The influence of nanostructure size on V₂O₅ electrochemical properties as cathode materials for lithium ion batteries

Abstract

In this paper, V₂O₅ nanostructures with a size depending on the annealing temperature are successfully synthesized by a sol–gel method. The crystal structure and morphology of the samples are characterized by X-ray diffraction (XRD), X-ray photoelectron spectroscopy (XPS), transmission electron microscopy (TEM), selected area electron diffraction (SEAD) and scanning electron microscopy (SEM), respectively. Electrochemical testing such as discharge–charge cycling (CD) and cyclic voltammetry (CV) and electrochemical impedance spectroscopy (EIS) are employed in evaluating their electrochemical properties as cathode materials for lithium ion batteries. One-dimensional nanostructures are successfully synthesized with the same structure, composition and similar shape. The results reveal that for one-dimensional nanostructures, next to the thickness which must be as small as possible, the length of the nanocrystals is crucial and should be above 2 μm. The longer nanostructures obtained at 650 °C deliver a discharge specific capacity of 281 mA h g⁻¹ at a current rate of C/5 which is over 95.5% of the theoretical capacity for two Li⁺ ion intercalation (294 mA h g⁻¹) within a voltage window of 2.0–4.0 V.

---

Introduction

Lithium-ion (Li-ion) batteries since their commercial introduction in the early 1990s¹ have become a widely used as rechargeable power sources for consumer electronics such as laptop computers, digital cameras and cellular phones. However, the development of electric vehicles (EVs) and hybrid electric vehicles (HEVs) during the past few years requires Li-ion batteries with higher energy density, long cycle life and high safety.^2,3 Therefore, the future use of electrical energy depends on the development of the next generation batteries. In particular, the search for new cathode materials is essential. Lithium cobalt oxide (LiCoO₂, LCO), which is currently used in commercial lithium ion batteries (LIBs), suffers from safety issues and being expensive and very toxic.^4,5 Among the alternative cathode material for wide application in Li-ion batteries seems to be vanadium pentoxide (V₂O₅). It is much less toxic and while maintaining precautions it is more environmentally.⁶ Additional advantages of V₂O₅ are low cost and high theoretical capacity, 294 mA h g⁻¹ for 2 Li⁺ ions intercalation.^1,2 However, the application of V₂O₅ electrodes in rechargeable Li-ion batteries has been deferred by its poor structural stability, low-diffusion coefficient of lithium ions (∼10⁻¹² cm² s⁻¹) and moderate electronic conductivity (10⁻² to 10⁻³ S cm⁻¹).^7–9

In order to overcome this limitations, it is important to design and fabricate nanostructured electrode materials that provide high surface area and short diffusion paths for ionic transport and electronic conduction. In the last decade, V₂O₅ nanostructured materials exhibited improved kinetics for Li⁺ ions intercalation and de-intercalation process compared to commercial materials. For instance, Pan et al.¹⁰ reported a series of studies on nano-sized and commercial micro-sized V₂O₅ particles. The V₂O₅ nanorods increase the capacity of 2–3 times in comparison with micro-sized V₂O₅. While Guo¹¹ et al. obtained porous V₂O₅–SnO₂/CNTs composite using a hydrothermal treatment and a heat treatment in air. They showed that the rate capabilities of the composites was much better than this of the commercial V₂O₅ materials. Especially, when the current density was higher than 500 mA g⁻¹, the specific capacities of the commercial V₂O₅ were almost zero, while the specific capacities of the composites were respectively 155 mA h g⁻¹. Summarizing, nanostructured vanadium pentoxide as electrode for lithium ion batteries exhibits better cycle stability, higher specific capacities (even at higher current density) and structural stability in comparison with commercial V₂O₅ materials. In addition, a great advantage of vanadium pentoxide is its layered structure¹² which allows for obtaining numerous one-dimensional structures (1D) such as nanodisks, nanotubes, nanoribbons and nanorods.^13–16 1D nanostructures are usually prepared by different techniques, such as hydrothermal method, electrospinning, magnetron sputtering, controllable self-assembly method and sol–gel method.^17–22

The most favorable for preparation of 1D structures may be especially sol–gel technique. Pioneering work on the synthesis and electrochemical properties of vanadium oxide nanotubes by combination of sol–gel reaction and hydrothermal treatment of vanadium oxide precursor was carried out by Spahr et al.^14,23 Most authors focus on investigating of the effect of particle size and shape on electrode rate capability.^19,20 Nevertheless, electrochemical properties of nanostructures mostly are compared to the commercial microsized V₂O₅.^10,25 In the literature one can find the sentence that the electrochemical behavior strongly depends on the particle size: the lower the particle size, the lower the cell polarization and the higher the cell capacity.²⁴ However, this statement is valid for individual and separated nanocrystals and ignores the fact, that one-dimensional nanostructure have a tendency to agglomerate which can affect the performance of the Li-ion batteries. In the case of 3D architecture, the micrometer size particles might prevent aggregation of primary nanoparticles.^26,27 For one-dimensional nanostructures, enlargement of one of the dimensions could prevent aggregation process as in the case of 3D architecture. To our best knowledge, there is a lack of information about how the size of nanostructures influence on aggregation. Moreover, there is a lack of works about the electrochemical properties of one-dimensional nanostructures prepared by the same method, having the same shape, structure and composition but with different size. In our previous work,²⁸ we have been investigated the influence of thermal conditions on V₂O₅ nanostructures prepared by sol–gel method. We demonstrated that by varying the annealing temperature we can obtain the different sizes of V₂O₅ nanostructures.

