Monodisperse Antimony Nanocrystals for High-Rate Li-ion and Na-ion

Jan 31, 2014 - Flexible P-Doped Carbon Cloth: Vacuum-Sealed Preparation and Enhanced Na-Storage Properties as Binder-Free Anode for Sodium Ion Batteri...
0 downloads 8 Views 4MB Size
Letter pubs.acs.org/NanoLett

Monodisperse Antimony Nanocrystals for High-Rate Li-ion and Naion Battery Anodes: Nano versus Bulk Meng He,†,‡,§ Kostiantyn Kravchyk,†,‡,§ Marc Walter,†,‡ and Maksym V. Kovalenko†,‡,* †

Institute of Inorganic Chemistry, Department of Chemistry and Applied Biosciences, ETH Zürich, CH-8093 Zürich, Switzerland Laboratory for Thin films and Photovoltaics, Empa − Swiss Federal Laboratories for Materials Science and Technology, CH-8600 Dübendorf, Switzerland



S Supporting Information *

ABSTRACT: We report colloidal synthesis of antimony (Sb) nanocrystals with mean size tunable in the 10−20 nm range and with narrow size distributions of 7−11%. In comparison to microcrystalline Sb, 10 and 20 nm Sb nanocrystals exhibit enhanced rate-capability and higher cycling stability as anode materials in rechargeable Li-ion and Na-ion batteries. All three particle sizes of Sb possess high and similar Li-ion and Na-ion charge storage capacities of 580−640 mAh g−1 at moderate charging/discharging current densities of 0.5−1C (1C-rate is 660 mA g−1). At all C-rates (0.5−20C, e.g. current densities of 0.33−13.2 Ag1−), capacities of 20 nm Sb particles are systematically better than for both 10 nm and bulk Sb. At 20C-rates, retention of charge storage capacities by 10 and 20 nm Sb nanocrystals can reach 78−85% of the low-rate value, indicating that rate capability of Sb nanostructures can be comparable to the best Li-ion intercalation anodes and is so far unprecedented for Na-ion storage. KEYWORDS: Antimony, nanocrystals, synthesis, Li-ion, Na-ion, batteries

L

nanocrystals (NPs and NCs), in order to mitigate the effects of volumetric changes and to enhance kinetics of the conversion and alloying reactions.6−13 With regard to SIBs, it is important to note an even greater need for efficient anode materials, because silicon does not reversibly store Na-ions at ambient conditions,14 Graphite shows negligible capacities of 30−35 mAh g−1,15 while other carbonaceous materials exhibit capacities of less than 300 mAh g−1 at rather low current rates and suffer from the low tap density.12 Fully contrary to the present situation with LIBs, there is much greater progress for cathodes than for anodes in SIBs.16,17 In the elemental form, antimony (Sb) has long been considered as a promising anode material for high-energy density LIBs owing to its high theoretical capacity of 660 mAh g−1 upon full lithiation to Li3Sb.3,4 Thus far, high-energy milled or chemically synthesized Sb nanocomposites,18−21 as well as bulk microcrystalline powders or vacuum evaporated thin-film materials, have been studied.22,23 Stable and reversible electrochemical alloying of bulk Sb with Na has also been recently demonstrated,22 pointing to the high utility of this element for SIBs as well. Several reports, published in 2012−2013, have demonstrated efficient Na-ion storage in Sb/C fibers,5 mechanically milled Sb/C nanocomposites,24 Sb/carbon nanotube nanocomposites,25 and in thin films.23 Notably, elemental

ithium(Li)-ion batteries (LIBs) remain the most prominent rechargeable, electrochemical energy storage technology1 with tremendous importance for the portable electronics as well as for rapidly growing sector of environmentally benign electrical mobility.2 Conceptually identical Sodium(Na)-ion batteries (SIBs) have recently received a revived interest as a viable alternative due to much greater natural abundance and more even distribution of Na reserves as compared to Li. When comparing to other electricity storage technologies, most important advantages of commercialized LIBs lie in their long operation life span over hundreds to thousands charge/discharge cycles and superior and broadly tunable balance between the energy density and the power density.3 This implies, inter alia, that in the search for alternative Li-ion anode materials not only theoretical chargestorage capacities must be higher than that of graphite (372 mAh g−1) but also satisfactory retention of capacity on a longterm and under fast charge/discharge cycling (high current densities) must be demonstrated. For instance, the transition from commercial graphite anodes to most intensely studied alternatives such as Si, Ge, Sn, and some metal oxides, possessing 2−10 times higher theoretical capacities (with 3579 mAh g−1 for Si being the highest),4 is primarily hampered by the structural instabilities caused by drastic volumetric changes of up to 150−300% upon full lithiation to, for example, Li3Sb, Li15Si9, Li15Ge4, Li22Sn5,5 or by slow reaction kinetics. Presently, great research efforts are focused on nanostructuring of the active material, by producing nanowires, nanoparticles and © 2014 American Chemical Society

