Journal of Cleaner Production 185 (2018) 646e652

Contents lists available at ScienceDirect

Journal of Cleaner Production

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

Recovery of valuable materials from spent lithium-ion batteries by

mechanical separation and thermal treatment
Fangfang Wang a, Tao Zhang b, c, *, Yaqun He b, c, Yuemin Zhao a, b, Shuai Wang b, c,
Guangwen Zhang b, Yu Zhang b, Yi Feng b
a
School of Environment Science and Spatial Informatics, China University of Mining and Technology, Xuzhou, 221116, China
b
School of Chemical Engineering and Technology, China University of Mining and Technology, Xuzhou, 221116, China
c
Advanced Analysis and Computation Centre, China University of Mining and Technology, Xuzhou, 221116, China

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

Article history: In this paper, a mechanical separation and thermal treatment process is developed to recover valuable
Received 17 August 2017 metals and graphite from the
0.25 mm crushed products of spent lithium-ion batteries (LiBs). Effect of
Received in revised form key parameters for roasting such as the temperature and roasting time are investigated to determine the
4 March 2018
most efficient conditions for surface modification of the mixed electrode materials by roasting. The
Accepted 7 March 2018
Available online 8 March 2018
roasted mixed electrode materials are separated by flotation operation to recover the cathode material
and anode materials respectively. The results show that most of the organic outer layer coated on the
surface of the mixed electrode materials can be removed at the temperature of 450  C for 15 min. After
Keywords:
Spent lithium-ion batteries
roasting treatment, the original wettability of LiCoO2 and graphite is regained. The
0.25 mm crushed
Recycling products of spent LiBs can be separated into LiCoO2 concentrate and graphite concentrate by flotation
Surface analysis process efficiently. The enrichment ratios of Co, Mn, Cu and Al are 1.35, 1.29, 1.25 and 1.19, their recovery
Thermal treatment rates are 97.66%, 93.66%, 90.14% and 86.29%, respectively. This process proposed for the recovery of
Flotation valuable materials is simple and of high efficient for the spent lithium-ion batteries recycling industry.
© 2018 Elsevier Ltd. All rights reserved.

1. Introduction
Lithium-ion batteries (LiBs) have been widely used in portable
electronic equipment such as mobile phones, cameras, laptops,
mobile power, etc., for the LiBs are of excellent performances
including safe, high-energy density, long serving life, without
hazardous elements (Zeng et al., 2015). In recent years, with the
rapid development of electric vehicles, LiBs have been chosen to
be the power sources (Hanisch et al., 2015). Therefore, the
worldwide consumption of LiBs has maintained rapid growth.
However, the service life of LiBs is in commonly 2e3 years, the
more consumption means the more amount of scrap. Typical LiBs
are generally composed of a metal shell, a cathode, an anode,
organic electrolyte, and a polymer diaphragm. The anode is a
copper foil coated with a mixture of graphite, conductor, poly-
vinylidene fluoride (PVDF) binder, and additives such as organic
electrode and LiPF6. Similarly, the cathode is an aluminium foil
* Corresponding author. School of Chemical Engineering and Technology, China
University of Mining and Technology, Xuzhou, 221116, Jiangsu. China.
E-mail address: taozhang@cumt.edu.cn (T. Zhang).
coated with a mixture of LiCoO2 (or some other kinds of cathode
materials) with the other things. It has been reported that in spent
LiCoO2 LiBs, there are 17.6 wt% Co, 7.2 wt% Cu, 21.6 wt% Al (Zhang
et al., 2014a, b). From the viewpoints of environmental preser-
vation (Choubey et al., 2017), recovery of major components or
valuable resources (Bahaloo-Horeh and Mousavi, 2017), the
recycling of spent LiBs is quite necessary.
Recycling of spent LiBs arouses many researchers’ interest, and
there are many useful achievements. In general, there are three
kinds of methods. They are physical method, chemical method and
biological method (Ordon ~ ez et al., 2016). In a recycling process of
spent LiBs, physical method is usually applied as a pre-treatment
(Chen et al., 2017) to treat the outer cases and shells and to
concentrate the metallic fraction, which will be conducted to
chemical method (Pinna et al., 2017) (a hydrometallurgical or a
pyrometallurgical recycling process) or biological method (Horeh
et al., 2016) (a biohydrometallurgy recycling process) to gain
purer metal products (Zhang et al., 2013). However, in either case,
the purification operation is very difficult because of the organic
matter used in LiBs, such as the graphite, polymer diaphragm,
electrolyte and PVDF binder, which not only increase the volume of

