Journal of Power Sources 343 (2017) 284e295

Contents lists available at ScienceDirect

Journal of Power Sources

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

Life cycle assessment of lithium sulfur battery for electric vehicles

Yelin Deng a, b, Jianyang Li b, Tonghui Li b, Xianfeng Gao b, Chris Yuan b, c, *
a
Department of Civil and Environmental Engineering, Soochow University, Suzhou 215131, China
b
Department of Mechanical Engineering, University of Wisconsin, Milwaukee, WI 53211, United States
c
Department of Mechanical and Aerospace Engineering, Case Western Reserve University, Cleveland, OH 44106, United States

h i g h l i g h t s

A hybrid LCA model is developed for lithium-sulfur battery for the first time.

The inventory data are based on laboratory experiment and literature.

The impacts are benchmarked with conventional lithium ion battery.

Potential improvements to meet the USABC targets are identified.

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

Article history: Lithium-sulfur (Li-S) battery is widely recognized as the most promising battery technology for future
Received 2 December 2016 electric vehicles (EV). To understand the environmental sustainability performance of Li-S battery on
Received in revised form future EVs, here a novel life cycle assessment (LCA) model is developed for comprehensive environ-
4 January 2017
mental impact assessment of a Li-S battery pack using a graphene sulfur composite cathode and a
Accepted 9 January 2017
lithium metal anode protected by a lithium-ion conductive layer, for actual EV applications. The Li-S
battery pack is configured with a 61.3 kWh capacity to power a mid-size EV for 320 km range. The
life cycle inventory model is developed with a hybrid approach, based on our lab-scale synthesis of the
Keywords:
Lithium sulfur battery
graphene sulfur composite, our lab fabrication of Li-S battery cell, and our industrial partner's battery
Graphene wrapped sulfur production processes. The impacts of the Li-S battery are assessed using the ReCiPe method and
Lithium metal anode benchmarked with those of a conventional Nickle-Cobalt-Manganese (NCM)-Graphite battery pack
Life cycle assessment under the same driving distance per charge. The environmental impact assessment results illustrate that
Environmental impact Li-S battery is more environmentally friendly than conventional NCM-Graphite battery, with 9%e90%
lower impact. Finally, the improvement pathways for the Li-S battery to meet the USABC (U.S. Advanced
Battery Consortium) targets are presented with the corresponding environmental impact changes.
© 2017 Elsevier B.V. All rights reserved.

1. Introduction density relative to other battery technologies. However, with

80e150 Wh kg1 energy density [6e8], current LIBs on board of EVs
The ground transportation sector generates approximately 27% are not able to power the EVs for a comparable driving range with
of the total greenhouse gas (GHG) emissions in the U.S [1]. To conventional vehicles. Meanwhile, current LIB technologies
address global warming concerns, the U.S. EPA has set a regulatory generate significant environmental impacts during their life cycle.
standard to reduce the average GHG emissions of the U.S. light For example, current LIBs consume 1141-1224 MJ-equivalent en-
passenger vehicles from 225 g mile1 in 2016 to 143 g mile1 in ergy and generate 57e85 kg CO2 equivalent emissions per kg bat-
2025 [2], mainly to be achieved through adoption of electric vehi- tery during their life cycle [9,10]. Recent years have witnessed rapid
cles (EVs) [3e5]. Current EVs are all powered by lithium ion bat- progress on the research and development of next-generation
teries (LIBs) due to their high energy density and higher power battery technologies with higher energy density and better envi-
ronmental performance.
Among various battery technologies under development,
lithium-sulfur (Li-S) battery is widely recognized among the most
* Corresponding author. Department of Mechanical and Aerospace Engineering,
Case Western Reserve University, Cleveland, OH 44106, United States. promising battery technologies for next generation EVs [11e14].
E-mail address: [email protected] (C. Yuan). Compared to the conventional Li-ion battery, Li-S battery offers a

http://dx.doi.org/10.1016/j.jpowsour.2017.01.036
0378-7753/© 2017 Elsevier B.V. All rights reserved.
 Y. Deng et al. / Journal of Power Sources 343 (2017) 284e295 285

much higher energy density, i.e., sulfur with a specific capacity of kg1 at the cell level and 200 Wh kg1 at the battery pack level,
1672 mAh g1 [14], corresponding to a high theoretical energy almost twice of conventional LIBs on EVs [40]. While Li-S battery is
density of ~2600 Wh kg1. Besides, sulfur is abundantly available in promising for next generation EV applications, the life cycle envi-
nature and often considered as a more environmentally benign ronmental impacts of Li-S batteries have never been studied and
material than those heavy metals used in conventional lithium-ion understood. The Li-S battery uses a wide variety of toxic chemicals
batteries [15]. in the electrode/electrolyte material synthesis and manufacturing
Li-S batteries have been facing some technical challenges, processes, including hydrogen fluoride, sulfuric acid, hydrogen
including rapid capacity fading due to the dissolution of sulfur chloride, etc. Besides, the Li-S battery requires energy-intensive
active materials from the polysulfide shuttle effect, low ionic and processes for the electrode materials and cell manufacturing. For
electronic conductivity of the sulfur material, and low coulombic instance, the synthesis of the GSC involves 1e2 h of sonication of
efficiency and dentrite growth of lithium metal anode, which all the GSC solution with an energy input of 140e300 kW per cubic
affect the battery life and cause safety hazard [12,14,16]. After meter [42,43]; battery cell manufacturing requires a 100e150  C
recent years of intensive research and development, these technical process heating to evaporate off the NMP solvent through around
challenges can now be effectively addressed. The dissolution of 8 h of continuous operation [44], as well as dry room facility to
sulfur problem in Li-S batteries can be suppressed by confining the maintain a moisture level less than 100 ppm. As the Li-S battery is
sulfur in thin film layered materials such as graphene [17e20]. rapidly moving toward commercial-scale production, it is impor-
Graphene including reduced graphene oxide has such functional tant to perform a comprehensive life cycle assessment (LCA) of Li-S
groups as hydroxyl, carboxyl and ester groups which can effectively battery technology to quantify and understand its potential envi-
bind S8 and polysulfide species, and hence to successfully trap ronmental impacts.
polysulfide from dissolution [21]. It has been reported that the In this paper, we developed a hybrid LCA model for an Li-S
graphene-sulfur composite (GSC) with nitrogen-doping can ach- battery pack for EV application, using a graphene-sulfur cathode,
ieve a stable cycling performance up to 1440 mAh g1 with a decay a lithium metal anode, and LiTFSI electrolyte. The graphene-sulfur
rate as low as 0.028% over 2000 cycles [22]. Another study dem- synthesis is modelled from our lab-scale fabrication process, and
onstrates that depositing nanosulfur on graphene layers can result the battery cell manufacturing processes are modelled with actual
in a low capacity fading of 0.039% per cycle over 1500 cycles in an data collected from our industrial partner's pilot scale production
ionic liquid electrolyte [23,24]. Recently, a water-based solution facilities. The life cycle environmental impacts are assessed by the
precipitation method has been developed for GSC production with ReCiPe method [45]. To understand the relative significance of the
a high yield rate, which eliminates the use of toxic solvents and is impacts, the life cycle environmental impacts of the Li-S battery are
regarded as a low-cost and environmental-friendly process, with benchmarked with those of a conventional LIB pack using a
promising potential for future industrial-scale productions Lithium-Nickel-Cobalt-Manganese oxide cathode and a graphite
[20,22,25]. The low conductivity problem of sulfur can be resolved anode (NCM-Graphite). The LCA results can be useful in sustainable
by adding such conductive additives as carbon black, carbon design and manufacturing of Li-S battery packs for future EV
nanotubes, etc. in the active material [14]. applications.
The low coulumbic efficiency and safety concerns of using a
lithium metal anode in the Li-S battery can be addressed by the
formation of a passivation layer on the lithium metal surface to 2. Method
isolate the anode from undesired side reactions to suppress
dendrite growth and improving coulumbic efficiency [16,26e28]. 2.1. Li-S battery pack configuration for a mid-size EV
The passivation layer in nature is a lithium ion conducting ceramic
membrane [29] that can be artificially produced by treating the The Li-S battery pack in this study is configured to power a mid-
lithium metal surface with such chemicals as tetraethoxysilane size EV for a 320 km driving distance (D) per charge with 120 kW
(TEOS) [30], chlorosilane [31], nitrogen [32], etc. In another way, a power (Fig. 1, Section 1 and calculations supplied in the electronic
thin layer of materials including interconnected hollow carbon supporting information (ESI) for details) [46]. For the battery sys-
spheres [28], polymer nanofibres [33] and metal oxides [34] can tem configuration, first, the battery energy for traction at wheel
also be coated onto the lithium metal surface prior to cell (Etraction, i) is calculated for the EV at various speeds under both
manufacturing to obtain the protective effect. Recently, the atomic highway and local driving conditions, respectively, using the EPA
layer deposition (ALD) technology has been applied for lithium Urban Dynamometer Driving Schedule (UDDS) and Highway Fuel
metal anode protection through depositing a thin layer of Al2O3 Economy Test (HWFET) patterns (Section 1.1 in the ESI) [47]. The
[34]. In addition, LiNO3 can be added into the lithium bistri- amount of energy discharged (Edischarged, i) from the Li-S battery
fluoromethanesulfonimidate (LiTFSI) electrolyte in the dimethyl pack is the sum of traction energy divided by the product of motor
glycol (DME) and Dioxolane (DOL) solution to enhance the forma- efficiency (hmotor ) and the EV transmission efficiency (htransmission ),
tion of the protective layer by reacting with the polysulfide species plus the energy consumed by the auxiliary system (Paux). The spe-
and lithium metal material [16,35]. It is also reported that adding cific energy consumption is the ratio of the discharged energy
LiNO3 into the LiTFSI electrolyte can greatly improve the safety (Edischarged, i) over the distance of the driving schedule (si). After-
performance of the lithium metal anode in Li-S batteries [23] and wards, the composite discharged energy (Edischarged) is the
also can substantially reduce the self-discharging of Li-S batteries weighted-sum of specific energy consumptions. The weight factors
from over 10% per day to less than 2% per month [36e38]. All these (a:b) is set to be 45%:55%, which is the distance ratio of local and
technical advances have prepared the Li-S battery as a promising highway driving under the U.S. national average [48]. Finally, the
alternative to the conventional Li-ion battery for next generation nominal capacity of the Li-S battery pack, with 85% accessible
EV applications. fraction by user (UR), is calculated as the ratio between the amount
Li-S batteries are ready to enter commercial production for next of energy discharged from the battery pack and the battery dis-
generation EV applications [39e41]. The European Union has charging efficiency ((hdischarging ), while the total energy drawn from
launched an Advanced Lithium Sulfur battery for EV (ALISE) project the wall outlet can be calculated as the ratio of the stored energy
to promote the applications of Li-S battery on EVs in future. Current (Estored) over the product of battery charging ((hcharing ) and charger
industrial prototypes of Li-S battery have already achieved 300 Wh efficiencies. The vehicle mass (M), as required for Etraction
286 Y. Deng et al. / Journal of Power Sources 343 (2017) 284e295