Received: November 10, 2013 Revised: January 23, 2014 Published: January 31, 2014 1255

dx.doi.org/10.1021/nl404165c | Nano Lett. 2014, 14, 1255−1262

Nano Letters

Letter

from the retention of 78−85% of Li-ion and Na-ion storage capacities at high current rates of 20C. Synthesis of Monodisperse Sb NCs. Uniform colloidal NPs and NCs represent a very convenient platform for ex situ or in situ electrochemical studies, ranging from macroscopic electrodes in conventional half-cells to in situ electron microscopy imaging of a single particle during lithiation/ delithiation.26,32 In the form of uniform NPs, Sb has so far remained underexplored due to very limited control over the size and the shape achieved so far.11,33−37 In this work, 10−20 nm Sb NCs with narrow size distributions of 7−11% (Figure 1; for details see Supporting Information, including Tables S1−S6 and Figures S1−S10) were obtained by injecting Sb precursor into a hot (150−200 °C) solution containing a mixture of trioctylphosphine (TOP), lithium diisopropylamide (LiN(iPr)2, LiDPA), and oleylamine (OLA, or OleylNH2). Two precursors, tris(dimethylamino)antimony(III) [Sb(NMe 2 ) 3 , or Sb(DMA)3] and inexpensive antimony(III) chloride, yielded NCs of very similar quality under same reaction conditions. Detailed structural characterization with high-resolution transmission electron microscopy (HR-TEM, Figure 2A), electron diffraction (Figure 2B) and powder X-ray diffraction (XRD, Figure 2C and Rietveld refinement in Supporting Information Figure S1), and energy-dispersive X-ray spectroscopy (EDX, Supporting Information Figure S2) confirmed the formation of chemically pure, highly crystalline rhombohedral Sb NCs (space group N166, R3m ̅ , a = b = 0.4307 nm, c = 1.1273 nm, JCPDS 35-0732). Figure 2C illustrates a finite size effect on the broadening of XRD reflections. The underlying basis of the proposed synthesis is a high reactivity of sterically encumbered Sb alkylamides (e.g., facile homolysis of Sb-nitrogen bonds), leading to the fast and controlled nucleation of colloidal NPs at temperatures sufficiently high for the formation of highly crystalline product but low enough for the conventional solution-phase chemistry (≤300 °C). Similarly to the present study, metal and metalloid alkylamides have been shown previously to yield monodisperse Sn,38 Bi,39 Pb,40 In,41 Co,42 CuTe,4 InSb,43Ag2Se,44 and SnTe45 NCs (nonexhaustive list). For a more comprehensive literature survey, we refer to our recent review article.46 The main role of secondary amide base (LiDPA) is to establish an acid−base equilibrium with the primary amine: OleylRNH2 + LiN(iPr)2

Figure 1. One-pot colloidal synthesis of monodisperse Sb nanocrystals with TEM images illustrating narrow size distributions of ≤11% and size-tunability in 10−20 nm range. See Supporting Information for detailed TEM survey of Sb NCs obtained under variable synthesis conditions.

Sb has not been reported in the form of highly uniform colloidal NPs and NCs, thus providing a formidable challenge for synthetic inorganic chemistry. At the same time, monodisperse nanoparticulate active materials are ideally suited for studying the effects of size and electrode morphology on the electrochemical performance.13,26−31 In this Letter, we draw readers’ attention to the first synthesis of narrowly sized colloidal Sb NCs, tunable in 10−20 nm size range, which allowed us to comprehensively study the effect of primary particle size on the electrochemical performance of Sb as anode material for LIBs and SIBs. We demonstrate and discuss the pros and cons of nano-Sb anodes in comparison to commercial microcrystalline Sb (200 mesh size, from here denoted as “bulk” Sb) and alternative alloying materials such as tin (Sn). We find that at moderate current rates of 0.5−1C (1C rate is 660 mA g−1), Li-ion and Na-ion storage capacities for all three sizes are very similar (580−640 mAh g−1, lowest for 10 nm Sb in SIBs). The major finding for both LIBs and, even more pronounced, for SIBs is that 20 nm Sb NCs exhibit considerably better overall performance than both 10 nm Sb NCs and “bulk” Sb. In particular, with the reduction of the primary Sb size to 20 nm significantly faster kinetics and more stable operation at higher current densities is achieved, while further downsizing to 10 nm or below may in fact be detrimental, as can be seen from the lowering of charge storage capacity by ca. 5% for Li-ions and by up to 20% for Na-ions. High rate-capability of both 10 and 20 nm Sb NCs is apparent

Figure 2. (A) HR-TEM image of an Sb NC, (B) selected-area electron diffraction pattern of an ensample of NCs collected from the area of ∼1 μm2, and (C) powder XRD pattern of 10, 14, and 20 nm Sb NCs along with the crystal structure of Sb. 1256

dx.doi.org/10.1021/nl404165c | Nano Lett. 2014, 14, 1255−1262

Nano Letters

Letter

Figure 3. Rate-capability tests of (A) Li-ion and (B) Na-ion half-cells employing Sb anodes made from monodisperse colloidal Sb NCs and from microcrystalline powders. The data for 10 nm Sn NCs are shown for comparison. All anodes had the same composition of Sb(64%)/CB(21%)/ CMC(15%) and were cycled at room temperature under variable current rate of 0.5−20C (1C = 0.66 Ah g−1, 9 cycles at each C-rate, first cycle at 0.1C). One molar LiPF6 in ethylenecarbonate/dimethyl carbonate mixture containing 3 wt % of FEC was used as electrolyte for Li-ion cells, whereas 1 M NaClO4 in propylenecarbonate containing 10 wt % of FEC was used for Na-ion batteries. All batteries were cycled in the 20 mV to 1.5 V potential range. The obtained capacities were normalized by the mass of Sb.

