Electrodeposition of adherent submicron to micron thick manganese dioxide films with optimized current collector interface for 3D Li-ion electrodes

Three-dimensional (3D) configuration of high-performance energy storage devices has been the subject of ongoing investigations targeting their integration in autonomous microelectronic systems. In this study we demonstrate a route toward the realization of high capacity cathode material for 3D thin-film lithium-ion (Li-ion) batteries. Electrolytic manganese dioxide (EMD) film can be applied as a Li-ion intercalation electrode upon its conversion to lithium manganese dioxide (LiMn2O4 or LMO) by solid-state reaction. The main challenges of depositing thicker EMD film directly on the current collector often lay in achieving a good film adhesion and preventing oxidation of non-noble current collectors such as TiN, Ni. To improve the adhesion of the EMD films we modify the surface of the current collector by means of thin-film or seed layer coatings, which also prevent the oxidation of the underlying current collector substrate during the anodic deposition process. As a result submicron to micron thick EMD films with good adhesion were deposited on various current collectors. The acidity of the electrolyte solutions was varied depending on the type of the surface coating or current collector used. The mechanism of the EMD film growth and morphology on different substrates was examined. Compatibility of the proposed current collector interface modification for the electrodeposition of conformal thick EMD films on high-aspect ratio microstructures was demonstrated. A method of EMD film conversion to LMO at low-temperature on different substrates was shown as the path toward their application in 3D Li-ion batteries.

Emerging electronic microsystems such as autonomous sensor systems for smart environments, body area networks, biomedical devices for lab-on-chip diagnosis and implants require onboard microstorage. For such applications the development of powerful, small and intrinsically safe batteries with sufficient storage capacity and long lifetime is critical. The key enabler to address the energy and power demands of footprint-limited autonomous devices is a three-dimensional (3D) configuration of energy storage devices such as lithium-ion (Li-ion) batteries and supercapacitors, where thin-film material is coated on a microstructured current collector. [1][2][3] The surface area enhancement associated with 3D structuring not only provides more material for the same footprint, but also increases the interface between the active electrode material and the electrolyte, thus reducing the cell resistance and providing more surface for reversible insertion and extraction of Li ions. 4 As opposed to nanostructured electrode configuration, 3D microstructuring of energy storage devices allows to increase the thickness of active layers for improved areal capacitance. However, conformal deposition of films with good adhesion on industrially-relevant current collectors is an essential requirement and still remains a challenge.
Manganese dioxide (MnO 2 ) has been widely explored for energy storage systems as a potential cathode material in rechargeable batteries and supercapacitors owing to its lower cost, environmental benignity and electrochemical properties. Electrodeposition is the primary method to prepare MnO 2 typically in the γ-MnO 2 form, which is commonly used today in primary alkaline and lithium batteries. 5 This phase of MnO 2 is referred to as electrolytic manganese dioxide or EMD. Typically, EMD is fabricated by anodic electrodeposition of several micrometer thick layers on titanium anodes from hot (80-98 • C) acidic electrolytes, followed by its mechanical removal from the anode and grounding it into powders to be used as electrodes. 6,7 Electrodeposition in one of the preferred techniques for MnO 2 preparation due to its processing simplicity, scalability and possibility to obtain conformal EMD deposition on high aspect ratio substrates. Therefore, electrodeposition of MnO 2 has also been explored for 3D thin-film storage devices, being electric double-layer capacitors or Li-ion batteries. In this work, thick MnO 2 films are deposited directly onto the surface of the current collector. The thickness of such films ranges from 250 nm to >1 μm for the optimized current collector interface. For 3D battery applications, the thickness of the EMD film on microstructures with an area enhancement of 20-40x (see Figure S1) should be close to 500 nm in order to match the capacity requirements of commercial Li-ion batteries for high-power applications (60-120 mAh/cm 3 ). 8,9 At the same time, the increase of the film thickness should not compromise the electrochemical performance of the material. 10 The main challenges of depositing thicker EMD film directly on the current collector often lay in achieving a good film adhesion, which can be challenged by the dissolution or passivation of the current collector in an acidic environment during the anodic EMD deposition. This typically results in delamination of the EMD film deposits or limitations in their electrochemical activity due to the oxidation of the underlying current collector. Various substrates have been used for the electrodeposition of MnO 2 material for their use in electrochemical capacitors or batteries. [11][12][13][14] The importance of substrate variation for the EMD deposition mechanism has been highlighted but not given much attention. 11 In this work we propose ways to obtain thicker and well-adherent EMD films on technologically-relevant but oxidizable materials such as titanium nitride (TiN) and nickel (Ni) which are often used as current collectors in energy storage applications. TiN is a common material used as a diffusion barrier on silicon. For thin film Li-ion batteries, in particular, TiN was shown to be a good Li-ion diffusion barrier and can also be used directly as an electrode material. 15 Therefore, TiN is the focus support material in this work, but the findings are relevant for other oxidizable substrates. We explore different substrate coatings including carbon nanostructures, ultra-thin MnO 2 seed layer and nickel (Ni) film for the deposition of thicker EMD films. The latter can also be used as a 3D current collector on its own. The physical properties of the deposited films and the deposition mechanism associated with the substrate and electrolyte acidity variation is discussed. The applicability of as-deposited EMD films as Li-ion intercalation electrodes in thin-film batteries is demonstrated by converting EMD to lithium manganese dioxide (LiMn 2 O 4 or LMO) films with higher capacity. Finally, conformal deposition of thick EMD films on 3D microstructures is shown paving a way toward their application in 3D energy storage devices.
