimated to accrue in the waste stream
current recycling efficiency of materials in LIBs: the potential value between 2015 and 2040 reduced from 37% to 23% (detailed analysis
from recycling one metric ton of LIBs ranged from $860 for LiMn2 O4 in SI section S7.3).
cathode batteries to $8,900 for LiCoO2 cathode batteries. Continued
development of advanced separation processes could increase the 3.4.2. Cell and battery energy storage and battery pack
recycling efficiencies of materials present in EV LIBs and hence the components
economic motivation for recovering materials from these batteries. The energy storage of the battery pack also plays an important
For example, a 10% improvement over current recycling efficiency role in determining the amount of cells per pack, and thus the mate-
for cobalt could raise the recycling revenue by 9% for cobalt based rials present in the EV battery waste stream. A sensitivity analysis
LIBs while a 10% improvement in copper recycling efficiency would was conducted wherein the baseline cumulative outflow of EOL EV
only improve revenue from LIB recycling by 1–5%, depending on the battery packs was held constant while the estimated EV battery
cathode chemistry (Wang et al., 2014). pack energy assumptions were varied between the low and high
The materials potentially recoverable by EV LIB recycling could scenarios (SI, section S7.2). The resulting estimates of total mate-
be used as inputs to the parent battery industry, as this sector is pre- rial mass of the EV battery waste stream decreased by up to 29%
dicted to become more resource intensive as vehicle deployment or increased by up to 40% when the low or high scenario cells per
increases. Increasing availability of secondary material sources pack assumptions were applied, as compared to baseline flows.
would reduce U.S. dependency on foreign resources in the long run. Hence, it follows that battery and EV technology (in terms of
Gaines and Nelson (2010) estimated that recycling LIBs could meet electric miles and vehicle mileage) would play a major role in gov-
almost 50% of the lithium required for battery production in the U.S. erning flows of EV batteries in the waste stream. Throughout the EV
by 2040. However, the recyclability of the EV battery waste stream sales forecast timeline considered in the model (2009–2034), the
and hence, the economic gains from battery recycling is likely to same battery chemistries were assumed and the energy storage by
depend on the battery technology prevalent in the future in terms individual cells was held constant. According to Srinivasan (2008),
of cathode chemistry as well as the form factor of the lithium-ion the energy density of lithium-ion batteries has been increasing
cells used in these batteries. At present there is significant uncer- at the rate of approximately 5% per year over the last one and a
tainty in this domain, which is analyzed in the subsequent sections, half decade. The average energy density of a typical 18650 cell is
along with the uncertainty due to differences in MFA parameters approximately 200 Wh/kg (Howard and Spotnitz, 2007). Accord-
such as EV sales, battery lifespan and number of cells per EV battery ing to Srinivasan and Lipp (2003), when lithium-ion batteries were
pack. introduced in the early 1990s, this number was around 90 Wh/kg.
 K. Richa et al. / Resources, Conservation and Recycling 83 (2014) 63–76 73

Fig. 8. Recyclability of EV battery waste stream under different scenarios of Li-ion cell cathode chemistry and form factor (cumulative flows from 2015 to 2040).

In other words, technological innovation has doubled the energy (A) Scenarios in which a single dominant lithium-ion cathode
density of these batteries. It is expected that this trend will con- chemistry (18650 form-factor) would be employed in all EV
tinue in the future, with ongoing research and development to batteries. The four candidate chemistries were Li2 CO3 (LCO),
introduce nano-materials and mixed-metal technology for higher LiMn2 O4 (LMO), LiFePO4 (LFP) and mixed metal (NCM), each
energy density (Ritchie and Howard, 2006; Howard and Spotnitz, modeled individually as a dominant cathode chemistry.
