DOI:10.30919/esee8c126

Received: 04 May 2018
Revised: 13 May 2018
Accepted: 15 May 2018
Published online: 16 May 2018

View PDF Version

View volume articles

Enhanced Hydrometallurgical Recovery of Valuable Metals from Spent Lithium-ion Batteries by Mechanical Activation Process

Jie Guan, Haiyang Xiao, Xiaoyi Lou, Yaoguang Guo, Xingmin Luo, Yingshun Li, Chao Yan, Xingru Yan, Guilan 1 1,3 1 1 7 6,* Gao, Hao Yuan, Jue Dai, Ruijng Su, Weixing Gu and Zhanhu Guo

1 Research Center of Resource Recycling Science and Engineering, School of Environmental and Materials Engineering, Shanghai Polytechnic University, Shanghai 201209, China
2 Key Laboratory of Control of Quality and Safety for Aquatic Products, Ministry of Agriculture, East China Sea Fisheries Research Institute, Chinese Academy of Fishery Sciences, Shanghai 200090, China
3 Shanghai Pudong Shuguang Research Center for Environmental Treatment Technologies, Shanghai 20209, China
4 Shanghai Xin Jinqiao Environmental Protection Co., Ltd., Shanghai 201201, China
5 School of Material Science and Engineering, Jiangsu University of Science and Technology, Zhenjiang, Jiangsu, China
6 Integrated Composites Laboratory (ICL), Department of Chemical & Bimolecular Engineering, University of Tennessee, Knoxville, TN37996, USA
7 Shanghai Julang Environmental Protection Co., Ltd., Shanghai 201712, China

*E-mail: ygguo@sspu.edu.cn (Y. G.); zguo10@utk.edu (Z. G.)

1.Introduction

Lithium-ion batteries (LIBs) have drawn great attention and been extensively applied in modern electronic devices due to their superior performance such as high energy density, long cycle life, high voltage and low self-discharge.1-4 China as a populous and fastdeveloping country has become a vital manufacturer, supplier and consumer of LIBs. As such, an increasing number of LIBs will be discarded annually. According to the statistics, the discarding quantity and weight of LIBs can exceed 25 billion units and 500 thousand metric tons by 2020.5 Thus, the spent LIBs would pose a

very specific threat, given that they contain a high percentage of dangerous electrolyte and metals.6-9 For instance, lithium cobalt oxide (LiCoO ) accounts for 27.5 w% in the cathode active materials of 2 LIBs. And 5-20 w% of cobalt (Co) and 5-7 w% of lithium (Li) contribute to the composition of the LIBs, which are higher than those in natural ores.10,11Therefore, it is highly desirable to recycle valuable metals from the spent LIBs to alleviate the potential environmental pressure and shortage of rare metals.

Among the reported recycling methods of Co and Li from the spent LIBs, such as pyrometallurgy, hydrometallurgy or bio- hydrometallurgy processes,5,12-15 hydrometallurgy process gained world-wide attention due to higher metal recovery with good purity, low energy consumption and minimal gas emission. Generally, the cathode active materials of the spent LIBs were firstly pretreated by discharge, mechanical separation, or thermal processes.10 The metals were then dissolved into solution via acid leaching process, and subsequently the target metals were recovered from the acid leachate and the final wastes via different processes, such as chemical precipitation, solvent extraction, electrochemical accumulation, etc. Acid leaching becomes a critical step for the recovery of valuable metals from the spent LIBs. The acid extraction of Co and Li from the spent LIBs is commonly conducted by using inorganic acids, such as sulfuric, hydrochloric and nitric acids,16-20 among which higher recoveries of Co and Li were obtained through the leaching processes with the hydrochloric and nitric acids.21 However, chlorine (Cl ) and NO are normally generated when hydrochloric and nitric 2 X acid were used as leaching agent for LiCoO , leading to much higher 2 recycling cost to acquire special antisepticising equipment for the treatment of secondary gaseous pollutants, and also more serious environmental problems could be caused.22,23 In comparison, the leaching media of sulfuric acid or a mixture of sulfuric acid and

hydrogen peroxide (H O ) can avoid the generation of toxic gases, 2 2 and also yield higher leaching rate of valuable metals.24-29 Usually, H O , as a reductant during the acid leaching, can easily reduce 2 2 Co(III) to Co(II) in the spent LIBs, which is more beneficially solubilized than the unreduced moieties.11,30 Nevertheless, for the leaching system of sulfuric acid, higher concentration of acid or its mixture with H O are necessary, as well as temperature higher than 2 2 60°C is needed,24-29 resulting in higher average costs and operational environmental risk, though. Consequently, an economic and environmental-friendly recycling process of Co and Li from spent LIBs is highly desirable.

