Recovery of Active Surface Sites of Shape-Controlled Platinum Nanoparticles Contaminated with Halide Ions and Its Effect on Surface-Structure

The interaction of halide ions (I − , Br − , Cl − ) with well-cleaned faceted platinum (nanocube, cuboctahedral) nanoparticles and platinum polycrystalline is investigated in 0.5 M H 2 SO 4 electrolyte. Under electrochemical conditions, the Pt surface gets poisoned with halide ad-atoms and it causes the attenuation of both hydrogen adsorption/desorption in the lower potential region (0.06–0.4 V) and electroxidation of Pt nanoparticles in the higher potential region (0.6–1.2 V). Above certain concentration (5 × 10 − 6 M), the strongly adsorbing I − ions mask the H upd features. On the other hand, Br − and Cl − ions alter the peak features in the H upd region, those are characteristic of different Pt surface sites. On excursion to higher potentials ( ∼ > 1.2 V), concurrent halogen evolution, Pt oxidation, and oxygen evolution are observed; the increase in peak intensity in the H upd region reﬂects the reconstruction of the Pt surface. To remove the adsorbed halide ions from the Pt surface, an in-situ potentiostatic method is employed, which involves holding the working electrode at ∼ 0.03 V in 0.1 M NaOH solution. The cleanliness and retention of surface-structure are conﬁrmed from the voltammograms recorded in the test electrolyte and the recovery of oxygen reduction reaction (ORR) activity after cleaning the Br − ion-contaminated Pt surface supports this conjecture.

Anions specifically adsorbed on precious metal surfaces adversely impact several electrochemical reactions including oxidation/ reduction, metal deposition, corrosion and dissolution. [1][2][3][4][5][6] Catalyst poisoning by the adsorbed anions (halides, sulfates and others) is a well-known phenomenon and it decreases the activity of electrocatalysts; for example, in fuel-cell, batteries and electrolyzers. 7,8 Especially platinum, a material for catalyzing a variety of electrochemical reactions, is prone to get poisoned in the presence of adatoms/adsorbates/solution anions under electrochemical conditions. Thus, oxygen reduction reaction (ORR) in fuel cells is affected by the adsorbed species or impurities and it is highly dependent on the cleanliness and surface structure of the catalyst. [9][10][11][12][13] In hydrogen-bromine fuel cells, bromides and bromine species migrate across the membrane and poison the hydrogen-electrode catalyst; moreover, Pt dissolves in Br − ion-containing solutions. 8 Effects of cations (Na + , K + , etc.), anions (chlorate, bi-sulfate, etc.) and impurities (halides, phosphates, ammonium, CO, SO 2 , H 2 S, etc.) on ORR and organic molecule oxidation reactions with platinum single crystal electrodes are well documented in the literature. [10][11][12][13][14][15][16][17][18][19][20][21] Feliu et al., Abruna et al., and others have extensively investigated the influence of halides (I − , Br − , Cl − ) on platinum single crystal electrodes. [22][23][24][25][26][27][28][29][30][31][32][33][34][35] Most of the available surface sensitive techniques were used to elucidate adsorbate coverage and structure. Such techniques, including auger electron spectroscopy (AES), low energy electron diffraction (LEED), second harmonic generation (SHG), surface X-ray scattering (SXS), electrochemical scanning tunneling microscope (ESTM) 31,34 and electrochemical quartz crystal microbalance (EQCM), used to investigate anion adsorption have offered significant insight on the dependence of adsorption process on exposed single crystal orientations. Thus, AES/LEED studies revealed that Cl − ion adsorption occurs on Pt(100) surface at lower potential than that with Pt(111) surfaces, indicating the higher susceptibility of the former to Cl − ion poisoning. 7 Bagotzky et al. reported that the chemisorbed I − , Br − , Cl − ions, and organic species present on smooth platinum electrode surface affect the distribution of adsorbed hydrogen and therefore the H upd area. 36 The adsorption of halides on platinum is a complex process and it is affected by several factors including the crystal facets, concentration of the halide species, pH, potential and the temperature. [37][38][39][40][41] Most z E-mail: nmanoj@iitb.ac.in of these investigations were conducted under ultra-high vacuum conditions using single crystal surfaces. Inferences obtained from those studies cannot be implied directly on the bulk polycrystalline catalyst surfaces.
