RESEARCH ARTICLE

www.advenergymat.de

Direct Upcycling of Leached FePO4 from Spent Lithium-Ion

Batteries toward Gradient-Doped LiMnxFe1−xPO4 Cathode
Material
Jian Zhou, Chunxian Xing, Jiawei Huang, Yucheng Zhang, Guowei Li, Long Chen,
Shuqiang Tao, Zhuoli Yang, Guangren Wang, and Linfeng Fei*

advantages of high stability, high safety,

Lithium-ion batteries (LIBs) with LiFePO4 (LFP) cathode materials have and low cost, which occupy over one-
occupied a significant market share in state-of-the-art power storage systems third of the market share.[2] However, a
and electric vehicles, yet the approaching “retiring wave” of these LIBs should significant issue associated with the in-
be appropriately treated (i.e., recycling of spent LIBs). Current recycling creasing production of LFP-based LIBs
is the huge amount of degraded batter-
strategies of degraded LFP materials largely direct on leaching of lithium ies at the end of their lifespan, which
elements, leaving the leached FePO4 residual as a waste. Inspiringly, requires them to be properly treated.[3]
LiMnx Fe1−x PO4 material is considered one of the potential candidates for Focusing on the core component of
next-generation cathode materials because of its high electrochemical these batteries, the LFP cathode mate-
performance, extraordinary stability, and low cost. Herein, the direct upcycling rial, conventional recycling methods (py-
rometallurgical/hydrometallurgical) are
of leached FePO4 toward LiMn0.25 Fe0.75 PO4 (LMFP) cathode material is
not economically applicable as it does not
demonstrated via a green, simple, and scalable mechanochemical pathway. contain valuable elements.[4] Therefore,
The product features a gradient-doping of Mn, as well as a uniform carbon state-of-the-art strategies for handling
coating for LMFP particles. As a result, the LMFP cathode material delivers degraded LFP materials are exclusively
superior electrochemical performance; the capacity (161.3 mA h g−1 at 0.1 C), aimed at the extraction of lithium.[5] Yue
rate capability (90.1 mA h g−1 at 5 C), cycling stability (95.6% capacity et al. used Na2 S2 O8 solutions to recover
Li from degraded LFP materials, which
retention after 800 cycles at 1 C), and energy density (≈15% increase as produced Li2 CO3 with a purity of over
compared to LFP) are almost comparable with those of fresh materials. The 99.5%.[6] Jin et al. developed an “air
as-established upcycling protocol offers a favorable reuse method for leached oxidation-water leaching” method to se-
FePO4 materials. lectively recover lithium from degraded
LFP with a leaching efficiency of up to
99.3%.[7] In this circumstance, it is worth
noting that the leached iron phosphate
1. Introduction (FePO4 ·2H2 O, abbreviated as FP) as a residue of lithium extrac-
tion from degraded LFP is directly discarded in most cases, which
Lithium-ion batteries (LIBs) are expanding their applications
is, in fact, not only a waste of resources but also an environmental
in miniaturized electronics, energy storage systems, and elec-
risk.
tric vehicles (EVs) because of their high capacity, lightweight,
Recently, a direct regeneration strategy, which proposes to re-
and flexible dimensions.[1] For example, LIBs based on lithium
store the electrochemical performance of degraded cathode ma-
iron phosphate (LiFePO4 , abbreviated as LFP) cathode materi-
terials by repairing their compositions without destroying the
als have become the technology of choice for EVs due to their
structures, has emerged as a preferred recycling approach for
degraded LFP due to its simplicity, efficiency, greenness, and
low cost.[8] Specifically, as the major degradation mechanisms
J. Zhou, C. Xing, J. Huang, Y. Zhang, G. Li, L. Chen, S. Tao, Z. Yang,
G. Wang, L. Fei for LFP materials are lithium loss and structural defects (mi-
School of Physics and Materials Science crocracks in LFP particles, fractures of carbon coating on their
Jiangxi Key Laboratory for Two-Dimensional Materials and Jiangxi surfaces, etc.),[9] the objectives for direct regeneration thus in-
Engineering Laboratory for Advanced Functional Thin Films clude lithium replenishment, structural restoration, and recov-
Nanchang University
Nanchang 330031, China ery of carbon coating.[10] Qiu et al. used hydrogen peroxide to
E-mail: fl[email protected] oxidize degraded LFP cathode material into FP, followed by cal-
cination for regenerating LFP material; their product demon-
The ORCID identification number(s) for the author(s) of this article strated excellent rate capability and long-term cycling stability
can be found under https://doi.org/10.1002/aenm.202302761 (initial capacity of 141 mA h g−1 and 99% capacity retention af-
DOI: 10.1002/aenm.202302761 ter 300 cycles at 1 C).[11] Ji et al. sintered an organic lithium salt

Adv. Energy Mater. 2024, 14, 2302761 2302761 (1 of 9) © 2023 Wiley-VCH GmbH
 16146840, 2024, 7, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/aenm.202302761 by Colorado State University, Wiley Online Library on [12/11/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.advancedsciencenews.com www.advenergymat.de

