Effects of phosphate additives on the stability of positive electrolytes for vanadium flow batteries

Accepted Manuscript Title: Effects of phosphate additives on the stability of positive electrolytes for vanadium flow batteries Author: Cong Ding Xiao Ni Xianfeng Li Xiaoli Xi Xiuwen Han Xinhe Bao Huamin Zhang PII: DOI: Reference:

S0013-4686(15)00508-3 http://dx.doi.org/doi:10.1016/j.electacta.2015.02.187 EA 24466

To appear in:

Electrochimica Acta

Received date: Revised date: Accepted date:

16-10-2014 17-12-2014 22-2-2015

Please cite this article as: Cong Ding, Xiao Ni, Xianfeng Li, Xiaoli Xi, Xiuwen Han, Xinhe Bao, Huamin Zhang, Effects of phosphate additives on the stability of positive electrolytes for vanadium flow batteries, Electrochimica Acta http://dx.doi.org/10.1016/j.electacta.2015.02.187 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Effects of phosphate additives on the stability of positive electrolytes for vanadium flow batteries Cong Ding

a,b,1

, Xiao Ni

b,c,1

, Xianfeng Li a, Xiaoli Xi

a,b

, Xiuwen Hanc, Xinhe Bao c*, Huamin

Zhang a** a

Division of Energy Storage, Dalian National Laboratory for Clean Energy, Dalian Institute of

Chemical Physics, Chinese Academy of Sciences, No.457 Zhongshan Road, Dalian, Liaoning 116023, PR China. b

University of Chinese Academy of Sciences, No. 19A Yuquan Road,Beijing 100039 , PR China.

c

State Key Laboratory of Catalysis, Dalian Institute of Chemical Physics, Chinese Academy of

Sciences, No.457Zhongshan Road, Dalian Liaoning 116023, PR China. *Corresponding author. Tel.: +86 411 8468 6637; Fax: +86 411 84379128 **Corresponding author. Tel.: +86 411 84379072; Fax: +86 411 84665057 1

These authors contributed equally to this work.

E-mail addresses: [email protected] (X. Bao), [email protected] (H. Zhang).

Highlights

A series of phosphates is investigated as additives for vanadium flow battery. Superior V(V) thermal stability and improved electrochemical performance. Enhanced battery efficiency and slower capacity fading. Mechanism for the stabilization and performance improvement is put forward. NH4H2PO4 indicates a promising candidate for additive of the positive electrolyte.

Abstract

A series of phosphates is investigated as additives to improve the stability of the electrolyte for vanadium flow battery (VFB). Two selected additives show positive effect on the stability of

electrolytes under ex-situ stability tests and in situ flow cell experiments. The effects of additives on electrolyte are studied by Nuclear magnetic resonance (NMR), X-ray diffraction (XRD), Raman spectroscopy, Cyclic voltammetry (CV), Electrochemical impedance spectroscopy (EIS) and charge–discharge

test.

The

results

show

that

a

VFB

using

the

electrolyte

with

NH4H2PO4additivedemonstrates significantly improved redox reaction reversibility and activity, and higher energy efficiency. In addition, the cell employing the electrolyte with NH4H2PO4 exhibits a charge capacity fading rate much slower than the cell without additives during the cycling at high temperature. These results indicate that the phosphate additives are highly beneficial to improving the stability and reliability of VFB.

Keywords: Energy storage; Vanadium flow batteries; Precipitate; Phosphate additives; Capacity retention.

1. Introduction Large-scale energy storage has attracted increasing attention due to its urgent need in load leveling, uninterruptible power supply systems and renewable energy storage [1-3]. Vanadium flow batteries (VFBs) initiated by M. Skyllas-Kazacos from UNSW in 1980s [4], have been widely regarded as one of the most suitable options for large scale energy storage due to their significant advantages such as high energy efficiency (>75%), deep discharge ability, fast response, long cycle life and most importantly, independent power and energy ratings [5]. VFBs can realize a reversible conversion between electrical energy and chemical energy through the reactions of two redox couples of V2+/V3+in a negative half-cell and VO2+/VO2+in a positive half-cell. By using the same element (vanadium) in both half-cell electrolytes, VFBs overcome the inherent issue of cross contamination caused by diffusion of different ions across the ion-exchange membrane [3]. In a VFB, the electrolyte serves not only as an ion conductor but also as an energy storage medium to store and release energy [6]. However, the poor stability of the electrolytes especially low

