Strengthening P–S bonding in TiO₂ for enhanced fuel cell startup/shutdown durability with an N, P, S–TiO₂/S–TiN catalyst

Abstract

Durability is crucial in polymer electrolyte fuel cells (PEFCs). The carbon supports currently employed in cathodes are oxidized during startup/shutdown, by increasing the cathode potential up to 1.5 V, causing the supported platinum–cobalt (PtCo/C) catalysts to lose activity. Therefore, system-level measures are currently used to control the potential below 1.0 V, which increases the cost of PEFCs. A recently reported S-doped TiN-supported N, P, S-tridoped TiO₂ catalyst is a promising candidate to replace currently available PtCo/C catalysts because, unlike other platinum group metal (PGM)-free catalysts, it is free from carbon supports. During the startup/shutdown cycles, the doped N³⁻ and S²⁻ anions substituted for O²⁻ in the TiO₂ lattice were stable, but some of the P⁵⁺ cations substituted for Ti⁴⁺ were removed from the TiO₂ surface, causing activity loss. Herein, P⁵⁺ dopants are stabilized by increasing the S²⁻ doping level, resulting in excellent startup/shutdown durability and enhanced intrinsic activity. The resulting reduction of half-wave potential after 5000 cycles between 1.0 and 1.5 V is the lowest of any reported PGM-free catalysts, at only 0.02 V. The P–S bonds formed in the TiO₂ lattice were found to be responsible for the durability of P⁵⁺, which provides a new strategy to accelerate the development of low-cost PGM-free catalysts with excellent durability.

---

1. Introduction

In efforts to ensure energy security, hydrogen is now receiving significant attention as a potential energy carrier in support of renewable energy sources such as solar and wind, which are associated with fluctuating electricity generation.¹ Polymer electrolyte fuel cells (PEFCs) are key energy conversion devices that are able to utilize hydrogen energy originating from renewable energy sources via water electrolysis. In a PEFC, the hydrogen oxidation reaction (HOR) takes place at the anode, and the oxygen reduction reaction (ORR) takes place at the cathode, as expressed by eqn (1) and (2), respectively.

Anode HOR: H₂ → 2H⁺ + 2e⁻ (1)

Cathode ORR: O₂ + 4H⁺ + 4e⁻ → 2H₂O (2)

As a source of emission-free clean energy, and because they display high power density, PEFCs are considered to be the next-generation power source for light duty vehicles traveling distances over 500 km and mid/heavy duty vehicles such as buses and trucks.^2–4 In the latest PEFC-powered vehicles, carbon black-supported platinum (Pt/C) catalysts and platinum–cobalt nanoparticle (PtCo/C) catalysts are used at the anode and cathode, respectively.⁵ However, around 20 g platinum is used in latest 128 kW-passenger vehicles,⁵ which is more than three-times higher than the target value for the widespread use of PEFC-powered vehicles.⁶ Furthermore, carbon black supports at the cathode are easily oxidized to form carbon dioxide during the startup and shutdown of the cell via a so-called reverse current decay mechanism,⁷ briefly described as follows. After the shutdown of PEFCs, the anode becomes contaminated with O₂ molecules from air or the counter cathode as they pass through the polymer electrolyte membrane (PEM) between electrodes. Next, the contaminated O₂ molecules are reduced to water because the anode Pt/C catalyzes the ORR (eqn (2)). The counter cathode provides protons during startup to maintain the undesired ORR at the anode, as the in-plane proton conductivity of the catalyst layers is insufficient. As a result, either the carbon oxidation reaction expressed in eqn (3) or the oxygen evolution reaction (OER; a reverse reaction of eqn (2)) takes place at the cathode. This raises the potential up to ∼1.5 V versus the reversible hydrogen electrode (RHE).⁷

C + 2H₂O → CO₂ + 4H⁺ + 4e⁻ (3)

The high platinum loading and instability of PtCo/C increases the cost of PEFCs markedly, and is therefore assumed to be a key barrier for the widespread adoption of PEFCs as power sources in next-generation vehicles.⁸ Hence, reducing platinum loading in PtCo/C without affecting the durability of the cell remains a pivotal challenge to date. The PtCo nanoparticles supported on carbon black do not catalyze the ORR after losing the carbon black supports by the oxidation process mentioned above. Therefore, system-level measures are usually undertaken to protect them; i.e., reducing the cathode air flow rate during shutdown to minimize the number of O₂ molecules crossing the PEM from the cathode to the anode, and introducing a small amount of H₂ gas to the anode during PEFC off-time to react with residual O₂ contaminants.⁹ However, these measures also increase the manufacturing/operating cost of PEFCs.

