NC 98,287- WO-I Inventors: Long et al.

Electroless Deposition of Nanoscale Manganese Oxide on Ultraporous Carbon Nanoarchitectures

[001] This application claims the benefit of U.S. provisional application No. 60/847,399 filed on September 11, 2006.

[002] Electrochemical capacitors (also denoted as supercapacitors or ultracapacitors) are a class of energy-storage materials that offer significant promise in bridging the performance gap between the high energy density of batteries and the high power density derived from dielectric capacitors. Energy storage in an electrochemical capacitor is accomplished by two principal mechanisms: double-layer capacitance and pseudocapacitance.

[003] Double-layer capacitance arises from the separation of charge that occurs at an electrified interface. With this mechanism the capacitance is related to the active electrode surface area, with practical capacitances in liquid electrolytes of 10-40 μF cm ^"2 . Electrochemical capacitors based on double-layer capacitance are typically designed with high-surface-area carbon electrodes, including carbon aerogels, foams, nanotube/nanofiber assemblies, and papers. Carbon aerogels and related porous carbons are particularly attractive due to their high surface areas, high porosities, and excellent conductivities (>40 S cm ^~ '). Although the high-quality porosity of such carbon nanoarchitectures supports rapid charge-discharge operation, the overall energy-storage capacities of carbon electrodes are ultimately limited by their reliance on the double-layer capacitance mechanism.

[004] Pseudocapacitance broadly describes faradaic reactions whose current— voltage profiles mimic those of double-layer capacitors. Because this mechanism involves true

0I

Isuαtwtskt

NC 98,287- WO-I Inventors: Long et al. electron-transfer reactions and is not strictly limited to the electrode surface, materials exhibiting pseudocapacitance often have higher energy densities relative to double-layer capacitors. The two main classes of materials being researched for their pseudocapacitance are transition metal oxides and conducting polymers. [005 J At present, some of the best candidates for electrochemical capacitors are based on nanoscale forms of mixed ion— electron conducting metal oxides, such as RuO ₂ , which store charge via a cation— electron insertion mechanism.

Ru ^IV O ₂ + x e ^" 4- x H ⁺ ^=* H _x Ru ⁱⁿ _x Ru ^lv , _-x O ₂ [1]

Electrodes based on disordered, hydrous Ruθ ₂ yield specific capacitances up to 720F g ^*1 . However, the application of RuO ₂ is limited by the high costs of the ruthenium precursors.

[006] Manganese oxides have recently gained attention as active materials for electrochemical capacitors, primarily due to their significantly lower cost relative to hydrous Ruθ2- A survey of recent publications, combined with research findings at the NRL, shows that when prepared in traditional electrode configurations, such as micron- thick films or in composite electrodes containing carbon and binders, MnO ₂ delivers a specific capacitance of ~200F g ^" ', which is competitive with carbon supercapacitors, but far short of the 720 F g ^"1 obtained with hydrous RuO ₂ . However, as reported independently by Pang et al. and Toupin et al., when MnO ₂ is produced as a very thin film (tens of nanometers or less) on a planar current collector, specific capacitances of 700 F g ^"1 and 138O F g ^"1 can be achieved, respectively. This disparity in measured capacitance can be attributed to poor long-range electronic and/or ionic conductivity for MnO ₂ , which can inhibit the charge-discharge process in conventional electrode designs. Although ultrathin films OfMnO ₂ deliver high specific capacitance, this configuration can limit the area-normalized capacitance for practical EC devices.

NC 98,287- WO-I Inventors: Long et al.

[007] Alternatives involve electrode structures incorporating carbon nanotubes (an expensive carbon substrate) or alternative MnO ₂ deposition methods (e.g., electrodeposition, sputtering), which are more complicated, costly and more difficult to control.

[008] Brief Description of the Figures

[009] Figure 1 shows a schematic of a hybrid electrode structure comprising a highly porous carbon nanostructure coated with nanoscopic Mnθ ₂ deposits.

[0010] Figure 2 shows a schematic of electrodeposition on an ultraporous electrode structure in (i) a poorly controlled manner in which the pores are ultimately blocked by the growing film and (ii) a controlled, self-limiting deposition.

[0011] Figure 3 shows scanning electron micrographs of (a and b) 4-h acid-deposited MnO ₂ - carbon, (c and d) 4-h neutral-deposited MnO ₂ - carbon, and (e and f) bare carbon nanofoam.