solubility of vanadium based electrolytes has affected the final VFB performance [7]. The precipitate of the negative electrolytes at lower temperature and positive electrolytes at higher temperature, especially when the electrolytes’ concentration exceeds 2 M, has limited the energy density of VFB (≤25Whkg-1), further increased the cost of the battery system [8, 9]. Therefore the solubility and stability of electrolytes are of great significance in the development of VFB systems. Specifically, the fully charged V5+ electrolyte solution suffers from precipitation at elevated temperatures (>310 K). This poor stability is witnessed as the irreversible formation of hydrated V2O5precipitates, which may cripple the pump circulation and lead to energy loss and the final failure of the battery [7, 8, 10]. In the past years, significant efforts have been devoted to improving the stability of the positive electrolyte, aiming at developing an electrolyte with high concentration and further improving the energy density for VFB systems [8, 9, 11]. The solubility of vanadium ions can be improved via the optimization of the supporting acid electrolyte. One of the effective methods was to enhance the solubility of the electrolyte. For example, higher concentration of sulfuric acid in electrolytes can effectively stabilize V(V) ions [8]. However, increasing the H2SO4 concentration will accelerate the precipitation of V(II), V(III) and V(IV) ions due to the common ion effect [9, 12]. Furthermore, a H2SO4 concentration of 3–4 M has been found to be more suitable, considering the cost and corrosive durability for materials. In addition, the employment of mixed acid as supporting electrolyte can considerably improve the thermal stability of V(V) ions [13, 14]. The mixed acid-based vanadium flow battery has been recently reported and displayed enhanced energy efficiency and charge/discharge capacities due to a higher vanadium concentration with excellent thermal stability, however, it also requires a high concentration of the mixed acid, which may result

in the increased risk of metal corrosion [15]. Another strategy to delay the precipitation of vanadium species is to add precipitation inhibitor [16], which is one of the most economic and effective methods to stabilize the vanadium electrolytes. Normally two types of additives, e.g., inorganic/organic can be used as stabilizer for VFB electrolytes [17-24]. Alcohols with ring or chain structures can increase the solubility of V(II)–V(V) ions in the solution, stabilize the electrolyte and reduce vanadium precipitates in the electrolyte [25]. However, these organic compounds suffer from low chemical stability in the strongly oxidative V(V) solution [19]. They could participate in the electrochemical reaction of the vanadium battery and subsequently result in capacity loss. Thus inorganic additives were widely investigated, e.g., phosphate based additive ((NaPO3)6, Na3PO4, Na4P2O7) [16, 17, 19], sulfate based additives (K2SO4, Na2SO4, Al2(SO4)3) [16, 19, 20], chloride based additives (BiCl3, CoCl3) [20] and metal ions (Gr3+, In3+) [26, 27], et al. Among the reported inorganic additives, phosphate based additives are a typical kind of efficient stabilizing agents due to the interaction between vanadate and pyrophosphate or phosphate, confirming the formation of the mixed anhydrides with vanadate analogous to pyrophosphate or triphosphate by vanadium

NMR

spectroscopy [28].

Skyllas-Kazacos

proposed

sodium

hexametaphosphate ((NaPO3)6) containing six phosphate groups in a ring as precipitation inhibitors for supersaturated VOSO4solutions, presumably by adsorbing on the surface of the nuclei and reducing the rate of crystal growth [16]. Zhang and co-workers also evaluated the influence of Na3PO4as stabilizing agents on both positive and negative electrolytes, indicating outstanding thermal stability but with deteriorated capacity retention through in situ flow cell test [19]. They deduced that phosphate and polyphosphate anions may have negative effects on the stability of

vanadium solutions due to the formation of insoluble VOPO4 with V(V) ions. Recently, Park employed sodium pyrophosphate tetrabasic (Na4P2O7, SPT) in the positive electrolyte to improve long-term stability of a non-flow VFB single cell [17]. In spite of these reports, studies of stabilization mechanism are very limited and the systemically work on phosphate additives is not clear, leading to very few relevant strategies for improving the solubility and stability of electrolytes. In this paper, we report our investigation of phosphate additives as an inorganic additive in positive electrolytes for VFBs and their effects on long-term stability and electrochemical performance in detail, including electrochemical properties and battery performance evaluation. We also provide an insight into the general and elementary stabilization mechanism of the phosphate additives.

2. Experimental 2.1. Electrolyte preparation and NMR study The V(IV) electrolyte solutions were prepared by dissolving VOSO4•xH2O in sulfuric acid solutions. The V(V) electrolyte solutions were prepared electrochemically by charging the V(IV) solutions in a flow cell. At the end of the electrolysis, the concentration of final solution was determined by using an Automatic Potentiometric Titration Instrument (Titrando 905, Metrohm, Switzerland).51V and

31

P NMR spectra were recorded on a Bruker AvanceⅢ400MHz spectrometer

operating at 105.27 MHz and 162.01 MHz, respectively, equipped with a 5 mm broad band probe. The

51

V and

31

P chemical shifts were relative to VOCl3 solutions and to 85% H3PO4, respectively.

