Highly active, bi-functional and metal-free B4C-nanoparticle-modifiedgraphite felt electrodes for vanadium redox flow batteries
H.R. Jiang, W. Shyy, M.C. Wu, L. Wei, T.S. Zhao*
Department of Mechanical and Aerospace Engineering, The Hong Kong University of Science and Technology, Clear Water Bay, Kowloon, Hong Kong, China
h i g h l i g h t s
� B4C-nanoparticle-modified graphite felt electrodes are developed for VRFBs.� First-principles study shows the central carbon atom of B4C can act as active site.� B4C shows bi-functional catalytic effect to V2þ/V3þ and VO2þ/VO2
þ redox reactions.� The battery with prepared electrodes shows enhanced performance than original ones.
a r t i c l e i n f o
Article history:Received 29 June 2017Received in revised form10 August 2017Accepted 18 August 2017
The potential of B4C as a metal-free catalyst for vanadium redox reactions is investigated by first-principles calculations. Results show that the central carbon atom of B4C can act as a highly active re-action site for redox reactions, due primarily to the abundant unpaired electrons around it. The catalyticeffect is then verified experimentally by cyclic voltammetry (CV) and electrochemical impedancespectroscopy (EIS) tests, both of which demonstrate that B4C nanoparticles can enhance the kinetics forboth V2þ/V3þ and VO2þ/VO2
þ redox reactions, indicating a bi-functional effect. The B4C-nanoparticle-modified graphite felt electrodes are finally prepared and tested in vanadium redox flow batteries(VRFBs). It is shown that the batteries with the prepared electrodes exhibit energy efficiencies of 88.9%and 80.0% at the current densities of 80 and 160 mA cm�2, which are 16.6% and 18.8% higher than thosewith the original graphite felt electrodes. With a further increase in current densities to 240 and320 mA cm�2, the batteries can still maintain energy efficiencies of 72.0% and 63.8%, respectively. Allthese results show that the B4C-nanoparticle-modified graphite felt electrode outperforms existingmetal-free catalyst modified electrodes, and thus can be promising electrodes for VRFBs.
To keep pace with the rapid development of renewable energiesbased on, e.g., solar, wind and water wave, which are fluctuated andintermittent, the development of reliable, efficient and cost-effective large-scale energy storage systems is in urgent demand[1e8]. However, the most matured electrochemical energy storagetechnology, lithium-ion battery (LIB), is not suitable for large-scaleapplication because of their high cost and safety concerns [9e20].Instead, in the past decades, redox flow batteries (RFBs) haveattracted increasing attention for large-scale energy storage due totheir advantages including high efficiency, high reliability, flexible
design, fast response and long cycle life [21e26]. Among the state-of-art RFBs, VRFBs, which adopt V2þ/V3þ and VO2þ/VO2
þ as redoxcouples, show the most promise because the same metal elementvanadium is employed in both negative and positive electrolytes,and thus the cross-contamination issuewhich greatly decreases thebattery performance is limited [27e35].
In spite of these compelling merits, the widespread applicationof VRFBs is still hindered by many issues, especially the high capitalcost. According to a recent cost and performance model, the stack-related cost is responsible to over 60% of the total cost for a 1 MW/0.25 MWh VRFB system [36]. Therefore, one effective way todecrease the high capital cost is to reduce the stack size, whichrequires to operate the VRFBs at high current densities withoutsacrificing the efficiencies. However, like what happens in otherbattery systems, the enhanced current density inevitably increases
Fig. 1. (a) The optimized atomic structure of B4C unit cell. The brown and blue ballsrepresent for boron and carbon atoms, respectively. (b) The electron density contour ofB4C (�11 0) surface. (For interpretation of the references to colour in this figure legend,the reader is referred to the web version of this article.)
H.R. Jiang et al. / Journal of Power Sources 365 (2017) 34e42 35
the cell polarization, and thus decreases the battery performance.In this regard, efforts should be taken to reduce the polarization ofVRFBs to obtain the high rate capability and high efficiencies.
