The Importance of Interphases in Energy Storage Devices: Methods and Strategies to Investigate and Control Interfacial Processes

: Extended interphases are playing an increasingly important role in electrochemical energy storage devices and, in particular, in lithium-ion and lithium metal batteries. With this in mind we initially address the differences between the concepts of interface and interphase. After that, we discuss in detail the mechanisms of solid electrolyte interphase (SEI) formation in Li-ion batteries. Then, we analyze the methods for interphase characterization, with emphasis put on in-situ and operando approaches. Finally, we look at the near future by addressing the issues underlying the lithium metal/electrolyte interface, and the emerging role played by the cathode electrolyte interphase when high voltage materials are employed.


Introduction
There is a growing discussion about the meaning and the role of interface/interphase in functional materials and devices [1]. Indeed, an interface is a two-dimensional boundary between phases in thermodynamic equilibrium. On the other hand, this could be seen just as a theoretical statement, as dynamic equilibrium will take place at the atomic scale. As an example, non-zero equilibrium exchange currents exist when electrodes are placed in ionic solutions [2]. Therefore, in the nanometer and sub-nanometer range, the concept of interface must be replaced by (or considered together with) that of interphase, a threedimensional layer which in the solid state can be characterized by a different crystalline structure, microstructure, physico-chemical and functional properties with respect to the parent phases in contact [3]. Obviously, the role of the interphase becomes more and more important as particle/phase dimensions decrease.
Extended interphases are of relevance in electrochemical devices such as batteries, fuel cells, sensors and supercapacitors. Concerning lithium-ion batteries (LIBs), which at present are the market choice for both automotive and grid storage [4], a fundamental role to date has been played by the solid electrolyte interphase (SEI) between the anode (graphite) and the liquid electrolyte [5,6]. SEI is a passivation layer in the range~0.01-0.1 µm, due to electrolyte degradation products which are formed at low electric potential during the first cycles of the battery. Mechanical and chemical SEI stability is of paramount importance for the life cycle of state-of-the-art LIBs (generation 3a) which operate below 4.3 V vs. Li + /Li [6]. Next generation batteries (generations 3b and 4) will work at higher potential (up to 4.8 V), which will cause a major role of the cathode/electrolyte interphase (CEI), formed by the high voltage decomposition of both cathode and electrolyte [7,8]. This will require a careful tailoring of the electrolyte/cathode interface. The higher the voltage, the more important the interfacial control.
In this perspective we will address the role of the interphases in lithium batteries. We will put emphasis on the mechanisms and reactions at the basis of SEI formation (Section 2).

Interphase Investigation Methods
LIBs active components and materials were successfully characterized with tradi tional ex-situ approaches (e.g.: X-rays diffraction, electrochemical impedance spectros copy (EIS)), which are easy to implement, generally inexpensive, and accessible in man laboratories. Ex-situ characterization of pure pristine components (e.g., cathode and an ode powders or electrodes deposited on the current collectors) led to reliable informatio on both the structure and functional properties of such materials, allowing researchers t infer their behavior when embodied in the final device and under working conditions. At the graphite interface, the interphase is made up of a dense layer of nanometric grains of inorganic salts (Li 2 CO 3 , LiF, Li 2 O) formed by the decomposition reaction of the electrolyte. This dense layer prevents the contact between the active material and the electrolyte itself, arresting its further decomposition but allowing, at the same time, Li ions to pass through. Between the dense inorganic layer and the solution bulk there is a secondary layer made by organic compounds such as poly(olefins) and carbonates, which are also produced by the decomposition of the solvent. The role of this external layer, having much higher ionic conductivity, is fundamental because it favors the formation of the dense inner layer, preventing the penetration of small molecules into the graphite structure which could lead to delamination. This can explain the contradictory behavior between EC and PC, two very similar molecules that differ just for the presence of a methyl group attached to the ring. In fact, upon reduction, EC generates three dimensional branched polymeric chains which passivate the internal shell, while PC forms ternary radicals which can give substituted olefins or linear polymeric chains unable to form a dense protecting layer [19,20].
The nature and the thickness of the SEI layer depends on two main factors: the surface morphology and the electrolyte composition. Battery-grade graphite for LIBs has usually spherical morphology with micrometric particle sizes. This guarantees the presence of an extended edge plane surface responsible for the lithium-ion intercalation, which does not take place through the denser basal planes [9]. The anisotropic nature of the graphite structure has also an effect on the nature of the SEI layers. The "real" SEI is indeed observed only at the edge plane interface, with a thickness of about 35 nm and the already described chemical composition [21]. In contrast, the basal SEI is thinner (7 nm), it is made mainly by organic decomposition products, it continues to grow during cycling, and it has the only role of surface passivation without ion transport properties [22,23].
Since 1990, the improvement of the performance of graphite anodes was obtained thanks to the more careful control of the SEI, obtained by working on both passivation and ion transport properties [24]. Thus, additives such as vinylene carbonate (VC) [25] and fluoroethylene carbonate (FEC) [26] were developed to improve the graphite first cycle Coulomb efficiency, the rate capability and the capacity retention.

