Journal of Power Sources 262 (2014) 255e262

Contents lists available at ScienceDirect

Journal of Power Sources

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

A closed loop process for recycling spent lithium ion batteries

Eric Gratz, Qina Sa, Diran Apelian, Yan Wang*
Department of Mechanical Engineering, Worcester Polytechnic Institute, 100 Institute Rd., Worcester, MA 01609, USA

h i g h l i g h t s g r a p h i c a l a b s t r a c t

Recovery of materials from spent

Lithium ion batteries via a low tem-
perature hydrometallurgical
approach.

The recovery efficiencies for mate-
rials in the batteries are determined.

The impurities are removed so that
cathode materials can be recovered
as new cathode materials.

Economics of the process are
analyzed.

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

Article history: As lithium ion (Li-ion) batteries continue to increase their market share, recycling Li-ion batteries will
Received 31 January 2014 become mandatory due to limited resources. We have previously demonstrated a new low temperature
Received in revised form methodology to separate and synthesize cathode materials from mixed cathode materials. In this study
1 March 2014
we take used Li-ion batteries from a recycling source and recover active cathode materials, copper, steel,
Accepted 25 March 2014
etc. To accomplish this the batteries are shredded and processed to separate the steel, copper and
Available online 2 April 2014
cathode materials; the cathode materials are then leached into solution; the concentrations of nickel,
manganese and cobalt ions are adjusted so NixMnyCoz(OH)2 is precipitated. The precipitated product can
Keywords:
Lithium ion batteries
then be reacted with lithium carbonate to form LiNixMnyCozO2. The results show that the developed
Recycling recycling process is practical with high recovery efficiencies (w90%), and 1 ton of Li-ion batteries has the
Efficiency potential to generate $5013 profit margin based on materials balance.
Cost Ó 2014 Elsevier B.V. All rights reserved.

1. Introduction [2]. One reason for the increased use of Li-ion batteries in the
vehicle market is the ever-stricter CO2 emission standards auto-
Due to their high energy density, long lifespan and light weight, mobile manufacturers must meet [3].
lithium ion (Li-ion) batteries accounted for $11.8 billion dollars in The working principle of the Li-ion battery is that of a galvanic
sales in 2012, that makes up 60% of total portable battery sales and cell. When charging, Li-ions transfer from the cathode to the anode
37% of total battery sales [1]. The same factors also make them the through the electrolyte, and electrons are transferred from the
best candidate for use in hybrid electric vehicles (HEVs) and electric cathode to the anode via the external circuit. When discharging, Li
vehicles (EVs). Currently many Li-ion batteries are already used in ions transfer from the anode to the cathode through the electrolyte,
HEVs and EVs; total all time worldwide HEV sales were 1.5 billion as and electrons are transferred from the anode to the cathode
of 2011 billion and are expected to reach nearly 3.5 billion by 2015 through the external circuit. With every cycle of charging and
discharging, Li ions traverse back and forth between anode and
cathode. There are 2 primary failure mechanisms in Li-ion batteries
(1) secondary reactions, and (2) dendrite growth [4]. Studies on Li-
* Corresponding author. Tel.: þ1 508 831 5453; fax: þ1 508 831 5178.
E-mail address: [email protected] (Y. Wang). ion battery cycling in HEVs predict Li-ion batteries can run for

http://dx.doi.org/10.1016/j.jpowsour.2014.03.126
0378-7753/Ó 2014 Elsevier B.V. All rights reserved.
256 E. Gratz et al. / Journal of Power Sources 262 (2014) 255e262