Recently, Devivaraprasad et al. established the adsorption of solution anions with lower potential cycling, which results in the decrease of H upd area and the reconstruction of faceted Pt nanoparticles with higher potential cycling. 42 The relative adsorption strength of anions (F − < ClO 4 − < HSO 4 − < OH ad or ad-OH speices < Cl − < Br − < I − ) decides whether the surface is regenerated in the H upd region in H 2 SO 4 or HClO 4 electrolyte during the potential excursion to 1.2 V. Or the reconstrucion takes place by the exchange of ad-layer of anions with the oxygenated species; all these affects the surface structure and electrochemical surface area (ESA) and thus the electrochemical reactions simultaneously. Thus, there is a need to redeem the platinum surface devoid of impurities for long-term durability and sustained catalytic activity. Therefore, it is imperative to investigate the interaction of halide ions along with the solution anions on Pt nanoparticle surfaces and further to develop an in-situ method to recover ESA from contaminated Pt surface.
In this work, Pt-NC (platinum nanocube), Pt-CO (platinum cuboctahedral) and Pt-PC (platinum polycrystalline) morphologies are confirmed using ex-situ high resolution transmission electron microscopy (HR-TEM) and the fraction of {100}/{111} sites are estimated through the irreversible adsorption of Ge and Bi. Subsequently, these nanoparticles are subjected to voltammetric investigation in halidefree acidic electrolyte (0.5 M H 2 SO 4 ) and that contaminated with different concentrations of halide ions (I − , Br − , Cl − ). The investigations conducted over a wide range of halide ion concentration and potential reveal the significance of specific adsorption. The results obtained with shape-controlled Pt nanoparticle surface are compared to that reported with single crystal well-defined surfaces. Thereafter, two methods are used to clean the surface: (i) potential cycling and (ii) NaOH cleaning. The former causes reconstruction of the surface and complete recovery of the active surface is too difficult. The latter removes the adsorbed halide ions without perturbing surface structure of the shape-controlled nanoparticles. It is found that the halide ions on Pt surface effectively retards the reconstruction when cycled in the potential region of 0.06-1.2 V and the ORR. These investigations with shape-controlled nanoparticles helps bridge the gap between well-defined single crystal surfaces and the polycrystalline Pt nanoparticles. The findings in the present work have Physical characterization.-The HR-TEM and STEM images of Pt nanoparticles were recorded using a JEOL JEM 2100 F field emission electron microscope operated at 200 kV. The Pt nanoparticles were dispersed in ethanol, drop-cast on a Cu-grid and dried under an IR lamp (to remove the solvent and moisture content) prior to recording the images.
Electrochemical characterization.-A measured amount of the cleaned nanoparticles (5 mg) was taken in a glass bottle and 5 mL of ultrapure water was added to it. The mixture was then ultrasonicated for 30 min. to get a fine free flowing ink. 46 A measured volume of the catalyst ink was drop-cast on the rotating disk electrode (RDE) (loading ∼30 μg cm −2 ) and it was dried at room temperature. The electrochemical measurements were performed in a conventional three-electrode electrochemical cell. A Pt wire was used as the counter electrode and the reference electrode was Ag/AgCl (double junction with outer tube filled with saturated KCl). The potential was controlled using a WaveDriver 20 Bipotentiostat/Galvanostat system from Pine Research Instrumentation, USA. Cyclic voltammograms (CVs) were recorded at a scan rate of 50 mV s −1 in argon-saturated 0.5 M H 2 SO 4 electrolyte. All the potentials were recorded vs. Ag/AgCl but are reported vs. reversible hydrogen electrode (RHE) throughout this work. The surface structure of the shape-controlled Pt nanoparticles was examined through in-situ irreversible Bi and Ge adsorption and by the ex-situ HR-TEM analysis. The fraction of {111} and {100} terrace sites were estimated by comparing the charge obtained from the irreversibly adsorbed Bi and Ge voltammogram with that of the H upd region of the respective blank voltammograms. [47][48][49] A potentiostatic method was used to recover the active surface sites, which involves holding the halide ion contaminated Pt electrode in strongly adsorbing electrolyte (i.e., in 0.1 M NaOH) for ∼15 min. at ∼0.03 V vs. RHE. The cleanliness and recovery of the active surface were confirmed from the voltammetric features recorded in the test electrolyte (0.5 M H 2 SO 4 ) and the same were compared with the results obtained from the working electrode cleaned with ultrapure water. ORR voltammograms were recorded in oxygen-saturated electrolyte with a rotating disk electrode (RDE) at a scan rate of 20 mV s -1 with 1600 rpm with and without Br − ion contamination.  Fig. 1a shows the HR-TEM image of Pt-NC nanoparticle, and from the image, it is clear that the cubic structure is formed with an inter-planar spacing of 0.191 nm corresponding to {100} facets. Fig. 1b shows the HR-TEM image of Pt-CO nanoparticle and it is evident that the particle has grown in both {111} and {100} directions with d-spacings of 0.220 and 0.193 nm, respectively. Figs. 1c and 1d show that the average particle size (obtained from the TEM images of the dispersed nanoparticles) is 12 and 11 nm for Pt-NC and Pt-CO nanoparticles, respectively. It is important to realize that the shape and size selectivity of the nanoparticles is less than 100%, and therefore, particles with other morphologies also coexist with the desired nanoparticles. Moreover, an impressive electrochemical and catalytic response characteristic of well-cleaned surfaces with the shape-controlled nanoparticles is obtained (see the discussion on voltammetric analysis).