As a critical component of VRFBs, the electrode provides theactive sites for redox reactions and thus plays a vitally importantrole on the battery performance. Presently, graphite felt is adoptedas a typical electrode material for VRFBs because of its wide oper-ating potentials, high corrosion resistance in acidic solution, goodelectrical conductivity, large porosity and low cost. Even though,the graphite felt electrode also suffers from crucial issues includinginferior hydrophilic and poor kinetics towards vanadium redoxreactions, leading to the poor battery performance. In this regard,the footstep trying to enhance the activity of graphite felt electrodehas never been stopped in the past decades. Generally, there aretwo widely applied approaches to prepare the high-performancegraphite felt electrode. The first approach is etching the graphitefelt surface to increase the oxygen functional groups and effectivesurface areas. For example, Chen et al. [37] used atmosphericpressure plasma jets to modify the graphite felt electrode andachieved an energy efficiency of 76% at the current density of80 mA cm�2; Kabtamu et al. [38] adopted water activation methodby exposing the electrode in water vapor for 5 min at 700 �C, andthe VRFBs with treated graphite felt electrode showed an energyefficiency of 83.10% at the current density of 50 mA cm�2; Zhanget al. [39] adopted KOH to etch the graphite felt electrode and ob-tained an energy efficiency of 64% at the high current density of250 mA cm�2. However, other works found that the excessivesurface oxidation of carbon materials would lead to the electrodecorrosion [40]. Even worse, the oxygen functional groups them-selves are not stable during cycling, leading to the loss of oxygencontent and the deteriorated battery performance [41,42]. Anotherpromising approach is depositing nanostructured electrocatalystson the graphite felt surface to increase the active sites for redoxreactions, including metal-based (metals and metal oxides) andmetal-free materials. The metal-based materials, such as Ir [43],PbO2 [44], Bi [45], Nb2O5 [46], TiC [47], TiN [48] and ZrO2 [49], aredeposited on the surface of electrode as the electrocatalysts toenhance the reaction kinetics, and thus lead to a reduced activationpolarization and increased battery performance. However, many ofthe metal-based materials trigger the undesirable side reactionsuch as hydrogen evolution reaction, which decreases the columbicefficiency and capacity retention rate during cycling. Meanwhile,the high price of metal-based materials would inevitably increasethe capital cost of the systems, limiting their further commercialapplications [50]. In this regard, the metal-free materials haveattracted great attention to be used to modify the graphite feltelectrode because of their high electrical conductivity, large specificsurface area and low cost. For example, Gonz�alez et al. [51] elec-trophoretic deposited graphene oxide on the graphite felt electrodeand the battery showed an energy efficiency of 95.8% at the currentdensity of 25 mA cm�2; He et al. [52] used carbon nanofibers tomodify the electrode and achieved an energy efficiency of 83.3% atthe current density of 30 mA cm�2; Wu et al. [53] grew N-dopedcarbon nanospheres on graphite felt electrode and the assembledbattery can be operated at the current density of 100 mA cm�2 withan energy efficiency of ~78%. Although some progresses have beenmade, the state-of-art metal-free electrocatalysts still exhibit muchlower reaction kinetics towards both V2þ/V3þ and VO2þ/VO2
þ redoxreactions than metal-based electrocatalysts, resulting in the poorerbattery performance, especially at elevated current densities.
In this work, the potential of B4C-nanoparticle-modifiedgraphite felt as a highly active, bi-functional and metal-free elec-trode for VRFBs is evaluated theoretically and experimentally. B4C, alightweight refractory material with low density, small thermalextension coefficient, high melting point, excellent resistance to
chemical attack and low cost [54,55], is adopted as a metal-freecatalyst in VRFBs for the first time. Our density functional theorybased first-principles calculations show that the electrons arenonuniform distributed around C-C bonds in B4C and the centralcarbon atom can act as a highly active reaction site for redox re-actions, primarily due to the abundant unpaired electrons aroundit. Then, CV and EIS tests demonstrate that B4C has high catalyticeffect and stability towards both V2þ/V3þ and VO2þ/VO2
þ redoxreactions, indicating a bi-functional effect. In the battery tests, theVRFBs assembled with B4C-nanoparticle-modified graphite feltelectrodes show an energy efficiency of 80.0% at the current densityof 160 mA cm�2, 18.8% higher than the original graphite felt elec-trode. In addition, the battery can also be operated at a high currentdensity of 320 mA cm�2 with an energy efficiency of 63.8%. Allthese superior results indicate that the B4C-nanoparticle-modifiedgraphite felt is a promising electrode for VRFBs.
2. Methods
2.1. Computational methods
All density functional theory (DFT) based first-principles cal-culations were carried out adopting ABINIT [56,57] code. Theexchange-correlation functional was dealt with by adoptinggeneralized gradient approximation (GGA) of Perdew-Burke-Ernzerhof (PBE) type [58]. The electron-ion interactions weremodeled by Projector-augmented-wave (PAW) potentials [59]. Toensure a satisfied convergence for wave-basis expansion, the en-ergy cutoff of 24 Ha was used. To calculate the lattice parameters ofB4C, a unit cell containing 15 atoms (12 boron atoms and 3 carbonatoms) was adopted and the Brillouin zone was sampled on a10 � 10 � 10 Monkhorst-Pack grid [60]. The convergence criteria ofthe electron self-consistent loop and structural optimization were4.0 � 10�5 Ha Bohr�1 and 4 � 10�4 Ha Bohr�1, respectively.