Experimental

Synthesis

The starting solution was prepared by mixing vanadium(V) oxytripropoxide (98%, Aldrich) in an anhydrous ethyl alcohol (98%, POCH) as solvent and acetylacetone (99.7%, Aldrich). By drying the sol at 50 °C for 48 h, it transforms via gel into a xerogel powder. To obtain nanostructures the xerogel powder was subsequently annealed at temperature from range between 500 and 650 °C under air atmosphere for 10 h and then cooled to the room temperature. Based on our previous work,²⁸ to study the effect of nanostructure size on the electrochemical properties, we have chosen two annealing temperatures 550 °C and 650 °C, named in this paper as the first and the second sample respectively.

Characterization

Phase composition of manufactured samples was examined by X-ray diffraction method (XRD) by Philips X'Pert diffractometer system using CuKα radiation in range of 10–80° of 2θ at room temperature. XPS analyses were carried out with X-ray photoelectron spectrometer (Omicron NanoTechnology) with 128-channel collector. The measurements were performed at room temperature in a ultra-high vacuum conditions (below 1.1 × 10⁻⁸ mBar). The photoelectrons were excited by an Mg-Kα X-ray source. The X-ray anode was operated at 15 keV and 150 W. Omicron Argus hemispherical electron analyzer with round aperture of 4 mm was used for analyzing of emitted photoelectrons. The binding energies were corrected using the background C1s line (285.0 eV) as a reference.²⁹ XPS spectra were analyzed with Casa-XPS software using a Shirley background subtraction and Gaussian–Lorentzian curve as a fitting algorithm. The surface morphology and fine grain structure of the samples were studied also at room temperature by a FEI Company Quanta FEG 250 scanning electron microscope (SEM) and a FEI TECNAI G2 F20 transition electron microscopy (TEM).

For the electrochemical measurement, the working cathode was composed of 70 wt% V₂O₅ powder, 20 wt% carbon black as conducting agent and 10 wt% vinylidene fluoride (PVDF) as binder. After being blended in N-methylpyrrolidinone, the mixed slurry was spread on a thin aluminum foil and dried at 80 °C for 24 h, then cut into discs. The discs were dried subsequently at 80 °C in vacuum before use. The cells were assembled in a glove box (MBraun, Germany) filled with ultrahigh purity argon. As a separator quartz filter paper MN GF-2 (Macherey-Nagel GmbH & Co. KG, Germany) was used. As an electrolyte 1 M LiPF₆ dissolved in ethylene carbonate (EC) and dimethyl carbonate of 1 : 1 weight ratio (LP30, Merck KGaA, Germany) was applied. All the electrochemical measurements were performed in two electrode Swagelok® type cells with active material as a working electrode and lithium foil (99.9% purity, 0.75 mm thickness, Alfa Aesar, Germany) as a counter/reference electrode. Galvanostatic charging (C) and discharging (D) were performed between 4.0 and 2.0 V at different current rates (C = D, C/10 = 29.4 mA g⁻¹, C/5 = 58.8 mA g⁻¹, C/2 = 147 mA g⁻¹, C = 294 mA g⁻¹, 2C = 588 mA g⁻¹, 5C = 1470 mA g⁻¹), using Atlas-Sollich 0961 (Atlas-Sollich, Poland). Cyclic voltammetry and electrochemical impedance spectroscopy (EIS) were measured using the AUTOLAB 302N potentiostat–galvanostat (AUTOLAB, Eco Chemie, B.V., Netherlands) under Nova software. Specific capacity is calculated based on the weight of the active materials.

Results and discussion

Structure and morphology

To investigate the phase purity and crystallinity after annealing of the samples, X-ray diffraction patterns were recorded. Fig. 1 illustrates the XRD pattern (10–80° of 2θ) of V₂O₅ nanostructures obtained by annealing of the xerogel powders at 550 °C and 650 °C. All observed diffraction peaks for both samples, match very well with the JCPDS card no. 41-1426 which corresponds to the orthorhombic structure with a space group of Pmmn (no. 59) with a lattice parameter values of a = 11.51 Å, b = 3.565 Å and c = 4.372 Å. Further, no signals of others impurity phases are detected, indicating that high purity of V₂O₅ products is obtained. XRD patterns reveal that the intensity of diffraction peak at 20.35° corresponding to the (001) crystal plane increases when the temperature rises from 550 °C up to 650 °C, thus indicating that the crystallization is enhanced and the crystallite size increased.

Fig. 1 XRD patterns of the xerogel powders after heat treatment for 10 h at 550 °C and 650 °C.

