**3. Results and Discussion**

#### *3.1. Effect of Additives on the Stability of the V(V) Electrolyte*

The effect of different additives (Table 1) on the thermal stability of the V(V) electrolyte was investigated by adding 0.5 wt% of additives at 50 ◦C. Table 3 shows the very time when V2O5 started to precipitate in the V(V) electrolyte samples with different additives and the V(V) concentration in the positive electrolyte after 30 days.

**Table 3.** Effect of several additives (dosage 0.5 wt%, 50 ◦C) on the thermal stability of the 2 M V(V)/3 M H2SO4 electrolyte.


It was observed that the selected additives, except for ATMPA, delayed the precipitation of V(V) in the electrolyte under the same experimental conditions. The blank sample started to precipitate after 5 days and the retarding effect for the additives follows the order: HEDP (30d) > HDTMPA (15d) > DTPMP (12d) > EDTMPS (10d) > ATMPA (4d). The remaining V(V) concentration in the electrolyte after 30 days showed the same variation pattern as the initial time of precipitation. The remaining V(V) concentration in the electrolyte with HEDP was 1.86 M, followed by 1.46 M for HDTMPA, and that of the blank electrolyte sample was 1.27 M after 30 days. As for ATMPA, it was 1.20 M, which had a negative effect in this test. The vanadium concentration in the electrolyte directly determines the energy density and capacity of the battery [20], and the experimental results show that the thermal stability of the V(V) electrolyte is improved by the additives (except ATMPA). This means that these additives facilitate the VRFB to improve its energy density and capacity.

The stabilizing mechanism of HEDP for the electrolyte might be attributed to the synergism of –OH and phosphate. –OH could clad the hydrated penta-coordinate V(IV) vanadate, which prevents it from being oxidized at a low concentration and inhibits its precipitation [9,21]. Phosphate could interact with V(V) monomers or dimers, forming a stable phosphate-containing substance, and thus retarding the precipitation [10]. Similarly, the stabilizing capability of HDTMPA and DTPMP is probably due to the presence of more phosphate. The EDTMPS, with good chemical stability and temperature resistance, is soluble in water, non-toxic, and environmentally friendly. It can dissociate into eight anions/cations in aqueous solution, and thus chelate with multiple V(V) ions, forming multiple monomeric structured reticular macromolecular complexes that are loosely dispersed in water, so the normal precipitation process of V(V) was disrupted [22]. Although ATMPA was reported to have low limit inhibition, good chelation, and lattice distortion effects [23], it exhibited the worst effect on the thermal stability, which was likely due to the formation of chelate, which is not conductive to solution stability.