Fig. 1. Energy flows and efficiencies of EV powered by Li-S battery pack. Detailed information is provided in Section 1 in ESI. The entire calculation is presented in the excel
spreadsheet.

calculation, consists of glider, transmission, motor, and battery our experiments. The anode is made of a lithium metal with a
masses [49]. Transmission and motor masses are calculated ac- surface protective layer. For the entire anode, we consider 50% ac-
cording to mass-power coefficients, namely, 0.585 kg transmission cess of the lithium metal (NP ¼ 1.5) with 3861 mAh g1 specific
kW1 and 0.9 kg motor kW1 [49]. Glider mass is derived based on capacity (cLi) compressed onto a 8mm copper collector [28,55]. A
a baseline electric vehicle and mass compounding coefficient of conventional trilayer separator (polypropylene-polyethylene-
0.57 kg change in glider mass per kg change in powertrain mass polypropylene) with 20 mm thickness and 39% porosity is used. The
[50]. Finally, the battery mass is calculated as the ratio between the cell container consists of a multilayer polymer pouch covering the
nominal battery pack and energy density (sLiS ) of the Li-S battery. cell layers, and two terminals made of aluminum and copper.
The battery discharging and charging efficiencies are deter- Through the BatPac calculation (Fig. 1) a 61.3 kWh Li-S battery
mined considering the heat loss induced by the battery resistance pack containing 12 modules each with 28 cells is configured. Each
(Ubattery), with Joule's law being used, considering the currents, Id cell has 87.5 Ah capacity and is weighted at 547 g. The mass of the
and Ic, and duration tdischarging and tcharging, for discharging and total battery pack is 279 kg, including 184 kg cells, 4 kg BMS, 27 kg
charging, respectively. For the charging process, besides energy loss cooling system, 23 kg module packaging, and 41 kg pack packaging.
on battery resistance, hysteresis loss is also included. A range of Meanwhile, the battery efficiency (hbattery ) is calculated at 86% at a
170mV-400mV open-circuit voltage (OCV) difference is recorded low 200 mV voltage hysteresis. This value is in line with the 78%e
for the Li-S battery cell [51e53]. Then, the hysteresis loss is calcu- 91% energy efficiency measured for the Li-S battery cell [56].
lated as the product of the overall battery voltage hysteresis For the reference battery system, the NCM-Graphite battery
(Uhysteresis) and the battery capacity (Cbattery). pack is modelled at 120 Wh kg1 gravimetric energy density, rep-
Determination of sLiS is based on the configuration of the Li-S resenting the average of the 105e140 Wh kg1 range as reported in
battery. The Li-S battery pack is configured with stacked battery literature [7,46]. Accordingly, this NCM-Graphite battery has a 63.8
cells, a battery management system (BMS), a battery cooling sys- kWh capacity to power the EV for 320 km per charge. This is in good
tem, and battery packaging. The BatPac, a battery design software agreement with previous industrial studies which states that a
developed by Argonne National Lab [46], is adapted to model the Li- conventional LIB battery pack needs 56e75 kWh capacity to power
S structure in this study (Section 1.2 and Section 1.3 in ESI) [25]. In an EV for 320 km driving distance per charge [57,58]. Table 1
the BatPac method, the area specific impedances (ASIs) of the summarizes the battery information and parameters for both the
electrode system at constant discharging for energy use and pulse configured Li-S battery pack and the reference NCM-Graphite bat-
power discharging for power use are two important parameters. tery pack in this study. Li-S and NCM-graphite battery packs are
The electrode ASI is calculated as a function of current density (i) estimated to represent 17% and 26% of the entire vehicle mass,
from the polarization behavior [54]. A battery kinetic model is built respectively. The fractions are in line with the current EV design
to quantify the ASIs caused by charge transfer and conduction
phenomenon in each step of the electrochemical cell under current
Table 1
densities at energy (iE) and power (ip) use. Other technical pa- Comparison of technical parameters between the Li-S and reference battery pack.
rameters of the Li-S chemistry in the BatPac modelling are based on
Parameters Li-S NCM-graphite
current lab-scale experimental data. The cathode is porous struc-
ture composed of graphene-sulfur composite (GSC), carbon black, Battery mass (kg) 279 531
Gravimetric energy density (Wh kg1 battery) 220 120
and PVDF binder at 7:2:1. The specific capacity of the sulfur (csulfur)
Volumetric energy density (Wh L1) 298 350
is 1200 mAh g1 S with 56 wt% sulfur loading in the cathode (S%) UR 85% 85%
[54]. For the Li-S battery, a volume fraction range of 50%e90% of Capacity (kWh) 61.3 63.8
electrolyte over the total cathode is normally recommended [21]. In Corresponding vehicle mass (kg) 1625 2021
Driving distance per charge (km) 320 320
this study, 60% volume fraction of the electrolyte is used based on
 Y. Deng et al. / Journal of Power Sources 343 (2017) 284e295 287

where battery consists of 15%e25% of total mass for a midsized pouch and terminals, cooling system, module and pack level
vehicle [59]. packaging, inventories are constructed based on the outputs from
the BatPac method (see Section 2 in the ESI). For inventory analysis
2.2. Life cycle assessment of the Li-S battery of the GSC, lithium film, electrolyte (LiTFSI, DOL, LiNO3) and Li-S cell
manufacturing, we have developed process-based inventory
A hybrid LCA model is developed in this study to analyze and models as presented below, with details provided in Section 2 of
understand the potential environmental impacts of Li-S battery the ESI.
technology for actual EV applications. An overview of the LCA
system boundaries is illustrated in Fig. 2. The life cycle of the Li-S 2.3.1. Inventory analysis for the Li-S battery
battery pack is divided into five stages: raw materials extraction, A hybrid inventory analysis model is developed in this study
materials processing, battery manufacturing, battery use, and end- integrating our lab experimentation and theoretical modelling
of-life (EoL). The functional unit of the LCA is set per kilometer methods to compile the inventory data for Li-S battery materials.
driving for a mid-sized EV powered by the Li-S battery pack. Based Production of GSC and graphene oxide are modelled based on our
on the current lab-scale performance of the Li-S battery and the lab scale experiments [25] and the Hummers' method [61],
measured capacity fading rate, the overall driving distance of the respectively (Section 2.2 and Section 2.3 in ESI). The material input
EV is set at 200,000 km during the life cycle of the Li-S battery. In inventories of the two materials production are linearly extrapo-
this LCA study, the environmental impacts of the material and lated from the lab scale data assuming a 98.5% efficiency from
energy consumption, as well as waste and emissions directly typical industrial tubular centrifuges [62]. The energy consumption
associated with each life cycle stage of the Li-S battery are quan- of materials production are modelled according to the specific
titatively evaluated (Fig. 1). At the final EoL stage, battery recycling energy consumptions (SEC) of relevant unit processes (baseline
is modelled based on a hydrometallurgical process for material consumption, stirring, reaction heating, centrifugation, and dry-
recovery, while recycling credit is not considered in this study ing). The SECs of these basic unit processes are calculated based on
because the recycled materials may not meet the high-quality conventional mechanical and thermal methods or directly used
standards of materials for battery manufacturing [10]. industrial data [63e66]. Finally, the absolute energy consumption
The ReCiPe method is adopted to derive both the midpoint of the unit processes is obtained by multiplying the SEC with the
environmental impacts and endpoint environmental impacts for mass/volume of the materials being processed. No heat recovery is
the configured Li-S battery pack. With the ReCiPe method, the considered for the two materials production since both reactions
environmental impacts are quantified into 13 midpoint impact are below 100  C [67].
categories, including global warming (GWP), fossil depletion (FDP), Inventories of the lithium metal anode and the LiTFSI electrolyte
ozone depletion (ODP), photochemical oxidant formation (POFP), were constructed by modelling their unit production processes and
particulate matter formation (PMFP), terrestrial acidification (TAP), relevant data are compiled from literature and the Ecoinvent
freshwater eutrophication (FEP), marine eutrophication (MEP), database (Sections 2.5 and 2.6 of the ESI). The inventory of the
freshwater ecotoxicity (FETP), marine ecotoxicity (METP), terres- lithium metal anode involves the unit processes of lithium film
trial ecotoxicity (TETP), human toxicity (HTP), metal depletion production and the following surface treatment for forming a
(MDP). Besides, in order to evaluate the environmental impacts of protective layer (Section 2.5 in the ESI). Commercial production of
the novel GO nanomaterial, which cannot be assessed through the lithium film is made by extruding lithium metal through a slit [67],
traditional ReCiPe method, we have developed a modified USEtox followed by a sheet rolling process [68]. In this study, the extrusion
method to characterize the freshwater ecotoxicity of GO after it is process is modelled based on the mechanical model, and the cold
released into the U.S. environment, based on recent toxicity studies sheet rolling of steel available in the Ecoinvent database was con-
of GO nanomaterials [60]. verted into lithium cold rolling based on the density ratio of the two
materials. The inventory of the protective layer on the surface of the
2.3. Inventory analysis for the Li-S battery lithium metal anode was modelled as a surface treatment process
using TEOS resulting in a 1000 nm lithium-ion conducting mem-
Tables S5 and S6 in the ESI provide an overview of the materials brane formed at a density of 2.96 g cm3 [28]. Production of LiTFSI
and energy inputs for inventory of the Li-S battery pack. For the was modelled according to the synthetic route from tri-
inventory analysis of such materials as current collector, multilayer fluorometansulfonyl fluoride (CF3SO2F) [69], which is produced

