Magnetic Li–M (M = Ni, Ni0.8Cu0.2, Cr) layered oxides nanoparticles for Li-ion batteries electrodes

Nanoparticles of Li–Ni, Li–(Ni, Cu) and Li–Cr layered oxides, with potential applications as cathode materials in lithium batteries, were prepared by solid-state reaction and sol-gel method. The combination of structural analysis and magnetic characterization allowed the clear identification of the phases present in the synthesized nanoparticles. The main component of Li–Ni oxide nanoparticles is the electrochemically active and ferrimagnetic phase Li1−z Ni1+z O2, whereas those of Li–Cr oxide are the antiferromagnetic phases LiCrO2 and Cr2O3. A small substitution of Cu for Ni in Li–Ni oxide determines the formation of nanoparticles in which the main phase is the antiferromagnetic phase Li1−z Ni1+z O2. Operation tests in lithium batteries and post-mortem analysis, aimed at assessing the potential of metal oxide nanoparticles as cathode materials, were performed on all samples.


Introduction
In the last two decades, Li-ion batteries (LIB) have experienced enormous growth due to their applications, among others, in the automotive and portable electronics fields. For a long time, the material utilized for the cathode of such batteries has been essentially LiCoO 2 , which however suffers of limited reversible charge capacity and high material cost. For this reason, the substitution of Co with other elements represents an important goal in the effort of increasing the LIB capacity and reducing the cost. In particular, LiNiO 2 was considered the most promising cathode active material (CAM) for future LIB applications. However, critical issues like capacity losses in the first cycle, structural and chemical instabilities hindered the commercialization of LiNiO 2 as CAM. The partial substitution of Ni by other metals led to the improvement of LiNiO 2 performance. Indeed, it was demonstrated that CAMs based on solid solutions strongly benefit from synergetic effects of the different metals [1]. The new generation of nickel-based layered oxides (ANi 1−x M x O 2 , A = Li, Na, K; M = Co, Mn, Al, Cr, etc) gained a great commercial success in recent years, not only in LIBs but also in sodium-ion and potassium-ion batteries [2][3][4], although their charge capacity turns out to be not even enough for the increasing demand for high energy density for electric vehicles.
Thus, a further step is represented by the research focus on LiNiO 2 -based layered oxides LiNi 1−x M x O 2 with x > 90%, although they suffer of severe oxygen alteration and capacity degradation during cycling [5]. Moreover, the continuous Ni dissolution, structural disordering, particle cracking, and instability of cathode electrolyte interphase during cycling hinder their practical applications [6]. Among the others, Cr-based cathode materials (i.e. LiCrO 2 ) have attracted significant interest due to possible multiple electron transfer pathways and stable cyclability, which results from the three-electron redox couple Cr 3+ /Cr 6+ [7,8]. On the other hand, in recent years the cathode chemistry resumed the Ni-rich compositions (>80% Ni), mainly aiming at exploring the upper limit of Ni content in layered oxide CAMs [9,10].
From a more fundamental point of view, the synthesis of two-dimensional (2D) transition metal oxides has attracted considerable attention due to their interesting physical, chemical, and electronic properties. In particular, from the magnetic point of view, the peculiar structure of LiCrO 2 , which is characterized by a layered triangular arrangement of Cr 3+ ions, results in a 2D magnetic frustrated structure together with an overall antiferromagnetic behavior [11]. In a geometrically frustrated system, the geometry of the lattice precludes the simultaneous minimization of energy contributions related to all the magnetic interactions. This gives rise to a degenerate manifold of ground states rather than a single stable ground-state configuration. As a consequence, even slight perturbations may induce instabilities in such systems and allow the emergence of further unusual phenomena, as for example the suppression of long-range magnetic ordering and the promotion of short-range magnetic correlations [12].
A further element that should be considered concerns the possibility of studying LiMO 2 systems at the nanoscale, in particular in form of nanoparticles [13,14]. Recent studies have confirmed that nanomaterials possess both high storage capacity and mechanical stability for long term cycling [15,16]. Among various advantages of the nano-size materials, one may recall the reduction of electron diffusion, enhancing rate capability, the high surface/volume ratio, promoting lithium absorption for high capacities, and the better accommodation of crystalline expansion during lithiation. Moreover, many nanomaterials, arranged in a network, can improve the mechanical properties and electrical percolation. In recent years, it has been shown that the particle size and shape are crucial in optimizing the cyclability of layered Ni-rich oxide CAMs in solid-state batteries because of the need for proper ionic and electronic transport [10,15]. LiCrO 2 samples with grain size below 20 nm showed much higher electrochemical activities and a higher degree of cation mixing with respect to the coarser grain counterpart [8], and often required a conductive matrix to fully exploit their capacity [17]. Recently, metal oxide cathode nanoparticles has been grown as hollow polyhedra, with a resulting improvement of their electrochemical properties [3,4].