TEM analysis of Pt-NC and Pt-CO nanoparticles.-
Voltammetric analysis of Pt nanoparticles in halide-free 0.5 M H 2 SO 4 electrolyte.-Surface-structure of the shape-controlled and polycrystalline Pt nanoparticles is investigated in acidic solution. Fig. 2 shows the CVs recorded in argon-saturated 0.5 M H 2 SO 4 electrolyte at 50 mV s -1 . The contributions from different facets, those constitute terraces and step sites, in the voltammogram at various potentials are assigned as follows. 50 With Pt-NC and Pt-CO nanoparticles, it can be observed that the peak at 0.125 V in the H upd region is relatively less intense than that at 0.27 V. The peak corresponding to weakly adsobed hydrogen (0.125 V) is assigned to the {110} sites and the other due to the strongly adsorbed hydrogen (0.27 V) is assinged to the {100} step sites on {111} terrace sites. In addition, there is a distinguishable voltammetric profile (shoulder peak) in the potential range of 0.32-0.37 V and it is assigned to {100} steps and terrace sites. The contribution of {111} bi-dimensional terrace sites appears at 0.5 V due to the adlayer adsorption of bi-sulfate anions. With Pt-PC nanoparticles, only two peaks are observed in H upd region as there is no surface faceting. Since the surface of the shape-controlled nanoparticles consists of facets with different surface energies, thus, the characteristic H upd features appear depending on the shape of the nanoparticles. On cleaned Pt nanoparticles, the relative percentages of the {100} and {111} ordered domains are estimated from the charge involved in the oxidation process of irreversibly adsorbed ad-atoms (Bi and Ge). The quantitative analysis of the site distribution on nanoparticles is presented in Table S1. The formation of oxide monolayer on Pt at higher potentials irreversibly roughens the ordered terrace surfaces with the intrusion of defects; thus, Pt surface reconstruction happens at a potential of 1.2 V. When the potential is increased to 1.4 V, the flat region observed is attributed to the formation of superoxides of platinum. Oxygen evolution reaction (OER) is evident at further higher potential of 1.6 V. Conway et al. documented these findings with Pt electrodes. 51 Recently, Savaleva et al. established three types of surface species using near ambient pressure X-ray photoelectron spectroscopy (NAP-XPS) 52  Reactions at the oxide formation region: Reaction at OER: However, anions such as halides (I − , Br − , Cl − ) not only adsorb irreversibly on the metal surface but also interfere in the hydrogen adsorption and Pt oxidation processes. The extent of hindrance depends on their adsorption strength, and it is in the order of I − > Br − > Cl − . Furthermore, there exists a multi-fold reduction in the ORR activity under the influence of halide ad-atoms due the active site blocking effect. Similar decrease in ORR activity is observed with the catalysts in an operando fuel cell due to the presence of Cl − and Br − ions. 6,14,15,18 Thus, it is imperative to investigate the influence of other adatoms/anions along with the solution anions on metal nanoparticles. In the following, the adsorption processes are elucidated with sequential potentiodynamic experiments carried out with different concentrations of anions and the investigation of site blocking effect of the halide ad-atoms is done with shape-controlled and polycrystalline Pt nanoparticles.