2007). As this technology advances, fewer cells per pack and/or (B) A scenario in which prismatic cells instead of 18650 cells
less material per cell may be achievable, which may reduce total were used. The chemistry mix of EV battery waste stream was
material flows to the EV battery waste stream. assumed to be same as in the 18650 scenario (10% LCO, 30% each
The mass and composition of the EV battery waste stream of LMO, LFP and NCM cells), however here only the LCO cells
modeled here has considered only the cells within EV battery packs. were of 18650 form factor (consistent with adoption by some
For a typical Tesla Roadster EV battery, with 6800 cells weigh- auto makers) while the remaining 90% of the LIBs consisted of
ing approximately 46 g each, the total cell weight is about 313 kg, prismatic cells. See SI. for scenario details.
but the entire battery mass is about 450 kg (Berdichevsky et al.,
2006). The difference in weight is attributed to the battery pack When compared to the baseline EV LIB outflows (a mix of poten-
casing, module components, electronic parts, thermal insulation, tial chemistries in 18650 cells), the number of lithium-ion cells in
etc. (Dunn et al., 2012), which can account for anywhere between the waste stream was roughly the same for the different dominant
10 and 30% of the EV battery weight. Considering this entire battery chemistry scenarios. The LFP scenario resulted in the highest num-
system, the actual material flow entering the waste stream would ber of waste cells, approximately 35 billion, cumulatively between
be even greater than estimates calculated here. However, including 2015 and 2040 (as compared to the baseline 30 billion cells).
other EV battery pack components would require further model- When considering prismatic cells, results showed an interesting
ing, as these components may have higher reuse potential than the dichotomy: the total number of cells in the EV battery waste stream
batteries themselves and may not enter the waste stream at the would reduce drastically if most EVs employed prismatic cells to
same time as the LIB cells (Cready et al., 2003). almost 4.4 billion cells in the waste stream between 2015 and
2040, but the net flow and type of materials into the waste stream
would remain relatively constant. This consistency held across all
3.4.3. Battery chemistry and form factors different scenarios of cathode chemistry and form factor, which
All results shown to this point have followed an assumed mix- ranged between 1 and 1.5 million metric tons of battery waste on
ture of different cathode chemistries, and are based only on 18650 a cumulative basis between 2015 and 2040 (Fig. 8).
(cylindrical) form factors. In the future, composition of the EV Further, the recyclability of the EV battery waste stream would
battery waste stream will depend heavily on the actual cathode also vary with the battery chemistry and form factor (Fig. 8). For
chemistries and form factors selected by auto manufacturers. For instance, if the LiMn2 O4 chemistry is predominantly used, then
example, new EV models such as the GM Volt and Nissan Leaf the EV battery waste stream would contain negligible amount of
employ prismatic cells, while the 18650 form factor continues currently recyclable materials and large quantities of low value
to be used in the Tesla roadster BEVs. Though this MFA model materials that are not recycled. However, its recycling can generate
enables a reasonable approximation of the material and economic value if the currently non-recycled high value materials like lithium
flows of the EV battery waste stream, there is a need to assess and manganese can be recovered (Fig. 9). Similarly, even though
the role that a single dominant chemistry and prismatic form both the form factor scenarios (Chemistry mix “Base Case” and
factor could play in determining the volume, composition, eco- 90% prismatic) consisted of the same distribution of LIBs belonging
nomic value, and recycling potential of this waste stream. Using to the four cathode chemistries in the waste stream, the fraction
the baseline scenario estimation for EOL LIB packs, the num- of recyclable materials is slightly higher in the case of cylindrical
ber of cells and material mass were estimated for the following cells, which require more metallic casing components (typically
cases: aluminum or steel).
74 K. Richa et al. / Resources, Conservation and Recycling 83 (2014) 63–76

Fig. 9. Cumulative material value of EV battery waste stream (2015–2040).

An important caveat to these findings is that the high percent- high as 340,000 metric tons by 2040. Because of the high volume,
age by weight of recycled materials in the EV battery waste stream complexity and variety of materials forecast in the EV battery waste
may not translate into high economic gains from LIB recycling. For stream, it is evident that multiple waste management routes must
instance, even though 58% of a LiFePO4 cell waste stream consists be developed for EOL LIBs from electric vehicles:
of currently recycled materials, the relative economic value of this
stream is lower than any other chemistry (Fig. 9). While the relative
(1) reuse avenues for battery cells and packs with remaining life,
volumes of recycled materials are higher in the case of all cylindri-
(2) recycling infrastructure capable of recovering high value mate-
cal cells, as compared to the prismatic form, there is no significant
rial from multiple battery chemistries, and
difference in the total commodity value of materials in the waste
(3) safe disposal routes for materials with minimal or no secondary
stream associated with these two scenarios.
value or recovery infrastructure.