Experimental EMD film deposition.-EMD films were anodically electrodeposited using a three electrode cell configuration with Pt mesh counter electrode and Ag/AgCl/3M NaCl reference electrode (BASi analytical, 0.22 V vs SHE). An aqueous solution of 0.3M MnSO 4 · H 2 O (98.0-101.0%, Alfa easer) and 0.3M H 2 SO 4 (96%, OMG) was used for the deposition of manganese dioxide on most of the studied substrates at room temperature (20-25 • C). For comparative electrochemical measurements also an indifferent electrolyte solution of 0.3M Na 2 SO 4 (99.0%, Sigma Aldrich) and 0.3M H 2 SO 4 was applied. EMD on Ni-coated TiN layer was prepared from an aqueous solution of 0.5M MnSO 4 · H 2 O (98.0-101.0%, Alfa easer) and 0.5M CH 3 COONa (ReagentPlus, 99.0% Sigma Aldrich) with an addition of ∼10 wt% ethanol for wetting improvement. 16 All the depositions were carried out galvanostatically at a constant current density of 0.5 mA/cm 2 for different time periods, obtaining variable film thickness. For the deposition of EMD on Ni using acetate based electrolyte an additional nucleation step of high current density of 2.5 mA/cm 2 for 1-8 sec was added prior to the film growth at 0.5 mA/cm 2 . In the present study we focus on film deposition carried out at room temperature. After the deposition EMD samples were annealed at 350 • C in N 2 atmosphere (ambient pressure) to remove the water content within the film structure. 5 Substrate materials.-Planar silicon wafers (n type) coated with a 60 nm TiN layer were used for most of the results presented here. TiN layer was sputtered from a Ti target in a N 2 atmosphere (Applied Endura Extensa TTN). Graphitic carbon nanosheet (CNS) layers (20-70 nm) were used as one of the TiN coating materials, further referred to as C/TiN substrate. CNS layers were grown in a capacitively coupled PECVD reactor with a 13.56 MHz RF generator (Oxford Instruments plasma technology UK NANOCVD) introducing C 2 H 2 -H 2 (1:10) for 2-3 minutes at 800 • C, 0.45 Torr and RF power of 300 W 9 . Atomic layer deposition (ALD) was employed for the growth of ultrathin MnO 2 seed layers (2.2-4.6 nm) on top of the TiN current collector. Tris(2,2,6,6-tetramethyl-3,5heptanedionato)manganese (Mn(thd) 3 ) was used as a manganese precursor, and ozone was used as an oxygen source. 17,18 ALD films were grown in a home-built stainless steel high-vacuum ALD reactor with a base pressure of 10 −7 mbar at a substrate temperature of 180 • C, which was chosen to be well within the ALD window. 19 The number of ALD cycles was optimized to the required thickness. A layer of Ni (100 nm) was deposited on TiN from a Ni target in a PVD advanced low pressure source chamber (Applied Endura Extensa TTN). For 3D silicon microstructures conformal TiN coating was deposited by means of ALD (thermal process at 390 • C using TiCl 4 as precursor) and Ni was electroplated (galvanostatic deposition at a fixed effective current density of −10 mA/cm 2 , using 0.62 M Ni sulfamate with H 3 BO 3 ).