LiNiO 2 is known to be difficult to synthesize as a well ordered material [18]. Many different chemical synthesis methods have been tried to obtain Ni-containing layered oxide CAMs. For example: solid-state reaction [19], by using hydroxides and carbonates as lithium and nickel sources, solution methods such as co-precipitation [20] and sol-gel [21], the emulsion method [22], the combustion method with calcination under O 2 flow [23], the mechanical milling process [24], the microwave assisted synthesis [25], and others. Stoichiometric LiNiO 2 nanoparticles have been successfully obtained with the 'Pechini method' , which consists of a sol-gel synthesis in citric acid, followed by calcination [26,27]. In a recent work, the Ni hydroxide Ni(OH) 2 precursor was prepared also by an electrolysis process, and further used to synthesize a LiNiO 2 CAM [28].
In summary, nanoparticles of layered 2D transition metal oxides like LiNiO 2 , LiNi 1−x M x O 2 and LiCrO 2 still deserve a thorough study both from the fundamental and the applicative point of view. For this reason, we prepared, using specific chemical synthesis methods, Li-Ni, Li-Ni 0.8 Cu 0.2 and Li-Cr oxide nanoparticles and performed structural and magnetic analysis to identify their phase content and basic physical properties, as well as operation tests in lithium batteries and post-mortem analysis to evaluate their applicative potential as cathode material.

Methods
Li-Ni oxide nanoparticles were prepared by solid-state reaction followed by grinding and calcination of hydroxides, carbonates or nitrates [29]. In a typical synthesis procedure, NiNO 3 and LiNO 3 were dissolved in deionized water to obtain a mixed metal nitrate aqueous solution at room temperature. Then, citric acid was added to ethylene glycol and, after a short rinsing, this solution was added to the first mixture. The total mixture was slowly heated up to 80 • C for several hours to evaporate the water solvent and finally produce a dark brown viscous gel. The gel was kept in a stove. In order to obtain the best nanoparticles from the structural point of view, the dried gel was then calcinated at 500 • C for 8 h. The obtained material was washed in ethanol and distilled water. Then, Li-Ni oxide nanoparticles were collected through centrifugation for 5 min.
A similar process, adding the Cu precursor Cu(NiO 3 ) 2 to Ni(NO 3 ) 2 and LiNO 3 , was adopted for obtaining Li-Ni 0.8 Cu 0.2 oxide nanoparticles.
Li-Cr oxide nanoparticles were synthesized via a sol-gel method using lithium acetate di-hydrate (C 2 H 3 LiO 2 ·2H 2 O) and chromium nitrate Cr(NO 3 ) 3 as reactants. Stoichiometric amounts of the reactants were dissolved in deionized water and urea (N 2 H 4 CO) was added to the solution as a chelating agent for the gel. The solution was stirred for different stirring times (from 6 to 12 h), followed by heating at 90 • C with vigorous stirring until a viscous gel was formed. The gel was dried in an oven at 80 • C before calcination in air with a tube furnace at 700 • C for 7 h.
The samples were analyzed by x-rays diffraction (XRD), magnetic measurements and galvanostatic chronopotentiometry (GCP) in lithium half cells.
XRD measurements on the as prepared samples were performed in Bragg-Brentano geometry with a SIEMENS D500 powder diffractometer, equipped with a copper anode (CuKα), and a SIEMENS scintillation counter detector.
The dimension of synthesized particles was obtained by dynamic light scattering (DLS) in a Brookhaven ZetaPlus instrument. The particles were dispersed in water and sonicated for 10 min to separate aggregates formed during the calcination of the samples.
Magnetic measurements were carried out by using a Quantum Design MPMS-XL5 SQUID magnetometer in the temperature range 5-300 K and with a maximum magnetic field of 5 T. Temperature dependent iso-field measurements were performed by following two protocols: zero-field-cooled (ZFC) and field-cooled-cooling (FCC) protocol. In the ZFC measurements, the sample is cooled to 5 K without an external magnetic field, then a magnetic field of 0.01 T is applied and the magnetization of the sample is measured during a temperature sweep on heating (temperature sweep rate of 2 K min −1 ). Instead, following the FCC protocol, the magnetic field of 0.01 T is applied with the sample at 300 K and then the magnetization is measured on cooling down to 5 K.