Iodide adsorption on shape-controlled and polycrystalline Pt
nanoparticles.-In 0.5 M H 2 SO 4 electrolyte without any contamination, presumably one hydrogen atom occupies a Pt site during hydrogen adsorption. The coverage of bulkier anions (sulfate/chlorate) is much less than that of hydrogen on the Pt surface, may be due to steric/mass-transport effects. However, if the potential range is limited to 0.06-0.6 V, the number of adsorbed anions on the Pt surface increases gradually with potential cycling, resulting in the reduction of H upd area (see the Fig. S2). 42 The surface is regenerated in the H upd region after potential excursion to 1.2 V (Fig. S3). Therefore, surface sites free of adsorbed species are regenerated on the reverse scan to form complete H upd feature. In addition, when oxygenated species (OH) is adsorbed on Pt (at potential above ∼0.7 V), an H + is also generated which combines with SO 4 −2 to form HSO 4 − and thus converts to H 2 SO 4 . Perhaps, during the lower potential cycling (below 0.65 V), once hydrogen is desorbed, HSO 4 − ions may accumulate in the electrolyte; anions might adsorb any available surface sites to attain more stable state. 42 Reactions at H upd region: These dynamic surface processes do get affected in the presence of halide ions (I − , Br − , Cl − ) or impurities. This observation is similar to that reported with the potentiodynamic studies conducted in 0.5 M H 2 SO 4 electrolyte (see Fig. S2) without ad-atoms in the lower potential range (0.06-0.6 V). It is to be noted that the decrease in H upd area with cycling in pure H 2 SO 4 is much slower as compared to that in contaminated H 2 SO 4 electrolyte. This is due to the higher adsorption strength of I − ions than that of the solution anions, which causes a greater ionic concentration at the Pt The strongly adsorbed I − ions do not allow the oxygenated species to adsorb and eventual Pt oxidation is hindered upto 1.2 V, unlike that observed in the absence of halides (Fig. 2) where the onset of the formation of oxygenated species is evident above 0.8 V.
Interestingly, when the upper potential limit is increased to 1.4 V and further to 1.6 V, the features corresponding to I 2 evolution (in the positive scan) from the respective Pt nanoparticles are observed.
The three peaks of different intensities observed during the anodic scan at ∼1.27, ∼1.37 and ∼1.47 V are due to the I 2 evolu-tion from the adsorbed I − ions on ordered terraces and step sites of Pt nanoparticles concurrent with Pt oxidation; oxygen evolution occurs at further higher potential. As the fraction of these sites varies from one shape to another, the relative intensities of I 2 evolution features also changes. These features are realized only when the electrode is cycled to 1.4 and 1.6 V since the I 2 evolution takes place beyond 1.2 V. In the reverse (cathodic) scan, concurrent Pt reduction and I − adsorption peaks are observed at potentials ∼1.17, ∼1.0 and ∼0.67 V corresponding to the Pt oxidation/ and I 2 evolution happening during forward scan (compare with Fig. 2, discussed  earlier).
These peaks are clearly distinguished from the bulk OER happening at 1.8 V (see Fig. S5 (a)). Furthermore, the reconstruction of shape-controlled Pt nanoparticles can also be realized from the decrease in the peak intensity of I 2 evolution at 1.47 V observed due to adsorption of I − on ordered terraces (see Fig. S5 (a)). Thus, the ordered terraces converts to step sites and eventually to Pt-PC nanoparticles while cycling to higher potentials (upto 1.8 V). nanoparticles.-Fig. 4 shows the CVs recorded in 0.5 M H 2 SO 4 It is well-established in the literature that there exists an anomalous feature in the region of 0.4-0.6 V on Pt (111) electrode in halide ion-free H 2 SO 4 electrolyte and it is attributed to the adsorption of bisulfate anion. 38 Interestingly, no change in the anomalous feature was observed with pH at constant bi-sulfate concentration. But, at constant pH, the anomalous feature moves toward lower potentials with bi-sulfate concentration. [38][39][40] Similar features are observed with shape-controlled Pt nanoparticles that, there is a shift in the unusual adsorption states due to the increase in concentration of specifically adsorbed anions wherein the hump present at 0.5 V moves toward less positive potentials (see Figs. S6 and S7).