As cathodes require four-times the platinum loading than their anode counterparts due to the slower kinetics of the ORR compared with the HOR,¹⁰ many studies have focused on developing catalysts for cathodes free of platinum group metals (PGMs). Most of these studies have reported so-called M/N/C catalysts;^11–21i.e., graphitic carbons (C) having a large number of defects with doped nitrogen (N) and metals (M). M/N/C catalysts are sometimes referred to as M–N–C catalysts. The M used is typically Fe,^11–19 Co,^19,20 or Mn.^18,19,21 Although many other metals have also been investigated for use in M/N/C catalysts, the highest ORR activity is achieved with Fe/N/C.¹⁹ Nevertheless, a major drawback of M/N/C catalysts is their insufficient durability, according to experts in the automotive industries.²² Carbon species occupying a substantial volume of the M/N/C catalysts are less graphitic in nature than carbon black used in commercial PtCo/C catalysts. Therefore, carbon species in M/N/C tend to be easily oxidized.¹⁹ As a result, formation of CO₂ gas from carbon species in Fe/N/C (ref. 12 and 15) and Co/N/C,²⁰ even below 1.0 V versus RHE, has been reported in several studies.^12,15,20 To suppress the carbon oxidation of M/N/C catalysts, a standardized protocol has been proposed that limits the upper potential of the cathode with M/N/C catalysts at 0.925 V,¹⁷ which is lower than the value of automotive Pt(Co)/C catalysts at 1.0 V (ref. 23) or 0.95 V.²⁴ The equilibrium potential of eqn (3) is 0.207 V versus standard hydrogen electrode, and thus the reaction rate in eqn (3) is accelerated at high potentials. Therefore, system-level measures to maintain the cathode potential below 1.0 V, as described in the previous paragraph, are necessary to protect the M/N/C catalysts.

To avoid inclusion of system-level measures and achieve low-cost PEFCs, some research groups have focused on different types of PGM-free catalysts, such as an oxide/oxynitride catalyst containing group IV or V metals. These metals have unique chemical stability in acidic media, mimicking PEFC environments.²⁵ The drawback of this type of catalyst is its low conductivity. To date, half-cells have been used to quickly screen new catalysts as they require (1) a much lower catalyst mass and (2) less time to evaluate the ORR activity compared with a single cell.²³ The thinner catalyst layer with the smaller catalyst mass in a half-cell requires a lower catalyst conductivity than the catalyst layer in a single cell. However, the conductivity of oxide/oxynitride catalysts is so low that it makes the evaluation of the ORR activity difficult, even in a half-cell.^25a To address the low conductivity issue without relying on carbon supports, we recently utilized a conductive nitride, TiN, as the support for an N-doped TiO₂ catalyst.^26–30 Several control experiments using different Ti precursors revealed that the ORR activity and conductivity of the catalysts originated from the TiN support, and not from the carbon residues from urea.^26,31 Among various Ti precursors, titanium oxysulfate was selected after the first report²⁶ as it is stable in air, thus simplifying the synthesis of TiN-based catalysts.^27–30 Among the oxide/oxynitride catalysts reported to date, P and N co-doped TiO₂ catalysts on S-doped TiN (N, P–TiO₂/S–TiN) displayed the highest ORR activity in a single cell.² The ORR activity was stable during load cycles of between 0.6 and 1.0 V versus the RHE, while it deteriorated significantly after startup/shutdown cycles between 1.0 and 1.5 V owing to the removal of surface P- and N-atoms.²⁸ Therefore, the system-level measures used to protect carbon supports in PtCo/C catalysts are still necessary for this carbon-support-free catalyst (N, P–TiO₂/S–TiN). The stabilization of the surface dopants has remained a challenge for 5 years.^27–30

Recently, we reported a solution to stabilize the dopants by oxidizing the catalyst before the oxidative high potential cycles to form N, P, S-tridoped TiO₂ supported on S-doped TiN, in which sulfur dopants are used from the titanium oxysulfate precursor.³² The resulting catalytic durability was the highest among the PGM-free catalysts at that time.³² Among the three dopants, anionic N- and S-atoms were stable, while some cationic P-atoms were removed from the surface after the high potential cycles between 1.0 and 1.5 V. Therefore, the stabilization of P-atoms on the N, P, S–TiO₂/S–TiN catalyst remained a critical issue in the development of durable PGM-free catalysts until now. In this study, the P-atoms were successfully stabilized by forming a new P–S bond in the TiO₂ lattice by simply adding thiourea as a new sulfur source in the precursor dispersion. Because the amount of anionic S²⁻ on the TiO₂ surface was the key to produce the ORR active anatase/rutile heterophase junctions of N, P, S–TiO₂/S–TiN,³² we selected thiourea with S²⁻ as the new sulfur source. The valence of sulfur atoms in titanium oxysulfate precursors is 6+. Most of the S⁶⁺ in titanium oxysulfate were reduced to S²⁻ in the surface TiO₂ during the synthesis whereas the precise control of the S²⁻ content remained a challenge.³² The S²⁻ content in TiO₂ was successfully increased by a factor of ∼1.4 after adding an optimum amount of thiourea. The intrinsic ORR activity also increased by a factor of ∼1.4 after the thiourea addition. The S²⁻ anions strongly bond to P⁵⁺ cations in the TiO₂ lattice to exhibit excellent startup/shutdown durability.