Material analysis.-Different techniques have been applied in this work for material analysis. The morphology and thickness of as-deposited EMD films was characterized by scanning electron microscopy (SEM) using NOVA 200 microscope for standard measurements (up to 100,000x magnification) and a Hitachi SU8000 for higher resolution measurements (up to 300,000x magnification). Rutherford Back Scattering (RBS) measurements were carried out using He + beam of 1.523 MeV, with scattering angle 170 • , and sample tilt angle of 11 • . Ex-situ X-ray diffraction (XRD) with X'pert Panalytical using Cu Kα radiation (8.04 keV) in grazing incidence configuration was applied for the analysis of thin film crystallinity. In-situ XRD was performed in a Bruker D8 Discovered using Cu Kα radiation and a linear detector. The film crystallinity was monitored in-situ, measuring the samples during annealing in a home-built annealing chamber with con-trolled ambient and temperatures. 20 X-ray photoelectron spectroscopy (XPS) was performed for elemental composition measurements using a monochromatized Al Kα X-ray source (1486.6 eV) and a spot size of 400 μm. For the chemical composition analysis, measurements were carried out in Angle Resolved mode using a Theta300 system from Thermo Instruments. 16 spectra were recorded at exit angles between 22 • and 78 • as measured from the normal of the sample.
Electrochemical characterization.-Electrochemical characterization of as-deposited EMD films and EMD films converted to LMO was performed in a three-electrode Teflon cell setup with Li metal ribbons as reference and counter electrodes. The reference electrode compartment was connected to the main cell compartment through a Luggin capillary (∼1 mm above surface). The Li-ion electrolyte solution consisted of 1M LiClO 4 (99.99%, Sigma Aldrich) dissolved in propylene carbonate (99.7%, Sigma Aldrich). During characterization the electrochemical cell was placed inside an argon filled glove box (Pure lab, PL-HE-4GB-1800, <1 ppm O 2 and H 2 O level). An AUTOLAB potentiostat PGSTAT204 controlled by Nova 1.9 software (Metrohm) was utilized for these measurements.

Results and Discussion
Improved adhesion of thick EMD on TiN with C and MnO 2 surface coating.-Electrodeposition of MnO 2 from acidic electrolytes, typically being aqueous baths of manganese sulfate (MnSO 4 ) and sulfuric acid (H 2 SO 4 ), has been widely reported. 10 In our previous work, EMD films were electrodeposited from acidic MnSO 4 baths on planar platinum (Pt) coated substrates. 5 EMD films grown from acidic electrolytes on different substrates were shown to be highly porous (∼50% porosity) consisting of a network of interconnected 5-10 nm nanoparticles. 5,9 However, a very poor adhesion of the EMD films electrodeposited on the noble metal (Pt) coated substrate was observed. The maximum thickness of well-adherent EMD films was limited to 150 nm, with thicker deposits delaminating due to the mechanical film stress. 5 Additionally Pt-group metals do not comply with cost-effective application targets. In that respect, electrically conductive non-noble TiN is a suitable current collector which, at the same time, is a good Li-ion diffusion barrier. 15 However, the application of EMD films directly on TiN is limited, since TiN oxidation occurs in approximately the same potential range as the anodic electrodeposition of EMD. A linear sweep voltammogram of MnO 2 electrodeposition from an acidic (solid red) and an indifferent (dashed blue) electrolyte solutions using a TiN substrate is shown in Figure 1a. The presence of an anodic current peak at +1.6 V in an indifferent electrolyte solution indicates the electro-oxidation of TiN to TiON x , which forms a passivating layer of ∼1.5 nm (as estimated from the charge under the current peak) and inhibits the oxygen evolution. 9 In the reverse scan the current becomes nearly zero and no cathodic current is observed due to the formation of an electronically blocking interface between the current collector and the EMD film. Interestingly, this interfacial oxide does not fully block the anodic oxidation reaction as even thicker EMD deposits could be achieved for longer deposition times. An example of 380 nm thick EMD film on TiN is shown in Figure S2a. However, these thicker films deemed inactive with no typical Li-ion intercalation peaks observed (see Figure S2b). In fact, on TiN substrate EMD films of only less than 50 nm thick were found to be electrochemically active as shown in our earlier studies. 5 This indicates that with the extended oxidation of TiN, the interfacial oxide starts blocking the cathodic Li-ion intercalation reaction, whereas for short deposition times the oxidized TiN surface still allowed sufficient electronic contact between TiN ad MnO 2 . This seems counter intuitive at first as TiO 2 is n-type semiconductor and one would expect the anodic reaction to be blocked instead. However, this can be explained by the blocking of electrons with the formation of a p-n junction at the TiON x /MnO 2 interface. This phenomenon was confirmed by plating thicker EMD films on thin ALD TiO 2 coated TiN films (an example of 250 nm thick EMD film is shown in Figure S3a), which exhibited no electrochemical performance ( Figure S3b).