The mechanochemical process, based on the triggered physicochemical changes including phase transformations, structural defects, amorphization, and even direct reaction under normal temperature and pressure, has captured more attention for potential applications on metal recovery.31 The recovery of valuable metals in recent decades has focused on the mechanochemical process treatment. For example, Yuan et al.32 obtained 92.5% of lead by 3 M nitric acid leaching heated at 95 °C from scrap cathode ray tube funnel glass activated by mechanochemical method. Lee et al.33,34 developed a process for metal recovery from LiCoO powders via co- 2 grinding with polyvinyl chloride and graphite effectively. Wang et al.35 reported that high recovery rates of Co and Li could be reached by mechanical activating LiCoO with ethylenediaminetetraacetic 2 acid as co-grinding reagent. These studies implied that mechanochemical processes could obviously simplify valuable metal leaching. However, either higher temperature or higher concentration of leachates is needed in the traditional hydrometallurgical recovery of metals via dissolving LiCoO with H SO and H O .22,23 The 2 2 4 2 2 mechanochemical pretreatment for spent LIBs without any cogrinding materials to promote the extraction of valuable metals of Co and Li with lower acid concentration under room temperature has not been reported.

In this study, the mechanochemical process without any cogrinding materials was used to improve the acid leaching of Co and Li from spent LIBs with lower acid and H O concentrations at room 2 2 temperature from spent LIBs. Lithium cobalt oxide (LiCoO ) 2 powders were selected to investigate the influences of mechanochemical parameters on the Co and Li extraction, including ball milling time, rotary speed, and ball-to-powder mass ratio. The physicochemical changes established by Microtrac particle size analyzer, scanning electron microscope (SEM), and X-ray diffraction (XRD) were used to give insights into the mechanism of the mechanical activation process. Furthermore, the effects of leaching factors on the recovery of Co and Li, such as the concentrations of leachates and the solid-to-liquid (S/L) ratio were also studied, based on which the influence of mechanical activation process on the leaching kinetics was examined. Ultimately, the actual scrap sample separated and concentrated from the cathode materials in the spent LIBs was also investigated.

2.Experiments

2.1 Experimental Materials

Lithium cobalt oxide (LiCoO , 99.8% metals basis) powders were 2 purchased from Aladdin Industrial Co. Ltd, Shanghai, China. Sulfuric acid (H SO , AR, 95~98%) and hydrogen peroxide (H O , 2 4 2 2 AR, 30%), Sodium chloride (AR, NaCl), hydrochloric acid (HCl, AR, 36.0~38.0%,) and nitric acid (HNO , AR, 65.0~68.0%,) were 3 both obtained from Sinopharm Chemical Reagent Co., Ltd, Shanghai, China Ultrapure water (18.2 MΩ·cm) was used for all the experiments.

2.2 Experimental procedures

The examined spent LIBs (with the model: BL-5J and the capacity: 1320 mAh) used in the NOKIA mobile phones were chosen in this study. The spent LIBs were first discharged completely by using sodium chloride aqueous solution to remove any remaining electrical potential capacity and then were dismantled to remove the plastic package and steel cases. During the dismantling, the cathode materials were collected after the removal of plastic films. To effectively dislodge the polyvinylidene fluoride (PVDF) binder between the cathode materials and aluminum foil recovered in its metallic form, the cathode materials were thermally treated at 600°C for 10 h. And then, the LiCoO powders were cooled down to room 2 temperature and then dissolved in aqua regia for nearly 60 min to obtain the metal contents which were measured using an inductively coupled plasma-atomic emission spectrometry (ICP-AES, iCAP- 6300, TH ERMO, U.S.). The results indicated that Co and Li accounted for 72.6 w% and 6.9 w% of the total cathode materials.

The mechanical activation experiments were carried out in a planetary ball mill (CDI-EM60, Delixi Hangzhou Inverter Co., Ltd, China), comprising of four stainless steel pots with a volume of 100 mL containing balls with a diameter of 10.0 mm as grinding media. Prior to each experiment, certain aliquots were transferred to the reactor pots to obtain the desired amounts. After the milling, all the samples were collected and preserved in a drying vessel for the following characterization, as well as acid leaching.

All the batch leaching experiments were carried out in 40 mL certain aliquots of H SO solution and H O , filled into a conical flask 2 4 2 2 with a capacity of 50 mL immersed in a water bath to control the reaction temperature at 25 ± 1°C under constant magnetic agitation. After leaching, the solution was filtered by polyether sulfone (PES) filter with a diameter of 0.22 μm for the determination of Co and Li by ICP-AES. The recovery rates of Co and Li were expressed in percentage according to Equation 1. Most of the batch experiments were conducted in duplicate, and the error bars were less than 3%.

where χ is the metal leaching efficiency; C is the mass 0 concentrations of metal ions in the leachate; V is the volume of 0 leachate; m is the mass of samples; and w% is the metal mass fraction.