Interphase Investigation Methods
LIBs active components and materials were successfully characterized with traditional ex-situ approaches (e.g., X-rays diffraction, electrochemical impedance spectroscopy (EIS)), which are easy to implement, generally inexpensive, and accessible in many laboratories. Ex-situ characterization of pure pristine components (e.g., cathode and anode powders or electrodes deposited on the current collectors) led to reliable information on both the structure and functional properties of such materials, allowing researchers to infer their behavior when embodied in the final device and under working conditions.
On the contrary, ex-situ investigation of LIB interfaces/interphases can be challenging, as this involves the disassembly of the cell to expose the layer of interest. This operation can lead to modification of the peculiar microstructure and texture of the surface itself and can also alter the composition of these layers, mainly due to H 2 O/O 2 exposure even at very low concentrations. For these reasons, the in-situ (full device, not working) and operando (full device under working conditions) approaches, even if more expensive, less accessible, and more challenging are better tools for the investigation of the formation, structure, and evolution of LIBs interphases.
The major limitation for the implementation of in-situ and operando methods is the need of dedicated experimental set-up, dedicated cell design, and cell assembly. A second relevant point to be considered for the investigation of the interphases is their intrinsic nature of 3D layers with variable thickness, complex microstructure and composition. Thus, when considering the investigation technique of choice, spatial resolution must be considered in addition to sensitivity and selectivity.
Moreover, as the interphase is a layer evolving under working conditions, also time resolution is a relevant parameter for dynamic characterization of the SEI and CEI. Indeed, both SEI and CEI may undergo significant evolution during the working cycles of the battery in terms of volumetric expansion/contraction, phase transitions, parasitic reactions and both reversible and irreversible phases formation. Finally, a thorough evaluation may require spanning over several orders of magnitude both in spatial and time domains, and thus inevitably resorts to the use of multi-techniques approach offering complementary information. In the following, the most relevant techniques are discussed with some examples of application in the study of LIBs interphases. Table 1 reports the main advantages and disadvantages of the different techniques with respect to their capability of investigating interfaces/interphases.