https://doi.org/10.1016/j.jclepro.2018.03.069
0959-6526/© 2018 Elsevier Ltd. All rights reserved.
 F. Wang et al. / Journal of Cleaner Production 185 (2018) 646e652
the feeding, hinder the chemical reactions but also pollute the
environment.
It has been proved that after crushing, the metal shell, copper
foil, aluminium foil and polymer diaphragm still remained in large
size, with air separation process or gravity separation, the value
materials in big size could be recovered efficiently. However, elec-
trolyte and PVDF binder following LiCoO2 and graphite particles
concentrated in small size fractions (Al-Thyabat et al., 2013). To
separate the LiCoO2 particles from graphite particles become a new
problem. As the electrode materials have a very small particles size
that is only 20e70 mm, though the density of LiCoO2 particle is
much higher than graphite particle, it is almost impossible to
separate the mixed electrode materials with density separation.
However, LiCoO2 is hydrophilic for it is ionic crystal with strong
polarity; and graphite is hydrophobic for it is nonpolar. Therefore,
LiCoO2 and graphite have opposite surface wettability (Huang et al.,
2016). In theory, flotation may be a useful method for the separa-
tion of LiCoO2 and graphite. The flotation test results showed that
LiCoO2 and graphite particles were coated by electrolyte and PVDF
binder, which makes flotation result bad, to remove the organic
outer layer can improve the flotation result (Zhang et al., 2014a, b).
The flotation results are totally depended on the removing of
electrolyte and PVDF binder from the surfaces of LiCoO2 and
graphite particles. It was reported that the electrolyte and PVDF
binder can be removed by Fenton solution (He et al., 2017), the
organic materials such as electrolyte and PVDF binder were
destroyed by $OH generated from Fenton solution. The mixed
LiCoO2 and graphite can be separated and recovered respectively
and efficiently. However, Fenton solution needs Fe2þ reacting with
H2O2 to produce $OH, the introducing of Fe will make the following
chemical purification more complex and expensive. In addition, the
Fenton solution reacting conditions must be controlled very care-
fully, or the removing rate of electrolyte and PVDF binder will be
very low, the flotation result would be bad. Therefore, the elec-
trolyte and PVDF binder removing method should be easy to con-
trol, reacted fast and no more impurities introduced.
In this work, we proposed to meet the above requirements by
roasting method and recover LiCoO2 and graphite from the mixed
electrode materials by flotation respectively. The thermostability of
both recovered LiCoO2 and graphite was analysed firstly, and the
effect of roasting time and temperature on the wettability of the
recovered electrode materials were studied and discussed. Under
the optimum condition, the recovered mixed electrode materials
were roasted, and by flotation, the LiCoO2 and graphite were
separated from each other and the mechanism was discussed, too.
2. Experimental
2.1. Sampling and preparation
The graphical presentation of the experimental approach was as
shown in Fig. 1. Spent LiBs used in this research were 18650 type,