2. Results and discussion

2.1 Catalyst synthesis and bulk crystal structure

The synthesis of the new N, P, S–TiO₂/S–TiN (S) catalysts is schematically illustrated in Fig. 1(a) and the details are described in S1, ESI.† We first converted the titanium oxysulfate powders (used as Ti and S sources) into S–TiN particles by pyrolysis after mixing with urea, hypophosphorous acid, and thiourea, which acted as sources of N, P, and S, respectively. In this instance, we included thiourea as the source of S, which is the key difference with our previously reported N, P, S–TiO₂/S–TiN (S) catalyst.³² Our previous study emphasized that increasing the substitutional S²⁻ doping level and stabilizing P⁵⁺-dopants of N, P, S–TiO₂/S–TiN are crucial to enhance the ORR activity and durability against startup/shutdown cycles.³² Hence, we chose thiourea as an additional S source in this study, as thiourea exhibits better stability under a humid environment and its crystal structural/thermophysical properties are not affected by the moisture content in air. Next, the S–TiN particles were annealed with NH₄F under N₂ gas to form N, P, S–TiO₂/S–TiN via a reaction between TiN and HF from NH₄F.³² The changes in the crystal structure during the synthesis steps 1 and 2 are displayed in Fig. 1(b). Although some unknown peaks attributed to contaminants are observed in the as-received titanium oxysulfate powders, most of them are assigned to TiOSO₄·2H₂O. A single TiN phase was obtained after step 1; i.e., mixing TiOSO₄·2H₂O with urea, hypophosphorous acid, thiourea, and pyrolysis at 1173 K. This was then oxidized to form a mixture of anatase/rutile TiO₂ phases and the remaining TiN phase after step 2; i.e., annealing S–TiN with NH₄F at 1223 K. When N, P, S–TiO₂/S–TiN catalysts were synthesized without using thiourea in our previous study,³² increasing the temperature to 1223 K during step 2 led to decomposition of NH₄F at ∼398 K to form NH₃ and HF gases,³³ after which TiN partially reacted with HF to produce TiF₃ at 473–573 K.³⁴ The trace O₂ molecules contaminating the N₂ gas during annealing reacted with TiF₃ to produce anatase/rutile TiO₂. The changes in the crystal structure during steps 1 and 2 displayed in Fig. 1(b) are similar to those of N, P, S–TiO₂/S–TiN synthesized without using thiourea (Fig. S1, ESI†).³² This indicates that thiourea did not block the formation of anatase/rutile TiO₂, which is necessary to enhance the activity of the ORR, although it did affect the number of doped S-atoms. From the energy dispersive X-ray spectroscopy (EDS) analyses, the atomic ratio of sulfur to titanium in N, P, S–TiO₂/S–TiN (S) was 0.15 ± 0.00, which is nearly double that of thiourea-free N, P, S–TiO₂/S–TiN (0.08 ± 0.00), indicating that S-atoms from thiourea were successfully doped into the bulk N, P, S–TiO₂/S–TiN (S).

Fig. 1 Synthesis and characterization of a new N, P, S-TiO₂/S-TiN (S) catalyst. (a) Synthesis scheme for N, P, S–TiO₂/S–TiN (S) in two steps. Step 1 includes S-doped TiN formation via pyrolysis of a mixture of titanium oxysulfate, urea, hypophosphorous acid, and thiourea as Ti/S, N, P and new S sources, respectively, at 1173 K for 2 h. Step 2 includes annealing at 1223 K for 1 h after mixing S–TiN with ammonium fluoride. (b) X-ray diffraction (XRD) patterns of (bottom) the titanium oxysulfate precursor and N, P, S–TiO₂/S–TiN (S) before (middle) and after (top) step 2. (c) Ultraviolet (UV)-Raman spectrum, (d) transmission electron microscopy (TEM) image, and (e) X-ray photoelectron (XP) spectra of N, P, S–TiO₂/S–TiN (S). The central part of the TEM image in (d)-left is enlarged in (d)-right.

2.2 Surface crystal structure and chemical states

The surface crystal structures were evaluated using ultraviolet (UV) Raman spectroscopy and transmission electron microscopy (TEM). It was decided to evaluate the UV-Raman spectra of N, P, S–TiO₂/S–TiN (S) in this study because the ORR proceeds at the surface, and UV-Raman spectroscopy excited by a 325 nm laser offers greater surface sensitivity than XRD or conventional visible Raman spectroscopy (e.g., 532 nm excitation), due to the strong UV absorption of TiO₂.³⁵Fig. 1(c) and (d) show the UV-Raman spectrum and TEM image of N, P, S–TiO₂/S–TiN (S), respectively. N, P, S–TiO₂/S–TiN (S) displays peaks at ∼143, 200, 395, 517 and 633 cm⁻¹, which are assigned to E_g, E_g, B_1g, B_1g and E_g vibration modes of anatase TiO₂, respectively.^35–37 A weak peak at ∼329 cm⁻¹ can be assigned to the second-order phonon of anatase³⁶ or rutile.³⁷ The other two weak peaks at ∼260 and 445 cm⁻¹ are assigned to the second-order phonon and E_g vibration modes of rutile, respectively.³⁷ These results indicate that the outermost surface of N, P, S–TiO₂/S–TiN (S) is anatase/rutile TiO₂, free from TiN. Consistent with the UV-Raman spectroscopy results, several lattice fringes from different crystal planes are observed from the TEM image shown in Fig. 1(d). Clear lattice fringes with interplanar distances of 0.35 nm, corresponding to the anatase (1 0 1) plane, are observed at two different spots and are directly connected to another plane, rutile (1 1 0), at an interplanar distance of 0.32 nm via a heterophase junction as shown by the circles. Furthermore, one of the anatase (1 0 1) spots is also connected to the rutile (1 0 1) plane with an interplanar distance of 0.24 nm to display crossed fringes, as shown by the dashed circles. These heterophase junctions were shown to be necessary to enhance both the ORR activity and startup/shutdown durability of N, P, S–TiO₂/S–TiN synthesized without using thiourea in our previous study.³²

