Pt Interfaces: Dopant

Article pubs.acs.org/cm

Computational Study of Nb-Doped-SnO2/Pt Interfaces: Dopant Segregation, Electronic Transport, and Catalytic Properties Qiang Fu,† Niels Bendtsen Halck, Heine Anton Hansen, Juan Maria García Lastra, and Tejs Vegge* Department of Energy Conversion and Storage, Technical University of Denmark, Fysikvej, Building 309, 2800 Kgs Lyngby, Denmark ABSTRACT: Carbon black, a state-of-the-art cathode material for proton exchange membrane fuel cells (PEMFCs), suffers from severe corrosion in practical applications. Niobium-doped tin dioxide (NTO) is a promising alternative to support the Pt catalysts at the cathodes. Here, through a combined density functional theory and nonequilibrium Green’s function study, we investigate the Nb segregation at Pt/NTO interfaces under operational electrochemical conditions, and reveal the resulting effects on the electronic transport, as well as the catalytic properties. We find that the Nb dopants tend to aggregate in the subsurface layers of the NTO substrate, whereas their transport across the Pt/NTO interface is hindered by a high thermodynamic barrier under the operating condition of PEMFCs. The interfacial transport of Sn is, however, more facile, indicating possible formations of Sn−Pt alloys and tin oxides. The electronic conductivities of the Pt/NTO systems are not particularly sensitive to the distance of the Nb dopants relative to the interface, but depend explicitly on the Nb concentration and configuration. Through a dopant induced ligand effect, the NTO substrates can improve the catalytic activity of the Pt adsorbate toward the oxygen reduction reaction. We also investigate the co-doped SnO2 substrates by both Nb and Sb elements, and find that a small amount of Nb dopants could further improve the electronic transport of the Pt/Sb-doped-SnO2 interface. The fundamental understanding generated here will help shed light on future applications of Nb-doping and Nb−Sb co-doping in Pt/SnO2 type cathodes for PEMFC applications.

■

INTRODUCTION Owing to advantages like high energy efficiency, ultralow pollutant emission, and low operation temperature, proton exchange membrane fuel cells (PEMFCs) become increasingly attractive for applications in the automotive industry. Nevertheless, PEMFCs are still too costly and have therefore not replaced internal combustion engines. While parts of the cost can be reduced by mass production, e.g., in the case of the Toyota Mirai fuel cell car, the demand for precious metals like Pt, used as electrocatalysts at both electrodes of the PEMFC, would not diminish due to economy of scale. The energy loss caused by the high overpotential required for the oxygen reduction reaction (ORR) at the cathode is another problem that impairs practical application,1 which could be addressed by alloying Pt with other elements, e.g., lanthanides2 or the 3rd row transition metals like Ni and Co.3,4 The Pt nanoparticles are typically deposited on carbon black (the state-of-the-art support material) due to its low cost, high surface area, and excellent electronic conductivity. This material is, however, not sufficiently stable and suffers from oxidation under operating conditions of PEMFCs, such as a high potential and a low pH.5−9 The corrosion not only deteriorates the catalytic performance of Pt due to a reduction of dispersion,10,11 but also may bring about a detachment of the Pt nanoparticles from the electrodes.12,13 Therefore, much effort has been placed on searching for alternative support © XXXX American Chemical Society

materials, which can improve stability at high potentials and under acidic environments. Among various materials,14,15 rutile SnO2 has been proposed as a potential candidate,16,17 which shows sufficient stability under operational conditions of PEMFCs. 18−20 Porous structures of SnO2 with high surface areas can be synthesized by using the sol−gel method.21 The electronic conductivity, which is intrinsically low due to the presence of a large band gap (3.59 eV),22,23 can be significantly increased through n-type doping, since the dopant can provide an extra electron as the charge carrier. In the past decades, Sb-doped SnO2 (ATO) has been investigated intensely, with the main focus on the electrochemical stability, dopant location, and electronic conductivity of the material.24−32 Our previous work has also explored the nonuniform distribution of Sb dopants and its effect on the conductance of Pt/ATO interfaces through a synergetic computational and experimental study.33 Recently, Nb-doped SnO2 (NTO)29,31,34−42 and TiO243,44 have been considered as the support materials. For the Pt/NTO systems, while a lot of investigations have been performed experimentally, ab initio studies, which can provide an in-depth understanding at the atomic scale, are still very limited. Received: November 15, 2016 Revised: January 7, 2017 Published: February 6, 2017 A