(3,4-dihydroxybenzonitrile dilithium) with degraded LFP cathode low), which is beneficial for Li+ transportation kinetics and struc-
material for regeneration, and the product showed good cycling tural stability of the material. Besides, the addition of glucose en-
stability (88% capacity retention after 400 cycles at 5 C) and rate dowed the controllable carbon coating for LMFP particles, which
performance.[12] Therefore, the direct regeneration is of remark- is essential for improving the apparent conductivity of this cath-
able potential in addressing the upcoming “retiring wave” of LFP- ode material. On determining the appropriate carbon contents
based LIBs.[13] We want to further mention here that the delithia- for the upcycled materials, please refer to Figure S2, Supporting
tion degree in degraded LFP materials can be evidently different Information, for details.
from batch to batch (due to the distinct cycling histories and the
accidental disassembly of spent batteries at their charged state),
which makes the direct regeneration of these materials prefer- 2.2. Phase and Compositional Analysis
ably start from leached FP rather than degraded LFP in practical
scenarios.[14] XRD patterns of FP (the leached FePO4 ), R-LFP (the upcycled
Besides, it is noticed that lithium manganese iron phos- LFP with 5 wt% of carbon), and R-LMFP-G12 (the upcycled
phate (LiMnx Fe1−x PO4 , 0 < x < 1; abbreviated as LMFP) ma- LMFP with 12 wt% of carbon) are shown in Figure 1b,c. The char-
terial, which is obtained by doping LFP with controllable man- acteristic peaks in the XRD pattern of FP are in good agreement
ganese content,[15] has been proposed as a cost-effective, high- with those in the standard PDF card (JCPDS No. 72–0471), yet
performance cathode material for next-generation power batter- the overall peak intensity is relatively weak, indicating the poor
ies due to its higher voltage platform over LFP (LMFP has its crystallinity of FP material (also see Figure S1, Supporting In-
energy density 10–20% higher than that of LFP).[16] Consider- formation). One can also see that the XRD patterns of R-LFP
ing that for leached FP, its long-term electrochemical cycling his- and R-LMFP-G12 present similar, intense diffraction peaks and
tory together with the lithium extraction process should lead to both are basically consistent with the standard PDF card (JCPDS
the presence of defects (such as microcracks) in the material;[17] No. 83–2092), which suggests that they possess high crystallinity
and in case hetero-elements (such as Mn) are supplemented and are in typical olivine structure with an orthorhombic Pmnb
during the upcycling of FP, the material is prone to form a space group. Meanwhile, it is worth noting that the fine-scanned
gradient-doped structure from outside to inside of the particles, XRD patterns of R-LFP and R-LMFP-G12 in the 2𝜃 range of 22°–
which should be advantageous in improving the electrochemical 38° in Figure 1c show slight shifts in the peak positions; while
performance.[18] In this context, we herein demonstrate a direct the (111), (211), and (311) peaks of R-LFP coincide well with the
upcycling scheme for leached FP material toward carbon-coated, standard LFP, the (021), (200), and (131) peaks of R-LMFP-G12
gradient-doped LMFP material via a green, simple, and scalable shift toward low-angle directions, which implies the successful
mechanochemical pathway. The direct upcycling protocol in this Mn doping-induced lattice expansion of R-LMFP-G12.[19] In or-
work contains a step of ball milling for materials mixing (ele- der to further explore the influence of Mn doping, Rietveld re-
mental supplementation) followed by another step of solid-state finements for the patterns of R-LFP and R-LMFP-G12 were per-
annealing to recover/modify the structure and achieve carbon- formed (see Figure S3, Supporting Information, for the results).
coating. As expected, the upcycled LiMn0.25 Fe0.75 PO4 (LMFP)/C The lattice parameters of a, b, and c for R-LMFP-G12 are all
material delivered excellent electrochemical performance (initial larger than those of R-LFP, which is attributed to the larger ra-
capacity of 161.3 mA h g−1 at 0.1 C, 95.6% capacity retention after dius of Mn2+ (0.83 Å) than Fe2+ (0.63 Å).[20] Besides, the absence
800 cycles at 1 C, enhanced energy density by ≈15%), surpass- of characteristic peaks corresponding to carbon in the XRD pat-
ing even those of fresh materials. The as-developed direct upcy- terns of R-LFP and R-LMFP-G12 suggests the amorphous nature
cling strategy is environmentally friendly, economical, and feasi- of the carbon content within the upcycled products. Therefore,
ble, and hence holds its significance in reshaping the recycling Raman spectroscopy (RM) was utilized to probe the presence of
system for spent LFP-based LIBs. carbon in R-LFP and R-LMFP-G12 materials (Figure 1d). Both D-
band (1340 cm−1 ) and G-band (1580 cm−1 ) of carbon have been
2. Results and Discussion observed in the upcycled products, representing the disorder-
induced phonon mode and E2g vibrations of graphite, respec-
2.1. Experimental Design tively. The ID /IG (the areal ratio of D-band over G-band) value of
R-LMFP-G12 is 3.73, significantly lower than the 4.60 of R-LFP
We used a two-step mechanochemical approach to achieve the (Table S1, Supporting Information), which represents a higher
direct upcycling of leached FP toward LMFP material. The degree of graphitization for carbon content in R-LMFP-G12 and
schematic diagram of the upcycling process is shown in Figure hence a better electronic conductivity for it.[21]
1a. Typically, FP particles exhibited a broad size distribution and To determine the exact chemical compositions for upcycled
contained surface defects (e.g., microcracks; refer to Figure S1, materials, an inductively coupled plasma-optical emission spec-
Supporting Information). The first step of ball-milling in the troscopy (ICP-OES) test was employed (refer to Table S2, Support-
upcycling process simultaneously reduced the particle size and ing Information, for the results). The compositions of R-LFP and
mixed the raw materials (FP, Li2 CO3 , MnCO3 , and glucose), and R-LMFP-G12 were found to be LiFe0.989 P0.980 O3.919 (with 4.96 wt%
the following solid-state sintering (which was a carbothermic re- carbon) and LiMn0.245 Fe0.739 P0.958 O3.832 (with 12.01 wt% carbon),
duction reaction) promoted the structural/compositional restora- respectively, which are both closely matched with the expected
tion and modification toward carbon-coated LMFP material (see values of our synthesis. Furthermore, a combination of thermo-
Experimental section for details). Especially, the manganese con- gravimetric analysis (TG) and differential scanning calorimetry
tent was gradient-doped within the LMFP particles (see text be- (DSC) was employed to study the mass changes of FP, R-LFP,

Adv. Energy Mater. 2024, 14, 2302761 2302761 (2 of 9) © 2023 Wiley-VCH GmbH
 16146840, 2024, 7, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/aenm.202302761 by Colorado State University, Wiley Online Library on [12/11/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.advancedsciencenews.com www.advenergymat.de

Figure 1. Experimental design and structural characterizations of materials. a) Schematic illustration for the upcycling process of leached FePO4 toward
LiMn0.25 Fe0.75 PO4 material. XRD patterns for FP, R-LFP, and R-LMFP-G12 materials in the 2 theta regions of b) 15°–65° and c) 22°–38°. d) Raman
spectra for FP, R-LFP, and R-LMFP-G12 materials.