2.3 Analytical methods

The concentrations of metals in leaching liquor were determined by an ICP-AES instrument. The particle sizes of the original and ball milled samples were respectively measured by a Microtrac particle size analyzer (MT3300, Tokyo, Japan). To figure out the variation of microstructures during milling, the samples were characterized by scanning electron microscope (SEM, JSM 6400, JEOL, Japan). Xray diffraction (XRD) was performed using an X-ray diffractometer (D8 ADVANCE, BRUKER-AXS Corporation, Germany) equipped with a Cu K radiation from 2θ =10-80° with a step of 0.02°. The full a width at half maximum (FWHM) of peaks, crystallite size and degree of disorder were calculated by Jade 6.0, and the degree of disorder developed during milling process was obtained from Eq. (2).36

Fig. 1 Extraction of (a) Co and (b) Li from the non-activated and activated samples (ball milling time = 120 min, rotation speed = 650 rpm, ball-to-powder mass ratio = 45 : 1, [H SO ] = 1 M, [H O ] = 2 vol.%, S/L ratio = 20 g/L, leaching temperature = 25 °C).

where η is the degree of disorder, I is intensity of each crystal raw plane diffraction peak for raw samples, and I activated is the intensity of each crystal plane diffraction peak for activated samples.

3. Results and discussion

3.1. Evaluation of mechanical activation process

To evaluate the mechanical activation process for the improved leaching efficiency of Co and Li, the powders pretreatment by mechanochemical activation were examined (Fig. 1). The results showed that the recovery rates of leaching for 15 min increase significantly from 11.7% to 93.4% for Co and 12.3% to 90.3% for Li, respectively, indicating that the mechanical activation process has an extraordinary effect on the hydrometallurgical extraction of the valuable metals. Above all, the mechanochemical activation process renders the hydrometallurgical leaching reaction conditions to be more moderate than the previous traditional researches, such as with lower concentrations of leaching agents, shorter leaching time, and at room temperature, etc.24-29

3.2. Effects of mechanochemical parameters

3.2.1. Effect of ball milling time

Fig. 2a shows the effects of ball milling time on the recovery rates of Co and Li. The recovery rates increased significantly from 30.8% to 97.8% for Co and 40.7% to 92.5% for Li, respectively, with prolonging the ball milling time from 0 min to 60 min, and then hardly changed when the ball milling time was extended from 60 to 180 min, however, which might attribute to the particle size distribution change of the samples after milling. The particle size decreased significantly as the ball milling time was prolonged from 0 to 60 min. Whilst, as the milling time gradually extended to 180 min, the fraction with smaller particle size of samples was reduced, and the fraction with large particle size showed an upward trend. This indicated an agglomeration of small particles (Fig. 2b), that commonly appeared during dry grinding. It could be explained by the agglomeration of structurally modified particles following the initial reduction of particle size.37 Furthermore, the SEM observations of the samples after mechanical activation for different milling time also illustrate the variation of particle size (Fig. S1). The morphology of LiCoO powder changed drastically after mechanical 2 activation, and much rougher surfaces of the samples and a significant reduction of the particle size could be observed with increasing the grinding time from 0 min to 60 min (Fig. S1a-d). However, a slight agglomeration of small primary particles can be seen after 120 min and is consistent with the aforementioned particle size analysis (Fig. S1e&f).

In addition, the results revealed that the peak intensity of the activated samples decreased, and the full width at half maximum (FWHM) of all peaks was also broadened dramatically with the increase of the ball-milling time (Fig. 3), indicating that the crystal structure of the particles was destroyed and gradually transformed to an amorphous state due to the friction and impact during grinding process.35,38 Furthermore, both the changes of the crystallite size and the degree of disorder also reflected that the LiCoO crystal was 2 subjected to crushing and became amorphous. Fig. 3a&b present that the crystallite size, whether 003 or 104 crystal planes, decreased with prolonging the activated time. After ball milling treatment for 60 min, the crystallite size of 003 and 104 crystal planes decreased from more than 1000 nm to 525 nm and 115 nm, respectively. Moreover, Fig. 3c illustrates that the degree of disorder became higher with increasing the milling time from 0 to 60 min, and hardly changed when the grinding time was over 120 min, which is consistent with the aforementioned results obtained from particle size distribution and SEM analyses. In short, the LiCoO crystal became amorphous 2 and had more crystal defects after the mechanical activation process, which might be beneficial for the leaching of Co and Li.

Above all, the particle size would decrease to a limit and the crystal structure would also suffer from crushing and abrasion continuously with prolonging the grinding time,39 which was the reason that the recovery rate increased distinctly as the activated time prolonged from 0 min to 60 min. Despite of the slight agglomeration after the ball milling treatment of 180 min, the

Fig. 2 The influence of ball milling time on (a) Co and Li leaching, and (b) particle size distribution of samples after mechanical activation (rotation speed = 650 rpm, ball-to-powder mass ratio = 45: 1, [H SO ] = 1 M, [H O ] = 2 vol.%, S/L ratio = 20 g/L, leaching time = 45 min, 2 4 0 2 2 0 leaching temperature = 250C).