#### *3.2. CV Test*

Figure 1 shows the CV curves of V(IV) electrolyte samples with additives and blank one, and it can be observed that all CV curves show the similar peak position and one pair of redox peaks with a similar shape. The additives slightly changed the shape and position of the peak, which means that these additives will affect the reversibility of V(IV)/V(V) redox pairs to some extent [19]. The relevant data derived from Figure 1 are summarized in Table 4. The effect of additives on the V(IV)/V(V) redox coupling is characterized by IpO/IpR (ratio of the oxidation peak current to reduction peak current) and ΔVp (separation between the oxidation and reduction peak potential). The HDTMPA just incurred a minor decrease in the ΔVp of the electrolyte and a small increment in the IpO/IpR as if it had little effect on the reversibility of the V(IV)/V(V) redox pair. In addition, it had a small effect on the oxidation peak current and reduction peak current as well as the overall peak shape of the curve, indicating that its effect on the electrode reaction kinetics of the electrolyte was not that significant either [24]. The addition of HEDP increased the ΔVp of the electrolyte while decreasing IpO/IpR significantly. The addition of HEDP, DTPMP, ATMPA, and EDTMPS had a greater effect on the reversibility of V(V)/V(IV) redox pairs, and the HEDP significantly increased the peak oxidation current and peak reduction current, indicating that it might enhance the electrode reaction kinetics of the electrolyte. The main reason for the improvement in electrode reaction kinetics by HEDP might be attributed to the fact that the –OH could complex with V(IV)/V(V) ions, which provide more available –OH to the stable electrode reaction of V(IV)/V(V) for ion exchange on the electrode surface, thus resulting in a higher oxidation peak current and reduction in peak current [25]. Among all additives, ATMPA best enhanced the electrode reaction kinetics of electrolyte, but it caused a decrease in electrolyte thermal stability. HDTMPA, EDTMPS, and DTPMP probably affect the cyclic reversibility performance and electrode reaction kinetics of the electrolyte by, firstly, phosphate and, secondly, according to calculations, the C atoms adjacent to N atoms have a high positive charge density, counteracting the strong electron affinity of N atoms [26], and the positively charged C atoms activated by the N atoms can work as an active site, affecting electron distribution, thereby improving the electrochemical performance. In addition, EDTMPS benefits from its Na<sup>+</sup> ions, increasing the number of ionizable cations in the solution, which enhances the electrode reaction kinetics [12]. As for HDTMPA, its large groups slightly hinder the ion exchange and charge transfer on the electrode surface owing to the steric hindrance, which is obviously unfavorable [18]. When ATMPA and HDTMPA are used in combination, ΔVp of the electrolyte declined compared with the blank sample and IpO/IpR displayed a small change. The combination of ATMPA and HEDP also showed the same performance. At the same time, compared with the CV curve of the blank sample, the peak current and the peak area of these two complex schemes are larger, which indicates that the electrolytes affected by these two schemes have better electrochemical performance.

A series of CV curves on graphite electrode for the blank electrolyte and the electrolyte with different additives at different scan rates are depicted in Figure 2, which further reveals the effect of additives on the electrode reaction kinetics. The peak potential of the anode and cathode varies gradually with the scan rate, presenting the typical characteristics of a quasi-reversible single-electron process [27]. The diffusion coefficient of the quasi-reversible reaction (*D*) is between that of the reversible reaction (*D*1) and irreversible reaction (*D*2) [28]. As for the reversible and irreversible one-step single-electron reactions, their peak current (*ip*) can be represented as follows [29]:

$$i\_p = 0.4463(F^3/RT)^{1/2} \text{Ca} n^{3/2} v^{1/2} D\_1^{1/2} \text{ (Reversible reaction)}\tag{6}$$

$$i\_p = 0.4958(F^3/RT)^{1/2} \,\text{CAa}^{1/2} n^{3/2} v^{1/2} D\_2^{-1/2} \,\text{(Irreversible reaction)}\tag{7}$$

where *R* is the universal gas constant; *F* is the Faraday constant; *T* is the Kelvin temperature; *n* is the amount of substance of transferred electrons during electrode reaction; *A* is the

surface area of working electrode; *C* is the bulk concentration of primary reactant; *v* is the scanning rate; *α* is the transfer coefficient for an irreversible reaction; and *D*<sup>1</sup> and *D*<sup>2</sup> are the diffusion coefficients of reversible and irreversible reactions, respectively.

For a single-electron reaction at room temperature, Equations (6) and (7) can be simplified as follows:

$$i\_p = 2.69 \times 10^5 ACD\_1^{1/2} v^{1/2} \tag{8}$$

$$i\_p = 2.99 \times 10^5 a^{1/2} A \text{C} D\_2^{1/2} v^{1/2} \tag{9}$$

According to the present experimental conditions, Equations (10) and (11) can be further derived from Equation (8) and Equation (9), respectively:

$$i\_p/A = 538D\_1v^{1/2} = v^{1/2}k\tag{10}$$

$$i\_p/A = 598D\_2^{1/2}v^{1/2} = v^{1/2}k\tag{11}$$

Equations (10) and (11) indicate that the current density (*ip*/*A*) is linearly related to the square root of scan rate (*v*1/2) and *k* denotes the slope of this line, illustrated in Figure 3, thus *D*<sup>1</sup> and *D*<sup>2</sup> were calculated. The values of *k* for the blank electrolyte and those with 0.5 wt% of different additives are concluded in Table 5.

$$D\_1 = 3.45 \times 10^{-6} k^2 \tag{12}$$

$$D\_2 = 2.80 \times 10^{-6} k^2 \tag{13}$$

**Figure 1.** CV curves of the electrolyte (2.0 M V(IV)/3.0 M H2SO4) with additives (0.5 wt%) and the blank one at a scan rate of 20 mV s−<sup>1</sup> at room temperature.