To investigate the chemical composition and the valence analysis of vanadium oxide nanoparticles the XPS measurements were performed.

The spectrum for V2p and O1s regions are given in Fig. 2a and b for V₂O₅ nanostructures obtained at 550 °C and 650 °C respectively. For both samples, the most intense peak at 530.24 eV corresponds to the O1s and the less intense peak at 524.94 eV is attributed to V2p_1/2, while the other intense peak around 517 eV is attributed to V2p_3/2. The V2p_3/2 peak for both samples shows up asymmetry shaped with a very weak shoulder line on the lower binding energy side. To determine the nature and oxidation state, the V2p_3/2 peak was convoluted and fitted by two Gaussian–Lorentzian curve. As shown in Fig. 2a and b, two different components of V2p_3/2 are apparent. The strong peak around 517 eV is associated with V⁵⁺ in V₂O₅ while the weak peak around 516 eV is assigned to V⁴⁺ in VO₂. Convolution was also done for V2p_1/2 peak and components corresponding to V⁵⁺ and V⁴⁺ are also present. The relative concentration of V₂O₅ and VO₂ in the both samples (in percentage) calculated based on the peak areas are roughly 94% and 6% respectively. Literature data confirm that depending on the preparation method, the amount of V⁴⁺ can reach up to 10% of the total amount of vanadium ions, especially when the synthesis is performed in an organic solvent.^7,30

Fig. 2 XPS spectra: O1s and V2p regions of V₂O₅ nanorods obtained at (a) 550 °C and (b) 650 °C.

The representative SEM images of the samples that were collected after annealing at 550 °C and 650 °C are shown on Fig. 3a and b respectively. Our previous work²⁸ has shown that an initial growth of V₂O₅ nanostructures starts at 500 °C and the dimension of the nanocrystals are dependent on the annealing temperature. Nanocrystals seen in Fig. 3a have uniform size, the width of crystals ranges between 200 nm and 300 nm, the length of 500–1500 nm and the thickness of nanocrystals ranges from 100 nm to 200 nm. With annealing temperature increasing up to 650 °C, the V₂O₅ nanocrystals become longer and wider (Fig. 3b). The width of crystals is range between 500 nm and 1500 nm, the length of 2–4 μm and the thickness of nanocrystals about 300 nm.

Fig. 3 SEM images of the nanostructures obtained at (a) 550 °C and (b) 650 °C, (c) TEM image of an elongated nanocrystal obtained at 650 °C, (d) SEAD pattern obtained from the region highlighted by the square in (c).

The atomic structure of individual nanocrystal was studied by transition electron microscopy (TEM) and selected area electron diffraction (SEAD). In order to obtain the best possible resolution for measuring, one of the smallest nanocrystals obtained at 650 °C was selected. Fig. 3c presents TEM image of this nanocrystal. The diffraction pattern obtained from the region highlighted by the square in Fig. 3c is presented in Fig. 3d and is indexed as orthorhombic V₂O₅ on a [001] zone axis. The pattern indicates that nanocrystals grown with the length along the [010] crystallographic direction (b-axis) and the width along the [100] crystallographic direction (a-axis). The surface of the nanocrystals is mainly formed by the (001) atomic planes. The orthorhombic crystal structure of V₂O₅ can be described as layers of VO₅ square pyramids that share edges and corners. The crystal plane of (001) direction corresponds to the sixth V–O bond in the c-direction which consists of weak electrostatic interaction and permits thus the facilitate insertion of various ions and molecules between the layers. Thus Li-ion insertion and electronic transport occur more easily along the a–b plane rather than through the layers of the c-axis.^15,31,32

In conclusion, to investigate V₂O₅ electrochemical properties, nanostructures with the same structure and composition were manufactured nanostructures in the sample annealed at 550 °C have rod-like structure, while in the sample annealed at 650 °C, nanorods were obtained. It can be stated that, nanostructures have similar shape but they differ in sizes.

Electrochemical performance

The electrochemical performance of V₂O₅ nanostructures as cathode materials for Li-ion batteries was investigated by cyclic voltammetry (CV). Fig. 4 shows the first thirty cyclic voltammetry curves of the V₂O₅ nanostructures obtained at 550 °C and 650 °C (Fig. 4a and b respectively). The measurements were carried out in a voltage range from 2.0 V to 4.0 V vs. Li/Li⁺ at a scan rate of 1 mV s⁻¹. Both materials exhibit five reduction peaks during the first cathodic scan (Fig. 4a and b). The peaks for the first sample (annealed at 550 °C) are located at 3.44 V (with shoulder at 3.54 V), 3.30 V, 3.10 V, 2.42 V and 2.19 V (with shoulder at 2.34 V) vs. Li/Li⁺, while for the second sample (650 °C), the peaks are located at 3.45 V (with shoulder at 3.55 V), 3.33 V, 3.12 V, 2.39 V and 2.21 V (with shoulder at 2.31 V) vs. Li/Li⁺. Odani et al.³⁴ reported that, when the electrodes are cycled in the 2.0–4.0 V potential range, an irreversible phase transition of the Li_xV₂O₅ occurs and this reaction process is characterized by five to six sets of CV peaks, part of which appear as shoulders. Additionally, the shoulders indicate that lithium ions are inserted into two kinds of intercalation sites in V₂O₅.⁹ The main reduction peaks located at 3.30 V, 3.10 V and 2.19 V vs. Li/Li⁺ for the first sample and located at 3.33 V, 3.12 V and 2.21 V vs. Li/Li⁺ for the second sample are typical reduction peaks of crystalline V₂O₅ and indicate a multi-step lithium ion intercalation process, and the phase change from α-V₂O₅ to ε-Li_0.5V₂O₅, δ-LiV₂O₅ and γ-Li₂V₂O₅ consequently (see Table 1). The reversible chemical reactions can be then described as:^10,20,35