In fact, the economic feasibility of EV lithium-ion battery
recycling in the future would not only depend on collection and
recycling efficiencies, but also on the chemistries selected for EV Results also indicate that high variability in the potential eco-
battery manufacturing and ultimately ending up in the waste nomic value associated with the projected LIB waste stream may
stream. A large scale use of LCO and NCM chemistries for EV bat- pose challenges for development of recycling infrastructure. At
teries would translate into high economic values of the EV battery present, profit from LIB recycling is constrained by high collec-
waste. In the LCO and NCM chemistry scenarios, currently recycled tion and processing costs (Wang et al., 2014). Currently, there is
materials would constitute about 50% of the materials in the battery no federal regulation that mandates LIB recycling, and only two
waste stream by mass, but could account for 86% of the economic states – California and New York – have passed regulations banning
value of that stream. However, as battery manufacturers shift to landfill of these batteries. To overcome potential economic con-
cheaper chemistries such as LMO and LFP, to improve performance straints of LIB recycling, particularly for less valuable, non-cobalt
and avoid high cost and scarcity of cobalt resources (Nishi, 2001), chemistries, the recovery process and infrastructure may require
the resulting value of the currently recycled portion of the waste policy intervention to reach economies of scale.
stream could be reduced to as low as 340 million USD, on a cumu- Apart from economics, environmental health and safety may
lative basis (2015–2040). also motivate policy attention to future EOL LIB management. The
Cost efficient recycling procedures for the recovery of high value absence of consistent infrastructure and regulations for lithium-ion
materials, like lithium and manganese, which are currently not battery recycling may increase the potential risk of environmental
recycled in the U.S. would add some incentive towards recycling of impact due to EOL EV batteries. Though the state of California clas-
economically unattractive LIB chemistries. In the LMO chemistry sifies them as hazardous due to the presence of cobalt, LIB wastes
scenario, currently non-recycled high value materials constitute are included under EPA’s Universal Waste Rule (Gaines and Cuenca,
22% of the waste flows by mass, but account for 87% by value. 2000) and are in general not considered to be hazardous for the
Even more extreme, in the LFP battery chemistry scenario, the cur- environment due to absence of toxic elements like lead, mercury or
rently non-recycled high value materials accounted for only 1% of cadmium. However, landfill of EV LIBs may introduce environmen-
the waste EV battery cells by mass, but could make up to 38% of tal risks due to leakage of organic electrolytes, presence of heavy
the total material value. Although Wang et al. (2014) conclude that metals such as copper and nickel (Shin et al., 2005), reactive lithium
for a LIB recycling facility to be profitable, the proportion of LiCoO2 salts, and large quantity of carbonaceous materials (graphite and
cathode batteries in the waste stream needs to be 21%, improved carbon black).
recycling processes in the future is expected to improve the overall A number of uncertainties still exist, and exact estimation of
profitability of recycling EOL LIBs. future waste flows will depend on the ability to further refine the
forecasts of EV sales, battery and EV lifespan, and trajectories of
4. Conclusions and implications battery technology deployment. Waiting until such refinements are
possible, though, presents a risk of not allowing sufficient time for
It is clear that EV batteries will emerge as a future waste man- domestic infrastructure and policies to react to the emergence of
agement challenge, with projected annual waste flows reaching as a full scale battery waste stream. Thus, proactive advancement of
 K. Richa et al. / Resources, Conservation and Recycling 83 (2014) 63–76 75

a robust EOL battery reuse, recycling, and disposal system will be EPA (n.d.). Hydrogen fluoride study. Final report to Congress: Section 112(n)
required to handle the variety and volume of materials expected. (6) Clean Air Act as amended. Retrieved from http://www.epa.gov/
osweroe1/docs/chem/hydro.pdf. 2013.
Moreover, the MFA model provided here can be adapted to extend Frost and Sullivan. World hybrid electric/electric vehicle battery market, 9836-27,
the analysis of LIB wastes as more definitive data become available. May 2009.
Fu H. Implementations of electric vehicle system based on solar energy in
Singapore assessment of lithium-ion batteries for automobiles. In: Doctoral
Acknowledgements dissertation. Massachusetts Institute of Technology; 2009, Retrieved from
http://dspace.mit.edu/handle/1721.1/54200.