Table 1 materials including LiCoO2 ($2946), LiNi0.33Mn0.33Co0.33O2 ($1684),

Li-ion battery components by weight percent [9]. LiMn2O4 ($636), LNiO2 ($523) and LiFePO4 ($312), and Cu ($654),
Battery component Wt% and steel ($212). These materials account for w90% of the total
Casing 25
intrinsic material value, which will be recycled in our proposed
Cathode material 25 process. Current processes mainly recycle Co and Ni as pure metals
Anode material 14 and therefore the potential value is significantly lower than our
Electrolyte 10 process.
Copper electrolyte foil 8
As the chemistry of cathode materials changes, it is important
Aluminium electrolyte foil 5
Separator 4 for recyclers to be aware that the target recycling stream of
Other 6 incoming batteries is going to be at least 3e4 years old. Currently,
Li-ion battery recycling efforts have focused on recovering cobalt
from LiCoO2 cathode material. LiCoO2 is the predominate material
between 1700e5300 cycles depending on operating conditions, in the recycling stream today [10e12]. In a recent review paper by
which puts the operating lifetime of the battery at 4.5e14.5 years Xu et al. [7], the existing recycling methods considered only LiCoO2
[5,6]. as the cathode material. LiCoO2 is not the only cathode material
Li-ion batteries are made of an anode, a cathode, current col- used in the commercial world, for example currently LiCoO2 is only
lectors, a separator, liquid electrolyte, container and sealing parts. used as the cathode material in w37.2% of the Li-ion batteries, other
Table 1 shows the contribution that each component makes to the battery chemistries are LiNi0.33Mn0.33Co0.33O2 is used in w29.0%,
total battery weight. The anode is made of graphite, carbon and LiMn2O4 is used in w21.4%, LiNiO2 is used in 7.2% and LiFePO4 is
polyvinylidene fluoride (PVDF) binder; the anode current collector used inw5.2% [1]. New battery chemistries with less Co and Ni are
is copper. The electrolyte is lithium hexaflourophospate (LiPF6) being used in order to lower battery prices. For example, LiMn2O4 is
with an organic solvent commonly ethylene carbonate (EC), diethyl used as the cathode material for GM Volt batteries, and LiFePO4 is
carbonate (DEC), dimethyl carbonate (DMC) or their mixture. The used as the cathode material in A123 batteries. Thus recovering all
separator is made of polypropylene (PP) or polyethylene (PE). The cathode materials regardless of battery composition is critical
cathode is made of carbon, the PVDF binder and a lithium com- because these batteries will be entering the recycling stream soon
pound and the cathode current collector is aluminium foil. and present technologies are not in place to handle them. Some
The first generation Li-ion batteries used LiCoO2 as the cathode work has been carried that focuses on recycling of Li-ion batteries
and recycling Li-ion batteries with LiCoO2 cathode materials has with varying cathode chemistries [13,14], however they are com-
been widely studied [7]. Today however a wide variety of lithium plex and commercially not feasible since those methods are trying
compounds are used as cathode materials mostly commonly to separate every element/material in the cathode.
LiCoO2, LiMn2O4, LiNi0.33Mn0.33Co0.33O2, and LiFePO4. The most Industrially, companies such as Umicore, Recupyl Battery Solu-
expensive part of the battery is the cathode materials and the tions, and Toxco do recycle Li-ion batteries. Umicore uses an ul-
composition of the cathode can vary greatly as new technologies trahigh temperature smelting technology [15,16] and does offer a
are developed to produce low-cost and high energy density Li-ion closed loop solution for high value metals such as Co and Ni. Less
batteries. Fig. 1 shows the value of each component in 1 ton of Li- valuable metals such as Mn, Fe, Li, Al are ‘downcycled’, meaning
ion batteries. The amount of cathode materials in 1 ton of batte- that they are not recovered and reused in battery applications. It is
ries was calculated using the forecasted 2012 Li-ion battery cathode believed that Toxco focuses on Li-ion batteries with LiCoO2 chem-
chemistry makeup [1], and the value of the cathode materials was istry [17].
obtained from Gaines et al. [8]. The amount of each battery material The main issue with recycling Li-ion batteries with different
in 1 ton of batteries was calculated using the wt% provided by cathode chemistries is that cobalt, manganese, and nickel are hard
Maschler et al. [9]. From the figure, the intrinsic material value for 1 to separate from one another, without using expensive organic
ton of Li-ion batteries is $7708. The greatest value is from cathode reagents for solvent extraction [13,14]. The proposed process ad-
dresses this issue by offering a closed loop recycling process
where new cathode materials can be made from spent lithium
ion batteries. We have previously shown that unused Li-ion
cathode powders consisting of LiFePO4, LiCoO2, LiMnO2, and
LiNi0.33Mn0.33Co0.33O2 can be recycled to recover LiNi0.33Mn0.33-
Co0.33O2 and Fe(OH)3 by precipitating Ni0.33Mn0.33Co0.33(OH)2 and
reacting the precipitate with lithium carbonate, which is produced
by reacting the remaining lithium hydroxide with sodium car-
bonate [18]. In this study we recover LiNi0.33Mn0.33Co0.33O2, copper
and steel from spent Li-ion batteries obtained from Worcester
Polytechnic Institute (WPI) recycling center. The batteries used in
this experiment are primarily composed of LiCoO2 cathode chem-
istry and in a second experiment, cathode materials are recovered
from spent Li-ion batteries with LiMn2O4, LiNi0.33Mn0.33Co0.33O2
and LiFePO4 powders added to the shredded used batteries. In both
cases the batteries are discharged, then shredded. The steel casing
is removed from the shredded material by magnetic separation.
The cathode materials are separated from the aluminium current
collectors by dissolving the aluminium in NaOH. The powders are
then separated from the rest of the material by sieving. The copper
current collectors are then separated from the remaining material,
Fig. 1. Cost breakdown of each component in 1 ton of Li-ion batteries [1,8,9]. Total mainly plastic, by density separation. The cathode materials are
value of all components is $7708. then leached into solution and the LiNi0.33Mn0.33Co0.33O2 product is
 E. Gratz et al. / Journal of Power Sources 262 (2014) 255e262 257