2.2. Electrochemical measurements
CV tests were carried out on the Autolab (PGSTAT30) worksta-tion by using a typical three-electrode electrochemical cell. The
H.R. Jiang et al. / Journal of Power Sources 365 (2017) 34e4236
glassy carbon electrodes (GCE) modified by B4C nanoparticles(Aladdin, 99%, average particle size: 50 nm) and XC-72 carbonnanoparticles (ETEK, average particle size: 20e40 nm)were used asthe working electrodes, a saturated calomel electrode (SCE) wasused as the reference electrode, and platinummeshwas used as thecounter electrode. The voltage ranges of �0.7 to 0 V (vs. SCE) and0e1.2 V (vs. SCE) were used for V2þ/V3þ and VO2þ/VO2
þ redox re-actions, respectively. The electrodes were all submerged by solu-tions containing 0.1MVO2þ þ 3MH2SO4. The EIS tests were carriedout on a potentiostat (EG&G Princeton, model 2273) using a typicalthree-electrode electrochemical cell. The graphite felt, treatedgraphite felt and B4C-nanoparticle-modified graphite felt wereused as the working electrodes, SCE was used as the referenceelectrode and platinum mesh was used as the counter electrode.The sweeping frequencies were ranging from 10 mHz to 100 kHz atthe fixed potential of �0.5 V and 0.9 V for V2þ/V3þ and VO2þ/VO2
þ
redox reactions.
2.3. Electrode preparations
The commercial graphite felt (Sigracell SGL carbon, GFA6 EA)with an uncompressed thickness of 6 mm was divided into twopieces along cross-section, washed by deionized water and dried ina vacuum oven. B4C nanoparticles and 5% Nafion solution (Dupont,
Fig. 2. CV curves of B4C nanoparticles and XC-72 carbon nanoparticles at a scan rate of 10 msweeping measurement of B4C nanoparticles at a scan rate of 10 mV s�1 with potential wi
USA) were dispersed in ethanol to form a mixed suspension, andwas sonicated for 60 min. The graphite felts were cut into propersize and the dip-withdraw-dry method [61] was repeatedly useduntil the loading of B4C nanoparticles reached 2mg cm�2. Then, theprepared electrode was dried in a vacuum oven for 12 h at thetemperature of 60 �C. For comparison, the graphite felts were alsomodified with Vulcan XC-72 carbon nanoparticles with the sameloading and procedure. The treated graphite felt was fabricated byannealing at 400 �C for 4 h under ambient air in a muffle furnacewith the heating rate of 5 �C/min.
2.4. Performance tests and materials characterizations
The battery tests were conducted using a homemade VRFBsetup with a zero-gap serpentine flow-field structure on the ArbinBT2000 (Arbin Instrument, Inc.) by adopting the charge anddischarge cutoff voltages of 0.8 and 1.65 V, respectively. Both thenegative and positive electrodes were one piece of graphite feltwith the uncompressed thickness of 3 mm and the active area of2.0 cm � 2.0 cm, which were separated by a Nafion® 212 (Dupont,USA) membrane. The compression ratio of graphite felt electrodewas set to be 67% for all the tests. 20 mL solutions containing1 M V3þ þ 3 M H2SO4 and 1 M VO2þ þ 3 M H2SO4 were used asnegative and positive electrolytes, respectively. The electrolytes
V s�1 with potential window of (a) �0.7-0 V vs. SCE and (b) 0e1.2 V vs. SCE. Repeatingndow of (c) �0.7-0 V vs. SCE and (d) 0e1.2 V vs. SCE.
Fig. 3. Nyquist plots of original graphite felt, treated graphite felt and B4C-nano-particle-modified graphite felt at the voltage of (a) �0.5 V vs. SCE and (b) 0.9V vs. SCE.
H.R. Jiang et al. / Journal of Power Sources 365 (2017) 34e42 37
were circulated by a 2-channel peristaltic pump (Longer pump,WT600-2J) with a fixed flow rate of 46mLmin�1. Before the batterytests, N2 was bubbled to exhaust the air in the electrolytes andtanks. The polarization curves were obtained by discharging for20s, resting for 5s, and charging for 20 s at the same current densityat the state of charge (SOC) of 50%. The surface morphology wasobserved by a scanning electron microscope (JEOL-6390 SEM). Thetransmission electron-microscopy (TEM) images of B4C nano-particles were acquired by a high-resolution JEOL 2010F TEM sys-tem with a LaB6 lament at 200 kV.
3. Results and discussion
The optimized atomic structures of B4C unit cell and 2 � 2 � 2supercell are shown in Fig.1a and Fig. S1. It can be seen that the unitcell of B4C shows a rhombohedral structure containing 12 boronatoms and 3 carbon atoms. The boron atoms in B4C form icosahe-dron with six each of the so-called equatorial and polar sites.Within the icosahedra, each boron atom has five nearest neighborbonds. In addition, the equatorial boron atom has one bond withthe carbon atom at the end of chain sites, and the polar boron atomhas one bond with a polar atom in neighboring icosahedron. In thechain sites, the carbon atoms at the end of the chain are fourfoldbond with three of the equatorial boron atoms in different icosa-hedra and the central carbon atom in the chain, while the carbonatom in the center of the chain is twofold bond with two carbonatoms at the end of the chain. The optimized lattice constants of B4Care a ¼ b ¼ c ¼ 5.19 Å, consistently with previous experimental andcomputational results [55,62]. Moreover, the calculated B-C and C-Cbond lengths are 1.66 Å and 1.34 Å, respectively, indicating thestructure is stiff along the chain direction. Fig. 1b shows the elec-tron density contour of B4C (�1 1 0) surface. Due to the electro-negativity of carbon is larger than that of boron, the electrons tendto accumulate around the carbon atoms, forming a high electrondensity area. It is also found that the electron density of the centralcarbon atom in the chain is larger than that of the carbon atoms atthe end of the chain, which is attributed to the fact that there areunpaired electrons in the central carbon atom. From previousworks, it is known that the nonuniform distribution of electrons isbeneficial to the redox reactions [63e69]. Therefore, it is predictedthat the B4C can effectively promote the vanadium redox reactionsand the central carbon atoms in the chain can act as the active sites.