[0012] Figure 4 shows a scanning electron micrograph (top) for the cross-section of a 4-h neutral-deposited MnO ₂ -carbon, and corresponding elemental mapping images (middle and bottom) of this same area of the sample for carbon and manganese content, respectively.

[0013] Figure 5 shows cyclic voltammograms at 2 and 20 mV s ^"1 in 1 M Na ₂ SO ₄ for bare carbon nanofoam ( — ) and 4-h MnO ₂ - carbon nanofoam deposited from acidic permanganate solution ( ) and neutral permanganate solution (• • •)•

[0014] Figure 6 shows a Nyquist plot and a capacitance vs. frequency profile for acid- deposited, neutral-deposited MnOa-carbon nanofoam and bare carbon nanofoam.

[0015] Figure 7 shows scanning electron micrographs of the surfaces of 4-h neutral-

PC17US200I;

NC 98,287- WO-I Inventors: Long et al. deposited MnO ₂ - carbon nanofoams, where AgMnO ₄ is the deposition precursor, both with (top) and without (bottom) the addition of a pH 6.9 phosphate buffer.

Disclosure [0016] Nanostructured Mnθ ₂ -carbon nanoarchitecture hybrids can be designed as electrode structures for high-energy-density electrochemical capacitors that retain high power density. Homogeneous, ultrathin coatings of nanoscale MnO ₂ can be incorporated within porous, high-surface-area carbon substrates (such as carbon nanofoams) via electroless deposition from aqueous permanganate under controlled pH conditions. The resulting hybrid structures exhibit enhanced gravimetric, volumetric, and area-normalized capacitance when electrochemically cycled in aqueous electrolytes. This design can be extended to other mesoporous and macroporous carbon forms possessing a continuous pore network.

[0017] The performance limitations of MnO ₂ for electrochemical capacitors can be addressed with a hybrid electrode design, by incorporating discrete nanoscale coatings or deposits of MnO ₂ onto porous, high-surface-area carbon structures (see Figure 1). In such a configuration, long-range electronic conduction is facilitated through the carbon backbone and solid-state transport distances for ions through the MnO ₂ phase can be minimized by maintaining a nanoscopic carbon 11 MnO ₂ 11 electrolyte interface throughout the macroscopic porous electrode. Such a design can be realized using various types of porous carbon substrates including but not limited to aerogels/nanofoams, templated mesoporous carbon, and nanotube/nanofiber assemblies.

[0018] The synthesis and electrochemical characterization of MnO ₂ -carbon composites has been reported and primarily focused on incorporating nanoscale MnO ₂ deposits onto carbon nanotubes using a variety of approaches including simple physical mixing of the

components, chemical deposition using such precursors as KMnO ₄ , and electrochemical deposition. In these cases, the incorporation of MnO ₂ improves the capacitance of the electrode structures that contain the Mnθ ₂ -modified carbon nanotubes; however, the overall specific capacitance for the composite structures is typically limited to <200 F g ^"1 , even for electrodes with high weight loadings of MnO ₂ . One exception was reported by Lee et al., who demonstrated specific capacitances of up to 415 F g ^"1 as normalized to the MnO ₂ of the composite structure. However, those results were achieved only for micron- thick electrode structures containing MnOa-modified carbon nanotubes, again a configuration that limits energy density.

[0019] Templated mesoporous carbon powders have also been used as a substrate for MnO ₂ deposition as demonstrated by Dong et al., who used the reaction of permanganate with the carbon substrate to generate nanoscale MnO ₂ deposits directly on the mesopore walls. Electrochemical testing of the resulting MnO ₂ - mesoporous carbon structures revealed that the Mnθ2 deposits themselves exhibited a specific capacitance of ~600 F g ^"1 , which approaches the 700 F g ^'1 reported by Pang et al. for nanometers-thick MnO ₂ films. Despite the high Mn0 ₂ -normalized capacitance, the overall specific capacitance of the hybrid Mn0 ₂ -mesoporous carbon structure was limited to 200 Fg ^"1 , due to the relatively low weight loading (up to 26%) Of MnO ₂ . The extent Of MnO ₂ deposition within the mesoporous carbon substrate can be limited by the inherently small pore size (~3 nm) of the carbon.