and R-LMFP-G12 upon heating in air (Figure S4, Supporting In- ticles generally featured a few micro/nano cracks, which are at-
formation). It is known that for FP, the mass loss primarily arises tributed to their cycling history together with lithium extraction
from the evaporation of crystalline water; whereas for R-LFP and processes.[17] However, these cracks would facilitate the subse-
R-LMFP-G12, the mass loss can be attributed to two main fac- quent elemental supplementations, which can act as diffusion
tors: the oxidation of transition metal elements (Fe/Mn) and the channels for Li and Mn sources and lead to a gradient-doping
removal of carbon content. Therefore, the estimated carbon con- of Mn in the product (Figure 2b and Figure S1e–g, Support-
tents in R-LFP and R-LMFP-G12 were determined as 5.28% and ing Information). Elemental mappings of energy dispersive spec-
11.99%, respectively, which are both in good agreement with the troscopy (EDS) suggest that the distributions of Fe, P, and O ele-
above ICP-OES results. ments in FP particles are basically uniform, while the C element
may originate from residual impurities (binders, carbon, etc.) af-
ter lithium extraction (Figure S1d, Supporting Information).[22]
2.3. Microstructural Characterizations R-LFP, R-LMFP-G8 (the upcycled LMFP with 8 wt% of carbon
as a control sample), and R-LMFP-G12 particles have consid-
Scanning electron microscopy (SEM) images for FP, R-LFP, and erably smaller particle sizes than that of FP (Figure 2c,d and
R-LMFP-G12 materials are shown in Figure 2a–d. FP sample Figure S5a–c, Supporting Information), which are ascribed to
mainly exhibits irregular microparticles with occasional pres- the ball-milling treatment during our upcycling; besides, there
ence of impurities such as binder residues (Figure 2a and Figure are all visible carbon coatings on the surfaces of the three
S1a, Supporting Information). Moreover, the surface of FP par- materials.

Adv. Energy Mater. 2024, 14, 2302761 2302761 (3 of 9) © 2023 Wiley-VCH GmbH
 16146840, 2024, 7, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/aenm.202302761 by Colorado State University, Wiley Online Library on [12/11/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.advancedsciencenews.com www.advenergymat.de

Figure 2. Microstructural characterizations of materials. a) SEM image for FP particles. b) Schematic diagram for the elemental replenishment process
of FP particles during upcycling; the insets are SEM and TEM images of FP particles featured with surface cracks. c) SEM image for R-LFP particles.
d) SEM image for R-LMFP-G12 particles. HAADF-STEM micrographs of e) R-LFP and f) R-LMFP-G12 surfaces and the corresponding lattice-resolved
iFFT images. g) EDS elemental mappings for R-LMFP-G12 particles. h) STEM image of an R-LMFP-G12 particle and the corresponding Fe and Mn
L-edge EELS spectra (a linescan with a depth ranging from the surface to the interior of ≈35 nm). i) High-resolution XPS spectra of Fe 2p for R-LFP and
R-LMFP-G12.

Subsequently, systematical scanning transmission electron (200) crystal plane. Moreover, compared to the encapsulated car-
microscopy (STEM) characterizations were performed on these bon layer for R-LFP particles (which is about 1.6 nm in thick-
materials. High-angle annular dark-field scanning transmission ness; Figure 2e), R-LMFP-G12 particles possess a thicker car-
electron microscopy (HAADF-STEM) was used to identify the bon coating of about 2 nm (Figure 2f). Microscopically, the Fe,
atomic structure of R-LFP and R-LMFP-G12, and the images P, and O elements in both R-LFP (Figure S5a, Supporting In-
are shown in Figure 2e,f (including their regional inverse fast formation) and R-LMFP-G12 (Figure 2g) particles show their
Fourier transform patterns, iFFTs). In the HAADF-STEM im- uniform distributions in the EDS elemental mappings. More-
age of R-LFP (Figure 2e), a well-defined olivine crystal struc- over, considering the geometrical influence of “spherical” parti-
ture can be observed. The interplanar spacing is determined as cles on the elemental concentrations, the EDS linescans of Fe &
0.431 nm in the iFFT, corresponding to the (101) crystal plane Mn across an R-LMFP-G12 particle (Figure S6, Supporting In-
of the Pmnb-type structure. Similarly, R-LMFP-G12 (Figure 2f) formation) evidently reveal a gradient distribution of Mn from
also shows a distinct olivine structure, indicating that Mn dop- the surface (higher concentration) to the interior (lower con-
ing has not altered the crystal structure. The interplanar spac- centration) of the R-LMFP-G12 particles (also see the following
ing is determined as 0.528 nm in the iFFT, corresponding to the text).[18b]

Adv. Energy Mater. 2024, 14, 2302761 2302761 (4 of 9) © 2023 Wiley-VCH GmbH
 16146840, 2024, 7, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/aenm.202302761 by Colorado State University, Wiley Online Library on [12/11/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.advancedsciencenews.com www.advenergymat.de