and the total weight was 1 kg. Before dismantling and crushing
steps, in order to make the experiment process safe, the spent LiBs
were discharged completely in 5 wt% sodium chloride solution for
24 h, and let them air-dry afterwards. Then, the spent LiBs were
dismantled by manual work, the anode and cathode were obtained
respectively.
In this research, three kinds of samples were prepared, they
were the
0.25 mm crushed product only come from the anode
which were the graphite concentrate, the one only come from the
cathode which was the LiCoO2 concentrate and the one come from
both the anode and cathode which was mixed LiCoO2 and graphite
concentrate.
647
Fig. 1. Graphical presentation of the experimental approach.
2.2. Method
2.2.1. Roasting test
Graphite concentrate and LiCoO2 concentrate were analysed by
combining TG（Retsch STA-409C, Germany）and FTIR (NICOLET-
Nenus 470, USA) to show ignition loss and the composition of the
generated waste air from 25  C to 900  C. The TG test was in an air
atmosphere, the airflow was 100 ml/min, the temperature range
was from 25  C to 900  C and the heating rate was 15  C/min. the
FTIR test range was from 400 to 4000 cm
1, the resolution was
better than 1 cm
1, the S/N was better than 6000.
All the two samples were thermally treated in muffle furnace,
and the roasted samples were analysed by SEM and XPS to show
the surface properties. Powder structure and particle configuration
of the
0.5 mm crushed products were analysed by a scanning
electron microscope (SEM, FEI quanta 250, America) equipped with
a tungsten filament and coupled with an energy dispersive spec-
trometer (EDS, Bruker QUANTAX400-10, Germany) using a silicon
drift detector.
The XPS test was carried out by an X-ray photoelectron spec-
trometer (XPS ESCALAB 250Xi, America) at room temperature in an
ultra-high vacuum (UHV) system. The base pressure of the analysis
chamber during measurements was lower than 1  10
9 mbar. Al
Ka radiation (hv ¼ 1486.6eV) from a monochromatized X-ray
source was used for the experiments. For all analyses, the take-off
angle of photoelectrons was 90 and the spot size was 900 mm.
Spectra were recorded with the pass energy of 20 eV, and the en-
ergy step size was 0.05 eV.
The contact angle was measured by a Contact Angle Meter
（DSA100, Kruss, Germany）to show the wettability changes. The
roasted sample was pressed into sheet, and a drop of distilled water
was put on the surface of the sheet in six different positions. Six
contact angles were got, and the average value was calculated to be
the last result.
2.2.2. Flotation
Pulp density, impeller speed, pH and aeration during flotation
were maintained at 40 g/L, 1800 rpm, 9 and 0.8 L/min, respectively.
Methyl isobutyl carbinol (MIBC) and n-dodecane were used as
frother and collector in the flotation with the dosage of 150 g/t and
300 g/t. The dry masses of the concentrates and tailings were
determined and samples assayed for their elemental compositions
by X-ray fluorescence (XRF, Bruker S8 Tiger, Germany) using a
powder pressed method to prepare the samples.
648 F. Wang et al. / Journal of Cleaner Production 185 (2018) 646e652