the desired 105 M level. Furthermore, previously researchers have

proposed that Co ions are removed with Cu(OH)2 at to the forma-
tion of a CueCo complex [25]. However, these experiments were
conducted with very concentrated basic solutions such that the
change in volume during pH adjustment was negligible. In order to
test this theory with a less concentrated base, two experiments
were conducted, one with copper and one without. In the experi-
ment with copper, 5.5 g NiSO4$7H20, 5.0 g CoSO4$6H2O, 4.23 g
MnSO4$H2O and 0.1 g copper foil were added to 50 ml 4 M H2SO4
with 10 wt% H2O2. In the second experiment, the same masses of
nickel, cobalt and manganese sulfate without copper foil were put
into the solution containing 50 ml 4 M H2SO4 with 10 wt% H2O2.
After allowing the solutions to react for 2 h the solutions were
diluted to 100 ml with DI water in a 100 ml volumetric flask. The
concentrations of Co, Ni, Mn and Cu present in solution were
measured with atomic adsorption spectroscopy (AAS) (Perkine
Elmer AA Analyst 300). The pH was then increased to 6.5 by adding
2 M NaOH. Once the pH reached 6.5, the solutions were allowed to
stir for 4 h, then the solutions were diluted to 250 ml with DI water
in a 250 ml volumetric flask. The concentrations of Co, Ni, Mn and
Cu present in solution were measured with AAS.

2.2. Processing spent Li-ion batteries

Fig. 2. Procedure for processing spent Li-ion batteries to recover steel and copper, as Before Li-ion batteries can go through the hydrometallurgical
well as separating cathode powders to begin hydrometallurgical procedure.
route to recover the cathode materials, the spent Li-ion batteries
recovered via our hydrometallurgical process, see Fig. 2. One major must be processed, so that the lithium compounds can be leached
advantage of this process is that concentrations of nickel, manga- into solution.
nese and cobalt in the recovered cathode material can be specif- All Li-ion batteries collected were used in cell phones or laptops.
ically tailored. A wide variety of LiNixMnyCozO2 compounds have For the laptop batteries, the plastic casing that holds the 18650 cells
been used in Li-ion batteries [19e24]. Fig. 3 shows the composi- was removed by hand. The cell phone batteries were left as is. The
tions of several compound LiNixMnyCozO2 compounds, all of which batteries were then processed so they can be recycled. Fig. 2 shows
can be made by this process by adding the metal sulphates to the the procedure for processing Li-ion batteries into a form that will
leaching liquor to the desired compositions. allow the cathode materials to be leached into solution, as well as
recovering the copper and steel.
2. Experimental
Step 1: Before Li-ion batteries can be safely shredded, they must
be discharged due to the remaining capacity in batteries and the
2.1. Testing Cu impurity removal
possibility of a short circuit in the presence of the flammable
electrolyte [26]. 18 18650 cells from laptop batteries and 6 cell
When the concentration of Ni in solution is high, it is possible to
phone batteries were discharged in a stainless steel container
remove Ni ions before the concentration of the Cu impurity reaches
with stainless steel chips for 4 h, this is similar to the procedure
proposed by Nan et al. [12]. After discharge the voltages were
measured with a voltmeter and all were below 2.0 V.
Step 2: After discharging, the batteries were shredded in a
Schutte Buffalo Hammer Mill (Laboratory Scale, Model-6-H).
The hammer mill had contained a screen with ¼00 holes and
was run at 2,000 rpm. After shredding, random 20 g samples
were used for the remaining experiments.
Step 3: The steel casings in the 20 g sample were removed with
a rare earth. The magnet was passed 100 above a monolayer of the
shredded batteries until no more pieces were attracted to the
magnet.
Step 4: The remaining material was exposed to 75 ml of 2 M
NaOH to dissolve the anode current collectors (aluminium) ac-
cording to Eq. (1). The solution was given 2 h to react.