[0020] The investigations of Dong et al. and Lee et al. demonstrate that nanoscopic deposits of MnO ₂ on high-surface-area substrates do deliver high specific capacitance. To further optimize the performance of MnO ₂ -carbon hybrid structures for electrochemical capacitor applications, at least three design parameters must be addressed: (i) achieving high weight loadings of MnO ₂ (>50 wt. %); (ii) fabricating

NC 98,287- WO-I Inventors: Long et al. electrode structures with macroscopic thickness (tens to hundreds of microns); and (iii) retaining a through-connected pore network in 3D and with pores sized at >5 nm.

[0021] The use of thick carbon substrates, as opposed to dispersed carbon powders, presents new challenges for achieving homogeneous MnO ₂ deposition throughout the electrode structure, while preserving the native pore structure of the carbon template. A high-quality pore structure is vital for high-rate EC operation, facilitating electrolyte infiltration and ion transport. These properties can be achieved by using coating methods that are inherently self-limiting as shown schematically in Figure 2. Described in this disclosure is the self-limiting electroless deposition of nanoscale MnO ₂ , based on the redox reaction of aqueous permanganate (MnO ₄ ^" ) and carbon aerogel/nanofoam substrates. The MnO ₂ prepared by this protocol is a complex structure incorporating cations and water; this material will be designated as MnO ₂ in the body of this application.

[0022] Example 1 Electroless deposition OfMnO ₂ on carbon nanofoams.

[0023] Carbon nanofoam papers were either purchased from a commercial source or prepared in-house. MnO ₂ -carbon nanoarchitecture hybrids were created based on the reductive decomposition of permanganate from aqueous solutions. The carbon nanoarchitecture surface can serve as a sacrificial reductant, converting the aqueous permanganate to insoluble MnO ₂ .

[0024] Carbon nanofoam substrates (~ 170-μm thick) were first wetted in an aqueous solution of controlled pH (0.1 M H ₂ SO ₄ , 0.1 M Na ₂ SO ₄ , or 0.1 M NaOH) by vacuum infiltration. The samples were then soaked in 0.1 M NaMnO ₄ at each respective pH for a period of 5 min to 4 h. The Mnθ ₂ -carbon nanofoam papers were rinsed thoroughly with ultrapure water and subsequently dried at ~50°C under N ₂ for 8 hours and then under vacuum overnight.

PC17US200ϊ

NC 98,287-WO- 1 Inventors: Long et al.

10025] Control of the permanganate reduction reaction can be required to achieve nanoscale Mnθ2 deposits throughout the carbon nanoarchitecture as well as to inhibit the formation of thick MnO ₂ coatings on the outer boundary of the carbon electrode. Preliminary results suggest that pH can be a critical factor in determining the quality of the MnO ₂ deposition.

[0026] As shown by the scanning electron micrographs (SEM) in Figure 3 a and 3b, under acidic conditions, permanganate reacts with carbon nanofoams to primarily form thick crusts of Mnθ2 on the outer boundary of the carbon electrode, presumably due to the autocatalytic decomposition of permanganate in acid. A cross-sectional image of the acid-deposited Mnθ2 crust, shown in the inset of Figure 3a, reveals that the crust thickness was ~4 μm for a 4-h deposition. By contrast, permanganate reduction in neutral or basic pH solutions results in homogeneous MnO ₂ deposits (neutral sample, Figures 3c and 3d) that are nearly indistinguishable from the bare carbon aerogel (Figures 3e and 3f) with no MnO ₂ crust formation at the outer boundary of the nanofoam electrode.

[0027] The MnO ₂ mass uptake (up to ~60% for a 24-h deposition) can be relatively independent of the solution pH. The SEM analysis further confirmed that the porous texture of the initial carbon nanofoam can be largely retained following MnO ₂ deposition (see Figure 3d and 3f). The retention of the nanofoam's high-quality pore structure can result in better electrochemical performance under high-rate charge— discharge operation.

[0028] The cross-sectional SEM and elemental mapping images of the Mnθ ₂ -carbon nanofoam synthesized under neutral conditions in Figure 4 show that the Mn can be evenly distributed throughout the thickness of the electrode structure. Incorporation of the MnC» ₂ domains within the porous carbon nanoarchitectures in such a homogeneous, conformal fashion can result in hybrid electrode structures with superior performance relative to the less ideal structures obtained under acidic deposition conditions. X-ray

NC 98,287-WO- 1 Inventors: Long et al. photoelectron spectroscopy was used to verify that Mn deposits were primarily in the form of Mn" ^1/lv O ₂ , with no residual MnO ₄ ^" .