The results from X-ray diffraction (XRD) patterns, Raman spectra, and TEM imaging indicate that the bulk of the N, P, S–TiO₂/S–TiN (S) catalyst is a mixture of TiN, anatase, and rutile TiO₂ phases, whereas the outermost surface is anatase/rutile TiO₂. The chemical states of the surface of the N, P, S–TiO₂/S–TiN (S) catalyst are investigated by X-ray photoelectron spectroscopy (XPS), as shown in Fig. 1(e). The Ti 2p level splits into Ti 2p_3/2 and 2p_1/2 sublevels by spin–orbit coupling to display doublets at three different binding energies. In the Ti 2p_3/2 region, the dominant peak at ∼459 eV is assigned to Ti⁴⁺ in the TiO₂ lattice whereas the small shoulder at ∼457 eV is assigned to Ti⁴⁺ in N-doped TiO₂ and the smallest one at ∼455 eV is Ti³⁺ in TiN.^38,39 The area fraction of TiN peaks in the Ti 2p spectra is only 0.09 ± 0.01, indicating that the surface is mostly TiO₂ phase, consistent with the results of UV-Raman analyses. The N 1s spectrum displays four peaks to exhibit four different chemical states from doped N-atoms. The most dominant peak at ∼401 eV is assigned to interstitial N-atoms in TiO₂ and two shoulders at ∼399 eV and ∼397 eV are assigned to substitutional N-atoms in TiO₂.^40–42 The smallest peak at ∼396 eV is assigned to N-atoms in TiN.^38,39 Therefore, most N-atoms on the surface were doped into the TiO₂ lattice. The chemical states of the N, P, S–TiO₂/S–TiN (S) catalyst in the Ti 2p and N 1s regions are similar to our previous N, P, S–TiO₂/S–TiN catalyst synthesized without using thiourea,³² although slight changes are observed in the P 2p region as shown in Fig. S2, ESI.† The asymmetric peak is deconvoluted into two peaks at ∼134 eV and ∼131 eV, of which the former is assigned to P⁵⁺ (which is substituted for Ti⁴⁺ in the TiO₂ lattice⁴³) and the latter is assigned to P–S bonding in some compounds, S-doped black phosphorous–TiO₂ composite,⁴⁴ and P_xS_y.⁴⁵ Thus, the 131 eV peak is assigned to a bonding between cationic phosphorus and anionic sulfur atoms in N, P, S-tridoped TiO₂. In the S 2p region, two peaks at ∼164 eV and ∼166 eV are the S 2p_3/2 and 2p_1/2 sublevels, respectively, of S²⁻ substituted for O-atoms in the TiO₂ lattice.^46,47 Two smaller peaks at ∼168 eV and ∼170 eV are the S 2p_3/2 and 2p_1/2 levels, respectively, of S⁴⁺ substituted for Ti-atoms in TiO₂.^46,47 The binding energy of the anionic sulfur species doped into TiN is less than the value depicted in Fig. 1(e), approximately 163 eV,⁴⁸ indicating their absence. No clear changes in the Ti 2p peak positions have been observed for TiO₂ doped with S²⁻ and S⁴⁺,⁴⁶ unlike the case of N-doped TiO₂. These results indicate that anionic N, S-atoms are substituted for O-atoms and cationic P, S-atoms are substituted for Ti-atoms in surface TiO₂. The chemical states on the N, P, S–TiO₂/S–TiN (S) developed in this study are similar to those of the previous N, P, S–TiO₂/S–TiN synthesized without using thiourea,³² although a slight change in the P 2p region is observed with the appearance of P–S bonding, which is critical for the durability of the phosphorus dopant, as shown in the next section. Although the chemical states of the doped sulfur species are almost the same in the N, P, S–TiO₂/S–TiN catalysts synthesized with and without thiourea (Fig. S2, ESI†), the amount of sulfur was different. Because the amount of S²⁻ is critical to startup/shutdown durability,³² the atomic ratio of S²⁻/Ti was evaluated from the Ti 2p and S 2p spectra as 0.62 ± 0.03 and 0.46 ± 0.03 for N, P, S–TiO₂/S–TiN (S) and N, P, S–TiO₂/S–TiN, respectively. Thus, thiourea successfully acted as a source of S²⁻ which was substituted for O-atoms on the TiO₂ surface. The F 1s spectrum shown in Fig. 1(e) is noisy, indicating that F-atoms from NH₄F were absent on the surface.