To further investigate the valence states and concentration pro- shows that the appropriate carbon content (refer to above SEM
files of Fe/Mn in the material, a series of electron energy loss observations) can lead to an excellent electrochemical perfor-
spectroscopy (EELS) analysis was conducted from the surface to mance. The discharge capacity of R-LMFP-G12 is approaching
the interior (points 1 to 7) on a random R-LMFP-G12 particle, as that of R-LFP, indicating that a 25% level of Mn doping does
shown in Figure 2h. The fixed positions of Fe L-edge (710.4 eV) not significantly affect the capacity of LFP; therefore, consider-
and Mn L3 -edge (640.6 eV) confirm the homogeneous oxidation ing the higher voltage plateaus of LMFP, the energy density of
states (+2) for Fe & Mn in both surface and bulk phases. It is R-LMFP-G12 (322.6 W h kg−1 ) increases ≈15% after gradient-
worth noting that the intensity of the Mn L3 -edge monotonously doping of Mn (R-LFP: 280.33 W h kg−1 ; R-LMFP-G8: 299 W h
weakens from the outermost surface to the interior of the parti- kg−1 ). Thereafter, cyclic voltammetry (CV) tests were also con-
cle, substantiating the aforementioned gradient distribution of ducted in a range of ≈2.5–4.2/4.5 V at a scan rate of 0.1 mV s−1
Mn concentration in R-LMFP-G12 particles. X-ray photoelec- (Figure S8, Supporting Information). The oxidation and reduc-
tron spectroscopy (XPS) analysis was also performed to observe tion peaks observed in the CV profiles are in complete agree-
the surface information of the upcycled samples, as shown in ment with the plateaus in the above charge–discharge profiles,
Figure 2i and Figure S7, Supporting Information. R-LFP and R- indicating good electrochemical reversibility for the upcycled ma-
LMFP-G12 clearly exhibit their expected characteristic peaks for terials. It is also worth noting that the potential differences for
elements. Figure 2i shows the high-resolution Fe 2p XPS spectra oxidation/reduction peaks are 0.31 and 0.33 V for R-LFP and R-
for the two samples, which display characteristic peaks at bind- LMFP-G8, respectively, whereas it is only 0.20 V for R-LMFP-
ing energies of 711 (Fe 2p3/2 ) and 724.2 (Fe 2p1/2 ) eV, indicat- G12, which corroborates that R-LMFP-G12 exhibits its lower po-
ing the presence of Fe2+ in both materials after the carbothermic larization together with better electrochemical reversibility (due
reduction process. Similarly, in the high-resolution Mn 2p XPS to its smaller particle size and preferable carbon coating).[25] Fur-
spectrum for R-LMFP-G12 material (Figure S7b, Supporting In- thermore, the dQ/dV profiles of the three materials are shown
formation), characteristic peaks at binding energies of 642.9 (Mn in Figure 3b. The profiles of R-LMFP-G8 and R-LMFP-G12 are
2p3/2 ) and 653.9 (Mn 2p1/2 ) eV can be observed; these peaks in- similar; and compared to R-LFP, R-LMFP-G12 shows minimal
dicate the dominant presence of Mn2+ in R-LMFP-G12. In ad- voltage polarization (0.01 V; see the zoom-in inset of Figure 3b),
dition, the XPS fine spectra of C 1s in R-LFP and R-LMFP-G12 revealing that the crystalline structure of the upcycled materials
materials (Figure S7c, Supporting Information) can both be de- has been restored.[8c] Electrochemical impedance spectroscopy
convoluted into three subpeaks with binding energies of 284.8 (EIS) results in Figure 3c also confirm the superior capacitive
(C─C), 286 (C─O─C), and 288.5 (O─C═O) eV.[23] The areal ratios performance of R-LMFP-G12. The high-frequency semicircular
of C─O─C subpeak to C─C subpeak in the XPS spectra (Table portions in Nyquist plots of the three materials correspond to
S3, Supporting Information) are 0.43 and 0.41 for R-LFP and R- the ohmic resistance (Rs ) and charge transfer resistance (Rct ),
LMFP-G12, respectively, indicating that R-LMFP-G12 has fewer reflecting the electronic transportation limitations of the mate-
carbon defects (which is in line with the previous RM results).[21] rials; the low-frequency linear portions denote the diffusion limi-
These high-resolution SEM, STEM, EDS, EELS, and XPS results tations. R-LFP, R-LMFP-G8, and R-LMFP-G12 exhibit similar Rs
jointly demonstrate that FP has been successfully transformed and Rct values in the high-frequency region, whereas R-LMFP-
into LFP/LMFP with an olivine structure after upcycling; and es- G12 shows less diffusion limitation. As displayed in Figure S9,
pecially, R-LMFP has achieved a gradient doping of Mn within its Supporting Information, the calculated diffusion coefficient of
particles. Li+ for R-LMFP-G12 is higher than that of R-LFP and R-LMFP-G8
(Table S4, Supporting Information), which can be a result of its
smaller particle size. Overall, it is confirmed that the Mn doping
2.4. Electrochemical Evaluations does not significantly increase the impedance of the material, yet
the carbon coating, the nano-sized particle size, and the gradient
The above structural analysis demonstrates that R-LMFP-G12 doping structure of R-LMFP-G12 rather promote the transporta-
material has great potential to exhibit excellent electrochemical tion kinetics of Li+ .
performance. Figure 3a shows the first charge–discharge profiles Figure 3d presents the comparison of rate capability for the
for R-LFP, R-LMFP-G8, and R-LMFP-G12 materials at 0.1 C (1 C three materials at different current densities from 0.1 to 5 C. Al-
= 170 mA h g−1 ) in a voltage range of ≈2.5–4.2/4.5 V (vs Li+ /Li). though R-LFP exhibits higher capacity in all measured ranges
In contrast with R-LFP (which exhibits a single voltage plateau at (see also Figure S10, Supporting Information), it is noticed when
3.4 V), R-LMFP-G8 and R-LMFP-G12 synchronously show two the current density is restored to 0.1 C after the charge–discharge
voltage plateaus at around 3.5 and 4 V, respectively, correspond- at 5 C, the specific discharge capacity of R-LMFP-G12 recovers
ing to the oxidation reactions of Fe2+ /Fe3+ and Mn2+ /Mn3+ . Com- from 90.1 (at 5 C) to 160.2 mA h g−1 (at 0.1 C; namely, a ca-
pared to pure LFP (≈3.4 V) and LiMnPO4 (≈4.1 V), the discharge pacity recovery of 70.1 mA h g−1 , and the values are 50.2 and
plateaus for Fe2+ /Fe3+ and Mn2+ /Mn3+ in R-LMFP-G8 and R- 65.5 mA h g−1 for R-LFP and R-LMFP-G8, respectively). This in-
LMFP-G12 materials display slight differences, which are mainly dicates that R-LMFP-G12 exhibits a better recovery of capacity,
caused by the ion super-exchange effect of transition metals in which can be attributed to the enhanced structural stability of
the LMFP solid solution (i.e., LMFP contains Fe2+ and Mn2+ , R-LMFP-G12 due to the gradient-doped Mn in the material.[20]
which may occupy the positions of other ions in the lattice).[24] Furthermore, the cycling performance of R-LFP, R-LMFP-G8,
At a discharge rate of 0.1 C, the specific capacities for R-LFP, and R-LMFP-G12 materials is summarized in Figure 3e and
R-LMFP-G8, and R-LMFP-G12 materials are 164.9, 149.5, and Table S5, Supporting Information. The cycling stability of all
161.3 mA h g−1 , respectively. The high capacity of R-LMFP-G12 three materials under 1 C for 300 cycles is basically good, with

Adv. Energy Mater. 2024, 14, 2302761 2302761 (5 of 9) © 2023 Wiley-VCH GmbH
 16146840, 2024, 7, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/aenm.202302761 by Colorado State University, Wiley Online Library on [12/11/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.advancedsciencenews.com www.advenergymat.de