To identify the catalytic effect of B4C experimentally, CV tests ofB4C nanoparticles and XC-72 carbon nanoparticles towards V2þ/V3þ
and VO2þ/VO2þ redox couples are carried out and compared, as
shown in Fig. 2a and b. It can be seen that there is no oxidation andreduction peaks formed on carbon nanoparticles modified glassycarbon electrode for both V2þ/V3þ and VO2þ/VO2
þ redox reactions,indicating the poor catalytic activity of carbon nanoparticles. Onthe contrary, obvious peaks with increased peak currents areobserved on B4C modified glassy carbon electrode, and the oxida-tion/reduction peak potential separations of V2þ/V3þ and VO2þ/VO2
þ redox couples are found to be 149 and 250 mV at the scan rateof 10 mV s�1, respectively, implying B4C nanoparticles have muchbetter catalytic activities than carbon nanoparticles towards vana-dium redox reactions and the bi-functional catalytic effects. Toverify the durability of B4C as the electrocatalyst for vanadiumredox reactions, repeating sweeping measurement of B4C nano-particles at the scan rate of 10 mV s�1 with potential windowof�0.7-0 V (vs. SCE) and 0e1.2 V (vs. SCE) are carried out, as shownin Fig. 2c and d. It is found that for both V2þ/V3þ and VO2þ/VO2
þ
redox couples, the CV curves exhibit good repeatability withoutobvious degradation and the peak currents remain the same valuesduring the whole test process, indicating the excellent chemicaland electrochemical stabilities of B4C towards highly reductive V2þ
and highly oxidative VO2þ.
To further investigate the electrochemical activity of B4C to-wards V2þ/V3þ and VO2þ/VO2
þ redox couples, EIS tests are con-ducted for original graphite felt, treated graphite felt and B4C-nanoparticle-modified graphite felt at the voltages of �0.5 V (vs.SCE) and 0.9 V (vs. SCE), as shown in Fig. 3a and b. The semicirclepart at high frequencies stands for the charge transfer process andthe smaller radius of the semicircle corresponds to the lower chargetransfer resistance. From the results, it is clearly seen that theoriginal graphite felt has large charge transfer resistances of 12.4and 8.66 U cm2 towards V2þ/V3þ and VO2þ/VO2
þ redox reactions,indicating the poor reaction kinetics on original graphite felt.Furthermore, the charge transfer resistance of V2þ/V3þ redoxcouple is found to be larger than that of VO2þ/VO2
þ redox couple,which implies the V2þ/V3þ instead of VO2þ/VO2
þ redox reaction isthe limiting half-cell reaction for VRFBs, consistently with previousresults [70,71]. On the contrary, the charge transfer resistancesdecrease greatly after modified with B4C nanoparticles to only 2.97and 2.14 U cm2 for V2þ/V3þ and VO2þ/VO2
þ redox reactions,respectively, indicating the excellent bi-functional catalytic activ-ities of B4C, in coordinate with the CV results. More importantly, itis also found that the B4C-nanoparticle-modified graphite feltoutperforms the typical thermally treated graphite felt with lower
H.R. Jiang et al. / Journal of Power Sources 365 (2017) 34e4238
charge transfer resistances on both negative and positive sides,further demonstrating the high catalytic activities of the preparedelectrode.
The TEM images of B4C nanoparticles at different magnificationare presented in Fig. 4a and b. It can be seen that the averageparticle size of the B4C nanoparticles ranges from 20 to 40 nm, andthe nanoparticles possess a high degree of crystallinity, as indicatedby the clear lattice fringe. The SEM images of original graphite feltin Fig. 4c and d revel that it is formed by a great deal of inter-connected carbon fibers with smooth and clean surfaces. Aftermodified with B4C, the smooth surface becomes much rougherwith many protrusions because a layer of nanoparticle is coated onit, as shown in Fig. 4e and f, leading to the increased active surfaceareas and active sites for vanadium redox reactions.