[0029] Example 2 Electrochemical characterization of hybrid structures.

[0030J The MnO ₂ -carbon nanofoam electrodes were wetted with 1 M Na ₂ SO ₄ under vacuum for electrochemical analysis and characterized in a conventional three-electrode electrochemical cell using techniques such as cyclic voltammetry, impedance spectroscopy, and galvanostatic charge-discharge measurements. Representative cyclic voltammograms of the bare carbon aerogel, 4-h acid-deposited, and 4-h neutral-deposited

MnO ₂ - carbon nanofoam electrodes in 1 M Na ₂ SO ₄ at 2 and 20 mV s ^"1 are presented in Figure 5.

[0031] A saturated calomel reference electrode (SCE) and reticulated vitreous carbon auxiliary electrode were used in all electrochemical measurements. At 2 mV s ^~ ', all curves exhibit a nearly symmetrical rectangular shape, indicative of relatively low uncompensated electrode or solution resistance. , The gravimetric (normalized to total sample mass), volumetric, and area-normalized capacitance values calculated from these curves between 0.1 and 0.6 V vs. SCE are presented in Table 1. Both the total gravimetric and volumetric capacitance values increase for the acid- and neutral- deposited samples. Notably, the gravimetric capacitance increases by a factor of 2 for the neutral-deposited sample, while the volumetric capacitance is over 4 times greater. It is important to note that in the case of homogeneous, nanoscopic MnO ₂ deposits like those in the neutral-deposited hybrid electrode, the incorporation of MnO ₂ can contribute additional capacitance without increasing the bulk volume of the electrode structure.

[0032] When pulse power is required in a footprint- or area-limited configuration, as in microelectromechanical (MEMS) based and on-chip devices, the area-normalized energy- storage capacity should also be considered. Although the area-normalized capacitance is

NC 98,287- WO-I Inventors: Long et al. often not reported for MnO _∑ /carbon composites, it is usually around 10-50 mF cm ^"2 . In contrast, the present hybrid electrode design maintains the advantages of a nanoscopic electrode/electrolyte interface while projecting the electrode structure in three dimensions with a limited footprint, such that the area-normalized capacitance for the neutral- deposited Mnθ ₂ -carbon hybrid electrodes is orders of magnitude greater at >2 F cm ^'2 .

[0033] The upper and lower limits of capacitance attributed to MnO ₂ in Table 1 were estimated using one of two assumptions: (1) all capacitance arises from the MnO ₂ phase (upper limit) or (2) the total sample capacitance was the sum of the carbon double-layer capacitance and the Mnθ ₂ capacitance (lower limit). Although the capacitance attributable to the MnO ₂ phase for the acid- and neutral-deposited samples likely falls within this range, the capacitance contribution from the carbon is expected to be different for the acid and neutral case because of the variation in the MnO ₂ spatial distribution. For example in the acid case, the double-layer capacitance contribution of carbon should be largely unaffected due to the limited MnO ₂ deposition in the electrode interior. Thus, the Mnθ ₂ -normalized capacitance is likely near the lower estimated limit, while that for the neutral sample is expected to be higher as a result of extensive MnO ₂ coating the carbon on the electrode interior.

[0034] Although the total capacitance enhancement for the acid and neutral-deposited MnO ₂ samples presented in Table 1 is similar, the difference in the spatial distribution of MnO ₂ for the two samples results in a sloping voltarnmetric curve for the boundary- crusted, acid-deposited MnO ₂ -carbon nanofoam at 20 mV s ^"1 due to increasing resistance that results from non-uniform MnO ₂ deposition. This increased resistance is confirmed by electrochemical impedance analysis (E _DC ⁼ 200 mV vs. SCE) presented in Figure 6 (similar results were observed at 0 and 600 mV).