Diffraction-Based Techniques
Diffraction techniques are among the most exploited methods to characterize battery materials, and chiefly the electrodes [27][28][29][30][31]. The main advantages offered by diffraction methods are: (i) the versatility related to use different radiation sources (Xrays, neutron, electrons, offering complementary information, penetration depth, and sensitivity), (ii) the non-destructive nature of the measurement, and (iii) the possibility to detect both crystalline and amorphous components. However, some major complications must be considered when the in-situ and operando approaches are considered. Indeed, dedicated cells with specific design are needed to minimize the undesired radiation interaction with battery components such as case, current collectors and separators [27]. Moreover, special radiation sources (e.g., synchrotron [30]) are required to allow for fast acquisitions. Finally, this is a bulk technique which does not allow spatial resolution, but for tomography and microtomography applications (see Section 3.4) [30]. Finally, depending on the selected radiation source (laboratory X-rays, synchrotron light, neutrons, electrons) the data acquisition time can vary significantly, highly impacting on the possible time resolution.
The structural and compositional evolution of the electrode was reported by several publications [27][28][29][30][31][32][33][34][35][36][37]. The investigation of SEI formation with diffraction techniques is still challenging and only few examples were reported based on XRD [32][33][34] and neutron diffraction [35][36][37]. The use of in-situ fast XRD measurements allowed for the detection of the formation of LiF crystallites on the graphene surface and for revealing their texture [32]; the compositional analysis of the SEI layer was further confirmed by XPS analysis. XRD allowed also for the monitoring of the degradation of the cathode material and the SEI formation mechanism [34]. Neutron diffraction experiments were exploited for the analysis of commercial 18,650-type cell and the structural evolution of the electrodes and the SEI layer were reported [37].
With respect to X-rays, neutron diffraction offers higher sensitivity to detect lithium and other light elements, and thus can be preferred for the investigation of SEI layers. At the same time, neutron diffraction measurements require big facilities and longer acquisition times due to the lower intensity of the incident radiation beams. The information that can be obtained on the interphases is related to the identification of the constituent phases and their crystal/amorphous structures. Many of the components in the SEI and CEI layers are organic, amorphous, or small inorganic crystals, thus the analysis of the total diffuse radiation, and in particular the use of pair distribution function (PDF) can be also exploited to better characterize these species [38,39].

X-rays/Electron Absorption Spectroscopies
Beside diffraction, X-rays are a versatile probe due to the wide range of interactions with matter, including absorption and reflection. X-rays absorption spectroscopy (XAS), is indeed a powerful tool for the investigation of the atomic structure of the LIBs materials and constituents, leading to the identification of local coordination and oxidation numbers of the different atomic species. It offers good sensitivity and selectivity, can be used for the characterization of crystalline and amorphous components, and the acquisitions are generally fast enough to allow for operando studies under cycling on the time scale of the electrochemical processes [28,[40][41][42][43]. As for XRD, high-intensity sources are preferred for in-situ analysis, thus synchrotron light can be required together with dedicated cell assembly including special precautions (e.g., free-standing electrodes, inert atmosphere, use of specific electrodes [44,45]). As for the diffraction studies, also in this case the investigation of the interphases is still challenging but some pioneering works have been reported [43][44][45][46][47].
Similar to XAS, X-rays photoelectron spectroscopy (XPS), was widely exploited for the ex-situ characterization of electrode surfaces. In-situ XPS is now receiving particular attention for the characterization of interfaces due to the ability: (i) to detect almost all the elements with very low detection limits, and (ii) to perform a surface-sensitive analysis. Both laboratory and synchrotron radiation can be exploited for such analyses [48]. One of the major limitations is the requirement of ultrahigh vacuum, which is not compatible with the commonly used liquid electrolytes and with the non-conductive nature of the SEI layer. While ex-situ SEI characterization has been widely exploited and reviewed [48,49], in-situ investigations are still facing the challenges of cell design and proper experimental setup [50][51][52][53]. However, it is exploited to monitor the composition of the species formed during the electrochemical reaction at the electrode surface with no spatial resolution. A significant example of the combined application of XRD, XAS, and XPS measurements was reported by Deng et al. who were able to follow the cathode surface degradation mechanism and products in an all-solid state LIBs system [43]. In particular, XPS analysis was exploited for the identification of the chemical species and compounds constituting the CEI layer, formed upon oxidation of the cathode reacting with the solid electrolyte, demonstrating the ability of this technique to address such issues.