Fig. 3 XRD patterns of samples activated at different ball milling time(a), the influence of milling time on the crystallite size(b), and variation of FWHM of different crystal planes and the crystal degree of disorder(c) (rotation speed = 650 rpm, ball-to-powder mass ratio = 45 : 1).

extraction of Co and Li was still improved, illustrating that both changes of particle size and crystal structure play dominant roles in the improved extraction of Co and Li under moderate conditions by the mechanical activation process. Considering the energy consumption and recovery rates, the milling time of 60 min was selected as an optimum condition for the following experiments, at which a higher recovery of 97.8% for Co and 92.5% for Li obtained.

3.2.2. Effect of rotation speed

As is known, the magnitude of the force depends on the milling speed during the mechanical activation process, that is a higher speed provides larger force to almost completely destroy the material particles.40 Fig. S2a shows that the recovery rates increased sharply from 30.8% to 97.8% for Co and from 40.7% to 92.5% for Li, respectively, as the rotary speed increased from 0 rpm to 650 rpm. Then, the recovery rates of Co and Li increased insignificantly with the ball milling rotary speed increased to 700 rpm. Fig. S2b represents the particle size distribution of the activated samples at different rotation speeds, which might be in accordance with the above experimental results. A significant decrease of particle size with increasing the rotation speed could be attributed to the fact that the kinetic energy generated by the series of collisions among balls was transferred to the samples.41 Nevertheless, with the continuous increase of the ball milling speed, the LiCoO particles appeared 2 aggregated to some extent, attributing to the localized plastic deformation at the contact area of the adjacent particle with the sharp increase of Van der Waals force, leading to the particle size decreased to a critical value.41,42 As such, higher rotation speed can result in smaller particles, promoting the improved leaching of Co and Li, and much higher speed can lead to aggregation of secondary smaller particles, weakening the leaching of valuable metals. Furthermore, the changes in morphology and crystalline structure of the samples were evaluated by SEM characterization (Fig. S3) and XRD (Fig. S4), respectively (The detailed analysis was shown in SI.); and the obtained results were in accordance with the above analysis of the particle size distribution (Fig. S2b), supporting the results that the recovery rates of Co and Li were higher at the rotation speed of 650 rpm (Fig. S2a).

3.2.3. Effect of ball-to-powder mass ratio

The ball-to-powder mass ratio plays an important role during the ball milling treatment process, for which an inappropriate ball-to-powder mass ratio would lead to an unnecessary energy loss and weaken the ball milling capacity.42 Fig. S5a shows that the recovery rates of Co and Li all reached maximum values of 100% and 93.8% for Co and Li, respectively, at the ball-to-powder mass ratio of 85 : 1, and the extraction of Co and Li increased insignificantly when the ball-topowder mass ratio varied from 85:1 to 125 : 1. Fig. S5b presents that the particle size decreased dramatically and the size distribution became wider with increasing the ball-to-powder mass ratio, owing to that the increase of ball-to-powder mass ratio contributed to the increase of frequency and intensity of sheer and collision between the grinding medium and samples.41 However, with the ball-topowder mass ratios enlarged from 85 : 1 to 125 : 1, the aggregation of activated sample particles appeared and the fine particles (0.1 μm to 1 μm) were decreased (Fig. S5b), supporting the above results (Fig. S5a). Besides, the detailed analysis of SEM (Fig. S6) and XRD (Fig. S7) presented in SI also support the aforementioned results (Fig. S5).

In conclusion, the enhancement of hydrometallurgical

Fig. 4 The influences of (a) H SO concentration, (b) H O 2 4 2 2 concentration, and (c) S/L ratio on the Co and Li leaching (ball milling time = 60 min, rotation speed = 650 rpm, ball-to-powder mass ratio = 85 : 1, [H SO ] = 1 M (except for a), [H O ] = 2 vol.% 2 4 0 2 2 0 (except for b), S/L ratio = 20 g/L (expect for c), leaching time = 45 min, leaching temperature = 25oC).

recovery of valuable metals from spent lithium-ion batteries could contribute to the significant reduction of particle size, morphology defects, and destruction of crystal structure of the activated samples induced by the mechanical activation process.