Gaines L, Nelson P. Lithium-ion batteries: examining material demand and recycling
The research team would like to acknowledge funding from the issues. Argonne, IL: Argonne National Laboratory; 2010.
Golisano Institute for Sustainability, the New York State Energy Gaines L, Cuenca R. Costs of lithium-ion batteries for vehicles. Argonne National
Laboratories; 2000, Retrieved from http://ntl.bts.gov/lib/14000/14700/14729/
Research and Development Authority (NYSERDA) under PON DE00761281.pdf.
18503, the National Science Foundation Environmental Sustaina- Giffi C, Gardner M, Hill R, Hasegawa M. Gaining traction – a customer view of electric
bility Directorate under Award No. CBET-1254688, the National vehicle mass adoption in the US automotive market. Deloitte Development LLC;
January, 2010, Retrieved from http://www.deloitte.com.br/publicacoes/2007/
Science Foundation Environmental Health and Safety of Nanoma- MFG.Gaining Traction customer view of electric vehicle mass adoption.pdf
terials Directorate under Award No. CBET-1133425, and the New Gondelach S. Current and future developments of batteries for electric cars—an
York State Pollution Prevention Institute. We would also like to analysis; November, 2010, Retrieved from http://igitur-archive.library.uu.nl/
student-theses/2010-1126-200244/UUindex.html.
acknowledge support from the New York Battery and Energy Stor- Graedel TE, Allenby BR. Industrial ecology and sustainable engineering. Upper Saddle
age Technology (NY-BEST) Consortium and Ultralife Corporation. River, NJ, USA: Prentice Hall; 2010.
Greene DL, DeCicco J. Engineering-economic analyses of automotive fuel economy
potential in the United States. Annual Review of Energy and the Environment
Appendix A. Supplementary data 2000;25(1):477–535.
Greene DL, Patterson PD, Singh M, Li J. Feebates, rebates and gas-guzzler
taxes: a study of incentives for increased fuel economy. Energy Policy
Supplementary data associated with this article can be found, 2005;33(6):757–75.
in the online version, at http://dx.doi.org/10.1016/j.resconrec. Grosjean C, Miranda PH, Perrin M, Poggi P. Assessment of world lithium resources
and consequences of their geographic distribution on the expected development
2013.11.008. of the electric vehicle industry. Renewable and Sustainable Energy Reviews
2012;16(3):1735–44.
Gruber P, Medina P, Keoleian G, Kesler S, Everson M, Wallington T. Global lithium
References availability a constraint for electric vehicles? Journal of Industrial Ecology
2011;15(5):760–75.
Alonso E, Sherman AM, Wallington TJ, Everson MP, Field FR, Roth R, et al. Evaluating Harper EM, Kavlak G, Graedel TE. Tracking the metal of the goblins: cobalt’s cycle of
rare earth element availability: a case with revolutionary demand from clean use. Environmental Science & Technology 2012;46(2):1079–86.
technologies. Environmental Science & Technology 2012;46(6):3406–14. Howard W, Spotnitz R. Theoretical evaluation of high-energy lithium metal
Anderman M. The 2007 advanced automotive battery and ultracapacitor industry phosphate cathode materials in Li-ion batteries. Journal of Power Sources
report. Advanced Automotive Batteries; 2007. 2007;165(2):887–91.
Archuleta MM. Toxicity of materials used in the manufacture of lithium batteries. Huang S, Hodge BMS, Taheripour F, Pekny JF, Reklaitis GV, Tyner WE. The effects
Journal of Power Sources 1995;54(1):138–42. of electricity pricing on PHEV competitiveness. Energy Policy 2011;39(3):
Argonne National Laboratory. BatPaC model; 2012, Retrieved from http:// 1552–61.
www.cse.anl.gov/batpac. Humphrey J, Sargent D, Schuster J, Marshall M, Omotoso M, Dunne T. Drive Green
Axsen J, Burke AF, Kurani KS. Batteries for plug-in hybrid electric Vehicles (PHEVs): 2020: more hope than reality. J.D. Power and Associates; 2010, Retrieved from
goals and the state of technology circa 2008. Davis: Institute of Transportation http://www.altomelbilen.dk/presentationer/DriveGreen2020.pdf.