Spectral width of about 105.27 kHz and pulse width of 13 µs have been used for

51

V NMR

measurements. Spectral width of about 8.1 kHz and pulse width of 38 µs have been used for

31

P

NMR measurements. The free induction decay (FID) signals were exponentially line broadened (50

Hz for 51V and 10 Hz for 31P) prior to Fourier transformation. 2.2. Thermal stability test A certain amount of several phosphates with different structures (sodium or ammonium based normal salt, hydro phosphate, dihydric phosphate) were added into the electrolyte solutions before starting the stability tests. Each sample was sealed in a polypropylene tube and then immersed in a temperature-controlled liquid bath, using approximately 10 mL solution per sample. During the stability tests, the samples were kept static without any agitation and were monitored twice a day by naked eye for the formation of precipitation. After filtration and drying in air, the precipitation samples were completely dried in air at 520 oC for 24 h. 2.3. Crystal structures and morphology characterization of precipitation In order to analyze the composition of the red precipitates extracted from the electrolytes, X-ray diffraction (XRD) was carried out to investigate the crystal structures of precipitates. The samples were analyzed and characterized by XRD using a X'Pert Prox (PANalytical) with Cu-Kα radiation (λ= 1.5406 Å) in the 2θ range from 5 o to 50 o. A scanning electron microscope (SEM) (JEOL JCM-6000, Japan) was used to investigate the morphology of precipitates. 2.4. Raman spectroscopy Raman spectroscopy experiments of vanadium species with and without additives were carried out on Renishaw inVia Raman Microscope. The samples were placed on a glass microscope slide, being both inert and weak Raman scatter, causing little or no interference to the Raman spectrum. One drop of sample can show reasonable Raman bands of the desired intensity. These vanadium samples were excited by using Renishawin Via Raman Microscope (514 nm, green) laser, which provides a strong monochromatic beam. Each experiment was repeated at least three times to ensure the

reproducibility of the results. 2.5. Electrochemical measurements Cyclic voltammetry Cyclic voltammetry (CV) measurements of V(V) solutions with and without the additive were carried out on CHI604E electrochemical workstation (Shanghai Chenhua Instrument, China) at a scan rate of 5 mV s−1 in a potential range of 0.5～1.2 V at room temperature. The curves of current density versus potential were recorded in a three-electrode electrochemical cell with a graphite plate as a counter electrode, an aqueous saturated calomel electrode (SCE) as a reference electrode, and a freshly-polished graphite bar as a working electrode (Φ8 mm). Electrochemical impedance spectroscopy For further understanding the impact on the cell inner resistance brought by the addition of inorganic agents, electrochemical impedance spectroscopy was investigated on CHI604E electrochemical workstation (Shanghai Chenhua Instrument, China). The sinusoidal excitation voltage applied to the cells was 5 mV with a frequency range in between 0.01 Hz and 100 kHz. The initial states of charge (SOC) of electrolytes were controlled at 50%. The results were simulated and analyzed by Zsimpwin software. 2.6. VFB single-cell test A VFB single cell was fabricated by sandwiching a membrane between two carbon felt electrodes, clamped by two graphite plates. All these components were fixed between two stainless steel plates. 1.6M V3+/VO2+ in 3.0 M H2SO4solutions were employed as negative electrolytes. The positive electrolytes were prepared by dissolving various additives with certain amount into the pristine electrolyte. The electrolytes in the two half-cell were cyclically pumped into the two half cells. The

active area of the electrode was 48 cm2, and the volume of electrolyte solution was 60 mL in each half-cell. Charge–discharge cycling tests of the VFB single cell were conducted by using a multi-channel potentiost at (Model BT 2000, Arbin Instruments Corp., USA). The upper limit for the voltage of charge was 1.55 V, equivalent to the state of charge (SOC) of 65%, which prevented the corrosion of the carbon felts and graphite plates. The lower limit for the voltage of discharge was 1.0 V. The operating current density of single cell test was 80 mA cm-2. The single cell test was carried out at room temperature to investigate the effect of the selected additives on the battery performance and longevity. Then the single cells with different additives were also carried out at relatively higher temperature (in 45 oC water bath) to investigate the capacity decay of the cell.