The real performance of B4C-nanoparticle-modified graphite feltelectrodes is then evaluated in a practical single cell and comparedwith those with original, treated and XC-72-nanoparticle-modifiedgraphite electrodes. Fig. 5aed presents the charge-discharge curvesof VRFBs with these four kinds of electrodes at the current densitiesof 80, 160, 240 and 320 mA cm�2, respectively. The thermallytreated graphite felt electrodes are used to represent the state-of-art typical electrodes, and the XC-72-nanoparticle-modifiedgraphite felt electrodes are used to exclude the influence ofincreased surface area on the battery performance. From the
Fig. 4. (a, b) TEM images of B4C nanoparticles. SEM images of (c, d) orig
results, it is found that the battery with B4C-nanoparticle-modifiedgraphite felt electrodes exhibits the lowest overpotentials, corre-sponding to the highest discharge voltage plateau and the lowestcharge voltage plateau, as well as the largest charge/discharge ca-pacities. More importantly, only the battery with B4C-nanoparticle-modified graphite felt electrodes can be operated at the high cur-rent density of 320 mA cm�2.
The columbic efficiency, voltage efficiency and energy efficiencyof VRFBs with original graphite felt electrodes, treated graphite feltelectrodes, XC-72-nanoparticle-modified graphite felt electrodesand B4C-nanoparticle-modified graphite felt electrodes are sum-marized in Fig. 6. It is shown that the voltage efficiency decreasesand the columbic efficiency increases with the increase of currentdensity, due to the fact that the increased current density wouldlead to the enhanced cell polarization and the reduced charge-discharge time. Additionally, when compared the VRFBs withdifferent electrodes, it is found that the voltage efficiency decreasesin the sequence of B4C-nanoparticle-modified graphite feltelectrodes > Treated graphite felt electrodes > XC-72-nanoparticle-modified graphite felt electrodes > original graphite felt electrodes.In this regard, the performance enhancement of VRFBs with B4C-nanoparticle-modified graphite felt electrodes is attributed to twofactors: one is the high catalytic effect of B4C leads to the faster V2þ/V3þ and VO2þ/VO2
þ redox reactions, and the other is the much
inal graphite felt and (e, f) B4C-nanoparticle-modified graphite felt.
Fig. 5. Charge-discharge curves of VRFBs with original graphite felt electrodes, treated graphite felt, XC-72-nanoparticles-modified graphite felt electrodes and B4C-nanoparticle-modified graphite felt electrodes at the current densities of (a) 80, (b) 160, (c) 240 and (d) 320 mA cm�2.
Fig. 6. (a) Columbic efficiency and voltage efficiency, and (b) energy efficiency of VRFBs with original graphite felt electrodes, treated graphite felt electrodes, XC-72-nanoparticles-modified graphite felt electrodes and B4C-nanoparticle-modified graphite felt electrodes at different current densities.
H.R. Jiang et al. / Journal of Power Sources 365 (2017) 34e42 39
Fig. 7. Polarization curves of VRFBs with original graphite felt electrodes, treatedgraphite felt electrodes, XC-72-nanoparticles-modified graphite felt electrodes andB4C-nanoparticle-modified graphite felt electrodes at the SOC of 50%.
H.R. Jiang et al. / Journal of Power Sources 365 (2017) 34e4240
enhanced active surface areas induced by the uniformly distributednanoparticles. Because the energy efficiency is the product ofcolumbic efficiency and voltage efficiency, and the difference in
Fig. 8. (a) Capacity retention rates of VRFBs with original graphite felt electrodes, treated grcurrent density of 160 mA cm�2. (b) Cycling performances of VRFBs with B4C-nanoparticle
columbic efficiency is very small, the tendency of energy efficiencychanges is similar as that of voltage efficiency changes. Therefore,the VRFBs with B4C-nanoparticle-modified graphite felt electrodesshow the highest performance among the samples studied, withthe energy efficiencies of 88.9% and 80.0% at current densities of 80and 160 mA cm�2, 16.6% and 18.8% higher than those with originalgraphite felt electrodes. Furthermore, it is demonstrated that theoperating current densities can be further raised up to 240 and320 mA cm�2 with the energy efficiencies of 72.0% and 63.8%,implying an excellent rate capability. Fig. 7 presents the polariza-tion curves of VRFBs with different electrodes at the SOC of 50%. Atthe low current density region where the activation loss is domi-nated and the mass transport loss can be ignored, it is clearly seenthat the battery with B4C-nanoparticle-modified graphite feltelectrodes has the lowest polarization, further demonstrating thehigh catalytic of prepared electrode, in coordinate with the resultsobtained from battery tests. Noting that cycling stability is also ofgreat significance for stationary storage application, the VRFBs withdifferent electrodes were subjected to constant current densitycycles. Fig. 8a exhibits the capacity retention rates of VRFBs withoriginal graphite felt electrodes, treated graphite felt electrodes andB4C-nanoparticle-modified graphite felt electrodes during cyclingat the current density of 160mA cm�2. It is found that the dischargecapacity of the assembled battery with B4C-nanoparticle-modifiedgraphite felt electrodes maintains ~60% of initial value after 200cycles, corresponding to an average discharge capacity decay rate ofonly 0.20% per cycle (Fig. S2). On the contrary, the battery with
aphite felt and B4C-nanoparticle-modified graphite felt electrodes during cycling at the-modified graphite electrodes at the current density of 160 mA cm�2.