V₂O₅ + 0.5Li⁺ + 0.5e ↔ Li_0.5V₂O₅ (1)

Li_0.5V₂O₅ + 0.5Li⁺ + 0.5e ↔ LiV₂O₅ (2)

LiV₂O₅ + 1Li⁺ + 1e ↔ Li₂V₂O₅ (3)

Fig. 4 The first thirty voltammetric cycles of the V₂O₅ nanostructures obtained at (a) 550 °C and (b) 650 °C; scan rate 1 mV s⁻¹.

Table 1 Comparison of CV peak position with the earlier literature^{2,10,20,25,33}

Investigated voltage range (V)	2.0–4.0		2.0–4.0	2.5–4.0	2.0–4.0	2.5–4.0	2.0–4.0
Phase	CV peak positions (cathodic scan) (V)/2, 10, 20, 25 and 33						CV peak positions (cathodic scan) (V)
							Sample V2O5 annealed at 550 °C	Sample V2O5 annealed at 650 °C
ε-Li0.5V2O5	3.35/2	3.32/10		3.4/20	3.33/25	3.38/33	3.30	3.33
δ-LiV2O5	3.15/2	3.12/10		3.2/20	3.13/25	3.17/33	3.10	3.12
γ-Li2V2O5	2.25/2	2.17/10			2.20/25		2.19	2.21

---

In the following first anodic scan, three anodic peaks for both samples are observed. Peak positions for the first sample are located at 2.59 V (with shoulder at 2.71 V), 3.45 V and 3.55 V (with shoulder at 3.71 V) vs. Li/Li⁺ while for the second sample are located at 2.56 V (with shoulder at 2.70 V), 3.55 V and 3.66 V (with shoulder at 3.66 V) vs. Li/Li⁺. In both cases, the first peak and the latter two peaks were ascribed to the deintercalation of the second and the first lithium ions respectively.^2,10,20 However, as can be seen in Fig. 4a and b, in the subsequent cycles the intensities and the positions of the peaks are changing gradually.

During the cathodic scans, the intensities of the two main peaks, ascribed to the intercalation of the first lithium ion, are decreasing. For the first sample, the peak located at 3.30 V disappeared after 15 cycles (Fig. S1 in the ESI†) while the peak located at 3.10 V, also decreasing, is still visible after thirty cycles. In the case of the second sample, during the subsequent cycles, the cathodic peak corresponding to phase transformation from α-V₂O₅ to ε-Li_0.5V₂O₅ overlaps gradually with the peak located at 3.45 V (with shoulder at 3.55 V), whereas, the peak located at 3.13 V disappears in the fifteenth cycle (Fig. S2 in the ESI†). In the case of the peak corresponding to the phase transformation from δ-LiV₂O₅ to γ-Li₂V₂O₅ for both samples, the shoulder of this peak begins to dominate with the subsequent cycles. This could mean, that lithium ions change the intercalation site during the next cycles. In the following anodic scans in both cases, the intensities and the peak positions are also changing, similarly to the cathodic peaks. The change in shape of the voltammogram with cycle number, for both samples, indicates a change of the material during cycling – the material loses electroactivity to some extent. The most likely reason is the amorphization and/or other structural changes of the material upon cycling.¹⁴ It was reported that γ-Li₂V₂O₅ phase transition related to the intercalation of the second lithium ion is not as reversible as the ε-Li_0.5V₂O₅ and δ-LiV₂O₅ phase transitions for the first lithium ion intercalation/deintercalation, according to eqn (1) and (2).

Additionally the multi-step transformation may destroy the crystal structure of V₂O₅ nanoparticles and increase cell resistance.³⁶ However, it may be noted that the structure of the sample obtained at 650 °C is stable after fifteen cycles and the shape of the voltammogram does not change any more, while for the sample obtained at 550 °C the shape of the voltammogram is still changing up to thirtieth cycle.

