Article pubs.acs.org/Langmuir
Influence of Mesoporosity on Lithium-Ion Storage Capacity and Rate Performance of Nanostructured TiO2(B) Anthony G. Dylla, Jonathan A. Lee, and Keith J. Stevenson* Department of Chemistry and Biochemistry, University of Texas at Austin, Austin, Texas 78712, United States S Supporting Information *
ABSTRACT: Here, we present the Li+ insertion behavior of mesoporous ordered TiO2(B) nanoparticles (meso-TiO2(B)). Using presynthesized 4 nm TiO2(B) nanoparticles as building blocks and a commercially available ethylene glycolpropylene glycol block copolymer (P123) as a structure-directing agent, we were able to produce mesoporous structures of high-purity TiO2(B) with nanocrystallinity and mesopore channels ranging from 10 to 20 nm in diameter. We compared the Li+ insertion properties of nontemplated TiO2(B) nanoparticles (nano-TiO2(B)) to meso-TiO2(B) via voltammetry and galvanostatic cycling and found significant increases in overall Li+ insertion capacity for the latter. While nano-TiO2(B) and meso-TiO2(B) both show surface charging (pseudocapacitive) Li+ insertion behavior, meso-TiO2(B) exhibits a higher overall capacity especially at high charge rates. We attribute this effect to higher electrode/electrolyte contact area as well as the improved electron and ion transport in meso-TiO2(B). In this study, we have demonstrated the influence of both nanostructuring and mesoporosity on Li+ insertion behavior by rationally controlling the overall architecture of the TiO2(B) materials.
■
charge/discharge rates, and long-term stability.7 Brezesinski and co-workers reported that mesoporous templated anatase stores Li+ by both Faradaic and pseudocapacitive mechanisms.8 By comparing mesoporous materials and nontemplated TiO2 nanocrystals, they showed that nanocrystallinity and an open pore structure is key to the observed increased charge storage capacity. Mesoporous templating increases both overall storage capacity and high rate capacity due to improved Li-ion coupled electron transfer kinetics, faster diffusion of Li+ through the interconnected TiO2 network, and increased electrode/electrolyte contact areas. One of the most common strategies to produce mesoporous TiO2 thin film electrodes uses inorganic titanium precursors such as alkoxides or chlorides dispersed in a water/ethanol/surfactant (cationic or nonionic) to form a sol− gel coating solution. The film is prepared by spin- or dipcoating onto conductive substrates followed by aging and calcination of the film. The morphology of the mesochannel structure is primarily determined by the configuration of the self-assembled organic surfactants in these systems, which may change during the sol−gel reaction according to the nature and volume ratio of inorganic, water, and surfactant components.9−11 Brezesinski and co-workers synthesized highly ordered anatase TiO2 films by this method and demonstrated that the resulting materials show high charge storage capacity (300 C/g) with contributions by both capacitive and Faradaic charge storage mechanisms.8 Similarly, by P-doping with
INTRODUCTION For electric vehicles and grid storage devices to become viable alternatives to current technology, further development of energy storage materials that offer high capacity and high rate capabilities is needed.1,2 Pseudocapacitors can be considered hybrids of traditional batteries (high energy density but typically poor power output) and double layer capacitors (high power output during short bursts but low energy density). The pseudocapacitive energy storage mechanism is different from batteries in that surface redox properties dominate the charge-transfer processes rather than normal Faradaic insertion processes.3 Metal oxides such as RuO2 and IrO2 act as pseudocapacitors offering exceptional power, fast charging, and long-term stability while also affording some of the advantages of traditional secondary batteries such as relatively high storage capacity.3,4 By nanostructuring certain electroactive materials, this surface charge-transfer process (pseudocapacitive effect) becomes the dominant storage mechanism and can offer 10−100 times the capacitance of a traditional carbon-based double layer capacitor.3 Here, we present the formation of mesoporous ordered TiO2(B) nanoparticles and show the importance of both nanostructuring and mesoporosity to the pseudocapacitive storage mechanism through voltammetric and galvanostatic Li+ insertion studies. There are many recent reports on nanostructuring metal oxides such as MnO2, SnO2, V2O5, and, in particular, TiO2.5,6 TiO2 is a potential replacement for graphite anodes in Li-ion batteries with applications for electric vehicles due to a higher operating potential (1.7 V), minimal surface electrolyte interface (SEI) layer formation, reasonable capacity at high © 2012 American Chemical Society
Received: September 23, 2011 Revised: January 5, 2012 Published: January 6, 2012 2897
dx.doi.org/10.1021/la2037229 | Langmuir 2012, 28, 2897−2903
Langmuir
Article
charge rate is greatly increased in meso-TiO2(B). We attribute this effect to higher electrode/electrolyte contact area as well as the improved electron and ion transport in meso-TiO2(B).