2Al þ 2NaOH ¼ 2NaAlðOHÞ4 þ 3H2 ðgÞ (1)

Step 5: The material was sieved with a 250 mm screen after

drying at 60  C. The sub 250 mm fraction was composed pri-
marily of cathode powders, carbon and a small amount of
Fig. 3. Composition of common Li-ion nickel, manganese, cobalt cathode materials. copper.
258 E. Gratz et al. / Journal of Power Sources 262 (2014) 255e262

Step 6: The fraction with larger than 250 mm consists primarily Fe(OH)3, Al(OH)3 and Cu(OH)2 for ideal solutions; Ni(OH)2, Co(OH)2
of copper and the plastic separator. The copper, density and Mn(OH)2 are included to determine if it is possible to precip-
8.96 g cm3, and plastic (polyethylene), density w0.9 g cm3 itate them out before the impurities are removed and reduce the
were separated by heavy media. The material was put into a recovery efficiency and the reactions shown below. This table ig-
solution of diiodomethane (Alfa Aesar) density 3.3 g cm3 and nores common ion effects. The reactions for precipitation of nickel,
the material was allowed to settle for 30 min. After 30 min the manganese and cobalt hydroxide are shown in Eqs. (2)e(4). If one
material on the surface of the diiodomethane was skimmed off of these reactions occur before NixMnyCoz(OH)2 is precipitated or
and separated from the sunken material. before Fe(OH)3, AL(OH)3 or Cu(OH)2 is precipitated, then recycling
efficiency will be reduced. The solubility constant for NixMnyCo-
z(OH)2 is unknown at this time.
2.3. Processing spent Li-ion batteries with added cathode powders
Ni2þ þ 2NaOH ¼ NiðOHÞ2 þ 2Naþ (2)
All steps follow the procedure for the previous section. Except in
step 5, 2.07 g of LiMn2O4, 2.73 g of LiNi0.33Mn0.33Co0.33O2 and 0.49 g Co2þ þ 2NaOH ¼ CoðOHÞ2 þ 2Naþ (3)
LiFePO4 powders were added to the sub 250 mm fraction. These
masses were chosen to make a composition of cathode powders
present representative of the 2012 battery market. Based on pre- Mn2þ þ 2NaOH ¼ MnðOHÞ2 þ 2Naþ (4)
vious experiments 75% of the mass of the sub 250 mm fraction was
assumed to be LiCoO2. Therefore based on the assumed mass of Step 3: The concentration of Co, Ni, Mn, Li, Al, Fe and Cu in so-
LiCoO2, the masses of LiMn2O4, LiNi0.33Mn0.33Co0.33O2 and LiFePO4 lution was measured with AAS. Then MnSO4 and NiSO4 were
powders were calculated. added to adjust the ratio of Co, Ni, and Mn to 1:1:1. This ratio can
be changed based on the target materials, which is one of the
2.4. Hydrometallurgical recovery of cathode materials biggest advantages of our process.
Step 4: The pH was increased to 11 to precipitate out
The sub 250 mm fraction from step 5 above and the light material Ni0.33Mn0.33Co0.33(OH)2. It was shown previously that at pH 11
from step 6 were taken through the hydrometallurgical process. the concentrations of CoSO4, NiSO4 and MnSO4 are less than
The plastics were included in the leaching solution because it was 107 M [18].
observed that some cathode material had attached to the plastic
separator after shredding. Since the cathode powders leach into 2.5. Synthesis of LiNi0.33Mn0.33Co0.33O2
solution and the plastics do not, they were easily separated by
filtration in the first step of the hydrometallurgical process. The precipitated Ni0.33Mn0.33Co0.33(OH)2 was mixed with Li2CO3
in molar ratio 1:1.1 and then ground in the mortar. The precursor
Step 1: The powders and plastics were leached into solution was subjected to ball-milling for 48 h and pressed into pellets
using 60 ml of 4 M sulphuric acid and 10 ml of 50wt% hydrogen (Pressure ¼ 15000lbs, 12.96 mm diameter). The pellets were then
peroxide at 65e70  C. After allowing the solutions to react for sintered at 900  C for 15 h in air, and the temperature was increased
2 h, the solutions were diluted to 100 ml with DI water in a at a rate of 9  Cmin1 from room temperature. The reaction product
100 ml volumetric flask. The concentrations of Co, Ni, Mn, Li, Al, was ground into powder using a mortar and pestle.
Fe and Cu present in solution were measured with AAS. Carbon
and residual LiFeO4 did not dissolve into solution and were
2.6. Characterization
filtered out.
Step 2: The pH of the solution was increased to 6.47 by adding
The crystal structure of the Ni0.33Mn0.33Co0.33(OH)2/
2 M NaOH. The precipitate was filtered from solution, washed
Ni0.33Mn0.33Co0.33O(OH) precipitate and LiNi0.33Mn0.33Co0.33O2 was
and dried. AAS was performed on the remaining solution. The
characterized using an X-ray diffractometer (Bruker D8 Focus; S/N
rational for increasing the pH to 6.45e6.5 was to remove all
203207 with copper Ka radiation, Cu target: S/N 07/03-1120). X-ray
the Fe, Al, and Cu impurities. Once the pH reached 6.45e6.5
scintillation detector SD 650 Nal (TI) was used. The concentration of
the solutions were allowed to stir for 4 h, then the solutions
metal ions in solution was tested by AAS. The scanning electron
were diluted to 250 ml with DI water in a 250 ml volumetric
microscopy (SEM) images were obtained with a JEOL JSM-7000F
flask.
electron microscope. The pH of the solution was tested with a
Milwaukee MW102 pH meter.
Because Fe, Al and Cu are present in Li-ion batteries and they
leach into solution when exposed to 4 M H2SO4 with H2O2, they are
therefore common impurities in the recovered cathode materials, if 3. Results and discussion
steps are not taken to remove them from solution. Shown in Table 2
is the pH versus concentration in solution of the impurities 3.1. Testing Cu impurity removal