Figure 3. Comparative electrochemical performances for R-LFP, R-LMFP-G8, and R-LMFP-G12 materials. a) Galvanic charge/discharge profiles at 0.1
C. b) The corresponding dQ/dV profiles for (a). c) Nyquist plots at 4.2 V. d) Rate performance. e) Cycling performance at 0.1 and 1 C in a voltage range
of 2.5–4.2/4.5 V. f) The corresponding dQ/dV profiles during cycling for R-LMFP-G12. g) Comparison of electrochemical performances for R-LMFP in
this work against representative LFP and Mn-doped LFP materials in literature (be noted that the LFP materials are recycled samples, while the doped
LFP materials are freshly synthesized ones).[11,12,25,27].

capacity retention rates of 98.6% (R-LFP), 96.4% (R-LMFP- tively, R-LMFP-G12 shows the smallest voltage polarization of
G8), and 98.8% (R-LMFP-G12). However, R-LMFP-G12 exhibits 0.02 V, implying its stable crystal structure (which can be a result
97.6% capacity retention after 600 cycles, significantly higher of the gradient-doped structure of R-LMFP-G12 that weakens the
than the values of 77.6% and 91.7% for R-LFP and R-LMFP-G8, reaction between the material surface and electrolyte).[8c] Addi-
respectively. Moreover, the difference in cycling stability among tionally, the gradient-doped structure of R-LMFP-G12 inhibits the
the three materials becomes more pronounced after 800 cy- dissolution and diffusion of Fe2+ in the Li layer and/or electrolyte,
cles; the specific discharge capacity of R-LMFP-G12 remains at thus maintaining a stable olivine-type crystal structure during
121.7 mA h g−1 , much higher than those of 26.8 and 105.7 mA h such long cycling (i.e., excellent cycling stability). It is notewor-
g−1 for R-LFP and R-LMFP-G8, respectively (the corresponding thy that the highest voltage polarization of R-LMFP-G8 together
capacity retentions are 18.2%, 87.3%, and 95.6% for R-LFP, R- with its poor cycling stability in this regime can be caused by the
LMFP-G8, and R-LMFP-G12, respectively). The excellent capacity insufficient carbon coating on the particles.
retention of R-LMFP-G12 confirms its outstanding structural sta- The electrochemical performance of R-LMFP-G12 materials in
bility due to the gradient doping of Mn and the sufficient carbon this work is then compared with representative LFP (regenerated)
coating.[26] Moreover, the voltage hysteresis profiles of the three and doped LFP (fresh) materials in recent reports (Figure 3g and
materials during 600 cycles at 1 C are shown in Figure 3f and Table S6, Supporting Information).[11,12,25,27] It can be easily con-
Figure S11, Supporting Information. Compared to the voltage po- cluded that the electrochemical performance of our R-LMFP-G12
larizations of 0.11 and 0.13 V for R-LFP and R-LMFP-G8, respec- material is nearly on par with or even slightly higher than those of

Adv. Energy Mater. 2024, 14, 2302761 2302761 (6 of 9) © 2023 Wiley-VCH GmbH
 16146840, 2024, 7, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/aenm.202302761 by Colorado State University, Wiley Online Library on [12/11/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.advancedsciencenews.com www.advenergymat.de

Figure 4. Postmortem structural characterizations for R-LMFP-G12 material after 600 cycles at 1 C. SEM images for R-LMFP-G12 a) before cycling and
b) after cycling. c) HAADF-STEM micrographs of an R-LMFP-G12 particle after cycling and the corresponding lattice-resolved STEM image (inset: FFT
pattern). d) STEM image and EDS elemental mappings for an R-LMFP-G12 particle after cycling. e) STEM image of an R-LMFP-G12 particle after cycling
and the corresponding Fe and Mn L-edge EELS spectra (a linescan with a depth ranging from the surface to the interior of ≈70 nm). High-resolution
XPS spectra of f) Fe 2p, g) Mn 2p, and h) F 1s for R-LFP and R-LMFP-G12 after cycling.

previous materials in terms of specific capacity (161.3 mA h g−1 ings of 0.24/0.25 and 0.28/0.31 nm, corresponding to the inter-
at 0.1 C), rate performance (90.1 mA h g−1 at 5 C), and capacity planar distances for (−22-1)/(221) and (301)/(2-11) crystal planes
retention (95.6% capacity retention after 800 cycles at 1 C). of olivine structure, respectively. Besides, no amorphous region
was found in our repeated observations, which means that the
Mn gradient-doping has brought additional stability for LMFP
2.5. Postmortem Structural Analysis lattices. It is also worth noting that the thicknesses of carbon
coating on R-LFP and R-LMFP-G8 particles after cycling are only
We have also performed systematical postmortem analysis on the about 3.2 and 3.6 nm, respectively, while the cycled R-LMFP-G12
cathode materials after cycling to verify their structural advan- particles possess a carbon layer thickness of about 4.9 nm. The
tages resulting from our upcycling strategy. From the XRD pat- larger thicknesses of carbon coating in the materials after cycling
terns of R-LFP and R-LMFP-G12 before and after cycling (Figure compared to those before cycling may be attributed to the growth
S12, Supporting Information), it can be seen that the long his- of CEI (i.e., cathode electrolyte interphase). Furthermore, EDS
tory of 600 cycles has not markedly changed the crystal structures elemental mappings of Fe (Mn), P, O, and F elements for R-
for the upcycled materials. SEM images of R-LFP (Figure S13, LFP (Figure S14b, Supporting Information), R-LMFP-G8 (Figure
Supporting Information) and R-LMFP-G12 (Figure 4a,b) materi- S15b, Supporting Information), and R-LMFP-G12 (Figure 4d) cy-
als before/after cycling also show unidentifiable changes in their cled particles still show a uniform and synchronized distributions
morphologies and sizes. Further careful analysis was performed of these elements. Again, the EDS linescans of Fe and Mn in an
by HAADF-STEM and EDS (elemental mappings and linescans) R-LMFP-G12 particle after 600 cycles (Figure S16, Supporting In-
to identify the morphology, structure, and elemental distribution formation) confirm the stable gradient doping in R-LMFP mate-
of R-LFP and R-LMFP-G12 materials after cycling, as shown in rial. EELS analysis was also performed again on an R-LMFP-G12
Figure 4c,d and Figure S14, Supporting Information. The lattice- particle from surface to interior (points 1 to 7) after cycling to
resolved STEM image for R-LFP material after cycling (Figure investigate the valence states and concentration distributions of
S14a, Supporting Information) shows lattice spacings of 0.28 and Fe and Mn (Figure 4e). Although there are slight shifts in the Fe
0.36 nm, corresponding to the interplanar distances for (021) and L-edge and Mn L3 -edge as compared to those before cycling, the
(1-11) crystal planes of olivine structure, respectively. It is inter- spectra still indicate that Fe and Mn in R-LMFP maintain their
esting to find that region 2 exhibits the characteristics of amor- oxidation states as well as the Mn gradient-doped structure after
phization, indicating that localized phase transition occurs in the cycling.
cycling of R-LFP material; it is possible that a portion of the LFP Furthermore, XPS analysis was conducted to further observe
phase has transformed into the FePO4 phase.[28] The STEM im- the surface information of the materials after cycling, as shown
ages of R-LMFP-G8 (Figure S15a, Supporting Information) and in Figure 4f–h and Figure S17, Supporting Information. R-
R-LMFP-G12 (Figure 4c) materials after cycling show lattice spac- LFP, R-LMFP-G8, and R-LMFP-G12 clearly exhibit their original