3. Results and Discussion 3.1. 51V NMR (Nuclear Magnetic Resonance) study To validate the interaction between the phosphates with different structures (sodium or ammonium based normal salt, hydro phosphate, dihydric phosphate) and positive electrolytes, as well as to clarify the stabilization mechanism of the additives, the electrolyte with and without phosphate additives were characterized by 51V NMR to confirm the chemical environment of the vanadium ions in the electrolyte. Fig. 1(a) shows the

51

V NMR spectra of V5+ solutions with

different kinds of additives at room temperature. As can be seen from the NMR profiles, blank 51V solution NMR showed the main peak at about -551 ppm, which could be attributed to the monomer species of VO2+ ions, as previous literature reported [10]. Fig. 1(b) displayed the concrete changes of the various additives. The

51

V NMRchemical shift and line width of V5+solutions with existence of

51

V chemical shifts slightly moved to higher field, whereas the line width

gradually increased with the addition of phosphate additives. In general, the 51V chemical shift and

line width are mostly dominated by the structure of vanadium in the electrolyte solutions. Vanadium is reported to form a complex with a phosphate in the phosphate acid media [28] (as shown in Scheme 1). Compared to the additive-free electrolyte, the electrolytes with phosphates all exhibited obvious line width change, indicating a structural change probably due to the interaction between the V(V) and phosphate ions. The small differences of the chemical shifts may be caused by the dissimilar ionization degree of phosphate ions with different structure in 3 M H2SO4supporting electrolyte. Thus, no obvious difference was observed between the two kinds of dihydric phosphates (H2PO4-), when compared with the chemical shift of NaH2PO4 (-556.7 ppm) and NH4H2PO4 (-557.0 ppm). It might indicate that the types of cations had no evident impact on the chemical environment of vanadium. Fig. 1 Scheme 1.

3.2. Thermal stability To study the inhibition effect of phosphate additives on the thermal precipitation of V(V), different amounts of phosphate additives (inorganic additives containing PO43-, HPO42-, H2PO4-) with different structures were added into V(V) solutions. All the sealed samples of each solution were placed in a water bath at 50 oC. The results of thermal stability of V(V) ion are displayed in Table 1. As shown in the table, the stability of 1.6 M V(V)/3 M H2SO4 solution at 50oC was improved by adding different types of phosphate additives. Obviously, the precipitation time was delayed efficiently when the amount of the additive exceeds 0.15 M. Among all the additives based electrolytes, the solutions with 0.15 M Na2HPO4and 0.2 M NH4H2PO4 exhibited excellent stability. No precipitation was observed at 50 oC for an extended period of test time for the two kinds of

solutions. Therefore, 0.15 M Na2HPO4 and 0.2 M NH4H2PO4 were selected to further investigate the effect on the electrochemical property of the electrolyte and battery performance. Table 1

3.3. Temperature-dependent 51V NMR spectra To understand the chemistry behind the thermal stability of V5+ electrolyte affected by different kinds of additives,

51

V NMR spectra of V(V) electrolyte solution with two typical phosphate

additives at variable temperature (from 250 K to 353 K) were measured. Fig. 2 shows the temperature-dependent chemical shift and line width of 51V NMR spectra of these two solutions. The 51

V chemical shift moved to lower field when the temperature increased from 298 K to 353 K for

both two electrolyte solutions, indicating a thermally induced structural change of the vanadium species in the V5+ electrolyte solutions [7]. There was no significant difference in chemical shift between the two samples when the temperature increased. However, the line width exhibits a clear distinction for the additive of Na3PO4 and Na2HPO4. The line width of the electrolyte with Na2HPO4 almost unchanged from 298 K to 353 K, while the line width for the electrolyte with Na3PO4 changed nearly 10 kHz within the temperature range, indicating the Na2HPO4 additive prevented the V5+ species from polymerization to be V2O5 precipitation. As indicated in the thermal stability test, the electrolyte solution with 0.15 M Na3PO4 shows relatively poor thermal stability, while the solution with 0.15 M Na2HPO4has an excellent thermal stability. The differences of the

51

V NMR

spectral line width of the two phosphate additives may associate with the distinct influence on the thermal stability of V5+electrolyte. Fig. 2

3.4. Phase and morphology characterization of the precipitation

As reported, at elevated temperatures, this positively charged V(V) species would be converted to insoluble V2O5•3H2O by the following reactions[7]: [VO2(H2O)3]+ → VO(OH)3+ [H3O]+ (1) 2VO(OH)3→ V2O5•3H2O ↓

(2)

The crystal structure of precipitates was determined by XRD. Fig. 3 shows the XRD pattern of the red precipitate which was thermally treated at 520 °C for 24 h. XRD patterns with a few sharp peaks were observed. These diffraction patterns of the calcined precipitates extracted from the electrolytes with and without a phosphate additive correspond well to the orthorhombic V2O5 (α-products), suggesting that the water has been successfully removed from the red precipitate after heating. The impurity phase showed in the Fig. 3 may be ascribed to the NaV6O15 according to the standard card (JCPDS card No.00-019-1259), which probably formed by the slightly residual Na+ during the extracting and rinsing process, and then participated in the proceeding calcination procedure at 520 o