↔ OleylNHLi + HN(iPr)2. In the second stage, transient and highly unstable Sb(III) oleylamide is formed in situ by the fast reaction between OleylNHLi and Sb(DMA)3 or SbCl3. The decomposition of Sb(III)-oleylamide at 150−200 °C occurs instantly leading to nucleation of Sb NCs. In agreement with this mechanism, the absence of LiDPA leads to much slower reaction, yielding poorly defined, sub-100 nm Sb precipitates formed by the reduction of Sb precursors by OLA (Supporting Information Figure S3). The selection of the surfactants, which control the growth kinetics and shape of NCs, is the next most important consideration. We find that addition of TOP significantly improved uniformity of Sb NCs as compared to pure OLA system (Supporting Information Figure S4). In the absence of OLA, for example, in pure TOP the reaction proceeds uncontrollably fast, presumably due to both the fast formation of highly unstable Sb(III)-DPA species and the lack of efficient surface passivation provided by OLA. We therefore conclude that both TOP and OLA are needed as cosurfactants during the formation of Sb NCs. The size-tuning in 10−20 nm range can be conveniently achieved by varying the reaction time, temperature, quantity of precursors and, as an option, by adding diisobutylaluminum hydride (DIBAH) as a reducing agent. As the reaction proceeds, nearly spherical 10 nm Sb NCs evolve into more faceted 20 nm large NCs. It should be noted that addition of DIBAH enables higher reaction yield for smallest 10 nm Sb NCs. For the subsequent tests in LIBs and in SIBs, we removed bulky and electrically insulating organic capping ligands by treating NCs with 1 M solution of hydrazine in acetonitrile.47−49 The removal of at least 93% of the initial organic ligands has been estimated by integrating the intensities of aliphatic C−H stretching modes in ATR-FTIR spectra (Supporting Information Figure S11). Sb Nanocrystals as Fast Charge/Discharge Anode Material in Li-Ion and Na-Ion Batteries. This work was initially motivated by the report of Darwiche et al.,22 who showed stable storage of Li- and Na- ions in bulk microcrystalline Sb with capacities of ca. 600 mAh g−1 at rates of C/6 (considering 1C = 0.66 Ag1−) and at a higher rate of 4/3C 80th-cycle capacities of 372 and 360 mAh g−1 for Na-ion and Li-ions, respectively. These findings are especially striking in the view of large crystallite size of the used Sb (up to tens of

micrometers). Si, Sn, and Ge had never been reported to perform satisfactory at such large primary particle sizes. Furthermore, in the case of SIBs there are no other anode materials with comparably stable charge storage capacities above 300 mAh g−1. In our work, the major aim was to obtain a comprehensive comparison between well-defined nanoscopic Sb and conventional, commercially available microscrystalline Sb (200 mesh, Supporting Information Figure S12). Nanostructuring has been so far the most efficient approach to enhance the ionic and electronic transport and to reduce the cycling instabilities caused by volumetric changes in the alloying anode materials.50−52 Furthermore, several reports contain evidence for the existence of critical size below which the fracture of an active material may not occur: ∼150 nm for Si53 and less than 50 nm for Sn.54,55 In order to distinguish size-effects from other factors, the following experimental parameters were prescreened and fixed same for all electrodes: (i) the choice and mass fraction of a polymer binder and conductive additive and (ii) electrolyte composition for obtaining stable solid electrolyte interface (SEI).13 In the footstep of the recent improvements for Sn, Si, and Ge anodes, we used carboxymethylcellulose (CMC) as a water-soluble binder and fluoroethylcarbonate (FEC) as an SEI-forming electrolyte additive.56 Amorphous carbon nanoparticles (Super C65 from TIMCAL) was used as conductive additive. It is also important to note that ligand-removal and mechanical mixing of Sb NCs with aqueous solution containing carbon additive and CMC binder did not lead to any noticeable changes of the size and morphology of Sb NCs, and did not cause sintering of NCs. This has been confirmed by TEM images (Supporting Information Figure S13) and by XRD patterns (Supporting Information Figure S14), taken before and after the preparation of slurry. Mean crystallite size estimated by Rietveld refinement of XRD patterns has remained within 5% of the initial value. Furthermore, same mass loading of ∼1 mg cm−2 (whole electrode mass) was employed for all anodes in order to minimize the effect of the electrode thickness on the rate-capability. Electrode thickness (10−15 μm) and surface morphology were assessed with scanning electron microscopy (SEM, Supporting Information Figure S15). Further details of electrode preparation and 1257

dx.doi.org/10.1021/nl404165c | Nano Lett. 2014, 14, 1255−1262

Nano Letters

Letter

Figure 4. Galvanostatic charge and discharge curves for Sb electrodes comprising 10 nm Sb NCs and microscrystalline Sb powder. First-cycle discharge curves indicate the amount of initial irreversible capacity loss, primarily caused by the consumption of Li-ions for the formation of SEI layer.