3.3. Effects of leaching factors

3.3.1. Effect of H SO concentration

From Fig. 4a, it can be noted that the increasing concentration of H SO could enhance the recovery rates of Co and Li from the non- 2 4 activated samples, which could be explained by the chemical reaction of dissolving LiCoO into the H SO and H O solution (Eq. 2 2 4 2 2 3). The addition of the reacting substances can facilitate the forward reaction, resulting in the increase of the leaching efficiency.8 It is noteworthy that the leaching rates of Co and Li from the activated samples increased dramatically with increasing the H SO 2 4 concentration. the extraction rates increased from 42.8% to 100% for Co and from 70.8% to 93.8% for Li, respectively, when the H SO 2 4 concentration increased from 0.2 to 1 M (Fig. 4a). The main reason for the significantly enhanced exaction of valuable metals from the activated sample might attribute to the decrease of the samples particle sizes, the variation of the morphology and the destruction of the crystalline structure, which were induced by the mechanical activation process.

3.3.2. Effect of H O concentration

The chemical bond between cobalt and oxygen is strictly strong, leading to difficult acid leaching of LiCoO powders. However, when 2 the added H O ranged from 0 to 6 vol.%, the leaching of Co from 2 2 the non-activated samples increased from 34.2% to 69.3%, owing to that the evolved oxygen from the decomposed H O could convert 2 2 Co(Ⅲ) to Co(Ⅱ) which could be easily dissolved into the acid solution (Eq. 3).10,43 In comparison, with the addition of same concentration H O , the leaching rates of Co and Li from the 2 2 activated samples were both distinctly higher than those of the nonactivated samples. As is shown in Fig. 4b, with the H O 2 2 concentration of 2 vol.%, the leaching rates of Co and Li from the activated samples sharply increase to 100% and 93.8%, respectively, which might be caused by the decline of the sample particle sizes and destruction of the crystal structure induced by the mechanical activation process. In addition, the extraction rate of Co increased nearly 1.3-fold from the non-activated samples and 2.7-fold from the activated samples when the added H O increased from 0 vol.% to 2 2 2 vol.%, whereas the leaching rate of Li was only about 1.0-fold from both the raw and the activated samples. The results are consistent with the facts that the leaching efficiency of Co was dependent on the H O concentration, and the leaching efficiency of Li was 2 2 independent of the H O concentration (Equation 3).44

3.3.3. Effect of S/L ratio

Fig. 4c presents the effect of the S/L ratio on the leaching rate. For the samples without mechanical activation, the leaching rates of Co and Li decreased gradually with the increase of S/L ratio, attributing to the decrease of the available surface area per unit volume of the solution. However, when the S/L ratio was between 5 and 20 g/L, no significant variations were observed for the extraction of Co and Li; and when the S/L ratio was 20 g/L, the leaching rates of Co and Li were around 100% and 93.8%, respectively. Nevertheless, with further increasing the S/L ratio, the leaching rates of Co and Li decreased sharply. In industrial applications, a higher S/L ratio was considered to be favorable for the recovery of valuable metals; however, even higher S/L ratio would result in a lower leaching efficiency. It is worth noting that the much higher extraction of Co and Li from the activated samples could be obtained with the same S/L ratio, probably due to the decrease of the particle size, variation of the morphology and disruption of the crystal structure. Considering the lower chemical consumption and relatively higher recovery rates, the optimal S/L ratio of 20 g/L was chosen for the leaching of Co and Li from the samples activated by mechanochemical treatment.

3.4. Leaching kinetics

The above studies suggested that higher recovery rates of Co and Li from the activated samples could be obtained under the moderate conditions of 1 M H SO , 2 vol.% H O and room temperature, 2 4 2 2 which proved that the mechanical activation process was an effective enhanced technology. Given that the H O concentration had slight 2 2 influence on the leaching of Li, the leaching kinetics were investigated based on the leaching results of Co and Li with various concentrations of H SO rather than H O under different leaching 2 4 2 2 time to reveal another new insight of the mechanical activation process (Fig. 5a-d).

The selection of a kinetic model for the leaching results of Co and Li from the non-activated samples was obtained using Sharp’s method of reduced half time of reaction (Fig. S8) .45

where χ is the recovery rates of Co and Li; t is the metal leaching time; t is the time to reach χ = 50%, while constant A depends on a 0.5 function F(χ).