The galvanostatic charge–discharge technique was used to study the cycling performance of the V₂O₅ nanostructures. The first thirty charge–discharge curves for V₂O₅ nanostructures obtained at 550 °C and 650 °C at a rate of C/5 (1C = 294 mA g⁻¹) are shown in Fig. 5a and b, respectively. For both samples, the discharge curves show clear voltage plateaux at the potentials corresponding well to the peak positions at the CV curves. The initial discharge capacity is equal to 135 mA h g⁻¹ for the first sample (Fig. 5a), when for the second sample (Fig. 5b), the initial discharge capacity is more than two times higher (281 mA h g⁻¹), which is over 95.5% of the theoretical capacity for two Li⁺ ions intercalation (294 mA h g⁻¹). The initial charge are 117 mA h g⁻¹ and 240 mA h g⁻¹ for the first and the second sample, respectively. In both cases, the generally overlapping profiles of first thirty cycles indicate a good reversibility of the electrochemical process upon cycling.

Fig. 5 Charge/discharge curves of the cell with V₂O₅ nanostructures obtained at (a) 550 °C and (b) 650 °C at a rate of C/5 (1C = 294 mA h g⁻¹).

Fig. 6 shows the rate capability of V₂O₅ nanostructures obtained at 550 °C and 650 °C. A specific discharge capacity of 258 mA h g⁻¹ at C/10 rate (29.4 mA g⁻¹) can be obtained for the sample annealed at 550 °C. The capacity values decrease upon increasing polarization currents and are equal to 211, 187, 169, 155, 128 mA h g⁻¹ at C/5, C/2, C, 2C and 5C rate, respectively. However, for the sample annealed at 650 °C, the specific discharge capacities are about 20–30% higher than for the first sample and are equal to 300, 227, 211, 187, 170 and 142 mA h g⁻¹ at C/10, C/5, C/2, C, 2C and 5C respectively.

Fig. 6 Rate performance of V₂O₅ nanostructures obtained at 550 °C and 650 °C in the voltage range between 2.0 V and 4.0 V vs. Li/Li⁺. Here 1C is 294 mA h g⁻¹.

The cycling performances of the V₂O₅ nanostructures obtained at 550 °C and 650 °C recorded in the voltage range between 2.0 V and 4.0 V vs. Li/Li⁺ at the rate of C/5 and 2C are presented in the Fig. 7a and b, respectively. The initial discharge capacity of the first sample (obtained at 550 °C) is 135 mA h g⁻¹ but after 100 cycles the discharge capacity drops to 33 mA h g⁻¹ and the retention is only about 24%. Moreover, the initial coulombic efficiency is 87% and after 15 cycles increases to 96%. One can see, that the capacity fading rate is non-uniform and for the first 15 cycles is about 1% per cycle, from 15 to 80 cycles about 0.3% per cycle and about 2% per cycle above 80 cycles. For the second sample (obtained at 650 °C), the initial discharge capacity is 281 mA h g⁻¹ and after 100 cycles the discharge capacity drops to 87 mA h g⁻¹ which corresponds to a loss of 69%. The initial coulombic efficiency is found to be 86% and after 15 cycles increases to 97%. However, in this case, there are two capacity fading rate. For the first 15 cycles is about 1.5% per cycle and above 15 cycle is 1% per cycle. This non-uniform capacity fading rate can be assigned to the changes of the material structure during cycling as shown on voltammetry curves. Better cycles stability, for both samples, is obtained at the rate of 2C. The initial discharge capacity for the first sample is 212 mA h g⁻¹ and after 100 cycles the discharge capacity drops to 121 mA h g⁻¹ and the retention is about 57%. The initial coulombic efficiency is 95% and after 8 cycles increases to 98%. The capacity fading rate above 5th cycle is equal to approx. 0.6%. In the case of the second sample, the electrode exhibits increasing capacity over the first 5 cycles, which may be attributed to some long-term activation during charge/discharge cycling. After the first 6 cycles, the specific capacity of the nanostructures obtained at 650 °C (203 mA h g⁻¹) begins to decline and is equal to 170 mA h g⁻¹ after 100 cycles, which corresponds to a loss of 27%. The initial coulombic efficiency is 85% but after 6 cycles increases to 99%. The specific fading rate after 6 cycles is about 0.2%. For both samples, after 100 cycles the current rate was reset to C/5. Smaller discharge current values give rise to the capacity value of the sample annealed at 650 °C (232 mA h g⁻¹) but does not influence much the capacity of the sample annealed at 550 °C (125 mA h g⁻¹), what indicates better stability of the second sample.

Fig. 7 Residual discharge capacity versus cycle number of the V₂O₅ nanostructures obtained at 550 °C and 650 °C at the rate of (a) C/5 and (b) 2C (1C = 294 mA h g⁻¹) in the voltage range between 2.0 V and 4.0 V vs. Li/Li⁺.