Adv. Energy Mater. 2024, 14, 2302761 2302761 (7 of 9) © 2023 Wiley-VCH GmbH
 16146840, 2024, 7, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/aenm.202302761 by Colorado State University, Wiley Online Library on [12/11/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.advancedsciencenews.com www.advenergymat.de

characteristic peaks for all elements after cycling. Figure 4f shows 3:1:3.15:0.6), glucose (LMFP:C mass ratio = 1:0.08/1:0.12), and
the high-resolution Fe 2p XPS spectra for the two samples, which appropriate amount of deionized water. The mixture was ball
exhibit a Fe3+ peak at 712.3 eV (yellow region), but the relative milled at a speed of 600 r min−1 for 6 h before being transferred
intensity of Fe3+ is lower in R-LMFP-G12 compared to that of to an evaporating dish and dried at 80 °C to obtain the precur-
R-LFP, indicating that the Mn gradient doping reduces the con- sor. In the second step, the precursor was transferred to an alu-
version of Fe2+ to Fe3+ (which is the reason for the better cycling minum oxide crucible and subjected to an annealing sequence of
stability of R-LMFP-G12).[20] In addition, the high-resolution F 450 °C/4 h (heating rate of 5 °C min−1 ) plus 700 °C/16 h (heating
1s (Figure 4h) and P 2p (Figure S17b, Supporting Information) rate of 10 °C min−1 ) under an argon atmosphere. After anneal-
XPS spectra for the two samples show that their relative intensi- ing, the powder was washed three times with deionized water to
ties are higher in the post-cycled R-LFP, suggesting that the CEI remove soluble salts and then dried at 80 °C for 4 h to obtain the
of R-LFP may be thicker than that of R-LMFP-G12; this not only upcycled LMFP powder. The synthesis of upcycled LFP powder
confirms the above SEM observations of an increased thickness was consistent with the above descriptions except for the addi-
of the surface layer for R-LFP after cycling but further states that tion of MnCO3 . To demonstrate the potential and feasibility of
the structural design of R-LMFP-G12 stabilizes the material dur- this direct upcycling strategy, a brief techno-economic analysis
ing electrochemical cycling. In brief, during long-term cycling, was performed to estimate the economic viability, which shows
R-LFP material likely undergoes local phase transformation (i.e., the strategy has significant profit ($ 1.346 per kg; refer to Figure
LFP transforms into FePO4 ), which leads to slight degradation S18 and Table S7, Supporting Information).
of material; in contrast, R-LMFP-G12 with Mn gradient-doping Materials Characterizations: Raman profiles were taken us-
suppresses the above phase transformation, stabilizes the olivine ing a WITec alpha300R system. ICP-OES tests were conducted
structure, and hence prevents the material degradation. on a SHIMADZU ICPE-9000 spectrometer. XRD measurements
were taken using a Rigaku D/Max diffractometer with filtered Cu
K𝛼 radiation (𝜆 = 1.5406 Å). TG-DSC analysis was performed
3. Conclusion on a PerkinElmer STA 6000 system. SEM observations were
made on a Thermos Scientific Scios2 dual-beam electron micro-
In summary, we developed a green, simple, and scalable two-step
scope. TEM and HAADF-STEM images, EDS elemental map-
direct upcycling strategy to efficiently transform leached FePO4
pings and linescans, and EELS spectra were obtained through a
material toward LMFP. Benefiting from the structural charac-
double aberration-corrected Titan Themis Z electron microscope
teristics of leached FePO4 and our upcycling process, the LMFP
equipped with a Gatan Continuum 1065 EELS device. XPS spec-
product featured an Mn gradient-doping together with a uniform
tra were collected on a Thermos Scientific ESCALAB Xi+ system.
carbon coating on the surface of particles. Consequently, the up-
Cell Assembly and Electrochemical Measurements: The cathode
cycled material delivered excellent electrochemical performance;
material was mixed with Ketjen Black and PVDF binder in an
specifically, the capacity of the material at 0.1 C reached as high as
NMP solvent at a mass ratio of 8:1:1 to form a uniform slurry.
161.3 mA h g−1 , and it still retained 95.6% of its initial capacity
The slurry was then cast on an Al foil and vacuum-dried at 90 °C.
after 800 cycles at 1 C. Although this study focuses on the pro-
The above electrode film was cut into round discs with a diame-
duction of LMFP with an Mn/Fe ratio of 1:3, we anticipated that
ter of 12 mm (the active material loading is around 2.7 mg cm−2 ).
the direct upcycling protocol involved in this work can be applied
The CR2025 coin cells were then assembled in an Ar-filled glove-
to produce LMFP materials with other Mn/Fe ratios (e.g., 1:1),
box using Celgard 2400 membrane as the separator, Li metal foil
as well as LFP materials doped with other elements. Therefore,
as the counter electrode, and 1 mol L−1 LiPF6 (EC: DEC: DMC
the direct upcycling approach provides an effective value-adding
= 1:1:1 as solvent, volume ratio) as the electrolyte. Subsequently,
treatment for leached FePO4 material and is of practical potential
galvanostatic charge/discharge tests were conducted on a LAND
in the recycling of spent LIBs.
CT2001A battery tester in the voltage window of 2.5–4.2/4.5 V
versus Li/Li+ . CV with a sweep rate of 0.1 mV S−1 and a voltage
4. Experimental Section range of 2.5–4.2/4.5 V and EIS measurements in the frequency
range of 0.01 Hz–100 kHz were carried out on a CHI660E elec-
Materials and Reagents: Lithium carbonate (Li2 CO3 , Aladdin, trochemical workstation.
AR ≥ 99.99%), manganese carbonate (MnCO3 , Aladdin, AR
≥ 99%), ammonium dihydrogen phosphate (NH4 H2 PO4 , Al-
addin, AR ≥ 99%), glucose (C6 H12 O6 , Aladdin, AR ≥ 99%), Supporting Information
polyvinylidene fluoride (PVDF, Arkema, Mw ≈ 110 000–130 000),
N-methyl-2-pyrrolidone (NMP, Aladdin, AR ≥ 99%), and Ketjen Supporting Information is available from the Wiley Online Library or from
Black (EC-300J, Suzhou Yilong Energy Technology Co., Ltd.) were the author.
used as received. Leached iron phosphate (FePO4 ·2H2 O) was pro-
vided by Zhejiang Tianneng Co., Ltd. after Li leaching of de-
graded lithium iron phosphate (LFP) from spent LIBs. Acknowledgements
Upcycling of Leached FePO4 Material: The upcycling pro-
The authors acknowledge the financial support from the National Natu-
cess included two steps. First, leached FePO4 ·2H2 O material ral Science Foundation of China (22375081, 51972159, and U21A20500)
was placed in a nylon ball milling jar along with manganese and Jiangxi Provincial Natural Science Foundation (20212ACB204016,
carbonate, lithium carbonate, ammonium dihydrogen phos- 20192ACB21018, and 20171ACB20006). L.F. also acknowledges the sup-
phate (FePO4 ·2H2 O: MnCO3 : Li2 CO3 : NH4 H2 PO4 molar ratio = port from Nanchang University.