most plausible reason for the lower capacities of 10 nm Sb NCs in both Li-ion and Na-ion batteries is a greater volume fraction of an amorphous surface oxide shell that leads to irreversible capacity loss due to the formation of Li2O and Na2O in the first-cycle discharge (reduction). With increasing current rate, a rate-capability very similar to that of Li-ion batteries was observed; 10 and 20 nm Sb NCs repeatedly retain 80−85% of their Na-ion storage capacity at 20C-rate, whereas lower retention of ∼50% is obtained for bulk Sb. Besides the effect of the active storage material, the rateperformance is a complex function of the electrode formulation (chemistry and amounts of binder and conductive additive), porosity, electrode thickness, electrolyte, temperature, and so forth. These parameters were fixed identical for all cells in our experiments, allowing us to focus on the lithiation and sodiation kinetics intrinsic to the electrode material. There are several factors that are most plausible contributors to the fast alkali-ion insertion and extraction in Sb electrodes. First factor is the crystal structure of hexagonal Sb (Figure 2), comprised of puckered layered planes and characterized by low atomic packing factor of just 39%.4 Such crystal structure contains fast diffusion channels for Li and Na ions in the interlayer space. Second factor is high electronic conductivity due to semimetallic nature of this element. Third important consideration includes the chemical transformations occurring during the insertion and extraction of alkali metal ions. In our opinion, multiple phase transformations, especially involving crystalline phases, may slow down the overall reaction kinetics and may add complexity due to differences in the electronic and ionic transport within each intermediate phase. For instance, relatively poor rate-performance of Sn can be explained by up to six intermediate crystalline phases reported in Li−Sn system.57 Contrary to Sn, in situ XRD studies by Wang et al. and Hewitt et al. pointed to only one intermediate crystalline phase (Li2Sb) upon lithiation, while delithiation proceeds directly from Li3Sb toward crystalline Sb (Sb thin films were used as samples).58,59 Correspondingly, galvanostatic charging and discharging curves exhibit one-plateau behavior at about 0.7−0.8 V, very similar between nano and bulk material (Figure

assembly of batteries can be found in the Supporting Information. Figure 3 compares Li-ion and Na-ion discharge capacities for Sb anodes containing 64 wt % of 10 nm, 20 nm, and microcrystalline Sb at variable charge/discharge current densities of 0.5−20C (1C = 0.66 A g−1). In the case of Liion cells, 10 and 20 nm Sb NCs exhibit near-theoretical capacities of >600 mAh g−1 at 0.5C and capacity retention of 78 and 85%, respectively, at 20C. At the same time, bulk Sb anodes retained only 40% of their initial charge-storage capacity. Ten nanometers Sb NCs systematically showed by at least 5% lower capacities than 20 nm NCs. We also compared Sb-based electrodes with those comprising monodisperse, naturally oxidized 10 nm tin NCs.38 We have previously shown that at 1C-rate (∼1000 mAh g−1 for Sn) 10 nm Sn/SnOx NCs exhibited significantly better cycling stability than 20 nm particles and commercial 50−150 nm Sn and SnO2 nanopowders but no rate-performance tests were carried out so far. In this study, we find that capacities up to 750 mAh g−1 can be obtained at 0.5C-rate for 10 nm Sn/SnOx NCs, but capacity drops by 80% (down to ∼150 mAh g−1) at a rate of 20C. The rate-modulations of capacities were fully reversible. Comparison with Sn indicates an inherently faster Li-ion insertion, diffusion, and extraction in elemental Sb. To the best of our knowledge, such rate-capability of nano-Sb anodes has no analogues among conversion materials and is comparable to the best-performing intercalation compounds. Taking 80% as a benchmark for capacity retention, well-optimized rate capability of standard graphite electrodes with the thickness of 30−50 μm is typically limited to ∼2C lithiation rate and ∼20C delithiation rate (1C = 372 mAh g−1).29,30 Thinner graphite electrodes (≤10 μm) can be lithiated at rates up to 6C and delithiated at 680C.29 Similar high-rate capabilities have also been reported for highly porous carbons 31 and for lithium titanate (Li4Ti5O12).28 In Na-ion cells, high storage capacities of ∼580−620 mAh g−1 were obtained at 0.5C rate for 20 nm Sb NCs and for bulk Sb (Figure 3B), whereas capacity for 10 nm Sb NCs was systematically lower by up to ∼100 mAh g−1 (15−20%). The 1258

dx.doi.org/10.1021/nl404165c | Nano Lett. 2014, 14, 1255−1262

Nano Letters

Letter

Figure 5. Differential capacitance plots obtained from galvanostatic discharge curves for nanoscopic and microcrystalline Sb-electrodes (shown in Figure 4) for (A) Li-ion cells at 0.5C-rate and for (B) Na-ion cells at 1C-rate. These date correspond to the 10th cycle of cycling stability tests (Figures 6A,C).