C. The morphology of the precipitation was investigated by SEM as shown in Fig. 4. Clearly,

compared with the pristine precipitates, the particle sizes of the calcined precipitates extracted from the electrolytes with phosphate additives were smaller and the particles were well-distributed, indicating a mild and even deposit process at the presence of the phosphate additives. The results also confirmed the excellent thermal stability of V(V) with the existence of reasonable amount of phosphates. Fig. 3 Fig. 4 3. 5. 31P NMR (Nuclear Magnetic Resonance) study To further validate the interaction between the vanadium ions and phosphate additives,

31

P NMR

experiments were also carried out on two groups of solutions containing the selected two additives (0.15 M Na2HPO4 and 0.2 M NH4H2PO4). Each group included three samples for comparison: merely phosphate additive solution, phosphate additive in H2SO4 electrolyte and V(V)/H2SO4 electrolytes with phosphate additive, of which chemical shifts were detected for comparison. The results are summarized in Fig. 5(a) and 5(b), respectively. Apparently, the three samples exhibited distinct chemical shifts to each other, implying the diversity of the chemical environment of the 31P. For the group of Na2HPO4 (as shown in Fig. 5 a), the main peak at about 3.2 ppm in the spectra of the sample (Ⅰ) can be ascribed to the typical peak of phosphates in aqueous solution [29].The sample (Ⅱ) showed a main peak at approximately 0 ppm, which belongs to the phosphoric acid (H3PO4) formed when the phosphate blending with the protons from the supporting electrolyte H2SO4. However, with the presence of V(V), the main peak that located between 3.2 ppm and 0 ppm was neither identical to that of the peak of phosphate nor the peak of H3PO4, confirming the interaction between the V(V) and phosphates. The results in group of NH4H2PO4 also showed the similar trend, further confirming that the phosphates can form complex with V(V) and change the chemical environment of 31P. Fig. 5

3.6. Raman spectra The vibrational Raman spectrum was carried out to further check the interaction between the additives and electrolytes, since the structure and symmetry of the complex molecule, the strength of its coordinate bonds, as well as its interaction with the environment (solvent, ions bound in the outer sphere, other molecules) can affect the Raman vibration spectrum of the V(V). The result of Raman spectra in the range of 500–1800 cm−1 is given in Fig. 6. The frequencies of V–O vibrations can be

divided into V=O stretch in VO2SO4− and VO3−(860–880 cm-1), VO2+ symmetrical stretch (935 cm−1),V=O symmetrical stretch (995 cm-1) [30]. Among these vibrations, critical structural information, pertaining to geometry and bond distance can usually be obtained from the analysis of the feature of V–O stretching. Obviously, with adding of phosphates, a sharp decrease in peak intensity (935 cm−1) was observed due to VO2+ symmetrical stretch, implying that VO2+ ions take part in forming a complex with phosphate additives. Notably, the peak intensity of the NH4H2PO4 is much lower than that of the additive of Na2HPO4, indicating a stronger interaction between the NH4H2PO4 and V(V) ions. Fig. 6

3.7. Cyclic voltammetry test Fig. 7 shows the CV behaviors of electrolytes with and without the two kinds of phosphate additives at a scan rate of 5 mV s-1. The changes of current and potential interval of anodic peak and cathodic peak are summarized in Table 2. As can be seen, two aspects could be possibly involved for the influence of Na2HPO4 and NH4H2PO4 on V(IV)/V(V) redox couple. Clearly, the ∆Ep (the difference between the oxidation and reduction peak potential) of the solutions with the two additives is much lower than that of the blank solution. The ratio values of anodic peak current to cathodic peak current (Ipa/Ipc) for V(IV)/V(V) redox reaction are closer to 1 for the electrolyte with the two additives, implying that both Na2HPO4 and NH4H2PO4 could improve the redox reversibility of V(IV)/V(V). In addition, the increase of the anodic and cathodic peak currents is observed when the two kinds of additives are added into the vanadium electrolyte, which indicates the improvement of electrochemical activity, resulting from the decrease in mass transportation polarization of the electrolyte with introduction of the additives. The results show that the electrolyte with NH4H2PO4

additive exhibits better electrochemical behavior, with a higher peak current compared to the pristine electrolyte. Fig. 7 Table 2