Fig. 2. Scope and boundary of the LCA for the Li-S battery.
288 Y. Deng et al. / Journal of Power Sources 343 (2017) 284e295

through electrochemical fluorination (ECF) of methanesulfonyl rate [22], and assuming a battery replacement needed when the
chloride (CH3SO2F) with liquid hydrofluoric acid (HF) [70]. capacity drops down to 70% of its original level, the Ts is calculated
around 290,000 km using Equation (1). However, it is noted that in
2.3.2. Inventory analysis of battery pack production process this calculation, calendar loss from Li-S battery aging is not
After the electrode materials are fabricated, the Li-S battery cell considered, which in fact causes additional capacity fading during
is manufactured through similar processes as the conventional the battery use phase and will shorten the total driving miles. As
NCM-Graphite battery cell (Section 2.8 in the ESI). In this study, we currently no study has been conducted on calendar loss of Li-S
have obtained the energy demand of the Li-S cell manufacturing battery pack for EVs, here we assume the decay rate from Li-S
through measuring and modelling the energy consumption of calendar loss is at the same level with that of its cycling loss.
battery manufacturing equipment in a 93.2 m2 pilot scale dry room Consequently, the actual drivable miles during the service life of the
facility of Johnson Controls Inc. (Fig. S11 and Section 2.8 in the ESI). Li-S battery pack is set at 200,000 miles for about 10 years of service
As calculated, the electricity consumption of a configured 87.5 Ah life under average U.S. driving conditions [72], according to Equa-
Li-S cell is 46.6 kWh kg1, including 27.2 kWh kg1 for the cell tion (2). In this case, a single Li-S battery pack can be used to power
manufacturing processes and 19.4 kWh kg1 for the supporting dry the EV during the designed service life, with no battery replace-
room. The subsequent battery pack assembly is mainly performed ment needed. The possible variations of battery decay rates and the
manually with 3.9 Wh kg1 consumed during the battery pack resulting battery life are investigated in the later sensitivity
assembly process [7]. analysis.
As Li-S batteries are now at an early stage of commercialization, Then the inventory analysis for the Li-S battery use phase is
the pilot scale production can represent the near future situation of based mainly on the electricity consumption during EV operations.
Li-S battery manufacturing. In regards to the future large-scale in- In this study, the U.S. average electric grid is considered for elec-
dustrial production of Li-S batteries, the energy consumption will tricity generation. The operation of the EV is 55% city and 45%
be even lower. To evaluate and benchmark the energy use of Li-S highway driving, based on the U.S. national average data [48]. The
battery production at both pilot scale and industrial scale, here electricity consumption of the Li-S battery during the use phase can
the Li-S battery manufacturing is divided into unit manufacturing be calculated through:
processes, including active material mixing, slurry coating, drying,
calendering, notching, stacking, filling, welding, degassing, and Edischarged
Euse ¼ T (2)
packaging. The specific energy consumptions of each unit Dhbattery hcharger
manufacturing process at an industrial scale are compiled based on
Calculation of the discharged energy ðEdischarged Þ is provided in
corresponding industrial production data in literature (Table S20 in
the ESI). Finally, the energy consumption of Li-S battery production Section 1 in the ESI. Charger efficiency ðhcharger Þ is 0.90 [73]. Battery
at industrial scale is estimated in the range of 11.3e22.8 kWh kg1 efficiency ðhbattery Þ equals hdischarging  hcharging , respectively. T is the
Li-S cell. total driving distance.
The material inventory of the Li-S battery production is analysed
based on the mass compositions of the 87.5 Ah Li-S battery cell and 2.3.4. Inventory analysis of battery EoL
the 61.3 kWh battery pack (Section 2.1 in the ESI). The material Recycling of Li-S battery at the EoL of the EV is modelled from EV
inputs are subsequently calculated based on the material effi- collection, battery removal, discharging, battery disassembly, ma-
ciencies of the cell manufacturing process. It is expected that effi- terials shredding/sorting, until materials refining/disposal. The
ciencies of Li-S battery cell manufacturing resemble those in collection rate of EVs is taken at a generalized rate of 91% from
conventional LIB cell manufacturing, namely, 92% for the positive literature [74]. After that, the Li-S battery packs are removed from
and negative electrodes, 98% for the separator, and 94% for the the vehicle and discharged in a salt solution to deplete the residual
electrolyte [71]. In addition, the PVDF binder needs to be dissolved energy [71]. Since Li-S batteries do not contain precious transit
in NMP solvent at 8 wt% to produce the cathode slurry, with the metals (e.g. cobalt), a hydrometallurgical recycling process is
NMP solvent recovered at a 98% efficiency. For the 61.3 kWh Li-S adopted in this study [71]. Detailed information on the inventory of
battery pack, 336 pouch cells are stacked and packaged in 12 the Li-S battery hydrometallurgical recycling process is provided in
modules with the BMS and battery cooling system. Finally, all of the Section 2.10 in the ESI.
modules are situated in the aluminium based pack housing and
closed with a steel strap. Detailed inventory data for cell container 2.3.5. Reference battery system
(including pouch and terminals) are provided in ESI Section 2.7, and In this study, a conventional 63.8 kWh NCM-Graphite battery is
inventory data for the module and pack packagings are provided in configured as a reference battery system for benchmarking and
ESI Section 2.9. Inventory data for BMS is based on the information validating our results on the 61.3 kWh Li-S battery. The LCA data
in Ellingsen et al. [7]. and results from Ellingsen et al. on the NCM-Graphite battery for
EVs is adapted, using the same ReCiPe method [45]. In this study,
2.3.3. Inventory analysis of battery use the unit LCA results from Ellingsen's study are linearly scaled up to
The inventory for the Li-S battery use phase is analysed for our configured NCM-Graphite battery (63.8 kWh), for comparison
actual EV operations under average U.S conditions, with the Li-S with our Li-S battery results.
battery life considered first. With the Li-S battery degradation,
the total driving distance, Ts, of the Li-S battery pack can be 3. Results and discussion
calculated through [7].
3.1. Life cycle inventory of the Li-S battery pack
 
rc2
Ts ¼ D0 c  (1)
2 The life cycle inventory analysis results for the configured Li-S
battery pack are compiled for both material and inventory data.
where D0 is the initial driving distance per charge (320 km), c is the The material inventory provides masses of all materials needed to
cycling number; and r is the capacity decay rate. Using an experi- produce the Li-S battery pack during its whole life cycle (Table 2).
mentally determined 0.028% capacity loss per cycle as the decay As calculated, a total of 471 kg of material inputs is required during
 Y. Deng et al. / Journal of Power Sources 343 (2017) 284e295 289

Table 2
Material inputs inventory for the 61.3 kWh Li-S battery pack production.