Figure 6. Cycling performance of Sb electrodes in Na- and Li-ion half-cells. All electrodes were prepared identically and had composition of Sb(64%)/CB(21%)/CMC(15%). FEC was added as an additive to the electrolyte (3 wt % for LIBs, 10 wt % for SIBs). 1C-rate corresponds to the current density of 0.66 A g−1. For 4C-rate tests in Li-ion cells (B), initial conditioning cycles were attempted (2 cycles at 0.1C + 8 cycles at 0.5C) to stabilize the performance. Nearly identical results were obtained without conditioning cycling as well (not shown here).

often contain sharper features than cyclic voltammograms (CVs, Supporting Information Figure S16) and therefore may be more informative. It is important to note that the number of electrochemical features seen in differential capacitance plots, as well as in CVs, might be higher than the number of involved crystalline phases, because of the possible formation of amorphous phases, not captured by in situ XRD measurements. For both Li-ion and Na-ion cells at moderate C-rate (0.5−1C), 10th cycle dQ/dV plots (Figure 5, corresponding to the regime

4A). Such voltage profiles are notably different for Na-ion cells (Figure 4B), both in the average voltage being close to more optimal 0.5 V, as well as in their more complex shape. As is commonly found in known Na-storage materials,60,61 the insertion/removal chemistries of Na-ions may significantly differ from those of Li-ions. With respect to the mechanism of conversion reaction, additional hints can be found from differential capacitance plots (dQ/dV vs V, Figure 5), obtained from the discharge curves shown in Figure 4. dQ/dV plots 1259

dx.doi.org/10.1021/nl404165c | Nano Lett. 2014, 14, 1255−1262

Nano Letters

Letter

%ΔV = 100% × [Vm(LixM) − Vm(M)]/Vm(M). In the case of Sb, the volume change of 135% is expected, much smaller than for Si (310%) and Sn (260%), and may indeed explain higher cycling stability of Li-ion cells. At the same time, much greater %ΔV of 290% is estimated for Na-ion cells upon full sodiation to hexagonal Na3Sb.62 As discussed above, electrochemical cycling with Na-ion involves only one crystalline (Na3Sb) and several amorphous phases, including amorphous Sb (contrary to Li-ion cells!). We speculate that more isotropic expansion/ contraction of amorphous phases can reduce the amount of anisotropic mechanical stress.23 Furthermore, a thinner or more stable SEI layer may be formed in Na-ion cells containing FEC additive.23 The first-cycle irreversible capacity generally arises from the consumption of Li(Na)-ions for the formation of SEI layer, which is composed of various Li carbonates and fluorides. In particular, the first-cycle Coulombic efficiencies were 50−60% for Sb NCs and 70−75% for bulk Sb in Na-ion cells and 30− 40% for Sb NCs and 50% for bulk Sb in Li-ion cells. These variations are consistent with higher specific area of nanomaterials, as well as with thinner SEI generally reported for Na-ion cells with FEC-containing electrolytes. During the subsequent cycles Coulombic efficiencies were in 98−99% range for 20− 100th cycles. The highest Coulombic efficiencies of 99.5− 99.7% were observed for Li-ion cells cycled at 4C-rate. Because this charge loss of 0.5−2% is considerably larger than the average decay of discharge capacity of ca. 0.05−0.1% per cycle during 20−80th cycle, it can be attributed to parasitic processes such as dendrite formations on Li-electrode and decomposition of electrolyte due to repetitive formation and decomposition of unstable SEI layer. Conclusions. We report a facile colloidal synthesis of monodipserse antimony NCs with mean sizes tunable in the range of 10−20 nm and narrow size distribution of 7−11%. The underlying chemistry is based on the reductive decomposition of in situ formed Sb(III) oleylamide in oleylamine as a coordinating solvent. After removal of the surface capping ligands, 10 and 20 nm Sb NCs were tested as electrode materials in Li-ion and Na-ion batteries under deep cycling conditions (0.02−1.5 V) and were compared to the microcrystalline bulk Sb. In terms of cycling stability and ratecapability, electrochemical performance of Na-ion cells is as good as for Li-ion cells, although sodiation leads to much greater volumetric expansion. As a main conclusion, 20 nm Sb NCs performed systematically better than bulk Sb and 10 nm Sb NCs. For instance, 20 nm Sb NCs exhibit up to 20% higher Na-ion storage capacities than 10 nm NCs. With capacity retention of ca. 80% at high current rates of 20C (13.2 A g−1), nanoscopic Sb is the best-performing Na-ion anode material identified so far and is comparable to fastest Li-ion intercalation materials such as graphite and Li titanates. Further work is needed to stabilize the long-term operation for 1000 or more cycles, presumably through the engineering of the SEI layer with higher chemical and mechanical stability.