3.8. EIS analysis EIS was carried out to further analyze the effect of Na2HPO4 and NH4H2PO4 on the electrochemical property of electrolyte. The Nyquist plots of the electrolytes with and without additives at ambient temperature are presented in Fig. 8.A single depressed semicircle in the high frequency region and a sloped line in the low frequency region can be observed in each plot, demonstrating that the V(Ⅳ)/V(Ⅴ) redox reaction should be mix-controlled by the charge transfer reaction at the electrolyte/electrode interface and diffusion processes associated with VO2+ and VO2+ ions through the solution. Consequently, the Nyquist plots could be fitted with the simplified equivalent circuit model as shown in Fig. 8. In the equivalent circuit, R1 (contact resistance) represents the resistance composed of electrolyte resistance and the electrode resistance. CPE represents the electric double-layer capacitance of electrode/electrolyte interface [31]. R2 and W stand for the charge transfer resistance across electrode/solution interface in the electrochemical process and Warburg diffusion impedance, respectively. The EIS results are obtained by fitting the impedance plots with the equivalent circuit, as shown in Table 3. Apparently, there are no significant changes in R1 for the three samples, indicating that phosphate additives hardly influence the conductivity of electrolyte, which was confirmed by the conductivity measurement (418 ms cm-1 for the pristine, 396.7 ms cm-1 for electrolyte with Na2HPO4, and 417.3 ms cm-1 for electrolyte with NH4H2PO4). However, R2 of the electrolyte with Na2HPO4 and NH4H2PO4 is 1.833 Ω cm2 and 1.406

Ω cm2 respectively, much lower than that of the pristine electrolyte (2.827 Ω cm2). The reduced R2 implies the transfer of electron is more feasible with the addition of the two phosphates. As can be seen from the Table 3, CPE increases slightly from 1.101×10-3 F cm-2 to 1.163×10-3 F cm-2 with adding Na2HPO4 into electrolyte, implying the enhanced the electric double-layer capacitance of electrode/electrolyte interface, which may be ascribed to the buffer action of the phosphate additive [31]. It is noted that the electrolyte with NH4H2PO4 has relatively low contact resistance and the lowest charge transfer resistance, which is corresponding to the CV results. Therefore, better electrochemical performance can be expected by adding NH4H2PO4into the positive electrolyte. Fig. 8 Table 3

3.9. Charging and discharging test The charge–discharge experiments were performed in the VFB single cells using positive electrolytes with the two additives. Fig. 9(a) shows typical charge and discharge curves for a vanadium flow battery. As can be seen in the profiles, the single cell with addition of phosphate in positive electrolyte possesses lower charge voltage plateau and higher discharge voltage plateau than the blank one, indicating that the introduction of phosphates to the positive electrolyte could reduce the over-potential. The sample with NH4H2PO4 exhibits the best electrochemical property with the lowest value of charge–discharge voltage plateau interval, which indicates the improvement of electrochemical activity and kinetic reversibility. The result is consistent with the CV and EIS results. Fig. 9(b) shows the comparison of the average efficiencies of the single cells in consecutive

hundreds of cycles. In general, CE (Coulombic Efficiency), VE (Voltage Efficiency) and EE (Energy Efficiency), are used to evaluate the performance of VFBs. As an index measuring electric capacity conversion, CE intimately relates to vanadium cross-mixing rate during cycle tests. From the results, it can be confirmed that CE of the three cells is almost the same with each other, implying that the additives have no evident impact on the cross-mixing of the vanadium ions through the membrane and the side reactions. As expected, single cells with additives have a higher VE value than the phosphate-free one, owing to a lower charge transfer resistance. The energy efficiency of the battery using the electrolyte with NH4H2PO4 exhibits above 83% under a current density of 80 mA cm−2, higher than that of battery without any additive (80.5%). The single cell test proved that the additive of NH4H2PO4 shows superior performance to that of the Na2HPO4, even though the ex-situ thermal test of the Na2HPO4 was more efficient. Fig. 10 shows the energy efficiency of the single cell employing the phosphate additives in the positive electrolyte in the long-term cycling process for 300 cycles. With cycle proceeding, the energy efficiency of the single cell with addition of NH4H2PO4 keeps stable and remains above 83% at the current density of 80 mA cm−2, implying excellent cycling stability. Fig. 9 Fig. 10

3.10. Capacity decay test The charge capacity losses of the single cells operating at 45 oC are presented in Fig. 11. The key reason for the capacity fading during charge/discharge cycling includes the imbalanced vanadium active species caused by cross-contamination through the membrane as well as the loss of vanadium active species by precipitation at relatively higher temperature. Although all the cells showed a

decreasing trend in charge capacity during cycling, the capacity decay rate of the cell with adding the phosphates in electrolytes was remarkably slower than that of the blank one. After 50 cycles, the cells with phosphate additives exhibit better charge capacity retention, approximately 10% higher than that of the pristine cell (69.2%). The improved charge capacity retention with phosphate additives can be ascribed to better thermal stability of V( Ⅴ ) at 45 oC and the decreased electrochemical polarization of electrochemical reaction. Overall, the cell using NH4H2PO4 in the positive electrolyte exhibits preferable battery performance compared with that of the pristine electrolyte. Fig. 11

4. Conclusion In this work, a series of phosphates was introduced as inorganic additives to the positive electrolyte of VFBs and their effects on long-term stability and electrochemical performance of VFB were investigated. The thermal stability test showed the additives of phosphates could effectively delay the precipitate time of V(V). The improvement of the thermal stability for the electrolyte with phosphate additives could be ascribed to the formation of the complex between the additive and vanadium ion, as confirmed by the