Reflectometry and Scattering
Apart from the chemical composition and structure of the constituents, also morphological features of the interphase such as roughness, density, and thickness, do play a crucial role in determining its behavior in the working device. Reflectometry is one of the best tools to address these aspects. It is generally used for the analysis of surfaces and thin films and can exploit, similarly to diffraction, different radiation sources (X-rays, neutrons). X-rays reflectometry (XRR) can monitor the surface in terms of structural information, density, thickness, roughness of a layer. This can be done during in-situ evolution of the layer and down to the atomic scale, thus XRR is really promising for the investigation of the electrodes/electrolyte interface/interphase. Again, up to now only few examples were reported in literature due to the complexity of the cell scheme and design [54,55]. Neutron reflectometry is similar to XRR as far as the working principles are considered, but offers higher penetration depth and higher resolution, thus it is particularly appealing for the SEI investigation under operando conditions [56][57][58]. Neutron reflectometry was exploited for the operando investigation of irreversible capacity losses correlated with SEI formation in a Si-based LIB [56]. Indeed, neutron reflectometry offers the possibility to obtain quantitative information on the Li concentration in the amorphous SEI layer. Neutrons are also exploited for the small angle neutron scattering (SANS) and inelastic neutron scattering (INS) techniques, providing information about the size, shape, volume of nanostructures of the interphase layers in the range of 5-200 nm [59][60][61][62][63][64].

Imaging and Microscopy
The morphology of the interfaces can be monitored by exploiting imaging techniques based on different radiation sources. X-rays imaging can be declined on several techniques, including tomography (XRT), and transmission microscopy (XTM), which offer spatial resolution up to~20 nm using synchrotron radiation, and can be used to collect 2D images, as well all to reconstruct 3D images, thus offering a unique insight into the visualization of interfaces/interphases [36,[65][66][67][68][69][70][71]. Neutron radiation can also be exploited for acquisition of 2D and 3D images, so allowing to monitor Li dynamics and distribution, gas evolution and SEI formation [72][73][74][75][76]. Neutron depth profiling (NDP) was used for the selective detection of 6 Li isotopes to monitor the lithium mass transport within the cell under working conditions [77][78][79]. This allows for the detection of both reversible and irreversible formation of interfacial layers and their evolution. At the same time, this analysis is more challenging due to: (i) low neutron flux (thus not allowing for fast time resolution), (ii) issues related to magnetic interactions and detection of magnetic structures, and (iii) strong background from hydrogenated species, such as polymer of the binder and separator and electrolyte components.
More traditional and less demanding microscopy techniques were also exploited for the investigation of interfaces/interphases. Traditional optical microscopy, OM, offers relatively low spatial resolution (~200 nm). Therefore, only microstructural features can be observed, whereas the formation and evolution of the nanometric SEI components is challenging. Generally, OM is used for the direct observation of dendrites and the follow up of the electrode evolution in terms of volume and stress cycling [80,81]. Higher spatial resolution (~10 nm) can be obtained with scanning electron microscopy, SEM. At the same time, high vacuum and electrically conductive samples are generally needed, which limit the possible applications for in-situ studies to electrodes and solid and semi-solid electrolytes. Dedicated cells and configurations were developed trying to overcome the major limitations [82,83] and pioneering works for the study of interface/interphase were already reported [82,84,85]. Transmission electron microscopy, TEM, suffers of similar and even greater problems related to sample preparation and cell configuration. Therefore, only a few setups have been developed to date [86][87][88]. Resolution at the atomic level and fast acquisition allow to follow the dynamics of the systems under study. Images and movies can be obtained, and impressive results were recently reported combining information coming from imaging and diffraction analysis [89][90][91][92][93][94][95]. Recently, other innovative techniques such as scanning electrochemical microscopy, SECM, and cryogenic electron microscopy, CEM, were proposed for in-situ SEI investigation [96,97]. The variety of experimental setups, possible techniques, and variety of information that can be obtained were reviewed, highlighting the versatility of TEM analysis [93]. In particular, TEM was widely exploited to detect the lithation of the electrode materials, to observe the volumetric expansion of the electrodes, to detect the growth of passivation layers/particles as it enables for direct visualization of these phenomena with nanometric resolution and, combining diffraction information, for the detection of the crystal phases.
Atomic force microscopy, AFM, was exploited as alternative to SEM and OM thanks to the higher resolution (sub-nanometer range), the possibility to obtain 3D images, and less limitations on the nature of the samples (no needs for conductive samples and full compatibility with liquids). AFM generates a tomographic image of the surface of the investigated sample. For these reasons it is particularly suitable for the study of SEI formation and evolution, particularly for the evaluation of the thickness and stability with time [98][99][100][101].