2.3 Startup/shutdown durability

The amount of thiourea was optimized to maximize the ORR activity and selectivity by maximizing the number of anatase/rutile heterophase junctions (Fig. S3, ESI†). The maximized activity was observed at a half-wave potential (E_1/2) of 0.72 V versus the RHE, which is the potential at which half of the limiting current is obtained. This represents the highest value ever reported for oxide/oxynitride catalysts (Fig. S4, ESI†). Next, the startup/shutdown durability was evaluated using a protocol proposed by the Fuel Cell Commercialization Conference of Japan (FCCJ), in which the potential was cycled between 1.0 and 1.5 V versus RHE.²³Fig. 2(a) shows the rotating disk electrode (RDE) voltammograms of the N, P, S–TiO₂/S–TiN (S) and previous N, P, S–TiO₂/S–TiN catalyst synthesized without using thiourea before and after 5000 startup/shutdown cycles.³² The decrease in E_1/2 during the 5000 cycles was successfully suppressed to only 0.02 V. It should be noted that the durability of most non-PGM catalysts has been evaluated at potentials below 1.0 V in acidic media. Moreover, startup/shutdown durability has rarely been investigated, largely due to the harsh oxidative conditions that degrade activity when carbon supports are utilized. The few available startup/shutdown durability test results for non-PGM catalysts are compared in Table 1. State-of-the-art Fe/N/C catalysts display significant E_1/2 losses, and often fail to reach limiting current plateaus. The startup/shutdown durability of N, P, S–TiO₂/S–TiN (S) is the highest among the non-PGM catalysts reported to date. After the 5000 cycles, the N, P, S–TiO₂/S–TiN (S) catalyst displays (1) a slightly higher ORR activity than the previous thiourea-free N, P, S–TiO₂/S–TiN catalyst and (2) an increased current density at potentials below 0.6 V, relative to its initial performance. To clarify the source of the enhanced durability, UV-Raman spectroscopy and XPS analyses were performed on the N, P, S–TiO₂/S–TiN (S) catalyst after the 5000 cycles, as shown in Fig. 2(b) and (c). The stability of the surface crystal structure of the PGM-free catalysts after startup/shutdown cycles has never been reported, and we evaluated it using UV-Raman spectroscopy. After 5000 cycles, the N, P, S–TiO₂/S–TiN (S) catalyst contains a Nafion ionomer in the catalyst layer, which reduced the signal-to-noise ratio; as a result, quantitative analyses are not possible. Nevertheless, the Raman spectrum remained unchanged during the 5000 cycles, as shown in Fig. 2(b). Thus, the surface TiO₂ phase displays excellent stability during cycling. The bulk crystal structure was also stable, as shown in Fig. S5, ESI.† Furthermore, no Ti 2p, N 1s, P 2p and low binding energy regions in the S 2p spectra originating from the anionic S²⁻ dopants changed significantly, as shown in Fig. 2(c). After 5000 cycles, the N, P, S–TiO₂/S–TiN (S) catalyst showed a stronger peak in the high binding energy region above 166 eV, primarily due to the sulfonate group of Nafion. The stability of S⁴⁺ dopants is thus not evaluated, as the contributions of the sulfonate group of Nafion and the N, P, S–TiO₂/S–TiN (S) catalyst cannot be separated. In our previous study reporting the startup/shutdown durability of the thiourea-free N, P, S–TiO₂/S–TiN catalysts, all Ti 2p, N 1s and S²⁻ species in the S 2p spectra were stable enough not to lose their initial chemical state during the 5000 cycles.³² In contrast, the P 2p spectra changed significantly through the loss of a number of P⁵⁺ dopants, mandating further efforts to enhance startup/shutdown durability.³² Removal of P⁵⁺ during the startup/shutdown cycles has also been a source of degradation in other S-free N, P–TiO₂/S–TiN catalysts.^27,28 In contrast, we have shown the P⁵⁺-dopants in N, P, S–TiO₂/S–TiN (S) to be highly stable during startup/shutdown cycles. XP spectra were acquired at three different locations, and the average atomic ratios along with their standard deviations were calculated. The calculated atomic ratios of P to Ti (R_P/Ti), before and after the 5000 cycles, are shown in Fig. 2(d). Because these two regions are not affected by Nafion, the changes in R_P/Ti during the 5000 cycles arise solely from the changes on the catalyst surface. The thiourea-free N, P, S–TiO₂/S–TiN catalyst lost approximately one-quarter of its initial R_P/Ti during the 5000 cycles, while N, P, S–TiO₂/S–TiN (S) showed no reduction in R_P/Ti. These results clearly demonstrate that the incorporation of an optimum amount of thiourea into the precursor dispersion enhances the startup/shutdown durability of the N, P, S–TiO₂/S–TiN (S) catalyst, primarily by stabilizing the P⁵⁺-dopants. As reported in sections 2.1 and 2.2, EDS and XPS analyses revealed that both the bulk sulfur content and surface S²⁻ doping level were higher in N, P, S–TiO₂/S–TiN (S) than those in N, P, S–TiO₂/S–TiN, largely due to the presence of sulfur atoms from thiourea used for N, P, S–TiO₂/S–TiN (S). A portion of the S²⁻ species was found to bond with cationic phosphorus, forming P–S bonds. Song et al. reported that oxidation of phosphorus was significantly suppressed by forming P–S bonding to produce a stable sulfur-doped black phosphorus–TiO₂ composite in sodium ion battery anodes.⁴⁴ Similarly, our results suggest that P⁵⁺ dopants in the TiO₂ lattice of N, P, S–TiO₂/S–TiN (S) were stabilized by P–S bonding to increase resistance against oxidation at high potentials, as illustrated in Fig. 2(e). The results shown in Fig. 2(b–d) demonstrate the excellent stability of surface TiO₂. Bulk TiN also displayed stability, as evidenced by the lack of change in the XRD pattern during the ORR (Fig. S5, ESI†). Consistent with these results, the amount of titanium dissolved in the electrolyte solution from the N, P, S–TiO₂/S–TiN (S) catalyst after 5000 startup/shutdown cycles was beneath the threshold detectable with an inductively coupled plasma (ICP) spectrometer.