51

V NMR and Raman measurement. The CV and EIS

measurements suggested that better electrochemical performances were achieved by adding phosphate additives. The enhanced activity of electrolyte with additives led to an improvement in VFB performance. Thus, long-term stability of the V(V) electrolyte was improved by phosphate additives, as demonstrated by capacity fading test of the cell at relatively higher temperature, which will probably help to reduce the cost of maintenance of the electrolyte in long-term cycling. Based on the above results, NH4H2PO4 appeared to be a prospective candidate as stabilizing agent for the

positive electrolyte.

Acknowledgements The authors greatly acknowledge financial support from the National Basic Research Program of China (973 program No. 2010CB227201) and the China Natural Science Foundation (No. 51361135701). The Raman work was carried out at the State Key Laboratory of Catalysis of Dalian Institute of Chemical Physics. We appreciate Professor Baokun Huang for the help with the Raman measurement. References [1] B. Dunn, H. Kamath, J.-M. Tarascon, Science 334 (2011) 928-935. [2] Z. Yang, J. Zhang, M.C. Kintner-Meyer, X. Lu, D. Choi, J. P. Lemmon, J. Liu, Chem. Rev.111 (2011) 3577-3613. [3] W. Wang, Q. Luo, B. Li, X. Wei, L. Li, Z. Yang, Adv. Funct. Mater. 23 (2013) 970-986. [4] M. Skyllas-Kazacos, F. Grossmith, Electrochimi. Sci. Technol. 134 (1987) 2950-2953. [5] M. Skyllas-Kazacos, M. Chakrabarti, S. Hajimolana, F. Mjalli, M. Saleem, J. Electrochem. Soc. 158 (2011) 55-79. [6] C. Ding, H. Zhang, X. Li, T. Liu, F. Xing, J. Phys. Chem. Lett. 4 (2013) 1281-1294. [7] M. Vijayakumar, L. Li, G. Graff, J. Liu, H. Zhang, Z. Yang, J. Z. Hu, J. Power Sources 196 (2011) 3669-3672. [8] M. Skyllas-Kazacos, C. Menictas, M. Kazacos, J. Electrochem. Soc. 143 (1996) 86-88. [9] F. Rahman, M. Skyllas-Kazacos, J. Power Sources 189 (2009) 1212-1219. [10] X. Lu, Electrochim. Acta 46 (2001) 4281-4287. [11] A. Parasuraman, T. Lim, C. Menictas, M. Skyllas-Kazacos, Electrochim. Acta 101 (2012)27-40. [12] F. Rahman, M. Skyllas-Kazacos, J. Power Sources 72 (1998) 105-110. [13] L. Li, S. Kim, W. Wang, M. Vijayakumar, Z. Nie, B. Chen, J. Zhang, G. Xia, J. Hu, G. Graff, J. Liu, Z. Yang, Adv. Energy mater. 1 (2011) 394-400. [14] S. Peng, N. Wang, X. Wu, S. Liu, D. Fang, Y. Liu, K. Huang, Int. J. Electrochem. Sci, 7 (2012) 643-649. [15] J. Lee, S. Park, Y. Cho, Y. Shul, RSC Adv. 3 (2013) 21347-21351. [16] M. Skyllas-Kazacos, C. Peng, M. Cheng, Electrochem. Solid-State Lett. 2 (1999) 121-122. [17] S.-K. Park, J. Shim, J. Yang, C.-S. Jin, B. Lee, Y.-S. Lee, K.-H. Shin, J.-D. Jeon, Electrochimi. Acta121 (2014) 321-327. [18] G. Wang, J. Chen, X. Wang, J. Tian, H. Kang, X. Zhu, Y. Zhang, X. Liu, R. Wang, J. Energy Chem. 23 (2014) 73-81. [19] J. Zhang, L. Li, Z. Nie, B. Chen, M. Vijayakumar, S. Kim, W. Wang, B. Schwenzer, J. Liu, Z. Yang, J. Appl. Electrochem. 41 (2011) 1215-1221. [20] G. Wang, J. Chen, X. Wang, J. Tian, H. Kang, X. Zhu, Y. Zhang, X. Liu, R. Wang, J. Electroanal. Chem. 709 (2013) 31-38. [21] S. Li, K. Huang, S. Liu, D. Fang, X. Wu, D. Lu, T. Wu, Electrochimi. Acta 56 (2011) 5483-5487. [22] F. Chang, C. Hu, X. Liu, L. Liu, J. Zhang, Electrochimi. Acta 60 (2012) 334-338. [23] Z. He, J. Liu, H. Han, Y. Chen, Z. Zhou, S. Zheng, W. Lu, S. Liu, Z. He, Electrochimi. Acta 106 (2013) 556-562. [24] X. Wu, S. Liu, N. Wang, S. Peng, Z. He, Electrochimi. Acta 78 (2012) 475-482. [25] M. S. Kazacos, M. Kazacos, in, Google Patents, 2000.