Adv. Energy Mater. 2024, 14, 2302761 2302761 (8 of 9) © 2023 Wiley-VCH GmbH
 16146840, 2024, 7, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/aenm.202302761 by Colorado State University, Wiley Online Library on [12/11/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.advancedsciencenews.com www.advenergymat.de

Conflict of Interest Xia, Ind. Eng. Chem. Res. 2022, 61, 3831; c) W. Song, J. Liu, L. You, S.
Wang, Q. Zhou, Y. Gao, R. Yin, W. Xu, Z. Guo, J. Power Sources 2019,
The authors declare no conflict of interest. 419, 192.
[11] X. Qiu, B. Zhang, Y. Xu, J. Hu, W. Deng, G. Zou, H. Hou, Y. Yang, W.
Sun, Y. Hu, X. Cao, X. Ji, Green Chem. 2022, 24, 2506.
[12] G. Ji, J. Wang, Z. Liang, K. Jia, J. Ma, Z. Zhuang, G. Zhou, H.-M. Cheng,
Data Availability Statement
Nat. Commun. 2023, 14, 584.
The data that support the findings of this study are available from the cor- [13] a) Q. Jing, J. Zhang, Y. Liu, W. Zhang, Y. Chen, C. Wang, A. C. S.
responding author upon reasonable request. Sustain, Chem. Eng. 2020, 8, 17622; b) J. Tan, Q. Wang, S. Chen, Z.
Li, J. Sun, W. Liu, W. Yang, X. Xiang, X. Sun, X. Duan, Energy Storage
Mater. 2021, 41, 380.
[14] M. Zhang, L. Wang, S. Wang, T. Ma, F. Jia, C. Zhan, Small Methods
Keywords 2023, 7, 2300125.
direct upcycling, gradient doping, leached FePO4 , LiMnx Fe1−x PO4 , spent [15] J. Hong, F. Wang, X. Wang, J. Graetz, J. Power Sources 2011, 196, 3659.
lithium-ion batteries [16] a) Y. Deng, C. Yang, K. Zou, X. Qin, Z. Zhao, G. Chen, Adv. Energy
Mater. 2017, 7, 1601958; b) J. Ding, Z. Su, H. Tian, Ceram. Int. 2016,
Received: August 21, 2023 42, 12435; c) J. Li, C. Guo, Y. Qin, X. Ning, Mater. Technol. 2020, 35,
Revised: November 16, 2023 565.
Published online: December 6, 2023 [17] L. He, H. Li, X. Ge, S. Li, X.u Wang, S. Wang, L. Zhang, Z. Zhang, Adv.
Mater. Interfaces 2022, 9, 2200515.
[18] a) I. M. Nwachukwu, A. C. Nwanya, A. B. C. Ekwealor, F. I. Ezema,
J. Energy Storage 2022, 54, 105248; b) J. Meng, L. Xu, Q. Ma, M.
Yang, Y. Fang, G. Wan, R. Li, J. Yuan, X. Zhang, H. Yu, L. Liu, T.
[1] a) J. Li, J. Fleetwood, W. B. Hawley, W. Kays, Chem. Rev. 2022, 122,
Liu, Adv. Funct. Mater. 2022, 32, 2113013; c) H. Yu, Y. Cao, L. Chen,
903; b) J. Xu, X. Cai, S. Cai, Y. Shao, C. Hu, S. Lu, S. Ding, Energy
Y. Hu, X. Duan, S. Dai, C. Li, H. Jiang, Nat. Commun. 2021, 12,
Environ. Mater. 2023, 6, 12450; c) A. Manthiram, Nat. Commun. 2020,
4564.
11, 1550; d) C. Zhu, R. E. Usiskin, Y. Yu, J. Maier, Science 2017, 358,
[19] B. Zhang, W. Meng, Y. Gong, G. Hu, Z. Peng, K. Du, B. Makuza, J.
1400.
Wu, X. Xie, Y. Cao, Mater. Today Energy 2022, 30, 101162.
[2] a) E. Fan, L. Li, Z. Wang, J. Lin, Y. Huang, Y. Yao, R. Chen, F. Wu, Chem.
[20] F. Jiang, K. Qu, M. Wang, J. Chen, Y. Liu, H. Xu, Y. Huang, J. Li, P. Gao,
Rev. 2020, 120, 7020; b) L. Xu, J. Li, L. Li, Z. Luo, Y. Xiang, W. Deng,
J. Zheng, M. Chen, X. Li, Sustainable Energy Fuels 2020, 4, 2741.
G. Zou, H. Hou, X. Ji, Small 2021, 17, 2102978; c) W. Mrozik, M. A.
[21] Y. Zou, J. Cao, H. Li, W. Wu, Y. Liang, J. Zhang, Ind. Chem. Mater. 2023,
Rajaeifar, O. Heidrich, P. Christensen, Energy Environ. Sci. 2021, 14,
1, 254.
6099.