The concentrations of Co, Ni, Mn and Cu after adding their sul-

Table 2 phates to water were measured as a function of pH shown in
pH versus concentration of desired elements. Table 3. The results show that even when copper is present in the
Material pH at 1 M pH at 105 M initial solution there is at least 96% of Ni, Co and Mn remaining in
solution at pH 6.5, while only 14.5 ppm copper remained in solu-
Fe(OH)3 1.14 2.81
Al(OH)3 2.82 4.49 tion. Thus this is a suitable method for recovering Co, Ni and Mn
Cu(OH)2 4.15 6.85 from spent Li-ion batteries while keeping the copper impurity
Ni(OH)2 6.37 8.87 concentration low. Based on the solubility constant for copper
Co(OH)2 6.9 9.38 shown in the appendix the concentration of copper in solution
Mn(OH)2 7.65 10.15
should be on the order of 1 ppm. However, this represents a
 E. Gratz et al. / Journal of Power Sources 262 (2014) 255e262 259

Table 3
Efficiency of recovery for Co, Ni, and Mn with and without Cu.

Concentration at Total mass at Concentration at Total mass at Difference Efficiency

pH > 1 (ppm) pH > 1 (g) pH > 6.5 (ppm) pH > 6.5 (g) (% of original)

Experiment with Cu
Co 12320 1.23 4739 1.18 0.05 96
Ni 13540 1.35 5329 1.33 0.02 99
Mn 12260 1.22 4697 1.17 0.05 96
Cu 989.5 0.099 14.54 3.63  103 0.095 3.6
Experiment without Cu
Co 12230 1.22 4881 1.22 0 100
Ni 13540 1.35 5528 1.38 þ0.03 102
Mn 12280 1.23 4892 1.22 0.01 99
Cu ND ND NA

simplistic view of the problem, as there are many ions in solution 3.2.4. Sieving
and common ion effects would need to be taken into consideration. The material was sieved in a 12 in diameter 250 mm screen. The
mass of the sub 250 mm fraction was 3.79 g. The fraction with mass
greater than 250 mm weighed 4.62 g.
3.2. Processing 20 g of Li-ion batteries
3.2.5. Density separation
3.2.1. Shredding
The heavy fraction weighed 1.62 g and the light fraction
Shredding is an important step in processing the batteries
weighed 3.01 g. The sunk materials are shown in Fig. 5. The sunk
because the batteries must be processed into a workable form. By
fraction was leached into H2SO4 and H2O2, and the amount of
using a hammer mill the batteries were processed into a workable
cathode materials were measured. The cathode materials weighed
size less than ¼00 dimensions, Fig. 4.
0.3 g which represents 6% of the total theoretical cathode materials
present in 20 g of shredded batteries.
3.2.2. Magnetic removal
A rare earth magnet was passed over the material in Fig. 4. This 3.2.6. Leaching recovered powders
removed 4.77 g of material, and the material was primarily steel The composition of the leached solution was measured with
pieces on the order of 3 mm in diameter. These steel pieces con- AAS before and after increasing the pH to 6.47, Table 4. After
tained non-ferrous attachments which were removed by hand and leaching, 4.58 g NiSO4$7H20 and 3.87 g MnSO4$H2O were added to
weighed to be 0.091 g. The removed material is shown in Fig. 5. Of the solution so that the Co:Ni:Mn atomic ratio was 1:1:1.
this 0.085 g was found to be cathode materials, this represents 1.7% The product was precipitated at pH 11. The mass of the recov-
of the total theoretical cathode materials present in 20 g of ered product was 9.5 g after drying in air at 75  C for 24 h. After
shredded batteries. crushing the product with a mortar and pestle, a 1.75 g sample was
removed from the product and leached into the sulphuric acid/

3.2.3. Aluminium dissolution

2 M NaOH solution was analyzed with AAS to determine if any
cathode materials were removed during the exposure to the caustic
solution. Aluminium was measured to be 9128 ppm, Fe, Cu, Co, Mn,
and Ni were all below the detection limit of AAS. The mass of
material remaining after reactions with the 2 M NaOH was 13.58 g.