H3PO4 during sol−gel formation, Procházka and co-workers created mixed anatase/TiO2(B) phase mesoporous materials that showed reversible Li+ insertion with good capacity retention over 1000 cycles, yet a specific gravimetric capacity was not stated.12 Another method for producing mesostructured electrodes is to begin with small nanocrystal building blocks and use structure-directing agents to control the threedimensional (3-D) architecture. An advantage of this method is better control over the phase purity and building block unit size of the desired metal oxide. Brezesinski et al. used this strategy to template 3 nm anatase nanocrystals into a common diblock copolymer to produce mesoporous metal oxide thin films.8 These materials showed an increase in total capacity and increased pseudocapacitive contributions to the overall capacity as compared to the sol−gel mesoporous method. As a control, they also prepared anodes with untemplated nanocrystals. The overall capacity of the untemplated anatase nanocrystals was lower than both sol−gel-derived anatase films and the templated anatase nanocrystals, although the contribution due to pseudocapacitive effects was similar (∼50%) to that of the templated anatase nanocrystals. These results show that threedimensional interconnected mesoporosity significantly enhances the Li+ insertion processes. Interestingly, when using the sol−gel method for synthesizing mesoporous TiO2, a second TiO2 phase is often introduced as a minor component. The hydrolysis conditions required for the formation of Ti−O−Ti oligomeric networks are difficult to control and can result in multiphase materials. Kavan and coworkers noted the impurity during electrochemical experiments showing unique Li+ insertion at 1.6 V.13 This phase was first believed to be an amorphous surface anatase species but was later confirmed to be a TiO2(B) impurity.14 TiO2(B) (bronze) has been the focus of much research regarding Li+ battery anode materials due to its unique open crystal structure allowing facile Li+ insertion. TiO2(B) was first synthesized in 1980 by Marchand and co-workers from the layered titanate K2Ti4O9, which was converted to H2Ti4O9 via acid washing and finally dehydrated to the layered TiO2(B) structure.15 The crystal structure of TiO2(B) (space group C2/m) is comprised of edge- and corner-sharing TiO6 octahedra that form open channels that can lead to facile lithium-ion intercalation.6 The maximum Li+ insertion capacity for TiO2(B) is 240 mAh/g, higher than that of both anatase (175 mAh/g) and rutile (150 mAh/g).7 The difference in observed maximum capacity for the titania polymorphs is due to both structural changes and lattice expansion upon lithiation that make the material unstable. Like other TiO2 polymorphs, TiO2(B) exhibits greatly increased Li+ storage capacity when confined to the nanosize domain.16 Armstrong and co-workers demonstrated this effect by showing that both nanorod and tube geometries of TiO2(B) further increase insertion capacity (305 mAh/g) both by storing charge at surface sites and through normal Faradaic insertion charge storage.17,18 Here, we present the Li+ insertion properties of mesoporous ordered TiO2(B) nanoparticles, hereafter referred to as mesoTiO2(B) as compared to untemplated TiO2(B) nanoparticles in an effort to understand fundamental and intrinsic electrochemical properties. By combining the advantages of inherently high capacity TiO2(B) nanocrystals with a mesoporous architecture, we show increased capacity at high charge rates as compared to nonmesoporous TiO2(B) nanoparticles (nanoTiO2(B)). While both meso-TiO2(B) and nano-TiO2(B) exhibit pseudocapacitive behavior, the overall capacity at high
■
EXPERIMENTAL SECTION