The prepared materials were also tested as electrodes in half-cells. Active materials were ground in a mortar along with carbon black ((CB), Timcal, Super 65) and polyvinylidene fluoride as a plastic binder ((PVDF), Solvay, Solef® 6010). The ratio among the components was kept at 80% active material, 10% CB and 10% PVDF. The powder mixture was transferred in a flat-bottomed flask and N-methyl-pyrrolidone (anhydrous 99.5%, Sigma-Aldrich Co.) was added dropwise until a thick slurry was obtained. The product was mixed by magnetic stirring overnight and successively cast onto aluminum foil with a notch bar (0.3 mm thick). The wet cast was dried in a 60 • C oven for a couple of hours, then transferred in a 70 • C vacuum oven and left overnight under dynamic vacuum. The dried film was successively put under a hydraulic press at about 7 MPa, then punched in 16 mm diameter electrodes and further dried in a 70 • C vacuum oven. The dried electrodes were then moved inside an Ar-filled glovebox (MBraunLabmaster 130, O 2 and H 2 O atmosphere below 0.1 ppm). The electrochemical devices were realized in a standard 2032 coin-cell case with a metallic lithium disc as the counter electrode. Lithium hexafluorophosphate 1 M in an ethyl carbonate and dimethyl carbonate 1:1 v v −1 solution (LiPF 6 , battery grade, Sigma Aldrich Co.) was used as the electrolyte. A Celgard® disc was employed as a separator.
The electrochemical measurements of the half-cells were performed with a Landt CT2001A battery testing system. A preliminary test was performed, on a sacrificial cell, with GCP, in a voltage window between 1 and 5 V, in order to study the electrochemical reaction plateaus. Then, GCP measurements were performed on half-cells at different C-rates, operating in a voltage window between 3.2 and 4.1 V.
Post-mortem XRD measurements of the materials were performed in a Debye-Scherrer geometry with a Bruker D8 Discover diffractometer, equipped with a copper anode (CuKα) coupled with a Göbel mirror and a 0.5 mm collimator and a Rayonix MX225 2D area detector (with 73.2 µm pixels size). Proper calibration of the detector was performed using a corundum standard. Images were processed with the software FIT2D. Samples were prepared as follows: for each sample, two identical half-cells were assembled and charged-discharged for a long (>50) number of cycles. One of the cells ended its cycles in a fully-charged state (cathode completely de-lithiated), while the other cell ended its cycles in a fully-discharged state (cathode completely lithiated). The cells were then opened with a hydraulic cell disassembler (TMAX) in an Ar-filled glove box. The cathodes were thoroughly washed with dimethyl carbonate (battery grade, Sigma Aldrich Co.) to remove the electrolyte and avoid the formation of salt crystals on the surface. Grazing incidence XDR was then performed on each electrode, with a grazing angle of 10 • . Refinements were performed with GSAS-II software.

Li-Ni oxide nanoparticles
The DLS analysis of the Li-Ni oxide sample shows a bimodal distribution of particles size. Most of the particles have a dimension around 150 nm. Whereas other particles have a size of about 500 nm and are probably aggregates of smaller particles. The XRD pattern at room temperature of as-prepared Li-Ni oxide nanoparticles, reported in figure 1 (blue curve), reveals the presence of only one main phase with a rhombohedral structure (space group R-3 m). This structure can be associated to the electrochemically active layered LiNiO 2 phase [26]. The small ratio between the intensity of the reflections (003) and (104), evidenced by the diffraction peaks at 18.6 • and 44.0 • , points out the concomitant presence of a Li 1−z Ni 1+z O 2 or a Ni-deficient Li 1−x NiO 2 phase, isostructural with LiNiO 2 , due to a phenomenon of cation mixing, with extra Ni 3+ ions occupying the Li interslabs sites (z > 0) [26]. The cell dimensions (a = 2.91 Å and c = 14.28 Å), obtained through a Rietveld refinement (figure S1), are similar to the ones of the stoichiometric LiNiO 2 phase (a = 2.90 Å and c = 14.20 Å) [26], with an elongation of the c-axis, probably caused by presence of Ni 2+ ions, characterized by a larger ionic size [30], induced by a Li deficiency (z > 0). From the Rietveld refinement, a stoichiometry of Li 0.52 Ni 1.48 O 2 was obtained, although a distribution of different stoichiometries of the rhombohedral phases (LiNiO 2 and Li 1−z Ni 1+z O 2 ) is probable.