Quantity (kg) Weight ratio (%) Comment

Positive electrode
Graphene oxide 5.3 1.12% Input materials for GSC synthesis
Sodium thiosulfate 132.9 28.25% Sulfur donor
PVP 2.1 0.45% Additive, controlling sulfur size
HCl 61.4 13.06%
Carbon black 9.5 2.02% 20 wt% of cathode
PVDF 4.7 1.01% 10 wt% of cathode, in 8 wt% NMP solution
Aluminium, ingot 14.2 3.02% Current collector, 15 mm thickness
NMP 1.1 0.23% Solvent for PVDF binder.
Negative electrode
Lithium, billet, primary 13.5 2.87%
TEOS 2.5 0.43% 20 wt% of TEOS are lost during surface treatment
Copper, billet 24.2 5.14% Current collector, 8 mm thickness
Separator
PP 4.8 1.03% 20 mm trilayer separator, 39 vt% porosity
PE 2.4 0.51%
Electrolyte
LiTFSI 10.8 2.29% 1 M LiFTSI in DOL/DME solvent
DOL 19.5 4.15% Organic solvent at 1:1 ratio by volume
DME 36.3 7.73%
LiNO3 0.7 0.14% Additive, at 1 wt% fraction of electrolyte system
Cell container
Aluminium, ingot 5.7 1.21% Production of the trilayer pouch
PP 1.7 0.37%
PE 1.1 0.24%
Aluminium, ingot 2.7 0.56% Positive electrode terminal
Copper, billet 8.8 1.87% Negative electrode terminal
Module packaging
Aluminium, ingot 13.8 2.93% Inner and outer frames of cells for heat conductance
Copper, wire 2.9 0.62% Connection between modules
Integrated circuit 1.3 0.29% Controller
ABS 6.1 1.30% Module lid
Polymer 3.2 0.68% Spacer between modules
BMS
BMS 4.0 0.85%
Cooling system
Aluminium, ingot 7.8 1.65% External heat radiator
ethylene glycol 20.0 4.26% coolant
Pack packaging
Aluminium, ingot 42.4 9.02% Pack housing
Steel, billet 2.5 0.53% Strap for battery pack
Copper wire 0.9 0.19%
Total 471 100%

the life cycle of the 279 kg Li-S battery pack. From the inventory aluminium (18% of total material inputs) is used for the Li-S battery
analysis, the GSC based positive electrode is the most material with a distribution of produce pack housing (49%), inner and outer
demanding component in the Li-S battery pack, which requires frames of cells in modules (16%), positive electrode current col-
49.2% of total material inputs. Each gram of GSC synthesis requires lector (16%), radiators in the cooling system (9%). The cooling sys-
0.158 g GO, 4.00 g sodium thiosulfate, 0.063 g PVP, and 1.85 g HCl tem contains 20 kg ethylene glycol, determined through heat
acid (Table 2), from which the sodium thiosulfate and HCl acid generation rate during battery operations. In the calculation, we set
combined together account for 41.3% of total material flow. Mean- a criteria that the coolant amount should be sufficient to enable no
while, synthesis of GSC results in 0.266 L of wastewater flow for more than 1  C increase of battery temperature when the EV is
each gram of GSC synthesized. In comparison, production of the driving at a fast mode of 85 mph.
lithium foil anode and the passivation layer only accounts for 6.1% Copper use, as the main toxicity concerns for conventional LIB
of the material flow due to the high efficiency of material utilization [7], is relatively limited in the Li-S battery. A total of 36.8 kg copper
during the process, with 97.5% for extrusion and 99% for cold including the mass of copper wire, is used in the Li-S battery pack,
rolling. representing a unit rate of 0.6 kg copper use per kWh Li-S battery
The electrolyte requires 14.3% of total material inputs. The capacity. When compared, the conventional LIB battery has a
electrolyte includes 10.8 kg LiTFSI salt and 55.9 kg DOL/DME sol- copper concentration of 1.33 kg per kWh capacity [7]. In addition, a
vent. During the Li-S battery life cycle. The LiTFSI takes up only 2.3% total of 2.5 kg steel is used in the battery pack, which are used to
of total battery material inputs, but it involves 0.57 g of HF toxic produce strap for pack packaging.
chemical and generates 3.57 g hazardous wastes per gram of LiTFSI The energy consumption throughout the life cycle of Li-S battery
production (Table S12 in ESI). pack is another major component of the LCI results. Fig. 3 (nu-
In the Li-S battery pack, common plastics including 3.6 kg PE merical results provided in Table S24 in ESI) presents detailed in-
(polyethylene) and 6.5 kg PP (polypropylene) are used, mainly for formation on primary energy embodied in the main materials,
the separator in the Li-S cell as well as the trilayer pouch. The in- components, and processes for the Li-S battery pack. Here the
jection moulding and compression moulding processes are primary energy refers to the non-renewable energy extracted from
modelled for their production. In addition, a total amount of 86.6 kg nature that has not been refined into any form of secondary energy
290 Y. Deng et al. / Journal of Power Sources 343 (2017) 284e295

Fig. 3. Distribution and primary energy consumption of Li-S battery pack during its life cycle. Numerical results are provided in Table S24.

(e.g. electricity or steam). As calculated, the life cycle primary en- battery manufacturing), battery use (divided into plug-to-output
ergy consumption is 2.36 MJ km1 while the battery use phase and output-to-wheel), and EoL; the colored bars illustrate the
takes a share of 71%. In this study, the primary energy consumption contributions of each component of the Li-S battery pack to the
in the Li-S battery use phase is further broken-down into the environmental impacts of the battery production stage.
electricity loss from the charger, during the battery charging and
discharging (plug-to-battery output) process, and from the energy 3.2.1. Battery production
delivery process for driving the EV and supporting auxiliary sys- As shown in Fig. 4, within the production phase, the Li-S battery
tems (battery-to-wheel). In this study, it is found that the electricity cell manufacturing produced around 60% of share in both GWP and
loss accounts for 21% of total energy consumption during the use FDP impact categories. This is mainly caused by the intensive
phase. Next to the battery use phase, battery manufacturing con- electricity consumption in the amount of 46.6 kWh kg1 during the
sumes 0.51 MJ km1 of primary energy, in which dry room condi- Li-S cell manufacturing. Electrode drying, which is used to evapo-
tioning and electrode drying account for 38% and 13% of the total rate NMP solvent from the cathode slurry, has been identified as the
manufacturing energy, respectively. GSC manufacturing is found primary contributor with 13% of share. Electrode drying requires
consuming 0.089 MJ km1, which can be traced mainly to the 192 min processing time per cell under 150  C temperature. For
sonication process. The primary energy embedded in all Li-S ma- each 87.5 Ah Li-S cell manufacturing, drying consumes 8.2 kWh
terials is 0.16 MJ km1, less than half of the primary energy demand electricity, while other battery manufacturing steps, including
for battery manufacturing. mixing, coating, calendaring, notching, and stacking, only consume
In this study, it is found that the Li-S battery is mainly con- 0.10e1.77 kWh. Besides, energy demand from the dry room facility
structed from such bulk materials as sulfur (thiosulfate form), hy- in cell manufacturing is substantial as well. The dry room of the
drochloric acid, aluminium and DME, which can be obtained with pilot scale facility in this study is operating at 64.8 kW demand, in
moderate environmental burdens. For example, the sodium thio- which 41% is required by the heating and dehumidification ma-
sulfate, the largest material flow in the Li-S battery production (see chines. The energy intensity of cell manufacturing is expected to be
Table 2), is a by-product of sodium sulfide production. Therefore, significantly reduced when large scale production of Li-S batteries
production of sodium thiosulfate consumes a limited amount of is implemented in the future. Through modelling industrial-scale
primary energy at 39.6 MJ kg1. In comparison, GO synthesis is processes, our study estimates that an energy consumption in the
highly energy-intensive at 448 MJ kg1, while the GO only takes range of 11.3e22.8 kWh kg1 cell can be achieved for the large scale
1.12 wt% share of the material flow in the Li-S cell. Li-S cell production. In comparison, Ellingsen et al. estimated an
energy consumption of 28e112 kWh kg1 cell for the conventional
3.2. Life cycle impact assessment results NCM-Graphite cell production [7]. However, some other studies
reported substantially lower energy consumption data for con-
The life cycle environmental impacts of the Li-S battery are ventional LIB pack production only in the range of 1e5 MJ kg1 pack
calculated using the ReCiPe method into 13 impact categories, and (around 1.6e8 MJ kg1 cell assuming 60% mass fraction of cells) [9],
the hotspots are revealed by a contribution analysis in Fig. 4. The mainly because of ignoring the dry room facility in their studies.
pattern bars of Fig. 4 represent the contributions of life cycle stages In this study, it is found that the positive electrode production
to the environmental impacts of the Li-S battery, including battery takes a share of 7%e23% in most impact categories, which is mainly
production (including all activities from raw material extraction to attributed to the considerable energy demand in GSC fabrication.
 Y. Deng et al. / Journal of Power Sources 343 (2017) 284e295 291

Fig. 4. Contributions of Life cycle environmental impacts of the Li-S battery. Numerical results are provided in Table S25.