Bromide adsorption on shape-controlled and polycrystalline Pt
Moreover, the peaks at 0.125 and 0.27 V shift toward the lower potentials with the increase in the concentration of Br − ions in the electrolyte (formed due to concurrent adsorption/desorption of H/Br − ). This trend is in accordance with the shift observed in case of single crystal electrode. 33,35 Figs. 4d, 4e and 4f show the CVs of Pt-NC, Pt-CO and Pt-PC, respectively, recorded with 1 × 10 −3 M Br − ions in the potential range of 0.06-1.2, 1.4, 1.6 V. The Pt oxidation and reduction features of the CVs are suppressed when compared to that observed with halide-free electrolyte (Fig. 2). At further higher concentration (1 × 10 −2 M Br − ), bromine evolution current starts appearing at ∼1.2 V. The evolution current is further enhanced when the upper potential limit is increased to 1.4 and 1.6 V due to concurrent Br 2 evolution from the adsorbed Br − and the Pt oxidation. In the reverse scan, Br − adsorption peaks are observed concomitant with Pt reduction at potentials 1.1 and 0.7 V, respectively.
[Br] − + [Br] − = [Br 2 ] [10] The relative intensity of the reduction feature at 0.7 V is higher than that of the peak present at 1.1 V with Pt-NC and Pt-CO nanoparticles in comparison to Pt-PC, and it reveals a higher extent of specific adsorption of Br − ions on the ordered terrace sites (Figs. 4d and 4e). This is further reinforced by the less intense peak in the H upd region at 0.15 V observed with Pt-PC nanoparticles (Fig. 4c) even at higher concentrations (1 × 10 −3 and 1 × 10 −2 M).
Moreover, after cycling to 1.4 and 1.6 V, the step density increases (roughness). Thus, the reconstruction of shape-controlled Pt nanoparticles can be evidenced with the increase in the peak intensity at 0.07 V (peak feature at less positive potential). Therefore, the CVs of shapecontrolled nanoparticles resemble the features of Pt-PC nanoparticles after cycling to 1.4 and 1.6 V (see the inset to Figs. 4d, 4e and 4f). nanoparticles.-Fig. 5 shows the CVs recorded in 0.5 M H 2 SO 4 with different concentrations of Cl − ions. Though the adsorption strength of Cl − is relatively lower than that of both Br − and I − ions, the H adsorption in the H upd region is significantly affected. However, the OH adsorption is observed above 0.8 V and the oxide formation at further higher potentials. Voltammetric features, with the different concentrations of Cl − ions and by varying the upper potential limits, seem to be similar to that of the Br − ion-contaminated electrolyte.

Chloride adsorption on shape-controlled and polycrystalline Pt
It can be observed that, with increase in concentration of Cl − ions from 5 × 10 −6 M to 1 × 10 −2 M, two distinguishable features develop in the H upd region at ∼0.15 and ∼0.23 V with shape-controlled Pt nanoparticles (Figs. 5a and 5b).
Step site contribution with Cl − observed at ∼0.15 V is significant with Pt-PC nanoparticles as was reported by Arruda et al. 7 (Fig. 5c). With Pt-NC and Pt-CO nanoparticles, Cl − ions suppress the features corresponding to {100} terraces at 0.32-0.37 V. This confirms the irreversible adsorption of Cl − ions on Pt nanoparticles. The absence of the peak at ∼0.23 V with Pt-PC reveals that this contribution is due to adsorption of Cl − ions on the terrace sites. As long as the potential limit is less than 0.8 V, even with increased concentration of Cl − ions in the electrolyte, the sulfate adlayer at 0.5 V is not disturbed. But, once the potential limit is increased to 1.2 V and further to 1.4 and 1.6 V, the sulfate ad-layer is replaced by Cl − ions. This is evidenced by the flat double layer observed with all the Pt nanoparticles at higher concentration of Cl − ions (1 × 10 −3 M). Here, first the sulfate ad-layer is exchanged with the oxygenated species (OH), those are eventually replaced by Cl − ions. Thus, {111} terraces are less susceptible to the Cl − poisoning than {100} terraces. With higher potential cycling to 1.4 V and 1.6 V after Cl 2 evolution, the concurrent Pt reduction and Cl − adsorption features are observed at ∼0.71 and ∼1.41 V, respectively, in the reverse scan. Also, the reconstruction of shape-controlled Pt nanoparticles can be observed, which is evidenced from the increase in the peak intensity at ∼0.15 V as observed in the case of Br − adsorption.