[26] F. Huang, Q. Zhao, C. Luo, G. Wang, K. Yan, D. Luo, Chin. Sci. Bull. 57 (2012) 4237-4243. [27] Z. He, L. Chen, Y. He, C. Chen, Y. Jiang, Z. He, S. Liu, Ionics 19 (2013) 1915-1920. [28] M. J. Gresser, A. S. Tracey, K.M. Parkinson, J. Am. Chem. Soc. 108 (1986) 6229-6234. [29] R. Ye, Y. Chen，Y. Liu，L. Zhu, Z. Li, Y. Sun, W. Pan, J. Instru. Anal. 30 (2011) 624-628. [30] N. Kausar, R. Howe, M.Skyllas-Kazacos, J. Appl. Electrochem. 31 (2001) 1327-1332. [31] S. Komaba, T. Tsuchikawa, M. Tomita, N. Yabuuchi, A. Ogata, J. Electrochemi. Soc.160 (2013) A1952-A1961.

Figure captions:

Scheme 1. The structures of V5+ complexes formed in phosphate acid media, as predicted by previous literature. Fig. 1 (a)

51

V NMR spectra of V(V) solutions with and without phosphate additives. (b) The

chemical shift and line width of 51V NMR of the electrolyte with different additives. Fig. 2. Temperature-dependent

51

V NMR spectra of 1.7 M V5+ electrolyte solution with two typical

phosphate additives: (a) chemical shift analysis; (b) line width analysis. Fig. 3. XRD patterns of precipitates extracted from the electrolytes with and without phosphate additives. Fig. 4. SEM images of precipitates extracted from the electrolytes with and without phosphate additives: (a) pristine electrolyte; (b) electrolyte with 0.15 M Na2HPO4; (c) electrolyte with 0.2 M NH4H2PO4. Fig. 5. 31P NMR spectra of a series of solutions containing phosphates of (a) 0.15 M Na2HPO4 and (b) 0.2 M NH4H2PO4. Each group consists of the following samples: (Ⅰ) phosphate; (Ⅱ) phosphate +

H2SO4; (Ⅲ) phosphate + H2SO4+ V(V). Fig. 6. Raman spectrum of V(V) electrolyte with and without phosphate additives. Fig. 7. Cyclic voltammograms for 0.05 M V(IV)/V(V)+ 3 M H2SO4 solution with different phosphate additives at scan rate of 5 mV s-1. Fig. 8. EIS spectroscopy of positive electrolyte solution with and without phosphate additives. Fig. 9 (a) Electrochemical performance of VFB employing phosphate additives in single cell at a current density of 80 mA cm-2; (b) Single cell performance of the electrolyte with and without phosphate additives at a current density of 80 mA cm-2. Fig. 10. Cycle life test of the single cells with and without phosphate additives at a current density of 80 mA cm-2 at room temperature. Fig.11. Charge capacity retention rate of VFB with and without phosphate additives at temperature of 45oC. Table 1. Effect of different amount of phosphate additives on the stability of 1.6 M V(V)/3 M H2SO4

solution at temperature of 50 oC.

Additive types

Additive amount /[M] 0 0.05 0.1 0.15 Na3PO4 17 h 35 h 41 h Na2HPO4 2h <17 h 17 h 182 h NaH2PO4 19 h 36 h 43 h NH4H2PO4 <36 h 36 h 97 h The symbol of / represents the insolubility of the additive in the electrolyte.

0.2 / 72 h 72 h 121 h

Table 2. Cyclic voltammograms data for 0.05 MV(IV)/V(V)+ 3 M H2SO4 with different phosphate additives on graphite electrode at scan rate of 5 mV s-1.

Sample

Anodic peak, Ipa[mA]

Cathodic peak, Ipc[mA]

ΔEp[mV]a

Ipa/ Ipc

Blank Na2HPO4

1.55 1.92

1.04 1.49

0.194 0.183

1.49 1.29

NH4H2PO4

2.15

1.91

0.137

1.13

a

Potential vs. SCE Table 3. Parameters resulting from fitting the impedance plots with the equivalent circuit model in Fig. 7. Sample Blank Na2HPO4 NH4H2PO4

R1[Ω cm2] 1.004 1.046 0.9676

CPE[F cm-2] 1.101×10-3 1.163×10-3 1.116×10-3

R2[Ω cm2] 2.827 1.833 1.406

W[S∙s−5∙cm−2] 0.3276 0.1994 0.2595