[0035] At high frequencies, the uncompensated solution resistance (R _ω ) of each

NC 98,287-WO- 1 Inventors: Long et al. electrode is similar, as shown in the Nyquist plot (Figure 6a). However, the large hemispherical component for the MnO ₂ -carbon nanofoam electrode deposited under acidic conditions is indicative of polarization as expressed by a charge-transfer resistance (R _p ) of about 15 ω. In contrast, the profile for the neutral-deposited sample is more similar to that of the bare carbon nanofoam, with an Rp of ~ 1 ω. The capacitance vs. frequency profile of the neutral-deposited sample in Figure 6b shows that from about 0.01 to 1 Hz, the Mnθ ₂ component can increase the capacitance of the bare carbon nanofoam. As the frequency increases, the capacitance for both electrodes begins to decrease, falling to below 1 F g ^"1 around 200 Hz. The initial capacitance increase for the acid-deposited sample at 0.01 Hz, with respect to the bare nanofoam, can be much lower than that for the neutral-deposited sample and begins to decrease between 0.1 and 1 Hz, falling below 1 Fg ^"1 at 30 Hz.

[0036] The higher resistance and lower capacitance for the acid-deposited sample is likely due to the thick MnO ₂ crust that forms on the electrode exterior, hindering electron and ion transport, while the more ideal homogeneous distribution OfMnO ₂ in the sample deposited under neutral conditions results in electrochemical characteristics more similar to the bare nanofoam.

[0037]

Table 1.

Specific MnO ₂ -specifϊc Volumetric Area-normalized capacitance capacitance range capacitance capacitance

(F g ^" 0+MnO ₂ ) (F g ^" MnO ₇ ) (F cm ^"3 ) (F cm ^"2 )"

Bare nanofoam 53 — 20 0.56

Acid-deposited ^* 92 150 - 220 81 1.4

Neutral-deposited ^* 110 170-230 90 1.5

mimmmatmmm m \ smiwwmMBmam

NC 98,287-WO-l Inventors: Long et al.

^* These data are derived for 4-h depositions from acidic or neutral permanganate solutions.

^** Normalized to the geometric area of one face of the nanofoam electrode.

[0038] The electroless deposition described herein can be a simple, cost-effective, and scaleable approach for synthesizing Mnθ ₂ -carbon hybrid nanoarchitectures with electrochemical capacitance that is superior to unmodified carbon substrates. This disclosure demonstrates that by controlling solution pH during the deposition process, homogeneous MnO ₂ deposits are achieved throughout macroscopically thick porous carbon templates.

[0039] There can be many benefits of homogenous MnO ₂ deposition as can be evident when such structures are electrochemically analyzed. For example, Mnθ ₂ -carbon hybrids that exhibit uniform MnO ₂ distribution (neutral-pH deposition) also exhibit higher overall gravimetric and volumetric capacitance, and higher MnO ₂ -specific capacitance than acid-deposited Mnθ ₂ -carbon hybrids, in which the MnO ₂ is primarily deposited as a crust on the outer boundaries of the electrode.

[0040] Uniform deposition within the interior of the carbon nanoarchitecture also can result in greater enhancement when the volumetric capacitance is considered, as the addition of the MnO ₂ component contributes additional capacitance without increasing the bulk volume of the electrode structure. For example, with a carbon nanofoam coated under neutral-pH conditions the gravimetric capacitance is increased by a factor of 3.3, while the volumetric capacitance is increased by a factor of 4.1. Even greater enhancements in electrochemical performance for these hybrids can be realized with further optimization of the electroless deposition conditions (e.g., varying the solution temperature, precursor concentration, permanganate counterion — including transition metal speciation, constituents that define the acidic or neutral medium including buffers)

ϋϋωii iiiiiiissiiϋ itiiϋ i PCT/US200 IL

NC 98,287-WO-l Inventors: Long et al. and also by varying the carbon template pore structure, particularly targeting larger pore sizes (100-200 tun) and higher overall porosity, which should result in higher mass loadings of MnO∑.

[0041] In an example of two modifications to the deposition protocol, carbon nanofoams were soaked in nominally neutral aqueous solutions of commercially available AgMnO ₄ (substituted for NaMnO ₄ ) under buffered and unbuffered conditions. The morphology of the resulting Ag _x Mn ^11I/IV i. _x θ2 deposits is affected by the presence or absence of buffering agents. Without buffer, the oxide coating is more nodular and preferentially deposited on the outer boundary, while from buffered medium, the deposit is more uniformly distributed, not nodular, and less thick (as seen in Figure 7).

[0042] Obviously, many modifications and variations of the present invention are possible in light of the above teachings. It is therefore to be understood that the claimed invention may be practiced otherwise than as specifically described. Any reference to claim elements in the singular, e.g., using the articles "a," "an," "the," or "said" is not construed as limiting the element to the singular.