It is worth noticing that the nanostructures obtained at 650 °C exhibit better initial discharge capacity and better discharge capacity retention. In the literature one can find several references about influence of morphology and, indirectly, size of nanostructures on their electrochemical properties. For example, Wang et al.²⁰ synthesized various vanadium pentoxide nanostructures such as porous nanotubes, hierarchical nanofibers and single crystalline nanobelts by using electrospinning technique and subsequent annealing. They have shown, that the single-crystalline V₂O₅ nanobelts with diameter about 300 nm and length of 11 μm, exhibited superior cyclic-capacity retention over the other investigated nanostructures. These phenomena they attribute to that the single crystalline nanobelts could efficiently inhibit the volume changes and aggregation of the V₂O₅ electrode materials, thus retaining their stability to some extent during the lithiation and delithiation processes. Mai et al.³⁷ obtained, using electrospinning method combined with annealing, hierarchical vanadium oxide nanowires constructed from attached single-crystalline vanadium oxide nanorods with diameter of 100–200 nm and length of several millimeters. They have shown that, at a current rate of 30 mA g⁻¹, these nanostructures exhibited the initial discharge capacity 275 mA h g⁻¹ and after 50 cycles, the capacity retention was 68%. Additionally, they have examined electrochemical properties of short V₂O₅ nanorods prepared by hydrothermal method. They have found, that these nanostructures with length of 1–2 μm were easily self-aggregated together, resulting in a low discharge capacity 110–130 mA h g⁻¹. Research on influence of the morphology of V₂O₅ particles on the electrochemical behavior was carried out also by Cocciantelli et al.²⁴ They showed, that the larger particles (plate-type shape of the crystallites with size about 2–50 μm) contribute to the premature appearance of the γ-Li₂V₂O₅ and the δ-phase is completely transformed into γ-phase. On the other hand, for smaller particles (size about 2 μm), the δ → γ transformation upon intercalation occurs gradually, which is an irreversible reaction involving at least one metastable phase.

In order to check whether the larger nanostructures prevent agglomeration process, the SEM images of fresh cell for both samples were performed (Fig. S3 in the ESI†). It is evident, that in the case of first sample (nanostructures obtained at 550 °C) the agglomeration of nanoparticles is large and heterogeneous. For the second sample (nanostructures obtained at 650 °C) the agglomeration is small. Additionally, to check the structural stability and agglomeration after cycling, the SEM images for both samples after 15 cycles were carried out. After cycles, for both samples obtained at 550 °C and obtained at 650 °C, the degree of agglomeration has not changed (Fig. S5 and S6 respectively in the ESI†). However, for both samples, after cycling the structural changes are visible. The nanostructures start stratification along the (001) plane. Nevertheless, this structural changes are smaller and less frequent for longer nanoparticles obtained at 650 °C and reach a maximum of 20% of the length. In the case nanostructures obtained at 550 °C, this changes are more frequent and reach even 80% of the length.

In our work, for the second sample, where the length of the nanocrystal is about 2–4 μm, the initial discharge capacity and discharge capacity retention are higher than for the first sample, where the nanocrystals have a length of 500–1500 nm. On the basis of the mentioned works and our analyses, one may conclude that for one-dimensional nanostructures the length of nanocrystals is crucial and should be above 2 μm. The longer nanostructures limit the aggregation process which influences the initial discharge capacity and capacity retention. It should be mentioned that the sample annealed at 650 °C shows better cyclic stability for all investigated current rates. This result can be related to the limited ion intercalation for the second lithium ion at 2.21 V vs. Li/Li⁺.¹⁰ Additionally, in the case of the longer particles, the δ-phase is completely transformed into γ-phase without irreversible reaction involving at least one metastable phase. The improved discharge capacity retention for both samples at the rate of 2C compared to the lower current rate (C/5) may be attributed to less irreversible side reactions of slower kinetics. Furthermore, when current rate is reset to C/5 after 100 cycles, nanostructures obtained at 650 °C deliver almost the same discharge capacity as after 5 initial cycles at the C/10 rate (232 mA h g⁻¹ (Fig. 7) compared to 227 mA h g⁻¹ (Fig. 6)), what proves the stability of this sample regardless the changes of the polarization rates. Contrary, the smaller nanostructures obtained at 550 °C, after 100 cycles recorded at 2C do not retrieve the capacity values obtained at C/5 rate during first few cycles (125 mA h g⁻¹ (Fig. 7) compared to 187 mA h g⁻¹ (Fig. 6)). In addition, heterogeneous agglomeration of these nanostructures obtained at 550 °C affects of the electrochemical performance. For example, capacitance performance of these nanostructures may vary slightly depending on the agglomeration degree of active material. It should be mentioned, that the most of the known results from the literature have been obtained at a low discharge current density (usually less than 50 mA g⁻¹). For example, Ng et al.³⁸ synthesized one-dimensional nanostructures by precipitation process. They received capacity retention of 74% after 50 cycles. Liang et al.²⁵ received even higher capacity retention of 90% after 50 cycles for nanosheets prepared by sol–gel method. In both cases, used current density was equal to 50 mA g⁻¹. In the case of this work, the V₂O₅ nanostructures obtained at 650 °C exhibit capacity retention of 90% after 50 cycles at the discharge current density as high as 588 mA g⁻¹.