Fig. 4. Hammer milled Li-ion batteries with ¼00 screen spacing (step 1). Fig. 5. Material separated from hammer milled batteries by rare earth magnet (step 2).
260 E. Gratz et al. / Journal of Power Sources 262 (2014) 255e262

Table 4
Impurity levels (ppm) in solution.

Cu Fe Al Mn Co Ni

pH < 1 2863 471.4 2986 608.3 16952 1160

Grams total 0.286 0.0471 0.2986 0.0608 1.69 0.116
Adjusted total 0.2574 0.0424 0.268 0.0547 1.526 0.104
Ph ¼ 6.47 ND ND ND 208.4 5633 296.8
Total grams ND 0.0528 1.408 0.074
Recovery 97% 92% 71%
efficiency

hydrogen peroxide solution. The composition of the product solu-

tion after leaching is shown in Table 5. Additionally, Table 5 shows
the impurity concentration of Fe, Al and Cu to be less than 50 ppm.
It is believed that Li-ion battery cathode materials need impurity
concentrations less than 100 ppm. The cathode materials are made
by solid state reaction of Ni0.33Mn0.33Co0.33(OH)2 with Li2CO3, if
both components have impurity levels less than 50 ppm then the Fig. 6. XRD of Ni0.33Mn0.33Co0.33O(OH) product precipitated out of solution at pH 11.
LiNi0.33Mn0.33Co0.33O2 product will have impurity levels less than
100 ppm. Therefore, increasing the pH to 6.47 is sufficient to
remove the impurities in order to produce precursors for Li-ion 9 ml of 50wt% H2O2. 1.95 g of cobalt was recovered; 1.71 g of
batteries. manganese and 1.83 g of nickel were also recovered (Table 7). The
The XRD spectrum of the recovered product is shown in Fig. 6. atomic ratio of recovered Mn, Ni, and Co was 1:1.02:1.08, respec-
The product is a mixture of Ni0.33Mn0.33Co0.33(OH)2 and tively (Table 7). The Fe, Al and Cu were all under 50 ppm. The ef-
Ni0.33Mn0.33Co0.33O(OH) phases. To precipitate the pure ficiency of recovery for all Co, Mn and Ni ions added from batteries/
Ni0.33Mn0.33Co0.33(OH)2 phase, the precipitation process needs to cathode powders and sulphate solutions was 86%, 98% and 95%
be carried out under nitrogen [27]. respectively. The total recovery efficiency was greater than 85% for
all of the target metals.
3.2.6.1. Final recovery efficiency. The recovery efficiency of all Co, Ni
and Mn in solution was 94%, 96% and 91% respectively when ac- 3.4. Characterization of LiNi0.33Mn0.33Co0.33O2
counting for the initial concentrations of Co, Mn and Ni and the
added metal sulphates. The recovery efficiency is greater than 90% The XRD of the synthesized LiNi0.33Mn0.33Co0.33O2 prepared
which is sufficient for recycling Li-ion batteries. The atomic ratio of from the recovered Ni0.33Mn0.33Co0.33(OH)2 is shown in Fig. 7. Lat-
the Mn:Co:Ni in the Ni0.33Mn0.33Co0.33O(OH) and Ni0.33Mn0.33- tice parameters of samples were calculated from the d-value of 10
Co0.33(OH)2 recovered product is 1.000:1.005:1.069. strong diffraction peaks. The sample peaks can be indexed as a-
NaFeO2 structures with a space group of R3 m. Distinct splitting of
(006)/(012) and (018)/(110) peaks indicates that the sample pos-
3.3. Processing 20 g of Li-ion batteries with added cathode powders sesses a well-developed layered-structure [28]. No impurity peaks
are present in either pattern, which suggests that LiNi0.33Mn0.33-
20 g of shredded Li-ion batteries were processed using the Co0.33O2 is of high purity.
procedure previously described. The pH was increased to 6.47 to The SEM image of the LiNi0.33Mn0.33Co0.33O2 product is shown
remove the impurities, over 99% of the impurities where removed, in Fig. 8. Homogenous particles are well distributed throughout
Table 6. At pH 6.47 over 90% of the manganese and nickel ions without any agglomeration; particle size ranges from 200 to
remained in solution and over 80% of the cobalt ions remained in 300 nm. Uniform particle size and homogenous particles are critical
solution. The decrease in the efficiency of cobalt recovery is factors for high quality materials. Smaller particle size allows a
believed to be due to the increase in cobalt concentration from the shorter lithium diffusion length, which in turn provides higher
added LiNi0.33Mn0.33Co0.33O2 powders. 5.59 g of NiSO4$6H2O, 1.12 g battery performance, which has been demonstrated in the previous
CoSO4$5H2O and 0.53 g MnSO4 H2O were added so the precipitated results [18].
product would have a Ni:Mn:Co atomic ratio of 1:1:1. The pH was
then increased to 11 with 2 M NaOH to precipitate out the
3.5. Waste management
Ni0.33Mn0.33CO.33(OH)2 product.
The process produces minimal waste. The residual basic solution
3.3.1. Final recovery efficiency can be neutralized and disposed. The amount of solution generated
11.6 g of product was recovered by precipitation, and 1.58 g of 3112 L ton1 of Li-ion batteries recycled based on the ratios of acid
the product after drying, was leached into 20 ml of 4 M H2SO4 and