DOI: 10.1021/acs.chemmater.6b04879 Chem. Mater. XXXX, XXX, XXX−XXX

Article

Chemistry of Materials

exchange-correlation interaction is described by the GGA-type PBE functional.48 The double-ζ plus single polarization basis sets (DZP) are used for Sn, O, Nb, and Sb, and the double-ζ basis sets (DZ) are used for Pt. A 3 × 5 Monkhorst−Pack k-mesh49 is used to sample the Brillouin zone within the interface plane in the DFT and the transmission spectrum calculations, and 100 k-points are used along the transport direction.

In this work, we computationally investigate the segregation of the Nb dopants at Pt/NTO interfaces under operational electrochemical conditions. The Sn segregation processes across the interface are also taken into account for a comparative purpose. We find that the Nb dopants tend to aggregate in the subsurface layers of the NTO substrate, and their transport across the Pt/NTO interface is hindered by a high thermodynamic barrier under the operating condition of PEMFCs. The interfacial transport of Sn is, however, more facile, indicating a possible formation of Sn−Pt alloys and tin oxides. We explore the effects of Nb segregation on the electronic transport and the catalytic properties of the Pt/NTO interfaces. It is found that the electronic conductivity is not particularly sensitive to the specific location of the Nb dopants, but depends explicitly on the Nb concentration and configuration. The catalytic activity of the Pt adsorbate toward the ORR can be improved by the NTO substrates through a dopant induced ligand effect, possibly due to charge transfer from the support. The simulational results are compared with available experimental findings. In addition, in the investigation of the codoped SnO2 substrates by both Nb and Sb elements, we find a promoting effect of Nb dopants on the electrical conductance of Pt/ATO interfaces.

■

■

RESULTS AND DISCUSSION Segregation of Nb Dopants at Pt/NTO Interfaces. We first investigate the driving force for transport of the Nb dopants within the NTO substrate, i.e., the likelihood of the Nb segregation, from a purely thermodynamic point of view. In these studies, the Nb dopants are placed at the preferred Snpositions in the SnO2(110) surface, as shown in Figure 2. Here,

COMPUTATIONAL DETAILS

The interface model in this study is similar to our previous work about the Pt/ATO interface.33 It contains six atomic trilayers of SnO2(110) and three layers of Pt(111). The Pt layers are rotated by 13.9° relative to the SnO2 substrate, in order to reduce strain (a compression of 2.3% on the Pt(111) surface) at the interface caused by the lattice mismatch. Each SnO2 trilayer has 6 Sn atoms, and each Pt layer contains 10 Pt atoms. The outermost bridging oxygen atoms of the SnO2 substrate are removed, to ensure an effective adhesion with the Pt adlayers.33,45 Geometry optimizations and energy calculations are performed using Vienna ab initio simulation package (VASP),46,47 with the energy cutoff of 500 eV and the PBE exchange-correlation functional.48 A 3 × 5 × 1 Monkhorst−Pack k-mesh49 is used for the optimizations and a 5 × 7 × 1 k-mesh for the single-point energy calculations. In the treatment of the valence electrons, the calculations include the Nb 4s, 4p, 5s, and 4d states; the Sn 5s, 4d, and 5p states; the O 2s and 2p states; the Sb 5s and 5p state; and the Pt 5p, 6s, and 5d states. A dipole correction to total energies is added to eliminate spurious interaction between image supercells along the vertical direction. During the optimizations, the bottom two trilayers of SnO2 are kept frozen, and all other atoms are allowed to relax until the maximum force is below 0.02 eV/Å. The optimized structures are then used to construct the Pt/NTO device model (shown in Figure 1) for the electronic transport studies. The central region contains eight SnO2 trilayers and nine Pt adlayers, by extending additional two SnO2 trilayers and six Pt layers. Four SnO2 trilayers and six Pt layers are used to model the left and the right electrodes, respectively. The transport calculations are performed using the density functional theory (DFT) combined with the nonequilibrium Green’s function (NEGF) method, as implemented in Atomistix ToolKit (ATK) version 2014.0.50−52 For consistency, the