Fig. 2 Startup/shutdown durability of the N, P, S–TiO₂/S–TiN (S) catalyst and the source of the durability. (a) Rotating disk electrode (RDE) voltammograms of N, P, S–TiO₂/S–TiN (S) before (solid curve) and after (dashed curve) 5000 potential cycles between 1.0 and 1.5 V vs. RHE at 0.5 V s⁻¹ in 0.1 mol dm⁻³ H₂SO₄ solution. Curves for the N, P, S–TiO₂/S–TiN catalyst before (dash-dotted curve) and after (dash double dotted curve) 5000 potential cycles are also shown for reference with permission.³² Copyright 2024, Royal Society of Chemistry. (b) UV-Raman spectra and (c) XP spectra of N, P, S–TiO₂/S–TiN (S) before (solid curves) and after (dashed curves) 5000 potential cycles between 1.0 and 1.5 V vs. RHE at 0.5 V s⁻¹ in 0.1 mol dm⁻³ H₂SO₄ solution. The N, P, S–TiO₂/S–TiN (S) after 5000 cycles contains a Nafion ionomer used in the catalyst layer. (d) Atomic ratios of phosphorus to titanium (R_P/Ti) versus the cycle number (N) curves of N, P, S–TiO₂/S–TiN (S) and N, P, S–TiO₂/S–TiN calculated from their XP spectra. (e) Schematic image of the durable active sites on N, P, S–TiO₂/S–TiN (S). RHE = reversible hydrogen electrode.

Table 1 Startup/shutdown durability of PGM-free catalysts in acidic media

Catalyst	Electrolyte	Accelerated degradation test (ADT)				Source
		Protocol	Cycle number	ΔE1/2^a	Limiting current plateau after ADT
a E 1/2 difference before and after the ADT cycles, E1/2 (initial) − E1/2 (after ADT cycles). b The scan rate was not described. ADT = accelerated degradation test.
N, P, S–TiO2/S–TiN (S)	0.1 mol dm^−3 H2SO4	1.0–1.5 V at 0.5 V s^−1	5000	0.02 V	Present	This work
N, P, S–TiO2/S–TiN	0.1 mol dm^−3 H2SO4	1.0–1.5 V at 0.5 V s^−1	5000	0.02 V	Present	32
N, P–TiO2/S–TiN	0.1 mol dm^−3 H2SO4	1.0–1.5 V at 0.5 V s^−1	5000	0.08 V	Absent	28
N–Ti0.8Zr0.2O2/S–TiN	0.1 mol dm^−3 H2SO4	1.0–1.5 V at 0.5 V s^−1	5000	0.04 V	Absent	29
Fe/N/C	0.5 mol dm^−3 H2SO4	1.0–1.5 V^b	5000	0.042 V	Present	18
Fe/N/C	0.1 mol dm^−3 HClO4	1.0–1.5 V at 0.1 V s^−1	5000	0.07 V	Absent	19
Fe/N/C	0.1 mol dm^−3 HClO4	1.2–1.5 V at 0.5 V s^−1	5000	0.099 V	Absent	12
Fe/N/C	0.1 mol dm^−3 HClO4	1.0–1.5 V at 0.1 V s^−1	6200	0.27 V	Absent	13
Fe/N/C	0.5 mol dm^−3 H2SO4	1.0–1.5 V at 0.5 V s^−1	5000	0.035 V	Present	16
Fe/N/C	0.5 mol dm^−3 H2SO4	1.0–1.5 V at 0.5 V s^−1	1110	0.040 V	N/A	14

---

It is also noted that replacing urea with thiourea significantly decreased the ORR activity (Fig. S6, ESI†); thus, urea is shown to be an indispensable nitrogen source. The combined use of urea and thiourea was found to enhance startup/shutdown stability by increasing the S²⁻ doping level, which facilitates the formation of stable P–S bonds. The higher S²⁻ doping level of N, P, S–TiO₂/S–TiN (S) is the source of the higher intrinsic ORR activity when compared with N, P, S–TiO₂/S–TiN (Fig. S7, ESI†). When the surface is free from S²⁻, the ORR activity decreases significantly and all the surface N, P, and S atoms and the bulk conductive TiN phase are necessary for this catalyst (Fig. S8, ESI†). Our next area of research will focus on increasing the S²⁻ doping level and the number of P–S bonds in the TiO₂ lattice, which will further enhance the intrinsic activity and durability of this type of catalyst.

Conclusions

In conclusion, to stabilize P⁵⁺ dopants on the recently developed N, P, S–TiO₂/S–TiN catalyst, thiourea was added to the precursor dispersion as a new sulfur source. After optimizing the thiourea content, both the bulk and surface sulfur concentrations increased, accompanied by the emergence of unique surface chemical states. The key findings of the study can be summarized as follows:

• The majority of surface sulfur was S²⁻, which substituted for O²⁻ in the TiO₂ lattice. The amount of S²⁻ was increased to enhance the intrinsic ORR activity of the N, P, S–TiO₂/S–TiN catalyst by using thiourea. Furthermore, the S²⁻ strongly bonded to P⁵⁺, which substituted for Ti⁴⁺ in TiO₂, successfully stabilizing the P⁵⁺ dopant during startup/shutdown cycles.