Additionally, the positive electrode is the main impact source (48%) high aspect ratio, which also reduces the risk of human impact [81].
of the MDP impact category, confirming the fact that positive During the EoL treatment, the GO can be degraded during the high-
electrode fabrication involves the majority of mass inputs in battery temperature process. Therefore, the toxicity impact of GO is ex-
production. pected to be negligible.
The negative electrode, cell container, and module packaging Electrolyte contributes to about 21% of environmental impacts
are identified to be the main contributors to FETP, HTP, METP, and in Li-S battery pack production, but it accounts for a substantial
TETP impact categories. An in-depth analysis reveals that these environmental impact share over 85% in MEP and ODP categories.
toxicity impacts mainly stem from the copper uses as current col- The significant contribution to the MEP impact category is found
lector and terminals, as well as the anodising process to produce mainly from the synthesis of DME, with the associated ammonia
the inner and outer aluminium frames in module packaging. This and nitrate emissions. The significant impact share in ODP is mostly
finding is largely consistent with the results of conventional NCM- attributed to the large amount of methylchloride solvent used
Graphite LIB production in Ellingsen et al. [7]. They found that the during the LiTFSI synthesis.
copper current collector is the dominant source (>60%) for impacts
in FETP, METP, TETP, and HTP. 3.2.2. Battery life cycle impacts
To assess the impact of the novel GO nanomaterial used in the For the total life cycle impacts of the 61.3 kWh Li-S battery pack,
Li-S battery, we have developed an ecotoxicity model based on the it is found that the battery production phase contributes 28%e36%
modified USEtox methodology incorporating colloidal theory to to such categories as GWP, FDP, PMFP, POFP, and TAP, while 67%e
reflect the planar structure of GO, and found that the ecotoxicity 98% to such categories as FETP, FEP, HTP, METP, MEP, MDP, ODP,
characterization factor for GO emission is only 7.77  102 TETP. However, the battery use phase contributes 70% to both the
PAF m3 day kg1 emitted [60]. In comparison, the ecotoxicity GWP and FDP impacts (Fig. 4). The battery use phase is also a major
characterization factors of cobalt, copper, nickel, and manganese contributor to the PMFP, POFP, and TAP, sharing 61%e65% of their
are several magnitudes higher, at 4.48  104, 3.63  106, 1.07  105, life cycle impacts. Since electricity use is the only considered impact
and 6.32  103 PAF m3 day kg1 emitted, respectively [75]. With the source during the battery use phase, the environmental burdens
derived characterization factor for GO, we employed the USEtox from battery use are mainly associated with the U.S. grid power
method to calculate the ecotoxicity impact of the Li-S battery. The generation and distribution. When compared, battery recycling
release rate of GO during the production of the Li-S battery is contributes about 4% in most impact categories. The only exception
assumed to be 0.2%, the same as the emission rate of carbon is the TETP impact where a considerable 21% impact share is
nanotubes (CNTs) [76]. It is found that the ecotoxicity impact of GO observed, because the battery pack is assumed to be treated with
is 1306 PAF m3 day, which accounts for only 2.9% of the ecotoxicity hydrometallurgical recycling process involving a significant
impact from the Li-S battery. For the 1306 PAF m3 day impact, the amount of liquid nitrogen which triggers large impact in TETP.
GO production produces 1070 PAF m3 day while the release of GO
only incurs a small 8.2 PAF m3 day impact on ecotoxicity. The low 3.2.3. Comparison between Li-S and NCM-Graphite battery packs
ecotoxicity of GO nanoparticles is from several aspects. First, GO is a To understand the relative significance of the environmental
photodegradable material [77]. Fragmented GO is highly immobile impacts of the Li-S battery, here the life cycle environmental im-
in water due to the formation of irreversible aggregates and can be pacts of the 61.3 kWh Li-S battery pack is compared with those of a
removed from water quickly with sediments [78]. Furthermore, conventional 63.8 kWh NCM-Graphite battery (Fig. 5, Table S26 in
humid acid, a widely present environmental substance, is an the ESI for numerical results).
effective natural antidote which can neutralize the toxicity of GO As shown in Fig. 5, the Li-S battery is more environmentally
[79]. On the other hand, past studies show that GO concentration in friendly than the conventional NCM-graphite battery technology.
water needs to reach 50 mg L1 level to generate effective cyto- The calculated life cycle environmental impacts of the Li-S battery
toxicity on human cells, indicating a limited human health impact pack are 9%e90% lower than those corresponding impacts of the
[80]. Moreover, the uptake of GO in a human cell is subtle due to its conventional NCM-Graphite battery, except for ODP where Li-S
292 Y. Deng et al. / Journal of Power Sources 343 (2017) 284e295

Fig. 5. Life cycle impact benchmarking between the Li-S and the NCM-Graphite battery packs.

battery is 71% higher than the NCM-Graphite battery. It must be more than 80% shares to FETP, HTP, METP and TETP during the
noted that the environmental impacts of the Li-S battery are eval- NCM-Graphite battery production phase. When compared, the 61.3
uated here in a mixed lab-scale and pilot production scale, while kWh Li-S battery contains only 36.8 kg copper. The substantial
the NCM-Graphite battery is assessed on a full industrial-scale reductions of the toxic metals in the Li-S battery cell make it an
production capacity with the optimized level of energy intensity environmentally sustainable energy storage technology over con-
for cell manufacturing. In a partial production capacity, the life ventional lithium-ion battery technologies.
cycle environmental impacts of the NCM-Graphite battery can be
further increased by 12%e40% in the GWP and FDP categories. As a 3.2.4. Sensitivity
result, the Li-S battery will have significantly lower impacts than To evaluate the reliability and viability of the LCA results, a
the conventional NCM-Graphite battery when it reaches the same sensitivity analysis is conducted in this study to check the in-
large industrial production scale in future. fluences of the intrinsic variabilities of six major LCA parameters by
For GHG emission, the production of the Li-S battery generates changing their baseline values across potential ranges, including:
44.8 CO2 eq km1, 18.4% lower than the 54.9 CO2 eq km1 emission driving range per charge (D), battery accessible capacity ratio (UR),
from the NCM-Graphite battery pack production. The lower impact total mileage (T), battery efficiency (hbattery ), energy consumption of
of the Li-S battery in GWP can be mainly ascribed to those low- cell manufacturing, and decay rates (r) of Li-S battery during actual
impact bulk materials such as sodium thiosulfate used in Li-S bat- EV operations. Detailed information is provided in Table 3.
tery production. As calculated, the GHG emission from the battery Detailed results of the sensitivity analysis are presented in
use phase can be reduced from 117 g CO2 eq km1 from NCM- Fig. S14 and Table S27 in ESI. Summarily, the sensitivity analysis
Graphite battery to 112 g CO2 eq km1 from the Li-S battery. provides numerical evidence that the life cycle impact assessments
Overall, a 9% of life cycle GHG emission reduction can be achieved conducted on the Li-S battery and the benchmarking made be-
from using the Li-S battery over the NCM-Graphite battery. tween the Li-S battery and the conventional NCM-Graphite battery
In addition, using a Li-S battery over a conventional NCM- are viable. Except for the battery decay parameter (Fig. S14(f)), the
Graphite battery can reduce 80e90% of impacts in such categories influences of other parameters on the LCA results are moderate,
as FETP, METP, HTP, and TETP from the life cycle perspective, within 24%-24% variations from the baseline level
because the Li-S battery does not use heavy metals such as Nickel, (Fig. S14(a)e(e)). When compared, the life cycle impacts of the Li-S
Cobalt, and Manganese, and it uses much less copper than a con- battery are 9%e90% lower than those of the conventional NCM-
ventional NCM-Graphite battery. As calculated, the 63.8 kWh NCM- Graphite battery except ODP (Fig. 5). On the other hand, the base-
Graphite battery contains 26.3 kg cobalt, 26.0 kg nickel, 24.1 kg line energy density of the NCM-Graphite battery applied in the
manganese and 84.8 kg copper. Ellingsen et al. [7] reported that main text is 120 Wh kg1. The state-of-the-art NCM-Graphite, as
productions of these heavy metal based components contribute applied in Tesla Model S, reaches 140 Wh kg1 nowadays [85]. A

Table 3
Parameter variations for sensitivity analysis.

Parameter Unit Baseline Variation Comment

Driving range km 320 250e400 a lower end value of 250 km and a higher range of 400 km are assumed for sensitivity analysis.
Accessible capacity % 85 75e90
Total mileage km 200000 150000 The boundary of the total mileage for EVs is between 150,000 km and 300,000 km, following the data reported
e300000 in literature [82].
Battery efficiency % 86 75e86 It is expected that battery degradation will induce a decrease in battery efficiency along time due to the rising
of internal resistance and self-discharging.
Energy consumption of cell kWh 46.6 11.3e46.6 Industrial scale energy consumption of cell manufacturing is estimated to be 11.3e22.8 kWh kg1.
manufacturing
Battery decay rate % 0.028 0.028e0.1 For the GSC structure, some other decay rates at 0.039% [21], 0.046% [83], and a higher estimate of 0.10% were
reported in the literature,
 Y. Deng et al. / Journal of Power Sources 343 (2017) 284e295 293