Figure 2. Potential energy surfaces for the segregation of a single Nb dopant (4.2 atom %) within (a) a stoichiometric, and (b) a partially reduced SnO2(110) substrate without Pt adlayers. The doping sites correspond to the labels on the left side. The big gray and small red spheres represent the Sn and O atoms, respectively.

the interstitial sites within SnO2 are not taken into account. Such sites may play a role as the intermediate states in the segregation pathway, and may also affect the transport kinetics owing to a change of the diffusion barriers and electronic structure.53 However, both aspects are beyond the scope of this paper, and therefore, only the Sn lattice positions are considered as the Nb locations. The segregation of a single Nb dopant within a bare SnO2(110) substrate, without any Pt layer, is investigated initially. The corresponding potential energy surfaces (PESs), i.e., the energy changes with respect to the different doping positions, are shown in Figure 2. It should be noted that the thermodynamic barrier in the PES is a lower bound, while a higher kinetic barrier would exist between adjacent doping positions. Here, the s4a doping site of the Nb atom is used as the energy reference, and both the stoichiometric (Figure 2a) and the partially reduced (Figure 2b) SnO2(110) substrates are considered. There are two types of Sn atoms at the outermost layer, labeled as s1a and s1b, respectively. For the stoichiometric SnO2 substrate, the s1b site is the most favorable location of the Nb dopant, with a thermodynamic driving force as high as 1.15

Figure 1. Configuration of the Pt/NTO device model for transport calculations. The gray, red, blue, and cyan spheres represent the Sn, O, Pt, and Nb atoms, respectively. B


Article

Chemistry of Materials eV. The PES of the other pathway, i.e., toward the s1a site, is rather flat, and the energy profile displays an increase of 0.12 eV. For the partially reduced SnO2 substrate, the top layer is still the most preferred location, but in this case, the s1a site is more stable, with a driving force of 0.60 eV, whereas the PES toward the s1b site becomes flat. Depending on the experimental conditions, the SnO2(110) substrate may exhibit different stoichiometries. Nevertheless, the Nb dopants could segregate to the surface layer according to the simulational results. We then place three Pt layers above the partially reduced SnO2 substrate and investigate their influence on the segregation of the single Nb dopant. Upon deposition of the Pt layers, all 24 doping sites of the Nb atom are nonequivalent and are fully considered, with the PES shown in Figure 3.

Figure 4. Structures and relative energies of six configurations containing two Nb atoms within a 72-atom SnO2 supercell. With the configurational entropy taken into account at 353.15 K, the relative free energies become 0.00, 0.25, 0.22, 0.26, 0.28, and 0.29 eV, respectively, for the six structures. The cyan spheres represent the Nb dopants.

nature of the Nb-pairs is beyond the scope of this paper. We note that the attractive interaction between the Nb atoms is different from that of the Sb dopants which is repulsive.33 The results agree with an inhomogeneous doping of Nb and a homogeneous distribution of Sb within the SnO2 host observed in recent experiments.42 When more Nb dopants are placed into the Pt/NTO interface, it is worth noting that identifying the most stable configuration is not trivial. The Nb atoms prefer to aggregate, as mentioned above, and the energy gain upon the aggregation is larger than the energy variation regarding the different locations of the Nb dopants. It is thus difficult to determine the most stable configurations merely from the PES in Figure 3. Therefore, we consider a large number of possible configurations, in order to find the most stable one. For example, 29 structures are investigated in the case of five Nb dopants. In Figure 5, we present the most stable configurations for each number of the dopants N (N from 1 to 5, corresponding to a dopant concentration from 4.2 at. % to 20.8 at. %), with the doping sites listed. As expected, the Nb atoms tend to aggregate, and mainly locate in the second and the third layers of the SnO2 substrate. The driving force of the Nb segregation is also given, which is defined as the energy difference between

Figure 3. Potential energy surface for the segregation of a single Nb dopant (4.2 atom %) within a partially reduced SnO2(110) substrate deposited by three Pt layers. The doping sites correspond to the labels on the left side.