• In the previously reported N, P, S–TiO₂/S–TiN catalyst synthesized without thiourea, some P⁵⁺ dopants were removed during 5000 1.0–1.5 V cycles in 0.1 mol dm⁻³ H₂SO₄. However, no loss of P⁵⁺ from the TiO₂ lattice was found in the current N, P, S–TiO₂/S–TiN (S) catalyst synthesized in the presence of thiourea. This catalyst exhibits the highest durability reported to date for PGM-free catalysts.

• The P–S bonds formed in TiO₂ were stable against repeated startup/shutdown cycles, providing a potential new strategy for the stabilization of cations in TiO₂ catalysts for use in fuel cells.

Data availability

Data discussing the results of this article are available as a part of the main text, and data supporting this article have been included as a part of the ESI.†

Author contributions

Mitsuharu Chisaka: conceptualization, investigation, data curation, formal analyses, writing – original draft, writing – review & editing, and funding acquisition. Jubair A. Shamim: investigation, validation, and writing – review & editing. Hirofumi Daiguji: investigation, validation, and writing – review & editing.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

The authors acknowledge Yusei Tsushima at Hirosaki University, Yuichi Kitagawa at Horiba Techno Service Co., and Daisuke Kadohama at Public Nuisance & Medical Research Institute for acquiring TEM images–ED spectra, Raman spectra and ICP spectra, respectively. This work was partially supported by a Grant-in-Aid for Scientific Research, Grant Number JP23K26042, JP23H01347 from the Ministry of Education, Culture, Sports, Science and Technology (MEXT) in Japan. Some XP spectra were acquired at the University of Tokyo with the support by the Advanced Research Infrastructure for Materials and Nanotechnology in Japan (ARIM) of the MEXT, Proposal Number JPMXP1223UT0203.