H.R. Jiang et al. / Journal of Power Sources 365 (2017) 34e42 41
original graphite felt electrodes can only be operated for 55 cyclesat the current density of 160 mA cm�2 with the discharge capacitydecay rate of 0.75% per cycle, indicating the much enhanced ca-pacity retention rate of B4C-nanoparticle-modified graphite feltelectrodes. Another thing needs to be claimed is that the periodicstep like fluctuation of the values is attributed to the temperaturedifference between daytime and nighttime instead of the stabilityissues of the electrodes. Fig. 8b shows the cycling performance ofVRFBs with B4C-nanoparticle-modified graphite felt electrodes atthe current density of 160mA cm�2. It is shown that the battery canbe cycled for more than 200 cycles, demonstrating the good sta-bility of prepared electrodes. In addition, a decay of voltage andenergy efficiencies is also found, attributed to two factors: one isthe degradation of graphite felt electrode itself, as shown in Fig. S3,which is the dominant reason and has been a long-standing issuefor the development of VRFBs [72,73]; and the other is a portion ofnanoparticles may be dislodged by the flowing electrolyte duringlong-term cycling, as shown in Fig. S4.
4. Conclusions
To conclude, the highly active, bi-functional and metal-free B4C-nanoparticle-modified graphite felt electrodes are prepared forvanadium redox flow batteries in this work. Firstly, our first-principles calculations reveal the nonuniform distribution of elec-trons in B4C, which is beneficial to the redox reactions, and predictsthe central carbon atoms in the chain can act as the active sites.Then, the CV and EIS tests are carried out and the results demon-strate that the B4C can enhance the reaction kinetics towards bothV2þ/V3þ and VO2þ/VO2
þ redox reactions, indicating a bi-functionaleffect. In the battery tests, the VRFBs assembled with B4C-nano-particle-modified graphite felt electrodes show energy efficienciesof 88.9% and 80.0% at the current densities of 80 and 160 mA cm�2,16.6% and 18.8% higher than those with the original graphite feltelectrodes. More remarkably, the assembled battery can also beoperated at high current densities of up to 240 and 320 mA cm�2
with the energy efficiencies of 72.0% and 63.8%, respectively,demonstrating an excellent rate capability. The performanceenhancement of prepared electrodes is attributed to two factors:one is the high catalytic effect of B4C leads to the faster V2þ/V3þ andVO2þ/VO2
þ redox reactions, and the other is the much enhancedactive surface area induced by the uniformly distributed nano-particles. Finally, the battery can be operated for more than 200cycles with an average discharge capacity decay rate of only 0.20%per cycle, which successfully indicates its good stability and ca-pacity retention rate. All these superior results prove that the B4C-nanoparticle-modified graphite felt is a promising electrode forVRFBs.
Acknowledgements
The work described in this paper was fully supported by a grantfrom the Research Grants Council of the Hong Kong SpecialAdministrative Region, China (Project No. 16213414).
Appendix A. Supplementary data
Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.jpowsour.2017.08.075.
References
[1] P. Leung, X. Li, C.P. De Le�on, L. Berlouis, C.J. Low, F.C. Walsh, RSC Adv. 2 (2012)10125e10156.
[2] Y. Wang, X. Hua, C. Zhao, T. Fu, W. Li, W. Wang, Int. J. Hydrogen Energy 42
(2017) 5667e5675.[3] H. Jiang, T. Zhao, Y. Ren, R. Zhang, M. Wu, Sci. Bull. 62 (2017) 572e578.[4] M. Ulaganathan, V. Aravindan, Q. Yan, S. Madhavi, M. Skyllas-Kazacos,
T.M. Lim, Adv. Mater. Interfaces 3 (2016) 1500309.[5] F.C. Walsh, C. Ponce de L�eon, L. Berlouis, G. Nikiforidis, L.F. Arenas-Martínez,
D. Hodgson, D. Hall, ChemPlusChem 80 (2015) 288e311.[6] L. Zhang, Z. Xu, Z. Li, S. Xia, X. Yao, J. Electroceram 21 (2008) 605e608.[7] M. Zhang, L. Zhang, M. Zhu, Y. Wang, N. Li, Z. Zhang, Q. Chen, L. An, Y. Lin,
C. Nan, J. Mater. Chem. A 4 (2016) 4797e4807.[8] L. Zhang, Z. Liu, X. Lu, G. Yang, X. Zhang, Z. Cheng, Nano Energy 26 (2016)
550e557.[9] J. Tarascon, M. Armand, Nature 414 (2001) 359e367.