Therefore, in order to protect graphite from being burnt down and

to recover graphite as much as possible, the roasting temperature
should be lower than 550  C.
As in the LiCoO2 concentrate, only the organic outer layer could
be burnt, the generated gas during roasting process was analysed
by IR. The results showed that at the temperature of 110  C, there
was some steam generated, and with the temperature rising, CO2
began to be detected at 355  C, which implied the organic outer
layer started being burnt. In addition, at the temperature of 490  C,
the CO2 signal was the strongest. When the temperature reached
535  C, the CO2 signal was gone. That proved the organic outer layer
could be removed at the temperature lower than 500  C. Therefore,
the roasting temperature should be in the range of 300e500  C.
This could make sure to remove the organic outer layer while the
graphite values could be protected.

3.2. The effect of roasting temperature on surface property

Contact angle can reflect materials’ wettability, small contact

Fig. 2. Thermal gravimetric analysis (2a-LiCoO2,2b-graphite).
3. Results and discussion
3.1. Thermostability analysis
The TG and DTG analysis results were as shown in Fig. 2. The
results showed that during the roasting process, the mass variation
of LiCoO2 concentrate (Fig. 2a) was gently, in the end, there was
89.5 wt% left; while the mass of graphite concentrate (Fig. 2b)
changed a lot, only 9.27 wt% was left in the end of the roasting
process. As we reported before (Zhang et al., 2014a, b), the LiCoO2
and graphite particles recovered from spent LiBs were coated by
electrolyte and PVDF binder. The structure of these particles was a
LiCoO2 or graphite inner core with an outer organic layer. Thus
during the roasting process, the outer organic layer was burnt
down firstly, and as LiCoO2 was inorganic material which could not
be burnt down, at high temperature, the mass of LiCoO2 concen-
trate (Fig. 2a) changed a little. However, the inner core of graphite
particle was also organic material that could be burnt down. Before
the temperature reached 550  C, the mass variation of graphite
concentrate was very small, there was only 14 wt% lost. However,
when the temperature was higher than 550  C, the mass of the
graphite concentrate declined dramatically. Based on this, it could
be deduced that the graphite starting burning at 550  C, and at the
temperature of 620  C, the mass loss ratio was the highest.
angle means hydrophilic and large contact angle means hydro-
phobic. For flotation, the bigger the difference between two kinds
of materials’ contact angles is, the better the flotation result would
be. The LiCoO2 and graphite concentrates were roasted in a muffle
furnace at the temperatures of 350, 400, 450 and 500  C for 1 h
respectively. And the contact angles of the recovered samples
before and after being roasted at different temperatures were
measured as listed in Table 1. The results showed that before
roasting, the contact angles of the original electrode materials were
almost the same, so flotation was impossible. With the roasting
temperature increasing, both of the contact angles of the LiCoO2
and graphite concentrates become smaller. D-value defined by
formula (1) was used to show the difference between the contact
angles of the LiCoO2 and graphite concentrates.
D
value ¼ aG
aL (1)
Where aG is the contact angle of graphite concentrate, aL is the
contact angle of LiCoO2 concentrate. The D-value was the highest at
400  C, which meant the difference of the wettability between the
LiCoO2 and graphite concentrates was the largest, and that was a
good condition for flotation.
SEM images of graphite concentrate roasted at 350, 400, 450
and 500  C for 1 h were as shown in Fig. 3. The surfaces of the 350
and 400  C roasted products were smooth, but when the roasted
temperature was higher than 400  C, the surface of the graphite
become rough, and at 500  C, the state of the surface was even
worse. At high roasting temperature, not only the organic outer
layer was removed, but also the graphite was oxidized badly, there
would be many oxygen-containing functional groups generated on
the surface of the graphite. As a result, the hydrophilia of the

Table 1
Contact angles of the recovered samples at different roasting temperatures.

Temp./ C Material Experimental results Results D-value/

1 2 3 4 5 6

0 LiCoO2/ 54.1 53.8 53.9 53.7 54.3 53.8 53.9 ± 0.2 1.9
Graphite/ 55.7 55.9 55.7 55.8 55.8 55.9 55.8 ± 0.1
350 LiCoO2/ 26.5 26.8 26.6 26.3 26.5 26.4 26.5 ± 0.2 28.5
Graphite/ 55.3 55.2 55.2 54.9 54.8 55.1 55.0 ± 0.2
400 LiCoO2/ 19.1 19.5 19.2 19.3 19.1 19.2 19.2 ± 0.1 33.8
Graphite/ 53.0 52.9 52.7 53.2 53.5 53.1 53.0 ± 0.2
450 LiCoO2/ 18.2 18.5 18.4 18.4 18.6 18.3 18.4 ± 0.1 30.7
Graphite/ 49.1 49.0 49.3 49.2 48.9 49.1 49.1 ± 0.1
500 LiCoO2/ 16.3 16.2 16.3 16.1 16.4 16.0 16.2 ± 0.1 24.2
Graphite/ 40.6 40.5 40.6 40.4 40.1 40.2 40.4 ± 0.2
 F. Wang et al. / Journal of Cleaner Production 185 (2018) 646e652 649

Fig. 3. SEM images of roasted graphite concentrate at different temperature (3a-

350  C, 3b-400  C, 3c-450  C, 3d-500  C).

Fig. 4. Mass loss rates of electrode materials after roasting for 1 h.