Studies, University of California; 2008. International Energy Agency. Technology roadmap. Electric and plug-in hybrid
Babbitt CW, Williams E, Kahhat R. Institutional disposition and management of end- electric vehicles; 2011, Retrieved from http://www.iea.org/publications/
of-life electronics. Environmental Science & Technology 2011;45(12):5366–72. freepublications/publication/EV PHEV Roadmap.pdf.
Berdichevsky G, Kelty K, Straubel JB, Toomre E. The Tesla Roadster bat- Jobin P, Yamaguchi J, Balter G, Ceraso CJ, Chan A, Eggers D, et al. Electric vehicles.
tery system. Tesla Motors; August, 2006, Retrieved from http:// Credit Suisse, Global Equity Research, Energy Technology/Auto Parts & Equip-
webarchive.teslamotors.com/display data/TeslaRoadsterBatterySystem.pdf. ment; 2009.
Brunner PH, Rechberger H. Practical handbook of material flow analysis. New York: Kalhammer FR, Kamath H, Duvall M, Alexander M, Jungers B. Plug-in hybrid elec-
Lewis; 2004. p. 317. tric vehicles: promise, issues and prospects. In: EVS24 international battery,
Burke A. Performance, charging, and second-use considerations for lithium hybrid and fuel cell electric vehicle symposium. Norway: Stavanger; May, 2009.
batteries for plug-in electric vehicles. Institute of Transportation p. 1–11.
Studies, University of California-Davis; 2009, Retrieved from http:// Kang HY, Schoenung JM. Estimation of future outflows and infrastructure needed to
escholarship.org/uc/item/2xf263qp - page-1. recycle personal computer systems in California. Journal of Hazardous Materials
Campanari S, Manzolini G, Garcia de la Iglesia F. Energy analysis of electric vehicles 2006;137(2):1165–74.
using batteries or fuel cells through well-to-wheel driving cycle simulations. Karden E, Ploumen S, Fricke B, Miller T, Snyder K. Energy storage devices
Journal of Power Sources 2009;186(2):464–77. for future hybrid electric vehicles. Journal of Power Sources 2007;168(1):
Chalk SG, Miller JF. Key challenges and recent progress in batteries, fuel cells, 2–11.
and hydrogen storage for clean energy systems. Journal of Power Sources Kumar V, Sutherland JW. Sustainability of the automotive recycling infrastructure:
2006;159(1):73–80. review of current research and identification of future challenges. International
Chang T, You S, Yu B, Yao K. A material flow of lithium batteries in Taiwan. Journal Journal of Sustainable Manufacturing 2008;1(1):145–67.
of Hazardous Materials 2009;163(2–3):910–5. Kushnir D, Sandén BA. The time dimension and lithium resource constraints for
Cready E, Lippert J, Pihl J, Weinstock I, Symons P, Jungst RG. Technical and economic electric vehicles. Resources Policy 2012;37(1):93–103.
feasibility of applying used EV batteries in stationary applications. SAND2002- Lemp JD, Kockelman KM. Quantifying the external costs of vehicle use: evidence
4084. Albuquerque, NM: Sandia National Laboratories; 2003. from America’s top-selling light-duty models. Transportation Research Part D:
Dahn J, Erlich GM. Lithium-ion batteries. In: Reddy TB, editor. Linden’s handbook of Transport and Environment 2008;13(8):491–504.
batteries. 4th edition 2011. p. 26. London Metal Exchange [LME]. Pricing and data; 2012, Accessed July 2012 from
Dinger A, Martin R, Mosquet X, Rabl M, Rizoulis D, Russo M, et al. Batter- http://www.lme.com/.
ies for electric cars: challenges, opportunities, and the outlook to 2020. Madani A. In: HEV, P-HEV& EV Market Trends 2008–2020: Battery is the
The Boston Consulting Group, Technical Report; 2010, Retrieved from key! Presentation by Avicenne Développment for BATTERIES 2009: the
http://www.bcg.com/documents/file36615.pdf. international power supply conference and exhibit; 2009, Retrieved from
Dunn JB, Gaines L, Barnes M, Sullivan J, Wang MQ. Material and energy flows in mate- http://www.cars21.com/web/assets/link/Ali Madani presentation.pdf.
rials production, assembly, and end-of-life stages of the life cycle of lithium-ion Marano V, Onori S, Guezennec Y, Rizzoni G, Madella N. Lithium-ion batteries life
batteries, ANL/ESD/12-3. Argonne National Laboratory; 2012. estimation for plug-in hybrid electric vehicles. In: 2009 IEEE vehicle power and
Eberhard M, Tarpenning M. The 21st century electric car. Tesla Motors Inc; propulsion conference, vols. 1–3, no. 1; 2009. p. 476–83.