[22] a) L. Yang, Z. Gao, T. Liu, M. Huang, G. Liu, Y. Feng, P. Shao, X. Luo,
[3] a) M. Li, J. Lu, Z. Chen, K. Amine, Adv. Mater. 2018, 30, 1800561; b)
Environ. Sci. Technol. 2023, 57, 4591; b) L. Yang, Y. Tu, H. Li, W. Zhan,
X. Zhang, L. Li, E. Fan, Q. Xue, Y. Bian, F. Wu, R. Chen, Chem. Soc.
H. Hu, Y. Wei, C. Chen, K. Liu, P. Shao, M. Li, G. Yang, X. Luo, Angew.
Rev. 2018, 47, 7239.
Chem., Int. Ed. 2023, 62, 8702.
[4] a) Y. E. Durmus, H. Zhang, F. Baakes, G. Desmaizieres, H. Hayun, L.
[23] A. V. Shchukarev, D. V. Korolkov, Cent. Eur. J. Chem. 2004, 2, 347.
Yang, M. Kolek, V. Küpers, J. Janek, D. Mandler, S. Passerini, Y. Ein-
[24] Y.-K. Hou, G.-L. Pan, Y.-Y. Sun, X.-P. Gao, ACS Appl. Mater. Interfaces
Eli, Adv. Energy Mater. 2020, 10, 2000089; b) M. M. Thackeray, Adv.
2018, 10, 16500.
Energy Mater. 2020, 11, 2001117; c) B. Huang, Z. Pan, X. Su, L. An, J.
[25] B. Chen, M. Liu, S. Cao, G. Chen, X. Guo, X. Wang, Mater. Chem. Phys.
Power Sources 2018, 399, 274.
2022, 279, 125750.
[5] Y. Li, W. Lv, H. Huang, W. Yan, X. Li, P. Ning, H. Cao, Z. Sun, Green
[26] Y. Zhao, J. Liu, S. Wang, R. Ji, Q. Xia, Z. Ding, W. Wei, Y. Liu, P. Wang,
Chem. 2021, 23, 6139.
D. G. Ivey, Adv. Funct. Mater. 2016, 26, 4760.
[6] X.-H. Yue, C.-C. Zhang, W.-B. Zhang, Y. Wang, F.-S. Zhang, Chem. Eng.
[27] a) Y. Song, B. Xie, S. Song, S. Lei, W. Sun, R. Xu, Y. Yang, Green Chem.
J. 2021, 426, 131311.
2021, 23, 3963; b) G. Hu, Y. Gong, Z. Peng, K. Du, M. Huang, J. Wu,
[7] H. Jin, J. Zhang, D. Wang, Q. Jing, Y. Chen, C. Wang, Green Chem.
D. Guan, J. Zeng, B. Zhang, Y. Cao, A. C. S. Sustain, Chem. Eng. 2022,
2022, 24, 152.
10, 11606; c) C. Qi, S. Wang, X. Zhu, T. Zhang, Y. Gou, Z. Xie, Y. Jin,
[8] a) J. Neumann, M. Petranikova, M. Meeus, J. D. Gamarra, R. Younesi,
Y. Wang, L. Song, M. Zhang, J. Alloys Compd. 2022, 924, 166612; d)
M. Winter, S. Nowak, Adv. Energy Mater. 2022, 12, 2102917; b) J. Wu,
P. Liu, Y. Zhang, P. Dong, Y. Zhang, Q. Meng, S. Zhou, X. Yang, M.
M. Zheng, T. Liu, Y. Wang, Y. Liu, J. Nai, L. Zhang, S. Zhang, X. Tao,
Zhang, X.i Yang, J. Alloys Compd. 2021, 860, 157909; e) W.-C. Chien,
Energy Storage Mater. 2023, 54, 120; c) C. Xing, H. Da, P. Yang, J.
Z.-M. Hsieh, Thin Solid Films 2020, 700, 137890; f) D. V. Trinh, M. T.
Huang, M. Gan, J. Zhou, Y. Li, H. Zhang, B. Ge, L. Fei, ACS Nano
T. Nguyen, H. T. M. Dang, D. T. Dang, H. T. T. Le, H. T. N. Le, H. V.
2023, 17, 3194; d) M. K. Tran, M.-T. F. Rodrigues, K. Kato, G. Babu, P.
Tran, C. D. Huynh, Sci. Rep. 2021, 11, 12280; g) H. Shi, Y. Zhang, P.
M. Ajayan, Nat. Energy 2019, 4, 339.
Dong, X. Huang, J. He, J. Duan, D. Wang, Y. Zhang, Ceram. Int. 2020,
[9] Y. Jin, T. Zhang, M. Zhang, Adv. Energy Mater. 2022, 12, 2201526.
46, 11698.
[10] a) P. Xu, Q. Dai, H. Gao, H. Liu, M. Zhang, M. Li, Y. Chen, K. An, Y.
[28] W. Wang, R. Wang, R. Zhan, J. Du, Z. Chen, R. Feng, Y. Tan, Y. Hu, Y.
S. Meng, P. Liu, Y. Li, J. S. Spangenberger, L. Gaines, J. Lu, Z. Chen,
Ou, Y. Yuan, C. Li, Y. Xiao, Y. Sun, Nano Lett. 2023, 23, 7485.
Joule 2020, 4, 2609; b) X. Liu, M. Wang, L. Deng, Y.-J. Cheng, J. Gao, Y.

Adv. Energy Mater. 2024, 14, 2302761 2302761 (9 of 9) © 2023 Wiley-VCH GmbH