The outcomes from XRD analysis are strengthened by results from the magnetometric characterization of the sample. Indeed, ZFC-FCC magnetization versus temperature measurements ( figure 2(a)) clearly indicate the onset of a ferro-magnetic (FM) or ferri-magnetic (FiM) behavior at T C ≈ 240 K, in agreement with literature works [31] on bulk Li 1−z Ni 1+z O 2 samples with z = 0.36. A more recent work on magnetic and structural behavior of LiNiO 2 -NiO solid solution [32] has shown that a bulk Li 1−z Ni 1+z O 2 sample with z = 0.26 turns out to be a (FiM, uncompensated antiferromagnet) with a T C ≈ 231 K, slightly different from that measured in our case. The stabilization of the ferromagnetic order is promoted by the extra Ni ions in the inter slab space, since they introduce a strong coupling between adjacent NiO 2 slabs.
The possible presence of other secondary phases with antiferromagnetic (AFM) order, like LiNiO 2 (A-type antiferromagnet with T N ≈ 9 K [26]), could be hardly detected by magnetization measurements, being masked likely by the overwhelming FiM contribution of Li 1−z Ni 1+z O 2 . It has to be noted that it is very common to find in LiNiO 2 bulk samples, together with the Néel temperature at 9 K, a further transition at ≈240 K, due to chemical inhomogeneity (Griffiths phase) [26]. In any case, no trace of a Néel transition at 9 K is visible in our measurements. This result can be explained considering the difficulty to prepare LiNiO 2 with the 1:1 Li/Ni stoichiometry, unlike the related NaNiO 2 . Indeed, most of reported LiNiO 2 samples contain a small amount of Ni in the Li + layer, thus resulting intrinsically off-stoichiometric [26]. Concerning the hysteresis loop at low temperature (10 K) reported in figure 2(b): its peculiar shape could correspond to the superposition of the main FiM signal of Li 1−z Ni 1+z O 2 phase and a superparamagnetic (SPM) behavior, due to the contribution of small SPM particles of LiNiO 2 or NiO 2 , hardly detectable by XRD. Indeed, small particles of an antiferromagnetic material should exhibit SPM and weak ferromagnetic behavior, as pointed out by Nèel (1962) [33], who attributed the permanent magnetic moment to uncompensated spins in the two sublattices. In this case, the possible detection of a blocking temperature, corresponding to a peak in the ZFC curve, is likely to be screened by the large FM contribution of Li 1−x Ni 1+z O 2 . For example, it has been reported for NiO NPs that the blocking temperature can be in the range 50-270 K, depending on the size of NPs [34].
LiNiO 2 was also tested as cathode material for Li-ion batteries. Electrochemical performances were tested with GCP analysis, within a potential window ranging from 3.2 to 4.1 V vs Li/Li + . This range was chosen because of the observed potential plateau on a sacrificial cell, in agreement with the previously reported mechanism of discharge-charge process of LiNiO 2 -based lithium batteries. GCP charge and discharge curves, measured with a C-rate of C/20 during the 1st cycle, are shown in figure 3(a). From the GCP curves, performed on a sacrificial cell, two different potential plateaus, due to two different ion intercalation/ extraction processes, can be observed at about 1.7 V and 3.7 V. The presence of two redox reactions can be explained by the presence of two different phases of Li-Ni oxides. The intercalation process at 3.7 V is ascribable to a Li 1−x NiO 2 phase. The 1.7 V plateau can be explained either with the intercalation/extraction of lithium in the Li 1−z Ni 1+z O 2 phase, or with an intercalation/extraction process occurring after a known phase transition in Li 1−x NiO 2 when x < 0.7 [35]. For both plateaus, a slight ohmic shift in the potential can be observed, hinting at a high resistivity of the Li-Ni oxide cathode, which could be due to the relatively large size of the nanoparticles (150-500 nm).