Table 6
Table 5 Impurity concentration at pH 6.47 for batteries with added cathode powders.
Composition of hydroxide product.
Cu Fe Al Mn Co Ni
Co Mn Ni Fe Cu Al Li
pH < 1 (ppm) 4274 2400 3358 15840 23270 7675
Product 2991 2852 3185 0.90 5.2.0 ND 23.00 Mass (g) 0.427 0.24 0.33 1.58 2.327 0.767
(ppm) Adjusted total (g) 0.363 0.204 0.281 1.343 1.978 0.652
Product (g) 0.299 0.285 0.319 9.0E-5 5.2E-4 2.3E-3 Ph ¼ 6.47 (ppm) 9.99 3.49 8.4 5395 6591 2381
Total in 1.615 1.514 1.72 4.8E-4 2.8E-3 0.0129 Mass (g) 2.5e-3 8.73e-4 2.1e-3 1.34 1.648 0.595
sample (g) (5 ppm) (29 ppm) Recovery efficiency 0.68% 0.4% 0.7% 100% 83.4% 91.2%
 E. Gratz et al. / Journal of Power Sources 262 (2014) 255e262 261

Table 7
Composition of hydroxide products.

Co Mn Ni Fe Cu Al Li

Product (ppm) 9181 7862 8630 1.3 2.6 1.45 287

Product (g) 0.266 0.227 0.25 3.77E-5 7.54E-5 4.21E-5 8.3E-3
Total in sample (g) 1.954 1.673 1.834 2.39E-5 (24 ppm) 4.77E-5 (47 ppm) 2.66E-5 (27 ppm) 0.061

and base used to battery materials used in this paper. However, 0.6% of 250 kg. The efficiency of getting cathode materials from the
since the solution is non-toxic the basic solution can be neutralized battery and into solution was measured to be 97.7% based on the
and disposed. The amount of acid required to neutralize 3112 L of amount of cathode material lost during magnetic and density
solution is less than $5 based on our calculation and therefore not separation (sections 3.2.2 and 3.2.5). The efficiency of hydromet-
included in the economic analysis section that follows. allurgical process is expected to be 90%. Thus the total recovery
efficiency of the cathode materials is expected to be 88%. Therefore
3.6. Economic analysis it is expected that 249.99 kg of steel will be recovered, 62 kg of
copper, 220 kg of cathode materials and 121 kg of sodium alumi-
By analyzing the recovery efficiency of each step in the 20 g nate based on the masses in 1 ton of Li-ion batteries (Table 1). The
process (Figs. 1 and 2) we can scale to 1 ton of batteries and adjust value of the recovered products with these efficiencies is expected
the economic analysis from previous research [18]. The amount of to be $6432/ton (Table 8), with over 90% of the recovered value
cathode materials removed during magnetic separation represents coming from the new cathode materials. The chemical cost to
4.25 kg or 1.7% of 250 kg and 1.5 kg during density separation or recycle the batteries is expected to be $1419/ton shown in Table 9.
This results in a profit margin of $5013. It is noted that we only
consider the material balance. In the future, we will add the labor
cost, equipment cost and energy cost in the economical analysis.

4. Conclusion

We have demonstrated the feasibility of recycling Li-ion batte-

ries regardless of cathode chemistry via a low temperature hy-
drometallurgical process. The key to this process is the recovery of a
specific targeted nickel manganese cobalt hydroxide, and in this
research Ni0.33Mn0.33Co0.33(OH)2 was targeted and recovered. The
Ni0.33Mn0.33Co0.33(OH)2 was used to produce new cathode
LiNi0.33Mn0.33Co0.33O2, other cathodes maybe produced depending
on the target nickel manganese cobalt ratio. The purity of the
Ni0.33Mn0.33Co0.33(OH)2 recovered was high, with less than 50 ppm
of Al, Cu and Fe impurities. It is important to have high purity so
that the performance of the LiNi0.33Mn0.33Co0.33O2 is not compro-
mised. The recovery efficiency of the valuable cathode metals, Ni,
Co and Mn, was w90%. Accounting for and scaling the efficiencies
measured by this process, resulted in a margin of $5013/ton be-
tween the value of materials recovered and the cost of materials to
Fig. 7. XRD of LiNi0.33Mn0.33Co0.33O2 is from made from recycled batteries.
recycle.