[10] M. Park, J. Ryu, W. Wang, J. Cho, Nat. Rev. Mater 2 (2016) 16080.[11] B.R. Chalamala, T. Soundappan, G.R. Fisher, M.R. Anstey, V.V. Viswanathan,
M.L. Perry, Proc. IEEE 102 (2014) 976e999.[12] H. Jiang, T. Zhao, M. Liu, M. Wu, X. Yan, J. Power Sources 331 (2016) 391e399.[13] H. Jiang, W. Shyy, M. Liu, L. Wei, M. Wu, T. Zhao, J. Mater. Chem. A 5 (2017)
672e679.[14] M. Liu, D. Zhou, H. Jiang, Y. Ren, F. Kang, T. Zhao, Nano Energy 28 (2016)
97e105.[15] Z. Wei, S. Meng, B. Xiong, D. Ji, K.J. Tseng, Appl. Energy 181 (2016) 332e341.[16] H. Jiang, Z. Lu, M. Wu, F. Ciucci, T. Zhao, Nano Energy 23 (2016) 97e104.[17] X. Yang, G. Zhang, C. Wang, J. Power Sources 328 (2016) 203e211.[18] Y. Liu, X. He, D. Hanlon, A. Harvey, U. Khan, Y. Li, J.N. Coleman, ACS Nano 10
(2016) 5980e5990.[19] A. Nie, L. Gan, Y. Cheng, Q. Li, Y. Yuan, F. Mashayek, H. Wang, R. Klie,
U. Schwingenschlogl, R. Shahbazian-Yassar, Nano Lett. 15 (2014) 610e615.[20] Y. Yuan, A. Nie, G.M. Odegard, R. Xu, D. Zhou, S. Santhanagopalan, K. He,
H. Asayesh-Ardakani, D.D. Meng, R.F. Klie, Nano Lett. 15 (2015) 2998e3007.[21] K. Lin, Q. Chen, M.R. Gerhardt, L. Tong, S.B. Kim, L. Eisenach, A.W. Valle,
[22] Z. Li, G. Weng, Q. Zou, G. Cong, Y. Lu, Nano Energy 30 (2016) 283e292.[23] Y. Zeng, T. Zhao, X. Zhou, L. Wei, H. Jiang, J. Power Sources 330 (2016) 55e60.[24] P. Leung, T. Martin, A. Shah, M. Anderson, J. Palma, Chem. Commun. 52 (2016)
14270e14273.[25] Q. Xu, T. Zhao, Prog. Energy Combust. Sci. 49 (2015) 40e58.[26] X. Yang, Q. Ye, P. Cheng, T.S. Zhao, Appl. Energy 145 (2015) 306e319.[27] M. Skyllas-Kazacos, M. Rychcik, R.G. Robins, A. Fane, M. Green, J. Electrochem.
Soc. 133 (1986) 1057.[28] P. Leung, Q. Xu, T. Zhao, L. Zeng, C. Zhang, Electrochim. Acta 105 (2013)
584e592.[29] Z. Wei, T.M. Lim, M. Skyllas-Kazacos, N. Wai, K.J. Tseng, Appl. Energy 172
(2016) 169e179.[30] Q. Xu, T. Zhao, Phys. Chem. Chem. Phys. 15 (2013) 10841e10848.[31] M. Park, J. Ryu, Y. Kim, J. Cho, Energy Environ. Sci. 7 (2014) 3727e3735.[32] X. Zhou, T. Zhao, Y. Zeng, L. An, L. Wei, J. Power Sources 329 (2016) 247e254.[33] Z. Wei, K.J. Tseng, N. Wai, T.M. Lim, M. Skyllas-Kazacos, J. Power Sources 332
(2016) 389e398.[34] X. Zhou, Y. Zeng, X. Zhu, L. Wei, T. Zhao, J. Power Sources 325 (2016) 329e336.[35] Z. Wei, S. Meng, K.J. Tseng, T.M. Lim, B.H. Soong, M. Skyllas-Kazacos, J. Power
Sources 344 (2017) 195e207.[36] V. Viswanathan, A. Crawford, D. Stephenson, S. Kim, W. Wang, B. Li, G. Coffey,
E. Thomsen, G. Graff, P. Balducci, J. Power Sources 247 (2014) 1040e1051.[37] J. Chen, W. Liao, W. Hsieh, C. Hsu, Y. Chen, J. Power Sources 274 (2015)
894e898.[38] D.M. Kabtamu, J. Chen, Y. Chang, C. Wang, J. Power Sources 341 (2017)
270e279.[39] Z. Zhang, J. Xi, H. Zhou, X. Qiu, Electrochim. Acta 218 (2016) 15e23.[40] A. Di Blasi, O. Di Blasi, N. Briguglio, A.S. Aric�o, D. Sebasti�an, M. L�azaro,
G. Monforte, V. Antonucci, J. Power Sources 227 (2013) 15e23.[41] S. Rudolph, U. Schr€oder, I. Bayanov, D. Hage, J. Electroanal. Chem. 728 (2014)