In order to circumvent the issue of the substrate oxidation, we have proposed to coat the TiN substrate with a layer of short carbon nanosheets (CNS) with a sheet thickness ranging from a few to tens of nanometers 9 ( Figure 2). We believe that the application of such nano-rough graphitic carbon improves the EMD adhesion due to the formation of covalent-like bonds between MnO 2 particles and surface groups of defective carbon (C-O bonds). Well-adherent EMD films exceeding 1 μm were electrodeposited on top of carbon (C)-coated TiN substrates (Figure 2b). However, we recognize that for certain applications where elevated temperatures are required, the use of a carbon coating might impose certain temperature limitations on further stack processing. Therefore, we suggest another approach to avoid the oxidation of the substrate during the electrochemical deposition. It involves coating an oxidizable substrate such as TiN with a very thin MnO 2 seed layer by e.g. ALD. It was observed that the presence of the so-called MnO 2 seed layer improves the mechanical adhesion of the EMD film due to the formation of native covalent bonds between Mn-O and Mn-O. Thick EMD films up to ∼1 μm were prepared on top of a planar substrates coated with 2-4 nm MnO 2 seed layer ( Figure 3). The ability of such ultrathin MnO 2 seed layer to protect TiN from oxidation was confirmed by in-situ XRD analysis ( Figure S4). It is interesting to note that oxygen evolution reaction occurs somewhat earlier at ALD MnO 2 -coated substrate (Figure 1c) where the MnO 2 seed layer is playing an important catalytic role which is a subject of our separate study. 21 The electrochemical behavior in a Mn-containing electrolyte solution is very similar for both C/TiN and ALD MnO 2 /TiN substrates. Clear anodic current peaks of 5.7 mA/cm 2 at 1.36 V vs Ag/AgCl for the C/TiN substrate and 4.1 mA/cm 2 at 1.25 V vs Ag/AgCl for the MnO 2 /TiN substrate are distinguishable. These peaks are associated with the formation of an EMD film, which inhibits further deposition kinetics due to nearly insulating nature of the film. However, since the EMD film is porous, growth of the film still continues on the substrate due to the steady-state diffusion of the reagent through the pores of the film. That is well visible in the reverse scan plateau of Figures 1b-1c. For the uncoated TiN substrate a continuous oxidation of TiN is occurring simultaneously with the growth of EMD on top, as visible from Figure 4a from the constantly growing potential with the plating time. Stable deposition potential is observed for the EMD growth on top of the coated TiN substrates (Figures 4a, 4b).
Mechanism of EMD deposition.-A number of fundamental studies have focused on understanding the EMD deposition mechanism in detail, typically on a Pt substrate. 10,11,22,23 The general mechanism is  agreed to go via the electrochemical oxidation of the solvated Mn 2+ forming soluble Mn 3+ intermediate. The electrolyte acidity influences the Mn 3+ stability, which can further either diffuse out from the electrode surface to the bulk of the solution (initially) or precipitate as a passivating MnOOH layer. In the latter case, oxidation of MnOOH to form MnO 2 needs to happen to open up a site for the reaction to continue, which was identified as the limiting factor in the growth process. 11 The MnO 2 formation mechanism is as follows: MnOO H(s) → MnO 2 (s) + H + + e − [3] This deposition pathway is characteristic for the case when the acid concentration in the electrolyte solution is lower and Mn 3+ (aq) ions have lower stability. 11 The free Mn 3+ ions can disproportionate in the solution according to: 2Mn 3+ (aq) + 2H 2 O ↔ Mn 2+ (aq) + MnO 2 (s) [ 4 ] which will be observed mostly under acidic conditions as under these conditions the formation of free Mn 3+ is more prevalent according to Reaction 2. The EMD film growth occurs predominantly at the substrate interface with the diffusion of ionic species through the porous EMD film, as schematically demonstrated in Figure 5. MnO 2 is a poor electronic conductor and would quickly passivate the electrode if it was formed as a dense film. However, the deposit is nanoporous and the electrode reaction can proceed near the original electrode surface with slower ionic transport through the pores. Transport includes Mn 2+ (aq) travelling toward the current collector surface, where reaction occurs, and hydrated protons diffusing out of the nanoporous EMD structure ( Figure 5). This bottom-up EMD growth mechanism enables the formation of thicker EMD films on poorly conducting or passivated current collector materials, like the earlier discussed oxidized TiN ( Figure S2a), ALD TiO 2 coated TiN substrate ( Figure S3a) as well as on ALD-grown MnO coated TiN substrate ( Figure S5). With the increase of the EMD film thickness, a trend of porosity decrease was noticed, pointing toward film densification with increasing thickness. 5 This is attributed to the disproportionation of Mn 3+ inside the pores as described in Equation 4, which leads to the densification of the EMD film over time (therefore, an increase in the deposition efficiency).