Table 8
Value of recovered materials.

Material Mass (lbs) $/lb Value ($)

Steel 550 0.25 138

Copper 136.4 3.40 464
Sodium aluminate 266 1.32 351
LiNi0.33Mn0.33Co0.33O2 670 8.8 5899
Total 6432

Table 9
Chemical cost to recycle lithium ion batteries.

Chemical Mass (lb) $/lb Cost

NaOH 584.8 0.16 93

H2SO4 675.4 0.07 46
H2O2 (30 wt%) 235 0.23 53.4
MnSO4*H2O 208.7 0.27 56.4
NiSO4*7H2O 379.7 2.73 1036.7
Li2CO3 251.4 3.63 9,14.18
Total cost 1419.4
Margin 5013
Fig. 8. SEM image of recovered LiNi0.33Mn0.33Co0.33O2.
262 E. Gratz et al. / Journal of Power Sources 262 (2014) 255e262

Acknowledgments [10] C.K. Lee, K.-I. Rhee, Hydrometallurgy 68 (1e3) (2003) 5e10.
[11] M. Contestabile, S. Panero, B. Scrosati, J. Power Sources 83 (1e2) (1999)
75e78.
This work was supported the Center for Resource Recovery & [12] J. Nan, D. Han, X. Zuo, J. Power Sources 152 (0) (2005) 278e284.
Recycling (CR3) and the National Science Foundation (NSF) under [13] R.-C. Wang, Y.-C. Lin, S.-H. Wu, Hydrometallurgy 99 (3e4) (2009) 194e201.
grants 1230675 and 1343439. [14] A. Chagnes, B. Pospiech, J. Chem. Technol. Biotechnol. (2013).
[15] D. Cheret, S. Santen, Battery Recycling, 2007. Google Patents.
[16] F. Tedjar, J.-C. Foudraz, Method for the Mixed Recycling of Lithium-Based
References Anode Batteries and Cells, 2010. Google Patents.
[17] C. Vadenbo, Prospective Environmental Assessment of Lithium Recovery in
[1] Lithium-ion Battery Market: Cell & Components, Yano Research Institute, Battery Recycling, ETH IED-NSSI. ETH Zürich, Zürich, 2009, p. 174.
2011. [18] H. Zou, et al., Green. Chem. 15 (5) (2013) 1183e1191.
[2] B. Scrosati, J. Garche, J. Power Sources 195 (9) (2010) 2419e2430. [19] L. Wang, et al., J. Solid State Electrochem. 13 (8) (2009) 1157e1164.
[3] M. Armand, J.M. Tarascon, Nature 451 (7179) (2008) 652e657. [20] M.-H. Kim, et al., J. Power Sources 159 (2) (2006) 1328e1333.
[4] J.M. Tarascon, M. Armand, Nature 414 (6861) (2001) 359e367. [21] Z. Liu, A. Yu, J.Y. Lee, J. Power Sources 81e82 (0) (1999) 416e419.
[5] S.B. Peterson, J. Apt, J.F. Whitacre, J. Power Sources 195 (8) (2010) 2385e2392. [22] J.-M. Kim, H.-T. Chung, Electrochim. Acta 49 (6) (2004) 937e944.
[6] A. Millner, Modeling lithium ion battery degradation in electric vehicles, in: [23] J. Choi, A. Manthiram, J. Power Sources 162 (1) (2006) 667e672.
Innovative Technologies for an Efficient and Reliable Electricity Supply (CIT- [24] S.-W. Cho, K.-S. Ryu, Mater. Chem. Phys. 135 (2e3) (2012) 533e540.
RES), 2010 IEEE Conference on, 2010. [25] T. Suzuki, et al., Sep. Purif. Technol. 98 (0) (2012) 396e401.
[7] J. Xu, et al., J. Power Sources 177 (2) (2008) 512e527. [26] T. Tanii, et al., Method for Crushing Cell, 2003. Google Patents.
[8] L. Gaines, A comparison of Li-ion battery recycling options, in: SAE Interna- [27] A. van Bommel, J.R. Dahn, J. Electrochem. Soc. 156 (5) (2009) A362eA365.
tion, SAE International, Detriot, 2012. [28] J. Liu, et al., J. Power Sources 174 (2) (2007) 701e704.
[9] T. Georgi-Maschler, et al., J. Power Sources 207 (0) (2012) 173e182.