**Table 4.** ΔVp and IpO/IpR of the electrolyte (2.0 M V(IV)/3.0 M H2SO4) with additives (0.5 wt%) and the blank one on the graphite electrode.


**Figure 2.** CV curves of the electrolyte (2.0 M V(IV)/3.0 M H2SO4) with/without additives ((**a**) blank sample; (**b**) HDTMPA; (**c**) HEDP; (**d**) DTPMP; (**e**) ATMPA; (**f**) EDTMPS; (**g**) ATMPA + HDTMPA; (**h**) ATMPA + HEDP) on the graphite electrode at different scan rates at room temperature.

**Figure 3.** Linear plot of *ip* versus *v*1/2.



In fact, the diffusion coefficient of the electrolyte is between *D*<sup>1</sup> and *D*<sup>2</sup> for the quasireversible process. Under the present experimental conditions, the diffusion coefficient is 1.88–2.20 × <sup>10</sup>−<sup>7</sup> for the blank electrolyte and 1.84–2.26 × <sup>10</sup>−<sup>7</sup> for the sample with HDTMPA, and the result of the latter is close to that of the blank electrolyte. When it comes to the rest, their diffusion coefficients are 3.15–3.89 × <sup>10</sup>−<sup>7</sup> (HEDP), 3.94–4.86 × <sup>10</sup>−<sup>7</sup> (DTPMP), 5.52–6.80 × <sup>10</sup>−<sup>7</sup> (ATMPA), and 2.83–3.48 × <sup>10</sup>−<sup>7</sup> (EDTMPS), respectively. These increased diffusion coefficients of the electrolyte with additives indicate that the additives (except HDTMPA) can effectively improve the diffusion of vanadium species at the electrodes and enhance the mass transfer and charge transfer of the V(V)/V(IV) redox pair, thus increasing the corresponding reactivity. Compared with the blank sample, the combination of ATMPA + HDTMPA and ATMPA + HEDP also have a larger diffusion coefficient, showing that the compounding scheme has played a positive role in mass transfer and charge transfer in the electrolyte.

#### *3.3. Steady-State Polarization Test*

The steady-state polarization curve of the V(IV) electrolyte allows the determination of the polarization resistance, the exchange current density, and the electrochemical reaction rate constant.

In the relatively-low-overvoltage region, the overvoltage and current density are approximately linearly correlated [6]. These parameters can be calculated by Equation (14).

$$R\_{ct} = \frac{\eta}{i}, \ i\_0 = \frac{RT}{nFR\_{ct}}, \ k\_0 = \frac{i\_0}{nF\mathcal{C}\_0} \tag{14}$$

where *Rct*, *i*0, and *k*<sup>0</sup> refer to the charge-transfer resistance, exchange current density, and rate constant, respectively; *R*, *T*, *n*, *F*, and *C*<sup>0</sup> are the universal gas constant, Kelvin temperature, amount of transferred electrons in the electrode reaction, Faraday constant, and solution concentration, respectively [30].