Electrochemical impedance spectra (EIS) for coin-cells (surface area 0.2 cm²) with different sizes (obtained at 550 °C and 650 °C) V₂O₅ nanostructures were recorded in the frequency range 100 kHz to 10 mHz. The measurement were performed for fresh cell and after 2, 10 and 15 cycles. The exemplary of the impedance spectra for fresh cell and after 10 cycles with equivalent electrical circuit and fitted data are shown in Fig. 8. The impedance spectra were analyzed using equivalent circuits with combination of R_e, R_sf‖CPE_sf, R_b‖CPE_b, R_ct‖CPE_ct and W, the Warburg finite length (open circuit terminus). At high characteristic frequency beginning of semicircle extended to Z′ axis is reflected by an ohmic resistance (R_e) which includes ionic resistance from the separator and resistance of the electrolyte. The R_sf and C_sf are resistance and capacitance of the solid-state interface layer formed on the surface of the electrodes, which correspond to the semicircle at high frequencies; R_b is bulk resistance of the cell, which reflects electric conductivity of the electrolyte, separator, and electrodes; R_ct and C_dl are faradic charge-transfer resistance and its relative double-layer capacitance, which correspond to the semicircle at medium frequencies; W is the Warburg impedance related to a combination of the diffusional effects of lithium ion on the interface between the active material particles and electrolyte, which is generally indicated by a straight sloping line at low frequency end.^39,40 In all measured cells constant value of ohmic resistance, R_e ∼ 15 Ω arising from the resistance of electrolyte was observed. The R_sf (resistance associated with high frequency semicircle) for fresh cell are changing from 70 to 150 Ω. The value of this resistance are depending on the agglomeration degree. For samples obtained at 550 °C (first sample) where the agglomeration is heterogeneous, the R_sf resistance are more variable than for samples obtained at 650 °C (second sample). After cycling, for both type of samples R_sf are slightly decreasing. The capacitance of this process is increasing during cycling from 4 μF to 6–8 μF. The charge transfer resistance (R_ct) show almost constant value of about 250–300 Ω, and corresponding capacitance (C_dl) was 2–3 mF. For the first sample the bulk resistance R_b, is growing twice during cycling from about 40 Ω to about 80 Ω. On the other hand, in the case of the second sample, these resistance change slightly from 30 Ω to 40 Ω. The bulk capacitance, before and after cycling for both sample has constant value around 0.5 μF. On the basis of the EIS analysis, one may conclude that, size of nanostructures has no significant effect on the obtained resistance values. However, significantly affecting the structural stability of nanostructures (which was seen on SEM images of cell after cycles).

Fig. 8 Nyquist plot of cell with active material obtained at 550 °C: fresh cell (green triangle) and cell after 10 cycling (blue circle). The inset shows the equivalent circuit used for fitting the impedance spectra. R_e is the resistance due to electrolyte and cell components. R_sf and CPE1 are respectively, the surface film resistance and associated capacitance (C_sf). R_b is the bulk resistance and CPE2 is associated capacitance (C_b). R_ct is the charge transfer resistance and CPE1 is double layer capacitance (C_dl), W Warburg resistance (open circuit terminus). Symbols present experimental spectra and continuous red lines represent fitted data using equivalent electrical circuit.

Conclusions

In summary, we have successfully synthesized one-dimensional V₂O₅ nanostructures by a sol–gel method. Obtained nanostructures have the same structure, composition, similar shape but they vary in size. Longer nanostructures obtained at 650 °C show better initial discharge capacity and better discharge capacity retention in comparison to the shorter nanostructures obtained at 550 °C. Additionally, for the second sample, the specific discharge capacities are about 20–30% higher than for the first sample for all current rates. Also, nanostructures synthesized at 650 °C, have better cycle stability at the high discharge current density (588 mA g⁻¹). We have shown that, as a cathode material for lithium ion batteries more suitable are one-dimensional nanostructures with the length above 2 μm. The longer nanostructures limit the aggregation process and prevent the formation of metastable phase during δ → γ phase transformation. Moreover, agglomeration degree influences more on the electrochemical performance nanostructures than their size. These results also demonstrate that nanorods obtained at 650 °C are a promising alternative cathode material for next-generation lithium-ion batteries.