graphite become strong and the contact angle become small.
The contents of the elements of C, O and Co on the surface of
that the organic outer layer could be mostly removed by roasting at
LiCoO2 concentrate were measured by XPS. As it was shown in
400  C for 1 h. However, the MLR graphite concentrate was 29.5% at
Table 2, with the roasting temperature increasing, the content of C
400  C, at 500  C, the MLR was even as high as 84.2%. That would
become lower and the contents of O and Co become higher. That
loss much graphite values. Although before the roasting tempera-
was because at higher roasting temperature, the oxidation reaction
ture reached 550  C, the oxidative decomposition rate of graphite
was even stronger, therefore, the organic outer layer coated on the
was low, with roasting time increasing, the graphite also could be
surface of LiCoO2 particles could be removed much more
oxidative decomposed badly. Therefore, the roasting time effect
completely. The hydrophobic group was removed and hydrophilic
should be studied.
group was generated, and as a result, the contact angle of LiCoO2
concentrate became smaller.
The mass loss rate (MLR) defined by formula (2), was used to
3.3. The effect of roasting time on surface property
show the mass loss during roasting.
The condition roasting at 400  C for 1 h could get a good surface
MLR ¼ m1 =m0 (2)
property for flotation, but it was not very efficiency for 1 h was too
Where m1 is the mass after roasting, m0 is the mass before long. Therefore, in this section, we chose to reach roasting time
roasting. The MLRs of both LiCoO2 and graphite concentrates after effect at 450  C.
roasting for 1 h were as shown in Fig. 4. With the roasting tem- The contact angles of the recovered samples roasted for 15, 30,
perature increasing, the MLR of LiCoO2 concentrate became higher. 45 and 60 min were measured as listed in Table 3. The results
But at 400, 450 and 500  C, the MLRs were 6.63%, 6.99% and 7.13%, showed that with roasting time increasing, the contact angle of
respectively, they were very close. And Table 2 showed the contents LiCoO2 concentrate become smaller. After roasted for 30 min, the
of C, O and Co were very close at these temperatures. It was proved contact angles were almost the same. It proved that the organic

Table 2
Surface elemental contents of roasted LiCoO2 concentrate.

Temp./ C Element contents/at.%

C O Co

0 42.76 42.68 42.79 27.34 27.19 27.16 5.53 5.62 5.35

42.74 ± 0.03 27.23 ± 0.06 5.50 ± 0.08
350 47.67 47.49 47.53 20.21 20.16 20.06 4.77 4.73 4.61
47.56 ± 0.05 20.14 ± 0.04 4.7 ± 0.05
400 16.77 16.93 16.74 34.17 34.26 34.14 8.56 8.62 8.66
16.81 ± 0.06 34.19 ± 0.04 8.61 ± 0.03
450 16.13 16.32 16.22 41.68 41.77 41.89 8.78 8.67 8.71
16.22 ± 0.05 41.78 ± 0.06 8.72 ± 0.03
500 15.93 15.85 16.17 44.65 44.76 44.82 8.66 8.75 8.91
15.98 ± 0.09 44.74 ± 0.05 8.77 ± 0.07
650 F. Wang et al. / Journal of Cleaner Production 185 (2018) 646e652

Table 3
Contact angles of samples being roasted for different time.

Time/min Material Experimental results Results D-value/

1 2 3 4 5 6

15 LiCoO2/ 20.1 20.2 19.9 19.6 19.8 20 19.9 ± 0.2 33.6
Graphite/ 53.2 53.6 53.6 53.5 53.5 53.7 53.5 ± 0.2
30 LiCoO2/ 18.6 18.9 18.6 18.3 18.5 18.7 18.6 ± 0.2 34.0
Graphite/ 52.9 52.5 52.4 52.8 52.5 52.5 52.6 ± 0.2
45 LiCoO2/ 18.5 18.5 18.6 18.3 18.1 18.4 18.4 ± 0.2 32.6
Graphite/ 50.9 50.8 50.9 51.3 51.1 51.1 51.0 ± 0.2
60 LiCoO2/ 18.4 18.4 18.6 18.5 18.2 18.3 18.4 ± 0.1 30.7
Graphite/ 49.0 49.5 49.3 49.3 48.8 48.7 49.1 ± 0.3