The morphology of the as-deposited EMD films is largely defined by the dominant growth mechanism which is directly related to the electrode surface composition and its catalytic activity toward Mn 2+ oxidation. Due to the high catalytic activity of Pt, Mn 2+ rapidly oxidizes forming an excess of Mn 3+ at the Pt electrode interface at a short time and at low overpotentials. 23 Partial chemical dissolution of Mn 3+ both lowers the deposition efficiency and induces the interface porosity. This link between the increased deposition efficiency and concurrent decrease of porosity as the EMD film thickness on Pt increases was experimentally observed in our earlier work. 5 High film porosity at the interface during the earlier deposition stage could lead to poor adhesion and observed delamination of EMD from a Pt substrate with the film thickness increase. 5 High EMD film porosity and delamination issues were also observed for the nickel (Ni) substrate using an acid-based electrolyte, as addressed in the next section. Figure 6a shows the deposition efficiency with respect to the  charge density for the EMD deposition at the same current density on different substrates.
The deposition efficiency was estimated as the ratio of the equivalent MnO 2 film thickness determined from RBS (calculated from the quantity of Mn atoms assuming a dense MnO 2 film where the MnO 2 density is equal to 5.02 g/cm 3 ) to the theoretically expected MnO 2 film thickness derived from the deposition charge (using Faraday's law). The presence of the initial MnO 2 dense layer for seed ALD MnO 2coated TiN substrates is taken into account in calculations. Lower deposition efficiency is related to the formation of soluble Mn 3+ intermediates. As the EMD film grows, the deposition efficiency is observed to slightly increase which can be attributed to the increase of the ratio of Mn(IV) to Mn(III). Interestingly, the deposition efficiencies of Pt coated TiN substrates were consistently about 10% lower than that for C-coated and ALD MnO 2 -coated TiN substrates. This indicates a larger partial charge transfer for the formation of Mn 3+ and MnOOH on a Pt substrate (Pt > C > MnO 2 ). Subsequent structural changes during the solid-state oxidation of MnOOH to MnO 2 could have also contributed to building up stress in the EMD material and causing the observed delamination from a Pt substrate. 23 Deposition efficiency for the MnO 2 seed layer-coated substrate was found to be the largest of all the substrates. Note that the estimated deposition efficiency appears to be overestimated for all the EMD films due to the co-existence of Mn 2 O 3 or MnOOH next to MnO 2 within the as-prepare EMD films, as confirmed by XRD analysis ( Figure S6). That is apparent from the deposition efficiency for the thinnest EMD film on ALD MnO 2 -coated substrate which exceeds the physically possible limit.
Since low deposition efficiencies were found to be linked to high EMD porosity, we further examine the porosity of EMD films on Pt/TiN, C/TiN and MnO 2 /TiN substrates. Porosity of the films was estimated by comparing the equivalent RBS thickness of the MnO 2 films and the actual film thickness measured by cross-sectional SEM, 5 i.e.