The steady-state polarization curves of the 2.0 M VOSO4/3.0 M H2SO4 electrolyte with different additives on graphite electrode are demonstrated in Figure 4, and the corresponding parameters derived from Equation (14) are listed in Table 6. One can see that the charge transfer resistance of electrolyte samples with additives decreased and the electrochemical reaction rate constant and the exchange current density increased compared with the blank sample. The charge transfer resistance of the electrolyte with EDTMPS and DTPMP, for example, decreased from 12.40 Ω cm2 (blank sample) to 8.15 Ω cm2 and 8.84 Ω cm2, respectively, at 25 ◦C, while the exchange current density of these two samples increased from 2.07 mA cm2 (blank sample) to 3.15 mA cm<sup>2</sup> and 2.91 mA cm2, respectively. The corresponding electrochemical reaction rate constant increased from 1.07 × 105 cm s−<sup>1</sup> (blank sample) to 1.63 × <sup>10</sup><sup>5</sup> cm s−<sup>1</sup> and 1.51 × <sup>10</sup><sup>5</sup> cm s−1, respectively, at 25 ◦C. The other additives also accelerated the chemical reaction process of V(IV) on the electrode surface to varying degrees. The rest of the selected additives also accelerated the kinetics process of V(IV) species on the electrode surface to a certain level. Compared with the blank sample, the combination of ATMPA + HDTMPA and ATMPA + HEDP also had lower charge transfer resistance and higher exchange current density and electrochemical reaction rate constant, which was consistent with the CV tests.

**Figure 4.** Steady-state polarization curves for the 2.0 M V(IV)/3.0 M H2SO4 blank electrolyte and those with 0.5 wt% additives on graphite electrode at a scan rate of 1 mV s<sup>−</sup>1.

**Table 6.** Kinetic parameters for the 2.0 M VOSO4/3.0 M H2SO4 experimental electrolyte with different additives on the graphite electrode.


#### *3.4. Electrochemical Impedance Spectroscopy Test*

For the further analysis of the electrode reaction diffusion kinetics of vanadium and the charge transfer and mass transfer of the V(IV)/V(V) redox pair, Nyquist plots of the eight (including two compound schemes) V(IV) electrolyte samples at room temperature were recorded by electrochemical impedance spectroscopy. Figure 5 shows that each plot consists of a semicircle in the high-frequency region and a diagonal line in the lowfrequency region, indicating that the redox reaction of the V(IV)/V(V) pair is controlled by both high-frequency charge transfer and low-frequency diffusion. The radius of the semicircle corresponds to the charge transfer resistance and the linear part relates to the diffusion of vanadium species on the electrode [31]. The equivalent circuits of these Nyquist plots were fitted, and the corresponding parameters were obtained using ZView software, which are listed in Table 7. In the equivalent circuit, R1 is the resistance consisting of the solution resistance, electrode resistance, and contact resistance, and R2 and W0 represent the charge transfer resistance and diffusion impedance, respectively. The constant phase element (CPE) represents the bilayer capacitance at the electrode–electrolyte interface.

**Figure 5.** Nyquist plots of the 2.0 M V(IV) electrolyte on the graphite plate and the corresponding equivalent circuit.


**Table 7.** Model parameters of equivalent circuits corresponding to Nyquist plots.

The additives slightly increased the contact resistance of the electrolyte, and all additives except HDTMPA and HEDP decreased the charge transfer resistance of the solution. The decrease in charge transfer resistance implies a faster charge transfer process, which is

consistent with the above study. All of the additives except HDTMPA decrease the diffusion resistance of the electrolyte, which facilitates the diffusion process on the electrode surface and enhances the electrochemical reaction kinetics. In addition, all additives lead to an increase in the CPE parameter Y0,1, indicating an enhanced bilayer capacitance at the electrode–electrolyte interface. *n* represents the index of CPE, ranging from 0 to 1. The larger the *n*, the higher the capacitive property and the lower the resistive property of CPE.

Compared with the samples using HDTMPA and HEDP additives alone, the charge transfer resistance of the two compounding schemes (ATMPA + HDTMPA and ATMPA + HEDP) is reduced, which accelerates the charge transfer in the solution, indicating that the compounding schemes (HDTMPA + ATMPA and HEDP + ATMPA) improve the electrochemical performance of the electrolyte.

After the above experimental investigation, it was found that the different additives selected could have a positive effect on the positive electrolyte of VRFB in terms of thermal stability and electrochemical performance. In future studies, it is expected that these additives may work better if used in combination, rendering these additives potential and promising for development.