Other Spectroscopies and Electrochemistry-Based Approaches
The chemical composition and identification of the species constituting the SEI and CEI layers may be investigated via the traditional spectroscopic techniques such as FTIR, Raman or electron energy loss spectroscopy (EELS) [102]. FTIR allows for implementation of several modes (transmission, reflectance, reflection absorption, attenuated total reflection) for both ex-situ and in-situ investigations [103,104], and it is generally preferred to Raman due to strongest signals and fastest acquisition. Moreover, FTIR is highly sensitive to organic components, thus it can be exploited for the follow up of processes such as electrolyte decompositions, gas evolution, solvent intercalation, SEI formation, identification of the role of additives, changes in concentration of different species, coordination of different species near the electrode surface. Similar to FTIR, Raman was exploited using different cells schemes [105][106][107]. Raman is generally used for the study of the carbon component in terms of structure (crystalline vs. amorphous), defect levels, crystallites size, changes upon lithiation, etc. also allowing for spatial resolution (microRaman setups). EELS, which exploits the inelastic scattering of electrons (typically provided by TEM) from the sample, gives information about chemical composition and structure down to the atomic level.
Recently, nuclear magnetic resonance (NMR) was also exploited for the in-situ investigation of LIBs, thanks to the introduction of high-resolution solid-state techniques, sequences for the investigation of paramagnetic compounds, and (micro)imaging methods (magnetic resonance imaging or MRI) [38,39,[108][109][110][111]. NMR offers nuclear selectivity, and the different interactions (magnetic dipolar, electric quadrupolar, chemical shift anisotropy) can be exploited to better characterize the different components in the electrode and electrolyte. Recently, NMR has been exploited for the study of interfaces, thanks to dynamic nuclear polarization (DNP) and imaging techniques [112,113].
Other spectroscopic techniques, such as mass spectrometry, MS, [114], Mössbauer spectroscopy [115], and electron paramagnetic resonance, EPR, [116] have been exploited. For a systematic classification and discussion on the state-of-the-art techniques and available procedures for interface/interphase analysis it is possible to refer to dedicated reviews [117][118][119].