Porosit y = 100% · (1 − R B S f ilm thickness/S E M f ilm thickness)
[5] The electrodeposited MnO 2 films were found to have similar porosity value of ∼50% on average for Pt/TiN, C/TiN and MnO 2 /TiN substrates ( Figure S7). This confirms that the surface chemistry and surface morphology are at least partially responsible for the difference in adhesion between the electrodeposited MnO 2 film and the substrate. Figure 6b compares the maximum achieved thickness of EMD films on different substrates. The thickest EMD films were deposited on Ccoated and seeded TiN substrates where the film deposition efficiency was the largest. This highlights the important role of the surface coating for the initial stages of EMD nucleation and growth. In this case MnO 2 is likely to form through the oxidation of MnOOH without significant Mn 3+ desorption. 11 The presence of an ultrathin MnO 2 layer favors the formation of crystallite growth on top of the already deposited MnO 2 from the initial stages of the EMD deposition. This is confirmed by observing the potential change during the first seconds of EMD nucleation, where the smallest nucleation peak was observed for MnO 2 -coated TiN (1.19V during the first 0.4 s of the deposition) as opposed to the uncoated TiN substrates exhibiting the largest nucleation peak (1.33V after 1.6 s of the deposition) as shown in Figure 4b.    (Ni) is an alternative current collector material which is stable in air owing to the surface-limited self-passivation and can be simply and controllably electrodeposited on high-aspect ratio structures for 3D device applications. 16 Ni is used as a current collector material itself and in this work it is a coating layer on top of TiN. Figure 7a shows cyclic voltammograms (CVs) in a sulfuric acid based MnSO 4 and indifferent electrolyte solutions. The presence of a small cathodic stripping peak confirms a small partial current for MnO 2 deposition. However, simultaneously to EMD deposition, partial Ni oxidation and concurrent slow dissolution of the NiO occurs in this process. Note that the surface of the Ni layer is already self-passivated at the start of the electrodeposition process. Indeed, if the potential is brought to −1 V vs. Ag/AgCl, the passivating NiO gets reduced and upon a positive scan, a strong anodic Ni dissolution current can be observed starting at −0.2 V vs. Ag/AgCl, as shown for example in the 2 nd CV scan of Figure S9. The deposition of EMD films on Ni using acid-based electrolyte is possible at room temperature at very low current densities (∼0.5 mA/cm 2 ) when the NiO dissolution is inhibited by the growing EMD layer, but the thickness of this layer is limited to <100 nm (Figure 7b). The low deposition efficiency vs charge shown in Figure  6a for Ni-coated TiN substrate demonstrates that the partial current of parasitic Ni oxidation is significant, leading to a quick delamination of the EMD film grown on top. The as-grown EMD films were also found to be highly porous ( Figure S7) which is partially responsible for their delamination together with the dissolution of NiO.

Thick EMD films on Ni-coated electrode.-Nickel
To overcome these problems, a neutral aqueous electrolyte of 0.5M MnSO 4 and 0.5M CH 3 COONa with 10 wt% ethanol was used. 16 CVs were recorded for Ni in Mn-containing and indifferent electrolytes (Figure 8a). No significant oxidation peaks for Ni in an indifferent electrolyte were present, while the anodic oxidation peaks of Mn 2+ to MnO 2 and cathodic reduction peaks of MnO 2 to MnOOH in a Mn-containing electrolyte are clear. Multiple anodic peaks are due to oscillations as a result of cyclic pH change related to buffering effect of acetate ions. Well-adherent EMD films up to 500 nm were deposited on a Ni substrate as shown in Figure 8b. Thicker deposits (>500 nm) could also be deposited but were observed to delaminate due to stress upon the subsequent EMD annealing step.
The morphology of the EMD film on Ni as a function of the deposition time was studied. Initially, a more open fibrous or sheet-like EMD morphology is obtained from the neutral bath which gradually densified with plating time. The densification is also apparent from the porosity of the films plated for different times, determined from RBS and SEM thickness (Figure 9a). The sheet like structure of the EMD film is shown in Figure 9b. Densification with plating time goes in line with the earlier discussed bottom-up EMD growth  mechanism. The MnO 2 electrodeposition reaction produces 4 H + cations per MnO 2 molecule. Similarly to the slow diffusion of Mn 2+ from the bulk of the solution to the electrode surface through the nanoporous MnO 2 network, also the H + diffusion out of the MnO 2 layer is slowed down. As a result, the acidity inside the porous MnO 2 increases as the layer thickness increases. The increased acidity affects the electrodeposition mechanism and the partial current for Mn 3+ (Reaction 4) will increase. The free Mn 3+ (aq) at its turn has to diffuse out of the layer and the disproportionation into MnO 2 and Mn 2+ (aq) (Reaction 4) increases with the film thickness. This Mn 3+ (aq) disproportionation inside the pores results in the densification of the EMD film. The gradual densification is also supported by the variation of the film porosity which decreases with the increase of the plating time (Figure 9a).
The variation of deposition efficiency with deposition charge follows a similar trend as observed earlier for the EMD deposition using acid electrolyte. The calculated efficiency values of 100% or larger are attributed to the presence of other Mn oxidation states in the deposit. Indeed, XPS analysis of the EMD films on Ni/TiN could not exclude confirmed the presence of MnOOH or Mn 2 O 3 or combination, since their peaks are overlapping in shape and position ( Figure S8).