of stable cycling seen in Figure 6A,C) show clear electrochemical signature of the particle size: broadening with decreasing the size. Li-ion dQ/dV plots for 20 nm and bulk Sb closely resemble the common data from literature, namely two main peaks at 0.88 and 0.84 V corresponding to the twostep lithiation (via Li2Sb) and the main delithiation feature located at 1.05 V. Partial amorphization of Li3Sb during Liinsertion can possibly take place, as suggested by the third, weak feature at 0.75 V for bulk Sb and by the merging of both peaks into a continuous wave for 20 nm Sb. The reason for splitting of the oxidation sweep (Li-removal) into three peaks for 10 nm Sb NCs remains unclear. Na-ion dQ/dV data for bulk Sb closely resemble the work of Darwiche et al., who for the first time correlated electrochemical data with in situ XRD on microcrystalline Sb samples and observed only one crystalline phase, Na3Sb (mixture of cubic and hexagonal polymorphs).22 Three well-separated Na-ion insertion features at 0.3−0.75 V correspond to the sequence amorphousSb → amorphousNa3Sb → Na3Sbhex/Na3Sbcub → Na3Sbhex.22 Deinsertion primarily occurs as Na3Sbhex → amorphousSb transition (peak at 0.8 V) with the additional broad feature at ca. 0.88 V, which may correspond to partial crystallization of Sb. Cycling Stability of Sb-Based Li-Ion and Na-Ion Cells. Cycling stability tests are presented for the first 100 cycles at moderate currents of 0.5−1C (Figure 6A,C) and at high current rates of 4C (Figure 6B,D). Even higher current densities of 13.2 A g −1 (20C-rate) could be sustained for at least 200 cycles, as shown exemplarily for LIBs in Supporting Information Figure S17. At 0.5 and 1C-rates, all sizes of Sb showed excellent cycling stabilities with capacities above 600 mAh g−1 (e.g., ≥90% of theoretical). At 4C-rate, nanoscopic Sb electrodes remained stable, while the capacity of bulk Sb quickly fades. Most important general observation is that charge storage capacities are repeatedly higher for 20 nm Sb NCs as compared to 10 nm and bulk Sb in both Li-ion and Na-ion cells. As indicated above, lower capacity of 10 nm Sb NCs most likely originates from higher proportion of surface oxides in the smallest Sb NCs that cause irreversible formation of Li2O and Na2O during the first discharge (reduction). The fact that the difference between capacities of 10 nm and 20 nm Sb electrodes is much higher for Na-ion cells (∼20%) than for Li-ion cells (∼5%) can be explained by different properties of Li2O and Na2O. We assume that Li2O acts as rather benign impurity around NCs because of its high Li-ion conductivity. In contrast, Na2O is much poorer conductor of Na ions and, therefore, may exclude some portion of Sb material from reversible charge/discharge. Although Sb has considerably higher theoretical gravimetric storage capacity (660 mAh g−1) than graphite, this material is falling well behind the major Li-ion storage alternatives, Si (3579 mAh g−1 for Li15Si9) and Sn (990 mAh g−1 for Li22Sn5). At the same time, these materials are very similar when theoretical volumetric charge-storage capacities are considered: 1890 mAh cm−3 for Sb, 2200 mAh cm−3 for Si, and 2000 mAh cm−3 for Sn, still much higher than that of graphite (843 mAh g−1).4 From an economic viewpoint based on the raw material costs, Sb, Si, and Sn are comparably viable for large-scale battery production. Drastic volumetric changes are generally considered as a major obstacle for obtaining stable cycling in alloy-based anode materials. Theoretical values for volumetric changes can be estimated from the difference in the molar volumes (%Vm) between the final (LixM) and the initial metallic (M) phases:



ASSOCIATED CONTENT

S Supporting Information *

Experimental details, TEM and SEM images, FTIR spectra, EDX spectrum, Rietveld refinement, and further electrochemical data. This material is available free of charge via the Internet at http://pubs.acs.org. 1260