The Li/Electrolyte Interface
The intrinsic instability of the Li metal/electrolyte interface is the main obstacle towards the development of the lithium metal secondary battery (LMB) which is seen as one of the main breakthroughs in battery business. The implementation of a safe and reversible lithium platting/stripping process is carried on two development lines: (i) in the rocking chair configuration (lithiated cathode) it will be possible to develop a quasi-"lithium less" device, where the anode material is formed in-situ during the first charge, thus increasing the gravimetric energy, (ii) by using non-lithiated cathodes, overcoming the intrinsic limitation of intercalation chemistries and enabling the use of high capacity materials (S, O 2 ).
The presence of a SEI layer on the surface of metallic lithium is again a key point to improve electrolytes stability, since these latter work outside their thermodynamic electrochemical window. However, this scenario is completely different with respect to intercalation materials as graphite. The stripping/plating reaction takes place with a deep change of the electrode surface: fresh lithium is generated during plating and consumed during stripping. Thus, new SEI is formed during each plating cycle, with the corrosion of the metal active material, while the SEI fractures during stripping due to the generated mechanical strain. These processes have an opposite effect on the cell stability/performances: the continuous growth of SEI in plating increases the interfacial impedance and thus the electrode overpotential, so affecting the rate capability of the material. At the same time, the fragility of the SEI during stripping boosters for the formation of lithium fractal structures (dendrites) in the next plating cycle, because the SEI microfractures become the space zone for fast development of such dendrites [127], which can propagate fast through the electrolyte, eventually giving internal short circuit and dramatic cell failures. Moreover, during the successive stripping, dentrites can be detached from the metal layer originating "death" lithium particles in the electrolyte.
The successful implementation of LMB passes through SEI stabilization, and several different strategies have been proposed so far (Figure 2) [128]. A first approach is the use of electrolyte additives to improve the SEI chemical and mechanical stability. One of the most investigated additives is LiNO 3 , because nitrates decompose around 1.7 V vs. Li + /Li promoting the uniform growth of lithium during plating [129]. Moreover, the addition of polysulfide increases the homogeneity of the fresh metal layer [130]. Other investigated additives are FEC [131], AlI 3 [132], and Cs salts to promote the electrostatic shield mechanisms [133].
A second way to improve the interfacial properties of the Li/electrolyte interphase is the pretreatment of the metal to produce an artificial SEI layer before assembling the cell. From the interfacial point of view, there are two possibilities in this regard: to build the artificial layer onto the Cu current collector and to use it as a scaffold for the plating of lithium, or to make the SEI directly on the lithium foil. The former approach was used with porous carbonaceous materials such as carbon nanospheres [134], or 2D materials as boron nitride and graphene [135]. During plating, lithium ions can penetrate the layer and the lithium is deposited between the artificial SEI and the Cu foil. These SEIs show better mechanical and chemical properties compared to layers formed onto the lithium foil by electrolyte decomposition, increasing the Coulomb efficiency and the lifetime of the lithium anodes. Similar results were obtained with a series of protective layers of inorganic (LiF, Li 3 N, LiPO 4 , Al 2 O 3 ) and organic (Nafion TM , polyacrylic acid, polyacetylene) materials directly attached on the top of the lithium foil. Lithium-containing inorganic salts can be produced by the reaction of gases on the surface of metallic lithium (freon or F 2 for LiF [136,137], N 2 for Li 3 N), or by chemical reactions in solution (CuF 2 for LiF [138]). The deposition of oxides can be performed by atomic layer deposition (ALD) [139]. In the case of organic materials, different strategies can be used to realize the protective layer: it can be laminated on the top, as in the case of a few µm Nafion TM soaked with a commercial electrolyte [140], directed formed by the chemical reaction between the polyacrylic acid and Li [141] or by the in-situ polymerization to obtain conducting polymers (polyacetylene) [142].

Beyond Graphite: Alloys for Negative Electrodes
High-capacity novel negative electrodes for LIBS include materials such as conversions oxides (SnO2, SiOx) or metalloid elements (Si, Sn) able to make alloys with lithium. Despite of the obvious chemical differences between the two classes, their working principle in LIBs is similar, since MOn oxides are converted to M/Li2O composites during the Figure 2. Examples of the different approaches to improve the interfacial properties of the Li/electrolyte [138]. panel (a) taken from [134]; panel (b) taken from [130].