October, 2006, Retrieved from http://www.stanford.edu/group/greendorm/ Matheys J, Van Mierlo J, Timmermans JM, Van den Bossche P. Life-cycle assessment of
participate/cee124/TeslaReading.pdf. batteries in the context of the EU Directive on end-of-life vehicles. International
ecoinvent Centre. ecoinvent data v2.0. ecoinvent reports no. 1-25. Dübendorf: Swiss Journal of Vehicle Design 2008;46(2):189–203.
Centre for Life Cycle Inventories; 2007. Michaelis P, Jackson T. Material and energy flow through the UK iron and steel
EIA. Annual Energy Outlook 2012. In: Light-duty vehicle sales by technology – United sector: part 2: 1994–2019. Resources, Conservation and Recycling 2000;29(3):
States; 2012, http://www.eia.gov/forecasts/archive/aeo12/. 209–30.
76 K. Richa et al. / Resources, Conservation and Recycling 83 (2014) 63–76

National Institute for Occupational Safety [NIOSH]. NIOSH pocket guide to processing in the informal sector in Delhi. Environmental Impact Assessment
chemical hazards; 2007, Retrieved from http://www.cdc.gov/niosh/docs/2005- Review 2005;25(5):472–91.
149/pdfs/2005-149.pdf. Tanaka T, Ohta K, Arai N. Year 2000 R&D status of large-scale lithium-ion secondary
Leduc G, Munoz A. Plug-in hybrid and battery-electric vehicles: state of the batteries in the national project of Japan. Journal of Power Sources 2001;97:2–6.
research and development and comparative analysis of energy and cost effi- U.S. DOE. The recovery act: transforming America’s transportation sector. Batteries
ciency. In: Technical note. Institute for Prospective Technological Studies, Joint and electric vehicles; July, 2010, Retrieved from http://www.whitehouse.gov/
Research Centre, European Union; 2009, Retrieved from ftp://ftp.jrc.es/pub/ files/documents/Battery-and-Electric-Vehicle-Report-FINAL.pdf.
EURdoc/JRC54699 TN.pdf. USGS. Lithium, U.S. Geological Survey, Mineral Commodity Summaries
Neubauer J, Pesaran A. The ability of battery second use strategies to impact plug-in 2009; 2009, Retrieved from http://minerals.usgs.gov/minerals/pubs/
electric vehicle prices and serve utility energy storage applications. Journal of commodity/lithium/mcs-2009-lithi.pdf.
Power Sources 2011;196(23):10351–8. USGS. Mineral Commodity Summaries 2012. U.S. Geological Survey 2012; 2012.
Nishi Y. Lithium-ion secondary batteries; past 10 years and the future. Journal of Van den Bossche P, Vergels F, Van Mierlo J, Matheys J, Van Autenboer W. SUBAT:
Power Sources 2001;100(1):101–6. an assessment of sustainable battery technology. Journal of Power Sources
Oguchi M, Kameya T, Yagi S, Urano K. Product flow analysis of various consumer 2006;162(2):913–9.
durables in Japan. Resources. Conservation and Recycling 2008;52(3):463–80. Vimmerstedt LJ, Ring S, Hammel CJ. Current status of environmental, health, and
Park J, Hong SJ, Kim I, Lee JY, Hur T. Dynamic material flow analysis of steel resources safety issues of lithium-ion electric vehicle batteries (No. NREL/TP–463-7673).
in Korea. Resources, Conservation and Recycling 2011;55(4):456–62. Golden, CO, United States: National Renewable Energy Lab; 1995, Retrieved from
Paulino JF, Busnardo NG, Afonso JC. Recovery of valuable elements from spent Li- http://www.osti.gov/energycitations/product.biblio.jsp?osti id=135056.
batteries. Journal of Hazardous Materials 2008;150(3):843–9. Wang X, Gaustad G. Prioritizing material recovery for end-of-life printed circuit
Pauliuk S, Wang T, Müller DB. Moving toward the circular economy: the role boards. Waste Management 2012;32(10):1903–13.