Obviously, Fig. S8 shows that the curve of non-activated samples approximates to the kinetic function(k is the reaction rate constant), which means that the equation could be used for linearization of the leaching results of the non-activated samples (Fig. S9). Just reported, this kinetic model is a mixed control one, i.e., chemical surface reaction and diffusion.46 The kinetic function is in fact the equation of Kazeev-Eofeev:  for the conditions of n = 1.47 Parameter n, determining the reaction rate controlling step, is close to 1, suggesting that the chemical surface reaction is a control step, whereas with medium and small value of n (i.e. n ≤ 0.5), the reaction rate is diffusion controlled.46 The parameter n was calculated via Eq.  based on the results in Fig. 5a&b. Fig. S10a&b shows that the plots of ln[-ln(1-χ)] versus lnt for the leaching of Co and Li from the non-activated samples under different concentrations of H SO , and the quality of the fit is 2 4 quite consistent with the average square of correlation coefficients (R2) of 0.9951 for Co and 0.9895 for Li. Then the parameter n was determined from the slope of these straight lines and provided in Table 1. It shows that the values of parameter n were all close to 1 with H SO concentration ranging from 0.5 to 3 M. This result 2 4 suggested that the leaching of Co and Li from the non-activated samples was dominantly chemically controlled within the leaching time of 60 min. Nevertheless, the value of parameter n dramatically decreased to 0.787 with a H SO concentration of 3 M. The 2 4 decreasing trend indicates that the reaction rates started to decrease, and controlled by diffusion through the product layers besides chemical surface reaction or even solely controlled by diffusion, with a continuous increase of the H SO concentration.46 Therefore, we 2 4 speculated that with increasing the H SO concentration, the 2 4 chemical surface reaction rate at the initial stages was faster and expedited the leaching reaction.

In order to figure out the variation of the leaching kinetics after the mechanical activation process, some parameters of a kinetics model were also calculated to identify the dominant reaction step of the leaching for activated samples. Fig. S11 shows that these nine kinetics models using Sharp’s method including the above selected model -ln(1-χ)=kt did not fit the leaching data of the activated samples well. Hence, it needs to choose another kinetic model to analyze the experimental results of the milled samples. As reported, Kondo et al.48 have classified a reaction process based on the parameter N using the Jander equation (Eq. 5).

Fig. 5 The recovery rates of Co and Li from the (a and b) non-activated and (c and d) activated samples under different concentrations of H SO 2 4 (ball milling time = 60 min, rotation speed = 650 rpm, ball-to-powder mass ratio = 85: 1, [H O ] = 2 vol.%, S/L ratio = 20 g/L, leaching 2 2 0 temperature = 25°C).

Table 1. Values of the kinetic parameters n for the non-activated samples and N for the activated samples under different concentrations of H2SO4 
Concentration of
H2SO4 (M)		 Non-activated samples Activated samples
n(Co) R2(Co) n(Li) R2(Li) N(Co) R2(Co) N(Li) R2(Li)
0.2 0.928 0.9993 0.859 0.9975 20.76 0.9705 35.61 0.9179
0.5 1.023 0.9961 1.095 0.9949 9.12 0.9225 30.28 0.9012
1 1.091 0.9981 1.158 0.9972 3.19 0.9305 14.91 0.9141
2 0.901 0.9991 0.956 0.9969 3.37 0.9116 36.86 0.9596
3 0.787 0.9830 0.827 0.9612 3.02 0.9103 29.95 0.9577

where N is an indicator of the reaction stage.

For N ≤ 1 (i.e. Phase Ⅰ), the reaction is controlled by the chemical surface reaction, and when reaction is controlled by the diffusion of reactants through a layer of porous reaction products for 1 < N ≤ 2 (i.e. Phase Ⅱ). And if N > 2 (i.e. Phase Ⅲ), the reaction would be controlled by the diffusion of reactants through a layer of dense reaction products.48 It is worth noting that one of the above nine kinetic models (i.e.[1-(1-χ)1/3]2=kt) obtained using Sharp’s method was given by Jander for the condition of N = 2, which did not fit the leaching data of the activated samples (Fig. S11). This result indicated that the leaching rate of the activated samples might be controlled by the chemical surface reaction at the initial stage or the diffusion at the mid-to-late stage through the dense products layer. In addition, Fig. 5c&d illustrated that the mechanical activation process contributed to the obvious dissolution of the samples, compared with the non-activated samples. For example, as the leaching time prolonged to 10 min, the recovery rates of Co from the non-activated and activated samples were 6.58% and 93.25%, respectively. Thus, we could speculate that the Phase Ⅰ and Phase Ⅱ had been completed within the leaching time of 10 min and the leaching reaction for Phase III was controlled by the diffusion at the mid-to-late stage through the dense products layer (N > 2) with the leaching time ranging from 10 to 60 min. This was different from the non-activated samples.

The parameter N was identified via Equation 5 based on the results in Fig. 5c&d. Fig. S12 presented that the plots of ln[1-(1-x)1/3] versus lnt for the leaching of Co and Li from the activated samples under different concentrations of H SO . The values of parameter N 2 4 were calculated and shown in Table 1. The values of parameter N were all more than 2, supporting the above conclusion that Equation [1-(1-χ)1/3]2=kt) did not fit the experimental data of activated samples using Sharp’s method. the speculation that Phase Ⅰ and Phase Ⅱ had ended after the 10-minute leaching and Phase Ⅲ started thereafter. In comparison, for the non-activated samples, the leaching rates of Co and Li were controlled by the chemical surface reaction during the entire leaching process within 60 min. As such, it proved that the mechanical activation process accelerated the dissolution of LiCoO samples via changing the leaching kinetics.