Conformal coating of thick MnO 2 films.-For electronically conductive coatings achieving conformal depositions is typically
a challenge due to ion depletion in deep features, as for example demonstrated. in copper through-silicon vias technology. 24,25 In case of MnO 2 electrodeposition, the self-limiting nature of the oxide growth provides excellent conformality. The 3D-compatible deposition of the ALD seed MnO 2 layer and Ni film electroplating enable conformal thick EMD coating on high-aspect ratio structures. Figure 10 shows thick conformal EMD deposition on top of high aspect ratio micro-pillars (50 μm high, 2 μm in diameter with 2 μm interpillar spacing) with a 22 nm TiN deposited by ALD and coated by either seed MnO 2 layer with ALD ( Figure 10a) or thin Ni layer by electroplating (Figure 10b). Similar deposition was also achieved with C-coated Si pillars as demonstrated in our earlier work. 26 Ability to conformally coat high-aspect ratio structures opens up a variety of new applications of EMD material, including Li-ion batteries with increased battery capacity due to surface microstructuring.

Application of thick EMD films for Li-ion intercalation
electrodes.-In principle, MnO 2 is by itself suitable as an electrode material in Li-ion batteries. EMD is indeed used in commercial primary lithium batteries, and can be made reversible for thin films. 5 Cyclic voltammetry of activated EMD thin films shows distinctive cathodic current peak for Mn(IV) reduction and Li-ion intercalation at 2.75 V vs. Li + /Li and an anodic current peak for the re-oxidation  MnO 2 + x Li + + xe ↔ Li x MnO 2 , with 0 < x < 1. [6] Activation of the EMD films can be easily done by cycling in the Li-ion electrolyte solution. The improvement in Li + intercalation and extraction kinetics of the EMD films upon cycling is attributed to several factors, e.g. changing of the MnO 2 crystal structure with cycling, increased contact between MnO 2 and the electrolyte, reduction of the remaining water content within the EMD structure, doping of the film by Li-ions and hence increased electronic conductivity. 5 This was shown for example for the 70 nm thick EMD films grown on C/TiN substrate. 9 Unfortunately, only thin (<50 nm) EMD films show good reversible kinetics with sharp Li-ion intercalation and extraction peaks. As opposed to thin films, thicker EMD films display only broad cathodic (at ∼3-3.4 V vs. Li + /Li) and anodic (at ∼2.4-2.6 V vs. Li + /Li) current peaks in their CVs as shown in Figure 11a for EMD films deposited on the different substrates. Such poor electrochemical response with no distinctive Li-ion intercalation peaks but mostly a capacitive behavior, results from the high electronic and ionic resistances of the EMD films which increase with the film thickness, as discussed in more details in our earlier work. 9 We have thus developed a process to convert electrochemically deposited MnO 2 layer to LiMn 2 O 4 (LMO) by solid-state reaction, allowing the application of thick EMD films for Li-ion intercalation electrodes. 27 This is achieved by spin-coating a Li 2 CO 3 layer on top of the as-prepared porous EMD film. The solid-state reaction occurs upon annealing with the thermal budget and the amount of MnO 2 and Li 2 CO 3 materials optimized in order to fabricate a stoichiometric LiMn 2 O 4 film according to the following reaction: The MnO 2 conversion to LiMn 2 O 4 using Li 2 CO 3 was first demonstrated using a Pt current collector which allowed such high annealing temperatures as 750 • C. In this case thin EMD films of 50-70 nm were used due to the limitation to deposit well-adherent thicker EMD films on Pt. The CV of as-fabricated LMO film shows the typical Li-ion insertion and extraction peaks in/from LiMn 2 O 4 at around 3 and 4 V vs. Li + /Li, as demonstrated in Figure 11b. The electrode reactions are described as follow: The capacity of the converted 50 nm film calculated from the CV in the voltage range of 2.4-4.4 V vs. Li + /Li was found to be equal to 75% of the LMO's theoretical capacity (theoretical capacity = 1.26 Ah/cm 3 ). The cyclic voltammetry measurement was done at 10 mV/s, thus each scan took 200 s which is equivalent to C-rates over 18 C. Additionally, the formation of a crystalline spinel LiMn 2 O 4 material was confirmed by XRD analysis where LiMn 2 O 4 diffraction peaks were observed together with the substrate peaks ( Figure 12).