sensitivity analysis has been conducted with the energy density of reduced by 20% in terms of the eco-points. The pathway analysis
140 Wh kg1 resulting 6%e12% decreases among various impact indicates that a high areal loading is significant for the Li-S battery
categories (Fig. S15). Hence, a conclusion can be reliably made that to meet the USABC target while maintaining a lower environmental
the Li-S battery is more environmental-friendly than the conven- impact. However, a thick electrode, triggered by such a high areal
tional NCM-Graphite battery. loading, will intensify dentride growth and sulfur shuttle effect in
From the sensitivity analysis, it is noted that increasing the both anode and cathode. Therefore, further research on non-
driving range from 320 km to 400 km per charge will increase the flammable electrolyte or solid electrolyte for the lithium sulfur
life cycle impacts in GWP and FDP by 9% (Fig. S14(a)). However, battery are needed to solve the safety issue in future. For cathode,
under such a case, the GWP and FDP of the conventional NCM- such approaches like constructing free-standing 3D electrodes and
Graphite battery will be increased by 11% (Fig. S16 in the ESI). cross-linking nanostructured sulfur hosts are promising for devel-
Meanwhile, the scale-up of the Li-S production, with a 11.3 kWh oping high performance lithium sulfur batteries.
kg1 pouch cell assembly process, to the industrial scale will reduce Second, it is found in this study that the production of the
the impacts in these two categories by 13% (Fig. S14(e)). As a result, nanostructured GSC cathode dominates the material flows during
using a Li-S battery to replace the conventional NCM-Graphite the life cycle of the Li-S battery. In particular, the sonication process
battery will increase the driving range and reduce the life cycle used to create a homogenous dispersion of GO for sulfur precipi-
impacts simultaneously in the future. tation is an energy intensive process. In the future, alternative
It must also be noted here that past LCA studies tend to select energy-efficient technologies are needed to replace the sonication
the same storage capacity among various batteries for bench- process for dispersion of GO nanoparticles. Besides, the cathode
marking [10,84]. Following this procotol, the same 61.3 kWh NCM- structure is a crucial factor determining the decay rate of Li-S bat-
Graphite battery pack will slightly lower the range to around tery, which in turn has a substantial influence on the environ-
300 km per charge (Fig. S17). Since previous studies show that the mental impacts of the Li-S battery. Future investigations are needed
driving range of EVs is the major concern of consumers in market on optimal control of the size and the sediment speed of sulfur
choice, and an acceptable driving range for an EV is expected to be particles during the precipitation process to ensure an improved
353 km and 437 km per charge, respectively, in Germany and the binding and wettability between sulfur and graphene, and then, to
U.S. [86e88], the same driving range at 320 km is selected alleviate the decay rate.
accordingly as the benchmarking basis in this study between the Li- Thirdly, the environmental impacts of the Li-S battery can be
S battery and the conventional NCM-Graphite battery. further reduced through an improved recycling process [6].
Currently, there is no established recycling system available for the
conventional LIBs due to economic viability [83]. However, the
3.3. Future improvement potentials
economic potential of the Li-S battery recycling could be quite
promising. Unlike conventional LIBs built on lithium oxides, pure
Current Li-S batteries are with a gravimetric energy density of
lithium metal is used in the Li-S battery. Automated disassembly
220 Wh kg1 and volumetric energy density of 298 Wh L1, which
methods such as eddy current and density separators can be used
are still below the targets of U.S. Advanced Battery Consortium
to recover lithium metals by taking advantage of its substantially
(USABC) set at 235 Wh kg1 and 500 Wh L1, respectively. In this
low density. Meanwhile, LiTFSI salt in the electrolyte can be another
study, we present a predictive study on the potential improvements
valuable substance for recovery because LiTFSI is a chemically and
of the Li-S battery performance via those key technical parameters
thermally stable compound with high economic value.
including NP ratio, electrolyte fraction in the cathode, sulfur weight
concentration in the cathode, and areal loading at three different
4. Conclusion
production scales, as shown in Fig. 6 [54]. The corresponding
changes of the total environmental impacts along with these
In this paper, a comprehensive LCA model is developed to assess
technical improvements are analysed, using the total ecopoints
and understand the life cycle environmental impacts of Li-S bat-
from the hierarchical perspective of the ReCiPe method. As pre-
teries for next generation EV applications, based on our lab
dicted, the gravimetric energy density and volumetric energy
experimentation, our industrial partner's battery manufacturing
density of the Li-S battery could be increased to 308 Wh kg1 and
processes, and literature data. The Li-S battery is configured with a
415 Wh L1, respectively, with the NP ratio being reduced to 1.1,
61.3 kWh capacity capable of powering a mid-size EV for 320 km
electrolyte fraction decreased to 50%, the sulfur weight concen-
per charge. From the LCA study, a total of 471 kg material inputs is
tration increased to 72%, and the areal loading increased to 10.6
needed to generate the 279 kg Li-S battery pack. Sodium thiosulfate
mAh cm1. Corresponding, the total environmental impacts can be
and hydrogen chloride, which are used for GSC synthesis, are the
two most significant material inputs, representing 28.0% and 13.0%
of mass flows during the life cycle of the Li-S battery, respectively.
Moreover, this study obtained a 2.36 MJ km1 life cycle energy
consumption for the configured 61.3 kWh Li-S battery. For the life
cycle energy consumption, the battery use phase takes a share of
71%, followed by battery manufacturing at 22%, and battery mate-
rials processing at 7%. The intensive energy consumption in battery
manufacturing is found mainly on the electrode drying process of
NMP evaporation and dry room conditioning for cell
manufacturing. The life cycle impact assessment reveals that bat-
tery use accounts for 70% of life cycle GWP and FDP impacts while
battery production represents 28%.
The relative significances of the environmental impacts of the
Li-S battery are compared with those of a conventional NCM-
Fig. 6. Pathways of the environmental impact changes of the Li-S battery in the pro- Graphite LIB at the same 320 km driving range. The results
duction phase. clearly show that the Li-S battery is a more environmentally
294 Y. Deng et al. / Journal of Power Sources 343 (2017) 284e295