Notes and references

1. J. L. Liu, J. Fu, S. S. Wong and S. Bashir, Energy Fuels, 2023, 37, 3315.
2. M. Chisaka, J. Mater. Chem. A, 2024, 12, 18636.
3. O. Gröger, H. A. Gasteiger and J. P. Suchsland, J. Electrochem. Soc., 2015, 162, A2605.
4. D. A. Cullen, K. C. Neyerlin, R. K. Ahluwalia, R. Mukundan, K. L. More, R. L. Borup, A. Z. Weber, D. J. Myers and A. Kusoglu, Nat. Energy, 2021, 6, 462.
5. T. Nakamichi, Nikkei Electronics, 2022, vol. 6, p. 73.
6. A. Kongkanand and M. F. Mathias, J. Phys. Chem. Lett., 2016, 7, 1127.
7. C. A. Reiser, L. Bregoli, T. W. Patterson, J. S. Yi, J. D. Yang, M. L. Perry and T. D. Jarvi, Electrochem. Solid-State Lett., 2005, 8, A273.
8. M. M. Whiston, I. L. Azevedo, S. Litster, K. S. Whitefoot, C. Samaras and J. F. Whitacre, Proc. Natl. Acad. Sci. U. S. A., 2019, 116, 4899.
9. U. Eberle, B. Müller and R. von Helmolt, Energy Environ. Sci., 2012, 5, 8780.
10. A. Kongkanand and M. F. Mathias, J. Phys. Chem. Lett., 2016, 7, 1127.
11. E. Proietti, F. Jaouen, M. Lefèvre, N. Larouche, J. Tian, J. Herranz and J. P. Dodelet, Nat. Commun., 2011, 2, 416.
12. C. H. Choi, C. Baldizzone, J. P. Grote, A. K. Schuppert, F. Jaouen and K. J. J. Mayrhofer, Angew. Chem., Int. Ed., 2015, 54, 12753.
13. K. Strickland, E. Miner, Q. Jia, U. Tylus, N. Ramaswamy, W. Liang, M. T. Sougrati, F. Jaouen and S. Mukerjee, Nat. Commun., 2015, 6, 7343.
14. A. Muthukrishnan, Y. Nabae, T. Hayakawa, T. Okajima and T. Ohsaka, Catal. Sci. Technol., 2015, 5, 475.
15. Y. Shao, J. P. Dodelet, G. Wu and P. Zelenay, Adv. Mater., 2019, 31, 1807615.
16. X. Zhao, X. Yang, M. Wang, S. Hwang, S. Karakalos, M. Chen, Z. Qiao, L. Wang, B. Liu, Q. Ma, D. A. Cullen, D. Su, H. Yang, H. Y. Zang, Z. Feng and G. Wu, Appl. Catal., B, 2020, 279, 119400.
17. H. Zhang, L. Osmieri, J. H. Park, H. T. Chung, D. A. Cullen, K. C. Neyerlin, D. J. Myers and P. Zelenay, Nat. Catal., 2022, 5, 455.
18. J. Zhu, Z. Fang, X. Yang, M. Chen, Z. Chen, F. Qiu, M. Wang, P. Liu, Q. Xu, X. Zhuang and G. Wu, ACS Catal., 2022, 12, 6409.
19. A. Ly, E. Murphy, H. Wang, Y. Huang, G. Ferro, S. Guo, T. Asset, Y. Liu, I. V. Zenyuk and P. Atanassov, EES Catal., 2024, 2, 624.
20. X. Xie, C. He, B. Li, Y. He, D. A. Cullen, E. C. Wegener, A. J. Kropf, U. Martinez, Y. Cheng, M. H. Engelhard, M. E. Bowden, M. Song, T. Lemmon, X. S. Li, Z. Nie, J. Liu, D. J. Myers, P. Zelenay, G. Wang, G. Wu, V. Ramani and Y. Shao, Nat. Catal., 2020, 3, 1044.
21. M. Chen, X. Li, F. Yang, B. Li, T. Stracensky, S. Karakalos, S. Mukerjee, Q. Jia, D. Su, G. Wang, G. Wu and H. Xu, ACS Catal., 2020, 10, 10523.
22. K. Kodama, T. Nagai, A. Kuwaki, R. Jinnouchi and Y. Morimoto, Nat. Nanotechnol., 2021, 16, 140.
23. A. Ohma, K. Shinohara, A. Iiyama, T. Yoshida and A. Daimaru, ECS Trans., 2011, 41, 775.
24. U.S. Department of Energy and Hydrogen and Fuel Cell Technology Office, Fuel Cell Technologies Office Multi-year Research, Development, and Demonstration Plan, updated in 2016, https://www.energy.gov/sites/prod/files/2016/10/f33/fcto_myrdd_fuel_cells.pdf, last accessed on Jan. 9, 2025.
25. (a) M. Chisaka, in Electrocatalysts for Low Temperature Fuel Cells: Fundamentals and Recent Trends, ed. T. Maiyalagan and V. S. Saji, Wiley-VCH, Weinheim, Germany, 2017, ch. 14; (b) A. Ishihara, S. Tominaka, S. Mitsushima, H. Imai, O. Sugino and K. Ota, Curr. Opin. Electrochem., 2020, 21, 234.
26. M. Chisaka, Y. Yamamoto, N. Itagaki and Y. Hattori, ACS Appl. Energy Mater., 2018, 1, 211.
27. M. Chisaka and H. Morioka, Catal. Sci. Technol., 2019, 9, 611.
28. M. Chisaka, R. Xiang, S. Maruyama and H. Daiguji, ACS Appl. Energy Mater., 2020, 3, 9866.
29. M. Chisaka, R. Xiang, S. Maruyama and H. Daiguji, Energy Fuels, 2022, 36, 539.
30. M. Chisaka, T. Abe, R. Xiang, S. Maruyama and H. Daiguji, Phys. Chem. Chem. Phys., 2022, 24, 29328.
31. M. Chisaka, Y. Ando, Y. Yamamoto and N. Itagaki, Electrochim. Acta, 2016, 214, 165.
32. M. Chisaka, J. A. Shamim, W. L. Hsu and H. Daiguji, J. Mater. Chem. A, 2024, 12, 11277.
33. J. S. Sanghera, P. Hart, M. G. Sachon, K. J. Ewing and I. Aggarwal, J. Am. Ceram. Soc., 1990, 73, 1339.
34. Y. Lee, C. Huffman and S. M. George, Chem. Mater., 2016, 28, 7657.
35. J. Zhang, M. Li, Z. Feng, J. Chen and C. Li, J. Phys. Chem. B, 2006, 110, 927.
36. R. J. Capwell, F. Spagnolo and M. A. DeSesa, Appl. Spectrosc., 1972, 26, 537.
37. R. J. Capwell, F. Spagnolo and M. A. DeSesa, J. Solid State Chem., 1982, 42, 276.
38. K. S. Robinson and P. M. A. Sherwood, Surf. Interface Anal., 1984, 6, 261.
39. N. C. Saha and H. G. Tompkins, J. Appl. Phys., 1992, 72, 3072.
40. C. D. Valentin, G. Pcchioni, A. Selloni, S. Livraghi and E. Giamello, J. Phys. Chem. B, 2005, 109, 11414.
41. J. Wang, W. Zhu, Y. Zhang and S. Liu, J. Phys. Chem. C, 2007, 111, 1010.
42. G. Wang, Z. Jiang, H. Shi, T. Xiao and Z. Yan, J. Mater. Chem., 2010, 20, 5301.
43. S. P. Madhusudanan, B. Gangaja, A. G. Shyla, A. S. Nair, S. V. Nair and D. Santhanagopalan, ACS Sustainable Chem. Eng., 2017, 5, 2393.
44. T. Song, H. Chen, Z. Li, Q. Xu, H. Liu, Y. Wang and Y. Xia, Adv. Funct. Mater., 2019, 29, 1900535.
45. T. K. Paul and D. N. Bose, J. Appl. Phys., 1991, 70, 7387.
46. J. Ni, S. Fu, C. Wu, J. Maier, Y. Yu and L. Li, Adv. Mater., 2016, 28, 2259.
47. W. Jiao, N. Li, L. Wang, F. Li, G. Liu and H. M. Cheng, Chem. Commun., 2013, 49, 3461.
48. C. Li, J. Shi, L. Zhu, Y. Zhao, J. Lu and L. Xu, Nano Res., 2018, 11, 4302.

---

Footnote

† Electronic supplementary information (ESI) available: Experimental details, XRD patterns, XP spectra, and R(R)DE voltammograms. See DOI: https://doi.org/10.1039/d5cy00601e

---