SEM images of graphite concentrate roasted at 450  C for 15, 30,

Fig. 5. SEM images of roasted graphite at different time (5a-15min, 5b-30min,
5c-45min, 5d-60min).
outer layer could be removed by roasting at 450  C for 30 min. The
contact angles of graphite concentrate also become smaller, but
they were not changed much, which implied that the temperature
was mild enough. And from 15 to 30 min, the contact angle changed
slightly, but from 30 to 45 min and 45e60 min, it was changed a lot.
This implied that the thermostability of the organic outer layer was
stronger than graphite. The D-value showed that 30 min was the
best roasting time at 450  C。However, it should be noticed that
the D-value of 15 min was very close to 30 min.
45 and 60 min were as shown in Fig. 5. When the roasting time was
longer than 30 min, the surfaces of the graphite were much
rougher. The graphite was oxidized quickly when the organic outer
layer was removed, and then the surface of graphite concentrate
turned to hydrophilic. Therefore, during the roasting process,
excessive oxidation of the graphite should be avoided.
XPS results of the roasted LiCoO2 concentrate listed the contents
of C, O and Co on the surface of the particles (Table 4). After being
roasted at 450  C for 15 min, the content of C was reduced from
42.74 at.% to 16.48 at.%, and the content of Co increased from
5.5 at.% to 8.09 at.%. And then, both of the contents of these two
elements almost remained the same, which proved that the organic
outer layer was removed, too.
The MLRs of electrode materials under different roasting time
were as shown in Fig. 6. With roasting time increasing, the MLRs of
the LiCoO2 concentrate was almost not changed, they were 6.78%,
6.87%, 6.92% and 6.99% respectively. That indicated after being
roasted at 450  C for 15 min, the organic outer layer on the surface
of LiCoO2 concentrate was completely removed. However, the MLRs
of graphite concentrate were very high, when the roasting time
reached 15 min, the MLR was 23.2%, and when the roasting time
reached 60 min, the MLR was as high as 54.75%. Therefore, in order
to avoid losing too much graphite, the roasting time should be
shorter. In this study, under the condition that the roasting time
was 15 min and the roasting temperature was 450  C, the surface
property of the electrode materials could be modified satisfactory
for flotation.
3.4. Floatation test
The
0.25 mm crushed products (mixed electrode materials)
from spent LiBs were roasted at 450  C for 15 min, and then, the
roasted products were separated by flotation. The yield of LiCoO2
concentrate was 72.29 wt% and the yield of graphite concentrate

Table 4
Surface elemental contents of roasted LiCoO2.

Time/min Element contents/at.%

C O Co

0 42.76 42.68 42.79 27.34 27.19 27.16 5.53 5.62 5.35

42.74 ± 0.03 27.23 ± 0.06 5.50 ± 0.08
15 16.56 16.43 16.45 37.21 37.14 37.04 8.13 8.11 8.04
16.48 ± 0.04 37.13 ± 0.05 8.09 ± 0.03
30 16.68 16.54 16.58 38.19 38.25 38.26 7.88 7.95 7.96
16.6 ± 0.04 38.23 ± 0.02 7.93 ± 0.03
45 16.21 16.17 16.11 39.57 39.49 39.51 7.87 7.79 7.77
16.16 ± 0.03 39.52 ± 0.02 7.81 ± 0.03
60 16.13 16.32 16.22 41.68 41.77 41.89 8.78 8.67 8.71
16.22 ± 0.05 41.78 ± 0.06 8.72 ± 0.03
 F. Wang et al. / Journal of Cleaner Production 185 (2018) 646e652 651

Fig. 7. SEM and EDS images of flotation products (7a-LiCoO2 products, 7b-graphite
products).