friendly technology than the conventional NCM-Graphite battery, aerogels as cathode for advanced lithium sulfur batteries, Nanoscale 8 (2016)
8228e8235, http://dx.doi.org/10.1039/C5NR09037G.
with 9%e90% lower impacts in most impact categories. Specifically,
[14] A. Manthiram, Y. Fu, S.-H. Chung, C. Zu, Y.-S. Su, Rechargeable lithiumesulfur
158 g CO2 eq km1 is generated from a Li-S battery during its life batteries, Chem. Rev. 114 (2014) 11751e11787, http://dx.doi.org/10.1021/
cycle, compared with 174 g CO2 eq km1 from an NCM-Graphite LIB. cr500062v.
The reliability of the LCA study is evaluated through a sensitivity [15] A. Manthiram, S.-H. Chung, C. Zu, Lithiumesulfur batteries: progress and
prospects, Adv. Mater. 27 (2015) 1980e2006, http://dx.doi.org/10.1002/
analysis on six key parameters, including driving distance per adma.201405115.
charge, the accessible ratio of Li-S battery capacity, total mileage, Li- [16] A. Rosenman, E. Markevich, G. Salitra, D. Aurbach, A. Garsuch, F.F. Chesneau,
S battery efficiency, energy consumption of Li-S cell manufacturing, Review on Li-Sulfur battery systems: an integral perspective, Adv. Energy
Mater. 5 (2015), http://dx.doi.org/10.1002/aenm.201500212.
and Li-S battery decay rate. [17] H. Kim, H.D. Lim, J. Kim, K. Kang, Graphene for advanced Li/S and Li/air bat-
Based on the findings from the LCA, the Li-S battery can be a teries, J. Mater. Chem. A 2 (2014) 33e47, http://dx.doi.org/10.1039/
more sustainable energy storage technology for next generation EV C3ta12522j.
[18] L. Ji, M. Rao, H. Zheng, L. Zhang, Y. Li, W. Duan, et al., Graphene oxide as a
applications. Further reduction of the life cycle environmental sulfur immobilizer in high performance lithium/sulfur cells, J. Am. Chem. Soc.
impacts of the Li-S battery can be made through employing a 133 (2011) 18522e18525, http://dx.doi.org/10.1021/ja206955k.
binder-free electrode, developing an energy-efficient process for [19] J.W. Kim, J.D. Ocon, D.W. Park, J. Lee, Functionalized graphene-based cathode
for highly reversible lithiumesulfur batteries, ChemSusChem 7 (2014)
nanoparticle dispersion, improving the decay rate of the Li-S bat- 1265e1273, http://dx.doi.org/10.1002/cssc.201300782.
tery, and recycling Li-S batteries from retired EVs. [20] X. Fang, H. Peng, A revolution in electrodes: recent progress in rechargeable
lithiumesulfur batteries, Small 11 (2015) 1488e1511, http://dx.doi.org/
10.1002/smll.201402354.
Notes [21] Q. Pang, X. Liang, C.Y. Kwok, L.F. Nazar, Advances in lithiumesulfur batteries
based on multifunctional cathodes and electrolytes, Nat. Energy 1 (2016)
The authors declare no competing financial interest. 16132, http://dx.doi.org/10.1038/nenergy.2016.132.
[22] Y. Qiu, W. Li, W. Zhao, G. Li, Y. Hou, M. Liu, et al., High-rate, ultralong cycle-life
lithium/sulfur batteries enabled by nitrogen-doped graphene, Nano Lett. 14
Acknowledgement (2014) 4821e4827, http://dx.doi.org/10.1021/nl5020475.
[23] M.-K. Song, Y. Zhang, E.J. Cairns, A Long-Life, High-rate lithium/sulfur cell: a
multifaceted approach to enhancing cell performance, Nano Lett. 13 (2013)
Financial support from National Science Foundation (CBET- 5891e5899, http://dx.doi.org/10.1021/nl402793z.
1351602) is gratefully acknowledged. The authors would also like to [24] G. Li, J. Sun, W. Hou, S. Jiang, Y. Huang, J. Geng, Three-dimensional porous
thank Johnson Control Inc. for technical support in the energy carbon composites containing high sulfur nanoparticle content for high-
performance lithiumesulfur batteries, Nat. Comms 7 (2016) 10601, http://
analysis of the battery cell manufacturing.
dx.doi.org/10.1038/ncomms10601.
[25] X. Gao, J. Li, D. Guan, C. Yuan, A scalable graphene sulfur composite synthesis
Appendix A. Supplementary data for rechargeable lithium batteries with good capacity and excellent columbic
efficiency, ACS Appl. Mater. Interfaces 6 (2014) 4154e4159, http://dx.doi.org/
10.1021/am4057979.
Supplementary data related to this article can be found at http:// [26] W. Xu, J. Wang, F. Ding, X. Chen, E. Nasybulin, Y. Zhang, et al., Lithium metal
dx.doi.org/10.1016/j.jpowsour.2017.01.036. anodes for rechargeable batteries, Energy Environ. Sci. 7 (2014) 513e537,
http://dx.doi.org/10.1039/C3EE40795K.
[27] D. Wang, W. Zhang, W. Zheng, X. Cui, T. Rojo, Q. Zhang, Towards high-safe
References lithium metal anodes: suppressing lithium dendrites via tuning surface en-
ergy, Adv. Sci. (2016) 1600168, http://dx.doi.org/10.1002/advs.201600168.
[1] EPA, Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2013, 2015. [28] G. Zheng, S.W. Lee, Z. Liang, H.-W. Lee, K. Yan, H. Yao, et al., Interconnected
[2] EPA, NHTSA, 2017 and Later Model Year Light-duty Vehicle Greenhouse Gas hollow carbon nanospheres for stable lithium metal anodes, Nat. Nano-
Emissions and Corporate Average Fuel Economy Standards, 2017. technol. 9 (2014) 618e623, http://dx.doi.org/10.1038/nnano.2014.152.
[3] P. Girardi, A. Gargiulo, P.C. Brambilla, A comparative LCA of an electric vehicle [29] M.S. Park, S.B. Ma, D.J. Lee, D. Im, S.-G. Doo, O. Yamamoto, A highly reversible
and an internal combustion engine vehicle using the appropriate power mix: lithium metal anode, Sci. Rep. 4 (2014), http://dx.doi.org/10.1038/srep03815.
the Italian case study, Int. J. LCA 20 (2015) 1127e1142, http://dx.doi.org/ [30] G.A. Umeda, E. Menke, M. Richard, K.L. Stamm, F. Wudl, B. Dunn, Protection of
10.1007/s11367-015-0903-x. lithium metal surfaces using tetraethoxysilane, J. Mater. Chem. 21 (2011)
[4] L.A.W. Ellingsen, B. Singh, A.H. Strømman, The size and range effect: lifecycle 1593e1599, http://dx.doi.org/10.1039/C0JM02305A.
greenhouse gas emissions of electric vehicles, Environ. Res. Lett. 11 (2016) [31] Filippo Marchioni, Kurt Star, Erik Menke, Thierry Buffeteau, Laurent Servant,
054010, http://dx.doi.org/10.1088/1748-9326/11/5/054010. A. Bruce Dunn, et al., Protection of lithium metal surfaces using chlorosilanes,
[5] H. Huo, H. Cai, Q. Zhang, F. Liu, K. He, Life-cycle assessment of greenhouse gas Langmuir 23 (2007) 11597e11602, http://dx.doi.org/10.1021/la701662r.
and air emissions of electric vehicles: a comparison between China and the [32] G.Q. Ma, Z.Y. Wen, M.F. Wu, C. Shen, Q.S. Wang, J. Jin, et al., A lithium anode
U.S, Atmos. Environ. 108 (2015) 107e116, http://dx.doi.org/10.1016/ protection guided highly-stable lithium-sulfur battery, Chem. Commun. 50
j.atmosenv.2015.02.073. (2014) 14209e14212, http://dx.doi.org/10.1039/C4cc05535g.
[6] J.B. Dunn, L. Gaines, J. Sullivan, M.Q. Wang, Impact of recycling on cradle-to- [33] Z. Liang, G. Zheng, C. Liu, N. Liu, W. Li, K. Yan, et al., Polymer nanofiber-guided
gate energy consumption and greenhouse gas emissions of automotive uniform lithium deposition for battery electrodes, Nano Lett. 15 (2015)
lithium-ion batteries, Environ. Sci. Technol. 46 (2012) 12704e12710, http:// 2910e2916, http://dx.doi.org/10.1021/nl5046318.
dx.doi.org/10.1021/es302420z. [34] A.C. Kozen, C.-F. Lin, A.J. Pearse, M.A. Schroeder, X. Han, L. Hu, et al., Next-
[7] L.A.W. Ellingsen, G. Majeau-Bettez, B. Singh, A.K. Srivastava, L.O. Valoen, generation lithium metal anode engineering via atomic layer deposition, ACS
A.H. Stromman, Life cycle assessment of a lithium-ion battery vehicle pack, Nano 9 (2015) 5884e5892, http://dx.doi.org/10.1021/acsnano.5b02166.
J. Ind. Ecol. 18 (2014) 113e124, http://dx.doi.org/10.1111/Jiec.12072. [35] S. Zhang, K. Ueno, K. Dokko, M. Watanabe, Recent advances in electrolytes for
[8] B. Scrosati, J. Garche, Lithium batteries: status, prospects and future, J. Power lithium-sulfur batteries, Adv. Energy Mater. 5 (2015), http://dx.doi.org/
Sources 195 (2010) 2419e2430, http://dx.doi.org/10.1016/ 10.1002/aenm.201500117 n/aen/a.
j.jpowsour.2009.11.048. [36] W.-T. Xu, H.-J. Peng, J.-Q. Huang, C.-Z. Zhao, X.-B. Cheng, Q. Zhang, Towards
[9] D.A. Notter, M. Gauch, R. Widmer, P. Wa €ger, A. Stamp, R. Zah, et al., Contri- stable lithium-sulfur batteries with a low self-discharge rate: ion diffusion
bution of Li-Ion batteries to the environmental impact of electric vehicles, modulation and anode protection, ChemSusChem 8 (2015) 2892e2901,
Environ. Sci. Technol. 44 (2010) 6550e6556, http://dx.doi.org/10.1021/ http://dx.doi.org/10.1002/cssc.201500428.
es903729a. [37] Y.V. Mikhaylik, I. Kovalev, R. Schock, K. Kumaresan, J. Xu, J. Affinito, High
[10] B. Li, X. Gao, J. Li, C. Yuan, Life cycle environmental impact of high-capacity energy rechargeable Li-S cells for EV application: status, remaining problems
lithium ion battery with silicon nanowires anode for electric vehicles, Envi- and solutions, ECS Trans. 25 (2010) 23e34, http://dx.doi.org/10.1149/
ron. Sci. Technol. 48 (2014) 3047e3055, http://dx.doi.org/10.1021/es4037786. 1.3414001.
[11] D. Larcher, J.-M. Tarascon, Towards greener and more sustainable batteries for [38] L. Wang, J. Liu, S. Yuan, Y. Wang, Y. Xia, To mitigate self-discharge of lith-
electrical energy storage, Nat. Chem. 7 (2014) 19e29, http://dx.doi.org/ iumesulfur batteries by optimizing ionic liquid electrolytes, Energy Environ.
10.1038/nchem.2085. Sci. 9 (2016) 224e231, http://dx.doi.org/10.1039/C5EE02837J.
[12] A. Manthiram, Y. Fu, Y.-S. Su, Challenges and prospects of lithiumesulfur [39] S. Wu, R. Ge, M. Lu, R. Xu, Z. Zhang, Graphene-based nano-materials for
batteries, Acc. Chem. Res. 46 (2012) 1125e1134, http://dx.doi.org/10.1021/ lithiumesulfur battery and sodium-ion battery, Nano Energy 15 (2015)
ar300179v. 379e405, http://dx.doi.org/10.1016/j.nanoen.2015.04.032.
[13] Y. Hou, J. Li, X. Gao, Z. Wen, C. Yuan, J. Chen, 3D dual-confined sulfur [40] Oxis Energy, http://www.oxisenergy.com/technology/product/(Accessed 29
encapsulated in porous carbon nanosheets and wrapped with graphene November 2016).
 Y. Deng et al. / Journal of Power Sources 343 (2017) 284e295 295