of stocks in the Chinese steel cycle. Environmental Science & Technology Wang X, Gaustad G, Babbitt CW, Richa K. Economies of scale for future lithium-
2011;46(1):148–54. ion battery recycling infrastructure. Resources, Conservation and Recycling
Ritchie A, Howard W. Recent developments and likely advances in lithium-ion bat- 2014;83:53–62.
teries. Journal of Power Sources 2006;162(2):809–12. Wang X, Gaustad G, Babbitt CW, Bailey C, Ganter M, Landi B. Economic and environ-
Rydh CJ, Sandén BA. Energy analysis of batteries in photovoltaic systems. Part I: mental characterization of an evolving Li-ion battery waste stream. Journal of
performance and energy requirements. Energy Conversion and Management Environmental Management 2013, in press.
2005;46(11):1957–79. Watabe T, Mori M. LIB materials industry. Automotive LIB materials get set
Shanghai Metals Market. Minor metals prices; 2012, Accessed July 2012 from for growth phase in 2011. In: Global markets research. Deutsche Bank
http://www.metal.com/metals/minor-metals/prices. Group; 2011, Retrieved from http://www.fullermoney.com/content/2012-02-
Shiau CSN, Samaras C, Hauffe R, Michalek JJ. Impact of battery weight and charging 02/LiB12611.pdf.
patterns on the economic and environmental benefits of plug-in hybrid vehicles. Widmer R, Oswald-Krapf H, Sinha-Khetriwal D, Schnellmann M, Böni H.
Energy Policy 2009;37(7):2653–63. Global perspectives on e-waste. Environmental Impact Assessment Review
Shin SM, Kim NH, Sohn JS, Yang DH, Kim YH. Development of a metal recovery 2005;25(5):436–58.
process from Li-ion battery wastes. Hydrometallurgy 2005;79(3):172–81. Williams BD, Lipman TE. Strategy for overcoming cost hurdles of plug-in-hybrid bat-
Shrestha S, Martin S, Zhang J. Industry note. Alternative energy and tery in California. Transportation Research Record: Journal of the Transportation
infrastructure. In: Lazard Capital Markets; 2010, Retrieved from Research Board 2010;2191(1):59–66.
http://www.eosenergystorage.com/articles/LazardMarketReport 2010-03.pdf. Williams E, Kahhat R, Allenby B, Kavazanjian E, Kim J, Xu M. Environmental, social,
Srinivasan V. Batteries for vehicular applications. AIP Conference Proceedings and economic implications of global reuse and recycling of personal computers.
2008;1044:283–96, Retrieved from http://scitation.aip.org/content/aip/ Environmental Science & Technology 2008;42(17):6446–54.
proceeding/aipcp/10.1063/1.2993726 Yaksic A, Tilton JE. Using the cumulative availability curve to assess the threat
Srinivasan V, Lipp L. Report on the electrolytic industries for the year 2002. Journal of mineral depletion: The case of lithium. Resources Policy 2009;34(4):
of the Electrochemical Society 2003;150(12):K15–38. 185–94.
Steinmetz J, Shankar R. Autos & auto related: plug in hybrids: the next automotive Yang Z, Mei Z, Xu F, Yao Y, Zhang W, Qi W, et al. Different types of MnO2 recov-
revolution. New York: Morgan Stanley Research, North America; March, 2008. ered from spent LiMn2 O4 batteries and their application in electrochemical
p. 13. capacitors. Journal of Materials Science 2013;48(6):2512–9.
Steubing B, Böni H, Schluep M, Silva U, Ludwig C. Assessing computer Yoshida A, Tasaki T, Terazono A. Material flow analysis of used personal computers
waste generation in Chile using material flow analysis. Waste Management in Japan. Waste Management 2009;29(5):1602–14.
2010;30(3):473–82. Zou H, Gratz E, Apelian D, Wang Y. A novel method to recycle mixed
Streicher-Porte M, Widmer R, Jain A, Bader HP, Scheidegger R, Kytzia S. Key cathode materials for lithium-ion batteries. Green Chemistry 2013;15(5):
drivers of the e-waste recycling system: assessing and modelling e-waste 1183–91.