3.5. Cobalt and lithium extraction from spent LIBs

Fig. 6 Extraction of Co and Li from different samples (ball milling time = 60 min, rotation speed = 650 rpm, ball-to-powder mass ratio = 85: 1, [H SO ] = 1 M, [H O ] = 2 vol.%, S/L ratio = 20 g/L, leaching 2 4 0 2 2 0 time = 45 min, leaching temperature = 25 °C).

Above all, the leaching rates of Co and Li from pure LiCoO 2 powders after mechanical activation achieved up to 100% and 93.8% under the conditions of 1 M of H SO and 2 vol% of H O at room 2 4 2 2 temperature, which were milder than the leaching conditions reported in the previous research.24-29 However, it is worth verifying the effectiveness of the enhanced process for the recovery of valuable metals from the spent LIBs. The results showed that the recovery rates of Co and Li were increased significantly from 58.1% to 93.4% and from 81.3% to 100%, respectively, after the mechanical activation process (Fig. 6). Therefore, it proved that the mechanical activation process was an effective approach to pretreat the spend LIBs for higher hydrometallurgical extractions of the valuable metals under mild conditions.

4. Conclusions

Mechanical activation process is feasible and meaningful to enhance the extraction of valuable metals from spent LIBs under mild conditions effectively instead of higher concentration of leaching agents and heating. The mechanism of increasing the leaching efficiency with increasing the ball milling time, rotation speed and ball to powder mass ratio was studied by characterizing the raw and activated samples. Both the decreased particle size and the destroyed morphology and crystal structure after mechanical activation contributed to the improved leaching efficiency of Co and Li. Besides, the high recovery rates of Co and Li could be obtained under mild conditions of 1 M H SO , 2 vol% H O and room 2 4 2 2 temperature. the leaching behavior of valuable metals was influenced by the mechanical activation process, resulting in an accelerated hydrometallurgical extraction of metals as analyzed by the leaching kinetics. Ultimately, about 93.4% Co, and 100% Li could be extracted from actual spent LIBs pretreated by mechanochemical process under mild conditions. Therefore, this mechanical activation process is considered to be an effective and environmental-friendly pretreatment approach for the recycling of spent LIBs.

Acknowledgments

We gratefully appreciate the financial support from Shanghai “Chenguang” Program (15CG60), Shanghai Sailing Program (18YF1429900, 15YF1404300), Natural Science Foundation of China (51678353), Cultivate discipline fund of Shanghai Polytechnic University (XXKPY1601) and Eastern Scholar Professorship Grant. The authors also acknowledge the Graduate Student Funding Program of Shanghai Polytechnic University (A01GY17F022), Shanghai Polytechnic University Leap Program (EGD18XQD24), Industry-Academia-Research Project of Qingpu District,Shanghai (Qing- Industry-Academia-Research 2018-48) and Gaoyuan Discipline of Shanghai-Environmental Science and Engineering (Resource Recycling Science and Engineering).

Appendix A. Supplementary data

Supplementary data associated with this article can be found in the online version.

Conflict of interest

The authors declare that they have no conflict of interest.