From the above discussion it is clearly evident that the H ad/des is drastically affected by the presence of halides (I − , Br − , Cl − ). With the detailed voltammetric knowledge obtained with Pt single crystal electrodes as reported by Markovic et al. and others, there exists a significant contribution from the solution anions on H ad/des even in halide-free electrolytes such as in pure H 2 SO 4 and HClO 4 . 53,54 Irresepective of the electrolyte or contaminants present in the solution, anions do have a role in the H upd region. Thus, anion with relatively higher adsorption strength can be used to remove the pre-adsorbed species and the electrocatalytic nature of Pt nanoparticles can be redeemed.  lines). Thereafter, the electrolyte cell was replaced with the one containing argon-saturated 0.1 M NaOH solution in which the working electrode was subjected to the potential holding at 0.03 V for 15 min. In order to confirm the cleanliness, the electrolyte is again replaced with fresh argon-saturated H 2 SO 4 solution. From the CVs recorded (see red and green lines in Figs. 7a, 7b and 7c), the rejuvenation of pristine Pt surface is evident and the surface obtained is devoid of any halide contamination (see H upd area comparison Table ST2)). Moreover, it should be emphasized that the clean Pt surface is achieved without the loss of surface structure, as indicated by the characteristic voltammetric features.

Recovery of active surface from halide ion-contaminated
Especially, the most unstable features corresponding to ordered {100} terraces at 0.32-0.37 V along with the relatively stable bidimensional {111} terrace sites at 0.5 V are intact. It means that the recovery procedure adopted does not perturb the surface-structure. To ensure that the adsorbed halide ions are not present on the Pt surface, the working electrode was rotated at 1600 rpm for few seconds during the chronoamperometry measurement. The extent of the removal of adsorbed halide ions from the Pt surface is evident from the ex-situ STEM mapping (Fig. S9). Similar results are observed with the Pt-NC and Pt-PC nanoparticles (see the Figs. S10 and S11).
To emphasize the effectiveness of this method, the above results are compared with the voltammetric features of electrodes cleaned with ultrapure water. Figs. 8a, 8b and 8c show the surface recovery of Pt-CO nanoparticles contaminated with I − , Br − , and Cl − ad-atoms, respectively. As explained above, CVs are recorded sequentially except for the cleaning procedure with NaOH electrolyte. After recording the CVs with the contaminated surfaces, the halide-contaminated electrolyte is removed and the working electrode is thoroughly washed with the ultrapure water. Then the electrolyte is replaced with fresh argon-saturated H 2 SO 4 solution.
Contrary to the above procedure, complete recovery of active surface is not achieved, even after cycling to higher potential in the range of 0.06-1.4 or 1.6 V. With potential cycling to 1.2 V, no removal of adsorbed halides is observed, which is evident from the subdued features of both H upd and the electroxidation of Pt nanoparticles. When the potential is increased to 1.4 V, the ad-atoms etch from the Pt surface. This is evidenced by the increase in the Pt reduction current at ∼0.8 V and by the increase in the peak feature at ∼0.125 V (see Figs. 8d, 8e and 8f). On increasing the upper potential limit to 1.6 V, the reduction current is further increased due to the formation of Pt superoxides, allowing the replacement of halide ad-atoms during oxide formation. There is a significant loss in area due to the unavoidable contribution from dissolution, and more importantly, shape-controlled Pt nanoparticles undergo reconstruction. All the characteristic features of faceted Pt nanoparticles are lost and it is evidenced by the increase in the peak at ∼0.125 V and with loss of features at 0.32 and 0.5 V (corresponding to the loss of {100}/{111} terraces). These findings are further confirmed with the ex-situ STEM and HR-TEM analysis as shown in Figs. S9 and S12. It can be observed that the Pt-CO nanoparticle morphology is retained since the surface-structure is unaltered with NaOH cleaning (Figs. S12 (a) and S12 (b)). On the other hand, with higher potential cycling Pt-CO nanoparticles are converted to polycrystalline as a consequence of reconstruction (Figs. S12 (c) and Figs. S12 (d)).