The GCP charge and discharge capacity at different C-rates for a half-cell with Li-Ni oxide-based electrode is shown in figure 3(b). The electrode capacity is characterized by a good reversibility: no significant capacity decrease is observed for a given C-rate and the starting capacity is retained even after the most extreme charge and discharge conditions (C-rate of 5 C). The capacity is stable at C/20 after 50 cycles and up to the 70th cycle, with a specific capacity of about 75 mA h g −1 . This value is lower of the nominal capacity of LiNiO 2 , and this could be due to the presence of two different phases, Li 1−x NiO 2 and Li 1−z Ni 1+z O 2 , the latter of which not contributing to the capacity due to the fact that the cell is operating at a higher potential. Another explanation could be that lithium ions only intercalate the surface of the crystallites due to the slow diffusion that hints the fully intercalation into the core of the Li-Ni oxide particles. The electrode capacity decreases for high C-rates, probably due to a low conductivity and/or to the slow lithium ions diffusion in the Li-Ni oxide particles. The slow intercalation process confirms the presence of big Li-Ni oxide particles, or the presence of nanoparticle aggregates.
Post-mortem XRD was performed on the fully-charged (de-lithiated) and fully-discharged (lithiated) electrodes. The XRD of a de-lithiated electrode can be observed in figure 1 (orange curve). From the XRD on the fully de-lithiated electrode (completely charged) we observe an asymmetric broadening of the peaks with an average shrinking of the unit cell (a = 2.91 Å, c = 14.20 Å). This can be explained considering a non-uniform de-lithiation of the sample and an overall decrease of the particle size, which was evaluated to be ≈41 nm. The decrease of the particle size can be explained with the cracking of the nanoparticles due to the repeated intercalation/extraction of lithium. Less intense peaks (2θ = 41.7 • and 44.7 • ) are ascribable to the presence of metallic nickel.
From the XRD of the fully lithiated electrode (completely discharged) we can appreciate the enlargement of the unit cell (a = 2.91 Å, c = 14.44 Å) due to the Li intercalation. The cell dimensions exceed the ones of starting sample suggesting the ability to intercalate more Li than the one intercalated during the synthesis. The extra Li is probably taken from the electrolyte, which would further explain the low initial capacity and slight increase of the capacity after many cycles. The peaks remain very broad, indicating a decrease in particle size, which was evaluated to be ≈37 nm, consistent with the size of the nanoparticles of the fully de-lithiated electrode. A slight asymmetry of the peaks is observed, indicating a large distribution of Li concentration. Less intense peaks, ascribable to the presence of metallic nickel, were observed, indicating that the formation of metallic nickel from LiNiO 2 is irreversible.

Li-Ni 0.8 Cu 0.2 oxide nanoparticles
The DLS analysis of the Li-Ni 0.8 Cu 0.2 oxide particles shows again a bimodal distribution of particles size, with peaks at about 130 nm and 440 nm.
The XRD analysis of the Li-Ni 0.8 Cu 0.2 oxide sample ( figure S4) reveals that the partial substitution of Ni by Cu results in the widening of the structure. Moreover, the absence of the (003) peak at 2θ = 18.8 • suggests that the structure becomes cubic, which is a poorly electrochemically active phase [26]. Moreover, we can observe from the diffraction pattern that the peaks are clearly split, probably due to the segregation of a Li-Ni oxide phase. This agrees with the previous report on the poor structural stability of Cu-substituted LiNiO 2 [36]. Moreover, from the diffractogram, minor peak doublets that can be related to the presence of a secondary Ni oxide or (Ni, Cu) oxide phase are evident.
The  it is evident that a small Cu substitution for Ni is enough to determine a completely different scenario, from the point of view of the magnetic behavior of the main phase, that can be explained considering the inability of Cu, differently from Ni, to couple the NiO 2 slabs and promote a ferromagnetic order.
Lithium half cells based on Li-Ni 0.8 Cu 0.2 oxide nanoparticles were studied with GCP and post-mortem XRD (figures S5 and S6). They showed a rather poor performance, most likely due to the high inhomogeneity of the starting powder, as seen from the XRD analysis.

Li-Cr oxide nanoparticles
The DLS analysis of the Li-Cr oxide particles shows a lower size (about 50 nm) as compared to the Li-Ni oxide particles. Moreover, it is evident a second peak in the size distribution, centered at 250 nm, ascribable to the presence of aggregates.