Fig. 6. Mass loss rates of electrode materials under different roasting time.
was 27.71 wt%. The metal values contents in the flotation products
were listed in Table 5. The result showed that metal values in
the
0.25 mm crushed products from spent LiBs were enriched in
LiCoO2 concentrate.
The enrichment ratio (ER) defined by formula (3) was used to
show concentrating level and the metal recovery rate (MRR)
defined by formula (4), was used to represent recovery efficiency.
industrial LiBs are graphite and LiCoO2 (or some other kinds of
materials such as LiMn2O4, LiMnxNiyCo1-x-yO2), they are of opposite
surface wettability and very small particle size. Therefore, flotation
is a very suitable method for the separation of the anode and
cathode active materials. However, surface modification is very
necessary to regain the original surface properties because the
organic electrolyte and PVDF binder used in LiBs makes the surfaces
of the anode and cathode active materials almost the same. Ther-
mal treatment is of high efficiency and low cost to remove the
organic outer layer to regain the original surfaces, for the reaction
time and temperature are only 15 min and 450  C. In the flotation
step, both of the graphite and metals such as Li, Co, Ni and Mn are
separated and recovered. Although the flotation products are wet,
the following purification step is hydrometallurgical process.
Therefore, drying step for metal concentration from flotation is
omitted. From this perspective mechanical separation, thermal
treatment and flotation are suitable for the total recycling process
of spent LiBs and all the materials used in LiBs can be recovered and
recycled to avoid environmental pollution.

ER ¼ m=M (3) 4. Conclusions

Mechanical separation and thermal treatment were applied to

MRR ¼ ER  Y
Where m is the metal content in the LiCoO2 concentrate, M is the
metal content in the
0.25 mm crushed products and Y is the yield
of the LiCoO2 concentrate. The ERs of Co, Mn, Cu and Al were 1.35,
1.29, 1.25 and 1.19, their recovery rates were 97.66%, 93.66%, 90.14%
and 86.29%, respectively.
The SEM and EDS images of flotation products were as shown in
Fig. 7. Most of the anode and cathode active materials were sepa-
rated by flotation, and flotation was an efficiency way to recovery
both graphite and LiCoO2. However, there were still some LiCoO2
left in the graphite concentrate and some graphite still left in the
LiCoO2 concentrate. Further research was needed to optimize the
flotation process for recovering much purer materials.
To recycle the spent LiBs, there must be many steps including
crushing, separation, purification and recovered materials’ reuse.
Until now, the anode and cathode active materials used in the
(4)
recover valuable materials from spent lithium-ion batteries. It was
found that the thermostability of the organic outer layer was
stronger than graphite, after the layer was removed, the graphite
would be destroyed very quickly. The organic outer layer coated on
the surface of the electrode materials could be removed by roasting
at 450  C for 15 min very efficiently and at the same time, graphite
was protected from being burnt down as much as possible. When
the organic outer layer was removed by roasting, the difference of
wettability between graphite and LiCoO2 recovered from spent LiBs
was enhanced and become suitable for flotation. By the roasting
and flotation method, both the metal values and graphite in
the
0.25 mm crushed products of spent LiBs were separated and
recovered. The content of Co in the LiCoO2 concentrate was as high
as 40.12 wt% and the recovery rate was 97.66%. However, further
research was still needed to optimize the flotation process for
recovering much purer materials. By using this process, all the

Table 5
The contents of flotation products/wt.%.