dx.doi.org/10.1021/nl404165c | Nano Lett. 2014, 14, 1255−1262

Nano Letters



Letter

(24) Qian, J.; Chen, Y.; Wu, L.; Cao, Y.; Ai, X.; Yang, H. Chem. Commun. 2012, 48, 7070. (25) Galland, C.; Ghosh, Y.; Steinbrück, A.; Hollingsworth, J. A.; Htoon, H.; Klimov, V. I. Nat. Commun. 2012, 3, 908. (26) Xu, L. P.; Kim, C.; Shukla, A. K.; Dong, A. G.; Mattox, T. M.; Milliron, D. J.; Cabana, J. Nano Lett. 2013, 13, 1800. (27) Dimitrijevic, B. J.; Aifantis, K. E.; Hackl, K. J. Power Sources 2012, 206, 343. (28) Nakahara, K.; Nakajima, R.; Matsushima, T.; Majima, H. J. Power Sources 2003, 117, 131. (29) Heß, M.; Novák, P. Electrochim. Acta 2013, 106, 149. (30) Buqa, H.; Goers, D.; Holzapfel, M.; Spahr, M. E.; Novák, P. J. Electrochem. Soc. 2005, 152, A474. (31) Hu, Y. S.; Adelhelm, P.; Smarsly, B. M.; Hore, S.; Antonietti, M.; Maier, J. Adv. Funct. Mater. 2007, 17, 1873. (32) Liu, X. H.; Huang, J. Y. Energy Environ. Sci. 2011, 4, 3844. (33) Kim, H.; Cho, J. Chem. Mater. 2008, 20, 1679. (34) Tarascon, J. M.; Armand, M. Nature 2001, 414, 359. (35) Wang, Y. W.; Hong, B. H.; Lee, J. Y.; Kim, J.-S.; Kim, G. H.; Kim, K. S. J. Phys. Chem. B 2004, 108, 16723. (36) Zhu, J.; Sun, T.; Chen, J.; Shi, W.; Zhang, X.; Lou, X.; Mhaisalkar, S.; Hng, H. H.; Boey, F.; Ma, J.; Yan, Q. Chem. Mater. 2010, 22, 5333. (37) Liu, P.; Zhong, K.; Liang, C.; Yang, Q.; Tong, Y.; Li, G.; Hope, G. A. Chem. Mater. 2008, 20, 7532. (38) Kravchyk, K.; Protesescu, L.; Bodnarchuk, M. I.; Krumeich, F.; Yarema, M.; Walter, M.; Guntlin, C.; Kovalenko, M. V. J. Am. Chem. Soc. 2013, 135, 4199. (39) Yarema, M.; Kovalenko, M. V.; Hesser, G.; Talapin, D. V.; Heiss, W. J. Am. Chem. Soc. 2010, 132, 15158. (40) Zolotavin, P.; Guyot-Sionnest, P. ACS Nano 2010, 4, 5599. (41) Yarema, M.; Pichler, S.; Kriegner, D.; Stangl, J.; Yarema, O.; Kirchschlager, R.; Tollabimazraehno, S.; Humer, M.; Häringer, D.; Kohl, M.; Chen, G.; Heiss, W. ACS Nano 2012, 6, 4113. (42) Margeat, O.; Amiens, C.; Chaudret, B.; Lecante, P.; Benfield, R. E. Chem. Mater. 2004, 17, 107. (43) Yarema, M.; Kovalenko, M. V. Chem. Mater. 2013, 25, 1788. (44) Yarema, M.; Pichler, S.; Sytnyk, M.; Seyrkammer, R.; Lechner, R. T.; Fritz-Popovski, G.; Jarzab, D.; Szendrei, K.; Resel, R.; Korovyanko, O.; Loi, M. A.; Paris, O.; Hesser, G.; Heiss, W. ACS Nano 2011, 5, 3758. (45) Kovalenko, M. V.; Heiss, W.; Shevchenko, E. V.; Lee, J.-S.; Schwinghammer, H.; Alivisatos, A. P.; Talapin, D. V. J. Am. Chem. Soc. 2007, 129, 11354. (46) Yarema, M.; Caputo, R.; Kovalenko, M. V. Nanoscale 2013, 5, 8398. (47) Talapin, D. V.; Murray, C. B. Science 2005, 310, 86. (48) Law, M.; Luther, J. M.; Song, Q.; Hughes, B. K.; Perkins, C. L.; Nozik, A. J. J. Am. Chem. Soc. 2008, 130, 5974. (49) Zhang, H.; Hu, B.; Sun, L.; Hovden, R.; Wise, F. W.; Muller, D. A.; Robinson, R. D. Nano Lett. 2011, 11, 5356. (50) Bruce, P. G.; Scrosati, B.; Tarascon, J.-M. Angew. Chem., Int. Ed. 2008, 47, 2930. (51) Li, N. C.; Martin, C. R. J. Electrochem. Soc. 2001, 148, A164. (52) Lee, K. T.; Cho, J. Nano Today 2011, 6, 28. (53) Liu, X. H.; Zhong, L.; Huang, S.; Mao, S. X.; Zhu, T.; Huang, J. Y. ACS Nano 2012, 6, 1522. (54) Aifantis, K. E.; Haycock, M.; Sanders, P.; Hackney, S. A. Mater. Sci. Eng.: A 2011, 529, 55. (55) Wolfenstine, J.; Foster, D.; Read, J.; Behl, W. K.; Luecke, W. J. Power Sources 2000, 87, 1. (56) Chockla, A. M.; Bogart, T. D.; Hessel, C. M.; Klavetter, K. C.; Mullins, C. B.; Korgel, B. A. J. Phys. Chem. C. 2012, 116, 18079. (57) Rhodes, K. J.; Meisner, R.; Kirkham, M.; Dudney, N.; Daniel, C. J. Electrochem. Soc. 2012, 159, A294. (58) Hewitt, K. C.; Beaulieu, L. Y.; Dahn, J. R. J. Electrochem. Soc. 2001, 148, A402. (59) Wang, J.; Raistrick, I. D.; Huggins, R. A. J. Electrochem. Soc. 1986, 133, 457.

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Author Contributions §

M.H. and K.K. contributed equally to this work. The manuscript was prepared through the contribution of all coauthors. All authors have given approval to the final version of the manuscript.

Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS This work was financially supported by the Swiss National Science Foundation (Project 200021_140245), by Swiss Federal Commission for Technology and Innovation (CTIProjekt Nr. 14698.2 PFIW-IW) and partially by the European Union through the FP7 (ERC Starting Grant NANOSOLID, contract number 306733). We thank Dr. Maksym Yarema for helpful discussions and Dr. Frank Krumeich for high-resolution TEM images. Electron microscopy was performed at ETH Zürich Electron Microscopy Center and at Empa Electron Microscopy Center.



REFERENCES

(1) Palacin, M. R. Chem. Soc. Rev. 2009, 38, 2565. (2) Goodenough, J. B.; Kim, Y. Chem. Mater. 2009, 22, 587. (3) Hayner, C. M.; Zhao, X.; Kung, H. H. Annu. Rev. Chem. Biomol. Eng. 2012, 3, 445. (4) Park, C.-M.; Kim, J.-H.; Kim, H.; Sohn, H.-J. Chem. Soc. Rev. 2010, 39, 3115. (5) Zhu, Y.; Han, X.; Xu, Y.; Liu, Y.; Zheng, S.; Xu, K.; Hu, L.; Wang, C. ACS Nano 2013, 7, 6378. (6) Magasinski, A.; Dixon, P.; Hertzberg, B.; Kvit, A.; Ayala, J.; Yushin, G. Nat. Mater. 2010, 9, 353. (7) Chockla, A. M.; Klavetter, K. C.; Mullins, C. B.; Korgel, B. A. Chem. Mater. 2012, 24, 3738. (8) Kovalenko, I.; Zdyrko, B.; Magasinski, A.; Hertzberg, B.; Milicev, Z.; Burtovyy, R.; Luzinov, I.; Yushin, G. Science 2011, 333, 75. (9) Beattie, S. D.; Larcher, D.; Morcrette, M.; Simon, B.; Tarascon, J. M. J. Electrochem. Soc. 2008, 155, 158. (10) Chan, C. K.; Peng, H.; Liu, G.; McIlwrath, K.; Zhang, X. F.; Huggins, R. A.; Cui, Y. Nat. Nanotechnol. 2008, 3, 31. (11) Mosby, J. M.; Prieto, A. L. J. Am. Chem. Soc. 2008, 130, 10656. (12) Alcántara, R.; Jiménez-Mateos, J. M.; Lavela, P.; Tirado, J. L. Electrochem. Commun. 2001, 3, 639. (13) Klavetter, K. C.; Wood, S. M.; Lin, Y.-M.; Snider, J. L.; Davy, N. C.; Chockla, A. M.; Romanovicz, D. K.; Korgel, B. A.; Lee, J.-W.; Heller, A.; Mullins, C. B. J. Power Sources 2013, 238, 123. (14) Komaba, S.; Matsuura, Y.; Ishikawa, T.; Yabuuchi, N.; Murata, W.; Kuze, S. Electrochem. Commun. 2012, 21, 65. (15) Ge, P.; Fouletier, M. Solid State Ionics 1988, 28−30 (Part 2), 1172. (16) Cao, Y.; Xiao, L.; Wang, W.; Choi, D.; Nie, Z.; Yu, J.; Saraf, L. V.; Yang, Z.; Liu, J. Adv. Mater. 2011, 23, 3155. (17) Wang, L.; Lu, Y.; Liu, J.; Xu, M.; Cheng, J.; Zhang, D.; Goodenough, J. B. Angew. Chem., Int. Ed. 2013, 52, 1964. (18) Park, C.-M.; Yoon, S.; Lee, S.-I.; Kim, J.-H.; Jung, J.-H.; Sohn, H.-J. J. Electrochem. Soc. 2007, 154, A917. (19) Caballero, Á .; Morales, J.; Sánchez, L. J. Power Sources 2008, 175, 553. (20) Sung, J. H.; Park, C.-M. J. Electroanal. Chem. 2013, 700, 12. (21) Park, C.-M.; Sohn, H.-J. Chem. Mater. 2008, 20, 3169. (22) Darwiche, A.; Marino, C.; Sougrati, M. T.; Fraisse, B.; Stievano, L.; Monconduit, L. J. Am. Chem. Soc. 2012, 134, 20805. (23) Baggetto, L.; Ganesh, P.; Sun, C.-N.; Meisner, R. A.; Zawodzinski, T. A.; Veith, G. M. J. Mater. Chem. A 2013, 1, 7985. 1261

dx.doi.org/10.1021/nl404165c | Nano Lett. 2014, 14, 1255−1262

Nano Letters

Letter

(60) Kim, S.-W.; Seo, D.-H.; Ma, X.; Ceder, G.; Kang, K. Adv. Energy Mater. 2012, 2, 710. (61) Hong, S. Y.; Kim, Y.; Park, Y.; Choi, A.; Choi, N.-S.; Lee, K. T. Energy Environ. Sci. 2013, 6, 2067. (62) Darwiche, A.; Marino, C.; Sougrati, M. T.; Fraisse, B.; Stievano, L.; Monconduit, L. J. Am. Chem. Soc. 2013, 135, 10179.

1262

dx.doi.org/10.1021/nl404165c | Nano Lett. 2014, 14, 1255−1262