References

1. M. J. Armstrong, C. O'Dwyer, W. J. Macklin and J. D. Holmes, Nano Res., 2014, 7, 1–62.
2. X. Zhang, M. Wu, S. Gao, Y. Xu, X. Cheng, H. Zhao and L. Huo, Mater. Res. Bull., 2014, 60, 659–664.
3. M.-H. Yang, B.-M. Lin, S.-F. Yeh and J.-S. Tsai, World Electric Vehicle Journal, 2008, 2, 107–112.
4. J.-M. Tarascon, Nat. Chem., 2010, 2, 510.
5. R. V. Chebiam, F. Prado and A. Manthiram, Chem. Mater., 2001, 13, 2951–2957.
6. J. L. Domingo, Biol. Trace Elem. Res., 2002, 88, 97–112.
7. J. Muster, G. T. Kim, V. Krstic, J. G. Park, Y. W. Park and S. Roth, Adv. Mater, 2000, 12, 420–424.
8. C. Ban, N. A. Chernova and M. S. Whittingham, Electrochem. Commun., 2009, 11, 522–525.
9. T. Watanabe, Y. Ikeda, T. Ono, M. Hibino, M. Hosoda, K. Sakai and T. Kudo, Solid State Ionics, 2002, 151, 313–320.
10. A. Pan, J.-G. Zhang, Z. Nie, G. Cao, B. W. Arey, G. Li, S. Liang and J. Liu, J. Mater. Chem., 2010, 20, 9193–9199.
11. Q. Guo, Z. Sun, M. Gao, Z. Tan, B. Zhang and D. S. Su, J. Energy Chem., 2013, 22, 347–355.
12. C. Delmas, H. Cognacauradou, J. Cocciantelli, M. Menetrier and J. Doumerc, Solid State Ionics, 1994, 69, 257–264.
13. L. Ren, M. Cao, S. Shi and C. Hu, Mater. Res. Bull., 2012, 47, 85–91.
14. M. E. Spahr, P. Stoschitzki-Bitterli, R. Nesper, O. Haas and P. Novák, J. Electrochem. Soc., 1999, 146, 2780–2783.
15. C. K. Chan, H. Peng, R. D. Twesten, K. Jarausch, X. F. Zhang and Y. Cui, Nano Lett., 2007, 7, 490–495.
16. D. Chen, H. Quan, S. Luo, X. Luo, F. Deng and H. Jiang, Phys. E, 2014, 56, 231–237.
17. I. Mjejri, N. Etteyeb and F. Sediri, Ceram. Int., 2014, 40, 1387–1397.
18. J. Pan, M. Li, Y. Luo, H. Wu, L. Zhong, Q. Wang and G. Li, Mater. Res. Bull., 2016, 74, 90–95.
19. C. J. Patrissi and C. R. Martin, J. Electrochem. Soc., 1999, 146, 3176–3180.
20. H. Wang, D. Ma, Y. Huang and X. Zhang, Chem.–Eur. J., 2012, 18, 8987–8993.
21. C. W. Zou, X. D. Yan, J. Han, R. Q. Chen and W. Gao, J. Phys. D: Appl. Phys., 2009, 42, 145402.
22. M. A. Zoubi, H. K. Farag and F. Endres, J. Mater. Sci., 2009, 44, 1363–1373.
23. R. Nesper, M. E. Spahr, M. Niederberger and P. Bittereli, Int. Pat. Appl. PCT/CH97/0, 1997.
24. J. M. Cocciantelli, M. Menetrier, C. Delmas, J. P. Doumerc, M. Pouchard, M. Broussely and J. Labat, Solid State Ionics, 1995, 78, 143–150.
25. S. Liang, M. Qin, Y. Tang, Q. Zhang, X. Li, X. Tan and A. Pan, Met. Mater. Int., 2014, 20, 983–988.
26. H. Bin Wu, A. Pan, H. H. Hng and X. W. Lou, Adv. Funct. Mater., 2013, 23, 5669–5674.
27. Y. L. Ding, Y. Wen, C. Wu, P. A. Van Aken, J. Maier and Y. Yu, Nano Lett., 2015, 15, 1388–1394.
28. M. Prześniak-Welenc, M. Łapiński, T. Lewandowski, B. Kościelska, L. Wicikowski and W. Sadowski, J. Nanomater., 2015, 2015, 418024.
29. B. V. Crist, Handbook of Monochromatic XPS Spectra, Wiley, Chichester, 2000.
30. J. Livage, Chem. Mater., 1991, 3, 578–593.
31. R. Enjalbert and J. Galy, Acta Crystallogr., Sect. C: Cryst. Struct. Commun., 1986, 42, 1467–1469.
32. J. Haber, M. Witko and R. Tokarz, Appl. Catal., A, 1997, 157, 3–22.
33. S. Liang, Y. Hu, Z. Nie, H. Huang, T. Chen, A. Pan and G. Cao, Nano Energy, 2015, 13, 58–66.
34. A. Odani, V. G. Pol, S. V. Pol, M. Koltypin, A. Gedanken and D. Aurbach, Adv. Mater., 2006, 18, 1431–1436.
35. M. M. Rahman, A. Z. Sadek, I. Sultana, X. J. Dai, M. R. Field, D. G. McCulloch, S. B. Ponraj and Y. Chen, Nano Res., 2015, 8, 3591–3603.
36. S.-H. Ng, T. J. Patey, R. Buchel, F. Krumeich, J.-Z. Wang, H.-K. Liu, S. E. Pratsinis and P. Novak, Phys. Chem. Chem. Phys., 2009, 11, 3748–3755.
37. L. Mai, L. Xu, C. Han, X. Xu, Y. Luo, S. Zhao and Y. Zhao, Nano Lett., 2010, 10, 4750–4755.
38. S. H. Ng, S. Y. Chew, J. Wang, D. Wexler, Y. Tournayre, K. Konstantinov and H. K. Liu, J. Power Sources, 2007, 174, 1032–1035.
39. K. M. Shaju, G. V. Subba Rao and B. V. R. Chowdari, Electrochim. Acta, 2004, 49, 1565–1576.
40. S. S. Zhang, K. Xu and T. R. Jow, Electrochim. Acta, 2004, 49, 1057–1061.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra05695d

---