Metals Co Mn Cu Al

Roasted products 29.76 29.68 29.66 3.12 3.17 3.26 0.86 0.8 0.77 2.17 2.21 2.28
29.7 ± 0.03 3.18 ± 0.04 0.81 ± 0.03 2.22 ± 0.03
LiCoO2 concentrate 40.09 40.11 40.16 4.07 4.11 4.19 1.05 0.97 1.02 2.66 2.69 2.61
40.12 ± 0.02 4.12 ± 0.04 1.01 ± 0.02 2.65 ± 0.02
Graphite concentrate 2.46 2.52 2.56 0.76 0.73 0.71 0.32 0.39 0.34 0.71 0.69 0.79
2.51 ± 0.03 0.74 ± 0.01 0.35 ± 0.02 0.73 ± 0.03
652 F. Wang et al. / Journal of Cleaner Production 185 (2018) 646e652
materials used in LiBs can be recovered and recycled to avoid
environmental pollution.
Acknowledgements
The financial support for this work was from National Natural
Science Foundation of China (51574234), Science Foundation of
Jiangsu province (BK20160267), Fundamental Research Funds for
the Central Universities (2015XKMS099), and program sponsored
for Scientific Innovation Research of College Graduate in Jiangsu
Province of China (KYLX15_1437). The authors would like to thank
Advanced Analysis and Computation Centre of China University of
Mining and Technology for their technical support.
References
Al-Thyabat, S., Nakamura, T., Shibata, E., Iizuka, A., 2013. Adaptation of minerals
processing operations for lithium-ion (LiBs) and nickel metal hydride (NiMH)
batteries recycling: critical review. Miner. Eng. 45, 4e17.
Bahaloo-Horeh, N., Mousavi, S.M., 2017. Enhanced recovery of valuable metals from
spent lithium-ion batteries through optimization of organic acids produced by
Aspergillus Niger. Waste Manag. 60, 666e679.
Chen, X., Ma, H., Luo, C., Zhou, T., 2017. Recovery of valuable metals from waste
cathode materials of spent lithium-ion batteries using mild phosphoric acid.
J. Hazard Mater. 326, 77e86.
Choubey, P.K., Chung, K., Kim, M., Lee, J., Srivastava, R.R., 2017. Advance review on
the exploitation of the prominent energy-storage element Lithium. Part II: from
sea water and spent lithium ion batteries (LIBs). Miner. Eng. 110, 104e121.
Hanisch, C., Loellhoeffel, T., Diekmann, J., Markley, K.J., Haselrieder, W., Kwade, A.,
2015. Recycling of lithium-ion batteries: a novel method to separate coating
and foil of electrodes. J. Clean. Prod. 108, 301e311.
He, Y., Zhang, T., Wang, F., Zhang, G., Zhang, W., Wang, J., 2017. Recovery of LiCoO2
and graphite from spent lithium-ion batteries by Fenton reagent-assisted
flotation. J. Clean. Prod. 143, 319e325.
Horeh, N.B., Mousavi, S.M., Shojaosadati, S.A., 2016. Bioleaching of valuable metals
from spent lithium-ion mobile phone batteries using Aspergillus Niger. J. Power
Sources 320, 257e266.
Huang, Y., Han, G., Liu, J., Chai, W., Wang, W., Yang, S., Su, S., 2016. A stepwise re-
covery of metals from hybrid cathodes of spent Li-ion batteries with leaching-
flotation-precipitation process. J. Power Sources 325, 555e564.
Ordon ~ ez, J., Gago, E.J., Girard, A., 2016. Processes and technologies for the recycling
and recovery of spent lithium-ion batteries. Renew. Sustain. Energy Rev. 60,
195e205.
Pinna, E.G., Ruiz, M.C., Ojeda, M.W., Rodriguez, M.H., 2017. Cathodes of spent Li-ion
batteries: dissolution with phosphoric acid and recovery of lithium and cobalt
from leach liquors. Hydrometallurgy 167, 66e71.
Zeng, X., Li, J., Liu, L., 2015. Solving spent lithium-ion battery problems in China:
opportunities and challenges. Renew. Sustain. Energy Rev. 52, 1759e1767.
Zhang, T., He, Y., Ge, L., Fu, R., Zhang, X., Huang, Y., 2013. Characteristics of wet and
dry crushing methods in the recycling process of spent lithium-ion batteries.
J. Power Sources 240, 766e771.
Zhang, T., He, Y., Wang, F., Ge, L., Zhu, X., Li, H., 2014a. Chemical and process
mineralogical characterizations of spent lithium-ion batteries: an approach by
multi-analytical techniques. Waste Manag. 34, 1051e1058.
Zhang, T., He, Y., Wang, F., Li, H., Duan, C., Wu, C., 2014b. Surface analysis of cobalt-
enriched crushed products of spent lithium-ion batteries by X-ray photoelec-
tron spectroscopy. Separ. Purif. Technol. 138, 21e27.