Beyond Graphite: Alloys for Negative Electrodes
High-capacity novel negative electrodes for LIB S include materials such as conversions oxides (SnO 2 , SiO x ) or metalloid elements (Si, Sn) able to make alloys with lithium. Despite of the obvious chemical differences between the two classes, their working principle in LIBs is similar, since MO n oxides are converted to M/Li 2 O composites during the first cathodic reaction and then the reversible alligation reaction takes place: xLi + + yM = Li x M y Thus, in both cases, during cell functioning we should take in consideration the SEI layers on the metalloid surface, also because the conversion reactions take place at potentials often higher than the electrolyte decomposition one. These SEIs have similar problems as in the LMB case, i.e., mechanical instability due to the large volume changes during alloying and de-alloying, making the mechanisms of SEI formation dynamic and thus affecting the Coulomb efficiency of the process. The most important example in this regard is the use of silicon, beside lithium the material with the highest theoretical capacity for Li ions storage, thanks to the possibility to make alloys up to Li 4.4 Si and observed specific capacity exceeding 3500 mAh g −1 .
Studies for the implementation of silicon-based negative electrodes initially focused on the preparation of nanostructures able to allocate volume changes during alloying and de-alloying reactions, while neglecting the chemical aspects at the surfaces. The result was the preparation of high specific capacity electrodes with low efficiencies (<99.5%) and therefore insufficient cycling capabilities [143], so the importance of the study and control of the interface has re-emerged. The earlier works on the nature of SEI layers on Si nanostructures revealed its chemical composition, made by LiF, Li 2 CO 3 , LiSiO x and organic compounds [144], as well as its dynamic nature, i.e., the SEI changes with the potential and the cycling [145]. One of the most promising ways to overcome this issue is to have a "non-filling" coating wrapping the porous silicon structures [146]. The non-filling coating is made by a carbon film able to form a stable SEI allowing the penetration of lithium ions in the ensemble of the porous nanoparticles, wherein they can expand and contract without mechanical damaging. The same approach was used in the case of micrometric pomegranate-shaped silicon particles covered by a thin layer of carbon [147]. With these systems, the life of silicon anodes was extended over 1000 cycles at high current density and a gravimetric capacity higher than 1000 mAh g −1 (Figure 3).

Towards High Voltage: The Emerging Role of the CEI
LIB positive electrodes work at potentials just below the oxidative decomposition potential of the electrolyte, thus the role of the CEI interphase between the cathode and the electrolyte is often undervalued. However, in the last years, it emerged as one of the improving fields to design high energy materials, as well as more sustainable processes in battery manufacturing. Indeed, the attempts to produce high-capacity positive materials are not just focused in increasing the capacity, which is limited by the intercalation reaction, but also in enabling the use of high potential phases, such as the high voltage, Ni-based spinels (LiNi 0.5 Mn 1.5 O 4 , LNMO) or olivines (LiCoPO 4 ) [148,149]. Moreover, the stability of Ni-rich materials (both layered oxides and spinels) in water is affected by corrosion with loss of the transition metal cation, a problem which is limiting the development of water-based procedure to formulate the electrode layers. Contrary to the graphite, indeed, the commercial processing of the cathode still requires the use of toxic 1-methyl-2pyrrolidone (NMP).
One of the most interesting compounds, the spinel LiNi 0.5 Mn 1.5 O 4 , has a plateau at 4.7 V vs. Li + /Li, which is above the decomposition limit and leads to the formation of CEI. Apart the stabilization of the bulk structure of the spinel, proper surface engineering can be used to control the interphase between the active material particles and the electrolyte [150]. Seminal works on the properties of the LiNi 0.5 Mn 1.5 O 4 /electrolyte interface highlighted that its stability depends on crystal orientation, cations surface distribution, and surface reconstruction upon cycling [151,152]. The two main strategies to improve the interfacial properties are: (i) the coating of the active particles, and (ii) the use of electrolyte additives (see Figure 4). In the former case, a core-shell approach is applied to avoid the dissolution of transition metal cations and to inhibit the reaction with the electrolyte [153]. The proposed shells are usually made by ZnO, TiO 2 , or Al 2 O 3 as nano-shells formed by chemical selfassembling [154] or ALD [155,156]. From the other part, a better design of the electrolytes can also enable the reversible operation of high voltage-spinel [157]. CEI-forming additives include: (i) phosphorous based compounds like (trimethylsilyl)phosphate [158], which exhibit higher HOMO energy and thus higher decomposition potentials, (ii) carbonyl molecules, like quercetin [159]; (iii) nitrile containing systems [160]), or different lithium salts and lithium borates [161,162]. Another possibility is to design electrolytes with high antioxidative ability, such as sulfone-based solutions [163], phosphate-based systems [164], fluorinated solvents [165], ionic liquids [166], and highly concentrated solutions [167].
Among the various cathode materials for future LIB generations, however, the most promising and investigated compounds in recent years have been the layered oxides of general composition LiNi x Mn y Co z O 2 (NMCXYZ), The general trend in the composition optimization towards higher energy and low cost is the increase of the Ni amount with respect to the more toxic Co, also because of the better electrochemical properties of the couple Ni 4+ /Ni 3+ compared to Co 4+ /Co 3+ . Thus, NMC evolution moved from NMC111 to NMC532 and NMC622 of the last commercial generation, aiming towards NMC811 which should be the cathode of the next LIBs. However, this comes with the increase of both the thermal instability and the solubility of the phase, affecting the capacity retention and inhibiting the processing in aqueous solution [168]. An intriguing solution to this problem is the active material surface control, with the preparation of microparticles of active material with a concentration gradient, which has an excess of Mn and Ni at the surface and in the bulk, respectively [169]. Despite of the presence of 80% on Ni in the phase, electrodes based on this kind of NMC811 showed better cyclability compared to the more stable bulk phase NMA [170]. thermal instability and the solubility of the phase, affecting the capacity retention and inhibiting the processing in aqueous solution [168]. An intriguing solution to this problem is the active material surface control, with the preparation of microparticles of active material with a concentration gradient, which has an excess of Mn and Ni at the surface and in the bulk, respectively [169]. Despite of the presence of 80% on Ni in the phase, electrodes based on this kind of NMC811 showed better cyclability compared to the more stable bulk phase NMA [170]. Finally, we mention that, in the case of CEI, the large molar amount of carbon additives used in electrode formulation has an important role in the decomposition of the electrolytes at high voltage, due to their nanometric size and good electrocatalytic activity, which is at least comparable with that of active oxides [171], without considering that carbon black with graphitic domains can intercalate anions at high voltage.