[41] A. Fotouhi, D.J. Auger, K. Propp, S. Longo, M. Wild, A review on electric vehicle [65] S. Thakur, G. Das, P.K. Raul, N. Karak, Green one-step approach to prepare
battery modelling: from Lithium-ion toward LithiumeSulphur, Renew. Sus- sulfur/reduced graphene oxide nanohybrid for effective mercury ions
tain. Energy Rev. 56 (2016) 1008e1021, http://dx.doi.org/10.1016/ removal, J. Phys. Chem. C 117 (2013) 7636e7642, http://dx.doi.org/10.1021/
j.rser.2015.12.009. jp400221k.
[42] T.R. Frømyr, F.K. Hansen, T.X. Olsenrn, The optimum dispersion of carbon [66] Hielscher-Ultrasound Technology, http://www.hielscher.com/i10000_p.htm
nanotubes for epoxy nanocomposites: evolution of the particle size distri- (Accessed 29 November 2016).
bution by ultrasonic treatment, J. Nanotechnol. 2012 (2012) 1e14, http:// [67] M. Wakihara, O. Yamamoto, Lithium Ion Batteries:Fundamentals and Perfor-
dx.doi.org/10.1155/2012/545930. mance, Wiley-VCH Verlag GmbH, Weinheim, 2008.
[43] R. Arvidsson, D. Kushnir, B.A. Sande n, S. Molander, Prospective life cycle [68] B Hovsepian, Rolling of lithium, US Pat. 3 721 113, 1971.
assessment of graphene production by ultrasonication and chemical reduc- [69] A. Chakrabarti, Design and Synthesis of New Electrolyte Systems for Lithium-
tion, Environ. Sci. Technol. 48 (2014) 4529e4536, http://dx.doi.org/10.1021/ ion Batteries, Ph.D thesis, Illinois Institute of Technology, 2008.
es405338k. [70] H.J. Emele us, The Chemistry of Fluorine and its Compounds, Elsevier, 2013.
[44] D.L.I. Wood, J. Li, C. Daniel, Prospects for reducing the processing cost of [71] J. Dunn, C. James, L. Gaines, K.G. Gallagher, Material and Energy Flows in the
lithium ion batteries, J. Power Sources 275 (2015) 234e242. Production of Cathode and Anode Materials for Lithium Ion Batteries, Chicago,
[45] M. Goedkoop, R. Heijungs, M. Huijbregts, A. De Schryver, J. Stuijs, R. van Zelm, 2014. http://www.ipd.anl.gov/anlpubs/2014/11/108520.pdf.
ReCiPe 2008: a Life Cycle Impact Assessment Method Which Comprises [72] Federal Highway Administration, Average Annual Miles per Driver by Age
Harmonised Category Indicators at the Midpoint and the Endpoint Level, Group, http://www.fhwa.dot.gov/ohim/onh00/bar8.htm (Accessed 29
Ruimtelijke Odening en Milieubeheer, 2009. November 2016).
[46] P.A. Nelson, K.G. Gallagher, I. Bloom, BatPaC (Battery Performance and Cost) [73] S. Campanari, G. Manzolini, F. Garcia de la Iglesia, Energy analysis of electric
Software, Argonne National Laboratory, Chicago, 2015. vehicles using batteries or fuel cells through well-to-wheel driving cycle
[47] US EPA, Dynamometer Drive Schedules, https://www.epa.gov/vehicle-and- simulations, J. Power Sources 186 (2009) 464e477, http://dx.doi.org/10.1016/
fuel-emissions-testing/dynamometer-drive-schedules. (Accessed 29 j.jpowsour.2008.09.115.
November 2016). €ser, M. Soulier, L.A.T. Espinoza, Dynamic analysis of global copper flows.
[74] S. Glo
[48] US EPA, Fuel Economy Guide 2008, 2008. Global stocks, postconsumer material flows, recycling indicators, and uncer-
[49] C. Bauer, J. Hofer, H.-J. Althaus, A. Del Duce, A. Simons, The environmental tainty evaluation, Environ. Sci. Technol. 47 (2013) 6564e6572, http://
performance of current and future passenger vehicles: life cycle assessment dx.doi.org/10.1021/es400069b.
based on a novel scenario analysis framework, Appl. Energy 157 (2015) [75] A.D. Henderson, M.Z. Hauschild, D. van de Meent, M.A.J. Huijbregts,
871e883, http://dx.doi.org/10.1016/j.apenergy.2015.01.019. H.F. Larsen, M. Margni, et al., USEtox fate and ecotoxicity factors for
[50] J. Neubauer, A. Pesaran, C. Bae, R. Elder, B. Cunningham, Updating United comparative assessment of toxic emissions in life cycle analysis: sensitivity to
States Advanced Battery Consortium and Department of Energy battery key chemical properties, Int. J. LCA 16 (2011) 701e709, http://dx.doi.org/
technology targets for battery electric vehicles, J. Power Sources 271 (2014) 10.1007/s11367-011-0294-6.
614e621, http://dx.doi.org/10.1016/j.jpowsour.2014.06.043. [76] M.J. Eckelman, M.S. Mauter, J.A. Isaacs, M. Elimelech, New perspectives on
[51] G. Zheng, Q. Zhang, J.J. Cha, Y. Yang, W. Li, Z.W. Seh, et al., Amphiphilic surface nanomaterial aquatic ecotoxicity: production impacts exceed direct exposure
modification of hollow carbon nanofibers for improved cycle life of lithium impacts for carbon nanotoubes, Environ. Sci. Technol. 46 (2012) 2902e2910,
sulfur batteries, Nano Lett. 13 (2013) 1265e1270, http://dx.doi.org/10.1021/ http://dx.doi.org/10.1021/es203409a.
nl304795g. [77] W.-C. Hou, I. Chowdhury, J. David G Goodwin, W.M. Henderson,
[52] M. Helen, M.A. Reddy, T. Diemant, U. Golla-Schindler, R.J. Behm, U. Kaiser, et D.H. Fairbrother, D. Bouchard, et al., Photochemical transformation of gra-
al., Single step transformation of sulphur to Li2S2/Li2S in Li-S batteries, Sci. phene oxide in sunlight, Environ. Sci. Technol. 49 (2015) 3435e3443, http://
Rep. 5 (2015) 12146, http://dx.doi.org/10.1038/srep12146. dx.doi.org/10.1021/es5047155.
[53] L. Qie, C. Zu, A. Manthiram, A high energy lithium-sulfur battery with [78] I. Chowdhury, W.-C. Hou, D. Goodwin, M. Henderson, R.G. Zepp, D. Bouchard,
ultrahigh-loading lithium polysulfide cathode and its failure mechanism, Adv. Sunlight affects aggregation and deposition of graphene oxide in the aquatic
Energy Mater. 6 (2016) 1502459, http://dx.doi.org/10.1002/aenm.201502459. environment, Water Res. 78 (2015) 37e46, http://dx.doi.org/10.1016/
[54] D. Eroglu, K.R. Zavadil, K.G. Gallagher, Critical link between materials chem- j.watres.2015.04.001.
istry and cell-level design for high energy density and low cost lithium-sulfur [79] Y. Chen, C. Ren, S. Ouyang, X. Hu, Q. Zhou, Mitigation in multiple effects of
transportation battery, J. Electrochem. Soc. 162 (2015) A982eA990, http:// graphene oxide toxicity in Zebrafish Embryogenesis driven by humic acid,
dx.doi.org/10.1149/2.0611506jes. Environ. Sci. Technol. 49 (2015) 10147e10154, http://dx.doi.org/10.1021/
[55] K.G. Gallagher, S. Goebel, T. Greszler, M. Mathias, W. Oelerich, D. Eroglu, et al., acs.est.5b02220.
Quantifying the promise of lithiumeair batteries for electric vehicles, Energy [80] L. Feng, Z. Liu, Graphene in biomedicine: opportunities and challenges,
Environ. Sci. 7 (2014) 1555e1563, http://dx.doi.org/10.1039/C3EE43870H. Nanomedicine 6 (2011) 317e324, http://dx.doi.org/10.2217/nnm.10.158.
[56] Z. Yuan, H.-J. Peng, T.-Z. Hou, J.-Q. Huang, C.-M. Chen, D.-W. Wang, et al., [81] X. Zhang, W. Hu, J. Li, L. Tao, Y. Wei, A comparative study of cellular uptake
Powering lithium-sulfur battery performance by propelling polysulfide redox and cytotoxicity of multi-walled carbon nanotubes, graphene oxide, and
at sulfiphilic hosts, Nano Lett. 16 (2016) 519e527, http://dx.doi.org/10.1021/ nanodiamond, Toxicol. Res. 1 (2012) 62e68, http://dx.doi.org/10.1039/
acs.nanolett.5b04166. C2TX20006F.
[57] Byd E6, http://www.byd.com/na/old/auto/e6.html (accessed 29.11.2016). [82] H.C. Kim, T.J. Wallington, Life-cycle energy and greenhouse gas emission
[58] O. Gro € ger, H.A. Gasteiger, J.-P. Suchsland, Reviewdelectromobility: batteries benefits of lightweighting in automobiles: review and harmonization, Envi-
or fuel cells? J. Electrochem. Soc. 162 (2015) A2605eA2622, http://dx.doi.org/ ron. Sci. Technol. 47 (2013) 6089e6097, http://dx.doi.org/10.1021/es3042115.
10.1149/2.0211514jes. [83] H. Chen, C.H. Wang, Y.F. Dai, S.Q. Qiu, J.L. Yang, W. Lu, L.W. Chen, Rational
[59] I. Besselink, P.F. Van Oorschot, E. Meinders, Design of an efficient, low weight design of cathode structure for high rate performance lithiumesulfur batte-
battery electric vehicle based on a VW Lupo 3L the 25th World Battery, in: ries, Nano Lett. 15 (2015) 5443e5448, http://dx.doi.org/10.1021/
Hybrid and Fuel Cell Electric Vehicle Symposium and Exhibition, Shenzhen, acs.nanolett.5b01837.
China, 2010. [84] D.A. Notter, K. Kouravelou, T. Karachalios, M.K. Daletou, N.T. Haberland, Life
[60] Y. Deng, J. Li, M. Qiu, Y. Fan, J. Zhang, C. Yuan, Deriving characterization factors cycle assessment of PEM FC applications: electric mobility and m-CHP, Energy
on freshwater ecotoxicity of graphene oxide nanomaterials for life cycle Environ. Sci. 8 (2015) 1969e1985, http://dx.doi.org/10.1039/C5EE01082A.
impact assessment, Int. J. LCA (2016), http://dx.doi.org/10.1007/s11367-016- [85] Tesla Model S Battery, http://enipedia.tudelft.nl/wiki/Tesla_Model_S_Battery,
1151-4. (Accessed 29 December 2016).
[61] W.S. Hummers Jr., R.E. Offeman, Preparation of graphitic oxide, J. Am. Chem. [86] N.S. Pearre, W. Kempton, R.L. Guensler, V.V. Elango, Electric vehicles: how
Soc. 80 (1958), http://dx.doi.org/10.1021/ja01539a017, 1339e1339. much range is required for a day's driving? Transp. Res. Part C Emerg. Technol.
[62] W.W.-F. Leung, Centrifugal Separations in Biotechnology, Elsevier, Oxford, 19 (2011) 1171e1184, http://dx.doi.org/10.1016/j.trc.2010.12.010.
2007. [87] T. Franke, I. Neumann, F. Bühler, P. Cocron, J.F. Krems, Experiencing range in
[63] A. Thekdi, S.U. Nimbalkar, Industrial Waste Heat Recovery - Potential Appli- an electric vehicle: understanding psychological barriers, Appl. Psychol. 61
cations, Available Technologies and Crosscutting R&D Opportunities, 2015, (2012) 368e391, http://dx.doi.org/10.1111/j.1464-0597.2011.00474.x.
http://dx.doi.org/10.2172/1185778. [88] T. Franke, J.F. Krems, What drives range preferences in electric vehicle users?
[64] S. De Meester, J. Dewulf, L. Roes, M. Patel, Development of Basic Engineering Transp. Policy 30 (2013) 56e62, http://dx.doi.org/10.1016/
Modules for Prospective Estimations of the Material Flows and Energy Re- j.tranpol.2013.07.005.
quirements, Ghent University, 2011.