Materials. Ti metal powder 100 mesh (99.9%), anatase TiO2 powder (99.9%), and glycolic acid (>97%) were purchased from Sigma-Aldrich. KNO3, tetrahydrofuran (99.9%), anhydrous ethanol, 28% NH4OH, 30% H2O2, and 18 M H2SO4 were purchased from Fisher. The P123 diblock copolymer was a gift from BASF. The electrolyte was 1 M LiPF6 in propylene carbonate (PC) and dimethyl carbonate (DMC) at a 1:1 v/v ratio purchased from Novolyte Technologies. Synthesis of Bulk TiO2(B). Bulk TiO2(B) was synthesized using a method previously reported by Marchand et al.15 Briefly, 2.0 g of anatase was mixed with 1.3 g of KNO3 using a mortar and pestle. This powder was transferred to a ceramic boat and heated to 1000 °C in air for 6 h to form potassium titanate. Excess 0.45 M HNO3 was added to the solid and stirred vigorously for 3 days. The solid was filtered and washed with ultrapure water five times. The hydrogen titanate solid was dried overnight in a 60 °C oven and heated in air at 5 °C/min to 500 °C and held for 18 h to dehydrate the hydrogen titanate into the final bulk-TiO2(B) product. Synthesis of TiO2(B) Nanoparticles. TiO2(B) nanoparticles were synthesized using a slight modification to a previously reported method.19 117 mg of fine Ti powder was added to a mixture of 2.5 mL of 28% NH4OH and 10 mL of 30% H2O2. This mixture was stirred for 30 min at room temperature followed by the addition of 290 mg of glycolic acid. This mixture was stirred and heated to 80 °C for 30 min. The solution turned pale yellow in color, indicating the oxidation of the Ti metal and formation of a titanium glycolate. Upon complete dissolution of the Ti metal, the solution was cooled to room temperature followed by the addition of 0.215 mL of 18 M H2SO4. The solution instantly turned dark red, indicating formation of hydrogen titanate. This solution was hydrothermally heated to 110 °C for 20 h to form TiO2(B) nanoparticles. The reaction was quenched to room temperature in an ice bath, and the product was washed three times in water and dried overnight in a 60 °C oven. Synthesis of Electrode Thin Films. A casting slurry containing 10 mg of TiO2(B) nanoparticle powder, and 10 mg of P123 diblock copolymer in 10 μL of THF in 100 μL of ethanol was prepared for meso-TiO2(B) thin film preparation. The mixture was stirred for 6 h before drop casting onto chemically cleaned indium tin oxide (ITO) coated glass slides. Scotch tape was used to create a 0.25 cm2 mask, and the excess slurry was doctor bladed off. The films were calcined by slowly heating to 300 °C (∼0.5 °C/min) to remove the P123 diblock copolymer template. Thin films of nano-TiO2(B) were prepared similarly except that no P123 was added as a structure-directing agent. Characterization. Dilute solutions of the materials were drop cast onto carbon-coated Cu transmission electron microscopy (TEM) grids and allowed to dry at room temperature. Scanning electrochemical microscopy (SEM) was performed on a Hitachi S-5500 operated at 30 kV, and high-resolution TEM (HR-TEM) was performed on a Jeol 2010F operated at 200 kV. X-ray diffraction (XRD) data were collected in transmission mode on a Rigaku R-Axis Spider diffractometer with an image plate detector using a graphite monochromator with Cu Kα radiation (λ = 1.5418 Å). The powder sample was mounted on a Hampton Research CryoLoop. In the case of the meso-TiO2(B), the thin film was first scraped off of the ITO current collector and mounted onto the loop similarly to the bulk- and nano-TiO2(B) samples. Raman spectra were acquired using a Renishaw InVia microscope equipped with a 514 nm Ar+ laser operating at 3 mW. Electrochemical measurements were performed using a CHI 660D potentiostat inside an MBraun glovebox with