Similar results are observed with the Pt-NC and Pt-PC nanoparticles (see the Figs. S13 and S14). Therefore, the adopted surface cleaning method not only removes the contaminants effectively but also retains the surface-structure of shape-controlled Pt nanoparticles, avoiding the reconstruction and dissolution of Pt nanoparticles.
Further, the utility of the recovery procedure discussed above is applied to the ORR, which is one of the fundamental and profound electrochemical reactions. The structure-sensitivity of ORR is wellestablished in the literature and its kinetics is significantly affected by the anion (HSO 4 − , ClO 4 − , Br − , Cl − , I − ) adsorption and by the presence of other adsorbates or impurities in the electrolyte medium. In the present context, we have verified the in-situ cleaning procedure for ORR with Pt-CO nanoparticles contaminated with Br − ions. Fig. 9 shows the ORR voltammograms recorded in oxygensaturated 0.5 M H 2 SO 4 electrolyte at 20 mV s −1 with 1600 rpm before and after contamination with 0.01 M KBr solution. In the presence of Br − ions, there is a multi-fold decrement in the ORR activity as was observed with the single crystal electrodes. 18 This is attributed to the active site blocking by the bromide species. ORR recorded again in fresh oxygen-saturated electrolyte with the contaminated electrode subjected to in-situ cleaning procedure is compared in Fig. 9 (red and blue lines). It is evident that the ORR activity is recoverd and it is close to that observed before contamination. The characteristic ORR voltammetric features of pristine Pt are clearly evident and the surface is devoid of any contamination in fresh acidic electrolyte; however, the minimal decrease observed in the ORR activity after cleaning is attributed to the unavoidable contribution due to dissolution in the presence of excess halide ions.
Therefore, recovery method adopted could be beneficial in resolving the durability issues of Pt nanoparticles in fuel cell and flow batteries. Further, this method can be extended for the removal of other contaminants and to rejuvenate the pristine Pt surface for the durable and sustained electrocatalytic activity. However, in-situ spectroscopic methods can reveal some more fundamental evidence on the nature of halide adsorption on different Pt surfaces, which is beyond the scope of the present work.  electrolyte due to the strong adsorption of I − on the platinum surface. Almost a featureless voltammogram is obtained with the increase in I − ion concentration with all the shape-controlled Pt nanoparticles. The facet-dependence of I − adsorption is realized at extremely higher potentials (∼1.6 V) wherein the I − oxidizes selectively from the platinum nanoparticle surfaces, which results in unique voltammetric features due to concurrent Pt oxidation and iodine (I 2 ) evolution.

Conclusions
Br − and Cl − ions, with relatively weak adsorption strength in comparison to I − , exhibit a concurrent H ad/des /(Br − des/ad /Cl − des/ad ) features in the H upd region (0.06-0.4 V). The voltammetric features are characteristic of the respective halide ions and the faceted Pt surface. Since the adsorption strength of surface oxygenated species is lower than that of halides, the electroxidation of Pt nanoparticles is delayed (or hindered) upto 1.2 V. Reconstruction of shape-controlled Pt nanoparticles in the presence of halide ions can be evidenced by the increase in peak intensity at less positive potentials with potential cycling to 1.4 or 1.6 V. This is due to the conversion of ordered terraces to disordered sites and it results in eventual Pt-PC voltammetric features. An in-situ potentiostatic method, which involves holding the working electrode at ∼0.03 V in 0.1 M NaOH solution, is employed to remove the halide ad-atoms from the Pt surface. The active surface is recovered without altering the facet-structure. The cleanliness of the Pt surface obtained by this method is confirmed from the voltammetric features obtained in the test electrolyte (0.5 M H 2 SO 4 ). The cleaning with ultrapure water does not recover the H upd area of nanoparticles completely. Importantly, the surface structure is not retained because of the potential cycling in the range of 0.06-1.4 and 0.06-1.6 V. Moreover, there is an unavoidable contribution from the dissolution of Pt nanoparticles in the presence of halide ad-atoms at extremely high potentials. The in-situ cleaning method is tested for ORR recorded with the Br − ion-contaminated Pt surface and it proves the effectiveness in recovering the contaminant-free active surface. These findings have prominent relevance in obtaining the rejuvenated platinum surfaces from the contaminated electrolytes with ad-atoms/impurities in various electrocatalytic applications.