Next, the solid-state conversion to LMO was carried out with thick EMD films on the C/TiN, ALD MnO 2 /TiN and Ni/TiN current collector substrates. The annealing temperature was reduced from 750 • C to 350 • C to prevent oxidation of the underlying materials by oxygen or MnO 2 . For example, in the case of C-coated TiN substrate lowering the applied thermal budget was important to prevent MnO 2 reduction by the carbon at higher temperatures, rendering the material inactive. The annealing time was varied in accordance with the thickness of the EMD film. Thus, solid-state conversion reaction was carried out for  10 min at 350 • C in air for a 95 nm starting EMD film on C/TiN, while for a 250 nm EMD film on Ni/TiN the annealing time was increased to 120 min in air. In case of converting EMD on ALD MnO 2 /TiN current collector, the annealing temperature of the Li 2 CO 3 /EMD/MnO 2 /TiN stack was further reduced to 250 • C in order to prevent reaction between MnO 2 and TiN with TiN oxidation and reduction of the MnO 2 seed layer. In this specific case, the annealing time was increased to 6 hrs in order to allow enough time for the solid-state reaction with the 250 nm EMD film to occur. Table I summarizes the test conditions of the conversion process on different current collector substrates. The CVs of as-converted EMD films show the Li-ion insertion/extraction peaks in/from LMO at around 3 and 4 V vs. Li + /Li, demonstrating a successful conversion on all the substrates (see Figure  11b). The formation of crystalline LMO material was confirmed by XRD ( Figure 12). The ERD results of all the samples after annealing indicate that the fabricated LMO film has a stoichiometry close to the theoretical stoichiometry of LiMn 2 O 4 (Li 14%; Mn 29%; O 57%), e.g. in case of as-fabricated LMO on Pt the stoichiometry was as follows: Li 14%; Mn 24%; O 62%, or in case of as-fabricated LMO on Ni the stoichiometry was Li 13%; Mn 33%; O 54%.
The performance of the LMO films was assessed from the capacity integrated under the CVs. The capacity of the converted EMD films on C/TiN with only 10% of the theoretical capacity was the lowest of all films (i.e. 0.13 Ah/cm 3 ). Even though the activity of the film converted at 350 • C was significantly better than those treated at higher temperatures, the reduction of MnO 2 by the underlying graphitic carbon still seems to be an issue. The highest total capacity was obtained for the films on ALD MnO 2 /TiN and Ni/TiN with 44% of the theoretical capacity (0.55 Ah/cm 3 ) and 34% of the theoretical capacity (0.43 Ah/cm 3 ), respectively. The lower accessible capacity compared to the converted LMO film on Pt is due to the difference in film thickness, i.e. 50 nm of EMD on Pt versus 250 nm of EMD on ALD MnO 2 /TiN and Ni/TiN. The longer heat-treatment of 6 h (at 250 • C) for ALD MnO 2 /TiN substrate compared to 2 h (at 350 • C) for Ni/TiN substrate is believed to be responsible for the larger capacitance on ALD MnO 2 /TiN substrate. These results demonstrate the feasibility of the conversion process of EMD films to adherent and active LMO films with submicron to micron thickness on different substrates. Further development is underway to fabricate conformal LMO films also on the microstructured current collectors for 3D thin-film Li-ion battery devices.

Conclusions
A methodology for the electrodeposition of MnO 2 films onto easily oxidizable and technologically-relevant current collectors for 3D energy storage applications is proposed. Anodic electrodeposition of thick and electrochemically active EMD films from conventional acidic MnSO 4 electrolyte solutions is not possible on easy to oxidize substrates like TiN due to its passivation or Ni due to its oxidation and dissolution. Optimization of the current collector interface by introducing a non-corrosive protection layer such as carbon film or a dense MnO 2 seed layer allowed to deposit thick EMD films up to 1.3 μm on TiN surfaces with a good adhesion. The use of neutral and complexing acetate bath allowed thick EMD deposition on a Ni surface. The mechanism of EMD growth and film morphology were examined. Compatibility of the proposed approach for the electrodeposition of conformal thick EMD films on 3D high-aspect ratio micropillars was demonstrated. As-grown thick and porous EMD films plated directly on a current collector were shown to be applicable as Li-ion intercalation electrodes upon their conversion to LMO films. The method of solid-state conversion of EMD films to high capacity LMO material on different substrates at low-temperature was demonstrated, paving the way toward their application in 3D Li-ion batteries.