Conclusions
Extended interphases play an increasingly important role in determining the functional properties of electrochemical storage devices, especially regarding the long-term performance stability. While in the past decade the attention was focused on the graphite anode (SEI), the most recent technological developments place at the center of attention the interphase between the lithium metal and the electrolyte (liquid or solid), and CEI. Indeed, the advances in instrumentation setups and cell design allow for the implementation of in-situ and operando approaches of increasing complexity. This will help to understand the reaction mechanisms and to develop more sophisticated approaches for interphases design and control. At the same time, due to the complexity of the considered systems, SEI and CEI, it is not possible to identify a single technique of choice for the (a) (b) (c) Figure 4. Examples of the different approaches to improve the properties of the positive electrode/electrolyte interface in LIBs. Panel (a) taken from. [154]; panel (b) taken from ref. [169]; panel (c) taken from [160].
Finally, we mention that, in the case of CEI, the large molar amount of carbon additives used in electrode formulation has an important role in the decomposition of the electrolytes at high voltage, due to their nanometric size and good electrocatalytic activity, which is at least comparable with that of active oxides [171], without considering that carbon black with graphitic domains can intercalate anions at high voltage.

Conclusions
Extended interphases play an increasingly important role in determining the functional properties of electrochemical storage devices, especially regarding the long-term performance stability. While in the past decade the attention was focused on the graphite anode (SEI), the most recent technological developments place at the center of attention the interphase between the lithium metal and the electrolyte (liquid or solid), and CEI. Indeed, the advances in instrumentation setups and cell design allow for the implementation of in-situ and operando approaches of increasing complexity. This will help to understand the reaction mechanisms and to develop more sophisticated approaches for interphases design and control. At the same time, due to the complexity of the considered systems, SEI and CEI, it is not possible to identify a single technique of choice for the investigation of the battery interphases and multi-technique approaches should be preferred.
Author Contributions: C.F. developed the section on characterization techniques; R.R. developed the section of battery interphases; P.M. followed introduction and conclusions, and the overall organization of the paper. All authors have read and agreed to the published version of the manuscript.

Conflicts of Interest:
The authors declare no conflict of interest.