References

  1. 1. J. Guan, Y. Li, Y. Guo, R. Su, G. Gao, H. Song, H. Yuan, B. Liang, Z. Guo, ACS. Sustain. Chem. Eng., 2017, 5, 1026-1032.
  2. X. Zhang, Y. Xie, H. Cao, F. Nawaz, Y. Zhang, Waste Manage., 2014, 34,1715-1724.
  3. B. Scrosati, J. Hassoun, Y.K. Sun, Energ. Environ. Sci., 2011, 4, 3287-3295.
  4. T. Zhang, Y. Q. He, F. F. Wang, L. H. Ge, X.N. Zhu, H. Li, Waste Manage., 2014, 34, 1051-1058.
  5. X. Zeng, J. Li, N. Singh, Crit. Rev. Env. Sci. Tec., 2014, 44, 1129-1165.
  6. M. Hayashi, M. Takahashi, T. Shodai, J. Power Sources, 2009, 189, 416-422.
  7. I. Rivera, A. Roca, M. Cruells, F. Patiño, E. Salinas, Hydrometallurgy,2007, 89, 89-98.
  8. S. Zhu, W. He, G. Li, X. Zhou, X. Zhang, J. Huang, Trans. Nonferrous Met. Soc. China, 2012, 22, 2274-2281.
  9. F. Chen, B. Yang, W. Zhang, J. Ma, J. Lv, Y. Yang, Waste Manage, 2017, 60, 660-665.
  10. C. K. Lee, K. I. Rhee, J. Power Sources, 2002, 109, 17-21.
  11. S. Shin, N. Kim, J. Sohn, D. Yang, Y. Kim, Hydrometallurgy, 2005, 79, 172-181.
  12. D. Mishra, D. Kim, D. Ralph, J. Ahn, Y. Rhee, Waste Manage., 2008, 28, 333-338.
  13. A. Chagnes, B. Pospiech, J. Chem. Technol. Biot., 2013, 88, 1191–1199
  14. J. Xu, H. Thomas, R. Francis, K. Lum, J. Wang, B. Liang, J. Power Sources, 2008, 177, 512-527.
  15. L. Li, J. Ge, R. Chen, F. Wu, S. Chen, X. Zhang, Waste Manage., 2010, 30, 2615-2621.
  16. P. Meshram, B. Pandey, T. Mankhand, Waste Manage., 2015, 45, 306-313.
  17. P. Meshram, B. Pandey, T. Mankhand , Chem. Eng. J., 2015, 281, 418-427.
  18. K. Tanong, L. Coudert, G. Mercier, J. Blais, J. Environ. Manage., 2016, 181, 95-107.
  19. M. Contestabile, S. Panero, B. Scrosati, J. Power Sources, 2001, 92, 65-69.
  20. Y. Guo, F. Li, H. Zhu, G. Li, J. Huang, W. He, Waste Manage., 2016, 51, 227-233.
  21. P. Zhang, T. Yokoyama, O. Itabashi, T. Suzuki, K. Inoue, Hydrometallurgy, 1998, 47, 259-271.
  22. P. Choubey, K. Chung, M. Kim, Miner. Eng., 2017, 110, 104-121.
  23. J. Li, P. Shi, Y. Chen, C. Chang, Chemosphere, 2009, 77, 1132-1136.
  24. L. Sun, K. Qiu, J. Hazard. Mater., 2011, 194, 378-384.
  25. J. Nan, D. Han, X. Zuo, J. Power Sources, 2005, 152, 278-284.
  26. J. Kang, G. Senanayake, J. Sohn, S. Shin, Hydrometallurgy, 2010, 100, 168-171.
  27. D. Kim, J. S. Sohn, J. Power Sources, 2004, 132, 145-149.
  28. J. Nan, D. Han, M. Yang, M. Cui, X. Hou, Hydrometallurgy, 2006, 84, 75-80.
  29. L. Chen, X. Tang, Y. Zhang, L.X. Li, Z. Zeng, Y. Zhang, Hydrometallurgy, 2011, 108, 80-86.
  30. D. Ferreira, L. Prados, D. Majuste, M. Mansur, J. Power Sources, 2009, 187, 238-246.
  31. D. Tromans, J. Meech, Miner. Eng., 2001, 14, 1359–1377.
  32. W. Yuan, J. Li, Q. Zhang, F. Satio, Environ. Sci. Technol., 2012, 46, 4109- 4114
  33. S. Saeki, J. Lee, Q. Zhang, F. Satio, Int. J. Miner. Process, 2004, 74, S373-S378.
  34. J. Lee, Q. Zhang, F. Satio, Shigen to Sozai, 2000, 116, 919-922.
  35. M. Wang, C. Zhang, F. Zhang, Waste Manage., 2016, 51, 239-244.
  36. P. Balaz, R. Kammel, Metall., 1996, 50, 345-347
  37. P. Pourghahramani, E. Forssberg, Int. J. Miner. Pro., 2007, 82, 96-105.
  38. V. Galakhov, E. Kurmaev, S. Uhlenbrock, M. Neumann, D. Kellerman, V. Gorshkov, Solid. State. Com., 1996, 99, 221-224.
  39. T. Yuan, Q. Cao, J. Li, Hydrometallurgy, 2010, 104, 136-141.
  40. A. Nasser, U. Mingelgrin, Appl. Clay. Sci., 2013, 44, 141-150.
  41. Y. Guo, Y. Li, X. Lou, et al., J. Mater. Sci., 2018, 53, 13790-13800.
  42. P. Baláž, Mechanochemistry in Nanoscience and Mminerals Engineering, Springer Berlin Heidelberg, 2008, 257-296.
  43. M. Lain, J. Power Sources, 2001, 97-98, 736-738.
  44. B. Swain, J. Jeong, J. Lee, G. Lee, J. Sohn, J. Power Sources, 2007, 167, 536-544.
  45. J. Sharp, J. Am. Ceramic Soc., 1966, 49, 379-382.
  46. M. Sokić, B. Marković, D. Živković, Hydrometallurgy, 2009, 95, 273–279.
  47. A. Rozovskij, Kinetika topohemičeskih reakcij, Himija, Moskva., 1974. (In Russian).
  48. R. Kondo, M. Lee, Diamon, J. Am. Ceram. Soc., 1976, 84, 573-578.

Copyright © 2025 Engineered Science Publisher, All Rights Reserved