The XRD analysis (figure 5) reveals the presence of a large amount of Cr 2 O 3 phase with a rhombohedral structure (spatial group: R-3 c), together with the desired LiCrO 2 phase (rhombohedral, spatial group R-3 m). From the relative intensity of the XRD peaks, the content of Cr 2 O 3 was estimated to be of 19%, with the remaining 81% being LiCrO 2 . The ZFC M(T) measurement (figure 6(a)) shows the presence of a magnetic transition at about 365 K (inset of figure 6(a)) and a second transition slightly above 60 K. The transition at 365 K could be related to CrO 2 , which is a known ferromagnet, although the Curie temperature of such phase is rather higher: T C = 396 K [37]. However, it has to be considered that a system of Cr 2 O 3 /CrO 2 nanoparticles is reported in literature to show a Curie temperature T C = 386 K [38], thus meaning that a size effect could reduce the transition temperature. The kink at about 300 K could represent instead a signal of the possible presence of Cr 2 O 3 , which is a canted (or frustrated) antiferromagnet with a Néel temperature T N = 306 K, although its contribution is expected to be hindered by the overwhelming role of the ferromagnetic phase. As regards the transition at low temperature, it can be ascribed to LiCrO 2 , which is an antiferromagnet with a Néel temperature reported at T N = 64 K [39] or T N = 62 K [40]. The steep increase of the low-field magnetization below 50 K could be due to the contribution of another (unknown) paramagnetic phase: the M(H) curve at 10 K (figure 6(b)) displays indeed a linear behavior, typical of a paramagnet. This paramagnetic phase may be the amorphous phase highlighted by the large background at small angle of XRD pattern (figure 5).
In summary, despite the presence of the expected antiferromagnetic phase LiCrO 2 , the synthesized Li-Cr oxide nanoparticles show a prevailing content of Cr 2 O 3 and CrO 2 phases, in which Li ions are absent.
Lithium half cells based on LiCrO 2 nanoparticles were studied with GCP and post-mortem XRD (figures S7 and S8). They showed a rather poor performance, most likely due to the rather large size of the LiCrO 2 particles, since the electrochemical activity of LiCrO 2 is hindered in the bulk material and can be fully exploited only in form of nanoparticles and/or in conjunction with a highly conductive carbon-based matrix [8,17].

Conclusion
The synthesis process based on solid-state reaction has allowed the obtainment of Li-Ni oxide nanoparticles characterized by the presence of the FiM phase Li 1−z Ni 1+z O 2 as the main component, together with a Ni-deficient antiferromagnetic secondary phase Li 1−x NiO 2 . This result confirms the difficulty of obtaining only the stoichiometric LiNiO 2 phase, because of instability of trivalent nickel species and disordering of cationic distribution at lithium sites. Although these difficulties arise in general from the high temperature synthesis methods, it has to be considered that in our case a low-temperature solid state reaction has been adopted to avoid this problem. Li-Ni oxide was tested as an electrode in lithium-based half cells, showing good reversibility but a lower-than-expected specific capacity, probably due to the initial deficiency of lithium in the Li 1−x NiO 2 phase, to the presence of a secondary Li 1−z Ni 1+z O 2 phase, activated at a lower potential, and to the large size of particles that hinders the lithium ions diffusion.
The partial substitution of Cu for Ni to obtain Li-Ni 0.8 Cu 0.2 oxide nanoparticles induces a widening of the crystallographic structure that becomes cubic. This structural change, together with the segregation of secondary Li-Ni oxide phases, due to the instability of Cu-substituted LiNiO 2 phase, worsens the electrochemical performance of the oxide. Moreover, the substitution of Ni with Cu reduces the magnetic coupling of NiO 2 slabs suppressing the FiM order.
As regards the Li-Cr oxide nanoparticles, which have been synthesized by the sol-gel method, they have been found to be constituted by the antiferromagnetic phases LiCrO 2 and Cr 2 O 3 , together with the ferromagnetic CrO 2 as a secondary phase. A further contribution from an amorphous phase has been also detected. The electrochemical performance of Li-Cr nanoparticles results poor due to large amount of Cr-oxide phases and to the size of the particles.
In summary, the combination of structural analysis and magnetic characterization turned out to be fundamental to obtain a clear identification of the phases that are present in the synthesized Li-M-oxide nanoparticles and strongly affect their electrochemical performance. Further work is necessary in order to get pure stoichiometric Li-Ni and Li-Cr oxide nanoparticles with dimensions that can promote their exploitation in efficient cathodes for Li-ion batteries.

Data availability statement
All data that support the findings of this study are included within the article (and any supplementary files).