J. Electrochem. Soc. 163 (2016) A5202eA5210.[43] W. Wang, X. Wang, Electrochim. Acta 52 (2007) 6755e6762.[44] X. Wu, H. Xu, L. Lu, H. Zhao, J. Fu, Y. Shen, P. Xu, Y. Dong, J. Power Sources 250
(2014) 274e278.[45] B. Li, M. Gu, Z. Nie, Y. Shao, Q. Luo, X. Wei, X. Li, J. Xiao, C. Wang, V. Sprenkle,
Nano Lett. 13 (2013) 1330e1335.[46] B. Li, M. Gu, Z. Nie, X. Wei, C. Wang, V. Sprenkle, W. Wang, Nano Lett. 14
(2013) 158e165.[47] L. Wei, T. Zhao, L. Zeng, X. Zhou, Y. Zeng, Energy Tech. 4 (2016) 990e996.[48] L. Wei, T. Zhao, L. Zeng, Y. Zeng, H. Jiang, J. Power Sources 341 (2017)
318e326.[49] H. Zhou, Y. Shen, J. Xi, X. Qiu, L. Chen, ACS Appl. Mat. Interfaces 8 (2016)
15369e15378.[50] G. Kear, A.A. Shah, F.C. Walsh, Int. J. Energy Res. 36 (2012) 1105e1120.[51] Z. Gonz�alez, C. Flox, C. Blanco, M. Granda, J.R. Morante, R. Men�endez,
R. Santamaría, J. Power Sources 338 (2017) 155e162.[52] Z. He, L. Liu, C. Gao, Z. Zhou, X. Liang, Y. Lei, Z. He, S. Liu, RSC Adv. 3 (2013)
19774e19777.[53] L. Wu, Y. Shen, L. Yu, J. Xi, X. Qiu, Nano Energy 28 (2016) 19e28.[54] F. Th�evenot, J. Eur. Ceram. Soc. 6 (1990) 205e225.[55] V. Domnich, S. Reynaud, R.A. Haber, M. Chhowalla, J. Am. Ceram. Soc. 94
H.R. Jiang et al. / Journal of Power Sources 365 (2017) 34e4242
(2011) 3605e3628.[56] X. Gonze, J. Beuken, R. Caracas, F. Detraux, M. Fuchs, G. Rignanese, L. Sindic,
M. Verstraete, G. Zerah, F. Jollet, Comp. Mater. Sci. 25 (2002) 478e492.[57] X. Gonze, Zeitschrift für Kristallographie 220 (2005) 558e562.[58] J.P. Perdew, K. Burke, M. Ernzerhof, Phys. Rev. Lett. 77 (1996) 3865.[59] P.E. Bl€ochl, Phys. Rev. B 50 (1994) 17953.[60] H.J. Monkhorst, J.D. Pack, Phys. Rev. B 13 (1976) 5188.[61] L. Wei, T. Zhao, G. Zhao, L. An, L. Zeng, Appl. Energy 176 (2016) 74e79.[62] S. Aryal, P. Rulis, W. Ching, Phys. Rev. B 84 (2011) 184112.[63] H. Jiang, T. Zhao, L. Shi, P. Tan, L. An, J. Phys. Chem. C 120 (2016) 6612e6618.[64] G. Li, D. Zhang, Q. Qiao, Y. Yu, D. Peterson, A. Zafar, R. Kumar, S. Curtarolo,
F. Hunte, S. Shannon, J. Am. Chem. Soc. 138 (2016) 16632e16638.[65] H. Jiang, P. Tan, M. Liu, Y. Zeng, T. Zhao, J. Phys. Chem. C 120 (2016)
18394e18402.
[66] H. Jiang, M. Wu, X. Zhou, X. Yan, T. Zhao, J. Power Sources 325 (2016) 91e97.[67] M. Fang, S. Baldelli, J. Phys. Chem. C 121 (2017) 1591e1601.[68] J. Deng, Y. Lei, S. Wen, Z. Chen, Int. J. Min. Process 140 (2015) 43e49.[69] R. Cao, C. Saracini, J.W. Ginsbach, M.T. Kieber-Emmons, M.A. Siegler,
E.I. Solomon, S. Fukuzumi, K.D. Karlin, J. Am. Chem. Soc. 138 (2016)7055e7066.
[70] D. Aaron, C. Sun, M. Bright, A.B. Papandrew, M.M. Mench, T.A. Zawodzinski,ECS Electrochem. Lett. 2 (2013) A29eA31.
[71] C. Sun, F.M. Delnick, D.S. Aaron, A.B. Papandrew, M.M. Mench,T.A. Zawodzinski, J. Electrochem. Soc. 161 (2014) A981eA988.
[72] S. Rudolph, U. Schr€oder, I. Bayanov, D. Hage, J. Electroanal. Chem. 728 (2014)72e80.
[73] I. Derr, D. Przyrembel, J. Schweer, A. Fetyan, J. Langner, J. Melke, M. Weinelt,C. Roth, Electrochim. Acta 246 (2017) 783e793.