Compared with that in Figure 2b, the features of this PES change significantly. The outermost layer of SnO2 is no longer a favorable doping site for the Nb atom, and the second layer becomes the most preferred location. In addition, with the Nb dopant being in different layers, the energy change of the Pt/ NTO system is rather small, only 0.2 eV. The results indicate that, at the Pt/NTO inferface, the Nb dopants mainly distribute in the subsurface layer of SnO2. This situation is quite different from that of Sb-doping at the Pt/ATO interface, where the outermost layer is the most favorable doping site with a large thermodynamic preference.33 In the next step, we increase the number of the Nb dopants in the unit-cell and investigate the influence of the concentration on the Nb segregation. Before such studies, it is necessary to understand how the Nb atoms interact with each other in bulk SnO2. Thus, we consider a few configurations containing two Nb dopants in a SnO2 supercell with a total number of 72 atoms. The structures and the relative energies are shown in Figure 4. It can be seen that when the two Nb dopants are linked by only one O atom, the configuration has the lowest energy. Otherwise, the energy of the system increases by 0.20−0.28 eV, depending on the relative positions of the two Nb atoms. The effect of configurational entropy is quite small and should not change the most favorable geometry. The results indicate that the Nb dopants prefer to aggregate within the SnO2 bulk, which may likely be due to the formation of Nb3+−Nb5+ or Nb4+−Nb4+ pairs, where the latter have been found experimentally in Nb-doped TiO2.44 Due to challenges associated with the assignment of specific valences of Nb and the small relative differences in unit-cell volumes of NTO upon Nb-doping, an assertive assignment of the specific

Figure 5. Thermodynamic driving force for the segregation of the Nb dopants toward the subsurface region. A negative value represents an exothermic process. The most stable site for each of the N Nb dopants is provided by listing their locations. With the configurational entropy taken into account at 353.15 K, the free energy differences become −0.14, −0.07, −0.23, 0.00, and −0.55 eV, respectively. C


Article

Chemistry of Materials two configurations. In one configuration, all the N Nb dopants are placed in their most stable geometry, while in the other one, we adopt the most stable structure of N − 1 Nb dopants, and place the remaining Nb dopant in the fourth layer. Since most of the Nb dopants are in the second and the third layers, such a definition is not ideal but, nevertheless, could provide a qualitative description of the driving force. Upon doping with Nb atoms, the configurational entropy from the mixed cation lattice should also be considered when analyzing the thermodynamic driving force. For the dopant concentrations from 4.2 to 20.8 at. % under investigation here, the configurational entropy changes are below 0.1 eV at 353.15 K54 and thus do not change the thermodynamic trends. From Figure 5, one can see that the values are rather irregular with respect to the number N. For instance, the driving force is largest in the case of five Nb dopants, but negligible when N is four. It originates from the mutual interaction between the Nb dopants and the resulting aggregation. Overall, the results indicate that the segregation of Nb atoms is favorable toward the Pt/NTO interface, although they do not prefer to locate at the outermost layer. Next, we investigate the transport of Nb atoms out of the SnO2 host, i.e., the likelihood of dopant depletion in the electrochemical environment of PEMFCs. We divide the whole process into two stages: The first corresponds to Nb alloying with the Pt layers, and the second step is the oxidation of the Nb atoms that have moved out from the Pt catalysts. Here, we choose Nb2O5 as the final oxidation products of the Nb species. In the oxidation step, a number of intermediate structures may exist but are omitted, since we mainly focus on the thermodynamic driving force rather than the detailed processes. The Nb atom transport and the Gibbs free energy profile are schematically shown in Figure 6.

formations from C1 to C2 and from C4 to C5, the number and type of atoms are no longer the same. For example, compared with C1, C2 contains one additional Sn atom and one less Pt atom. Therefore, the free energy change in this transformation is expressed in eq 1. ΔGC1→ C2 = ECDFT − μSn + μPt − ECDFT 2 1

EDFT C2

(1)

EDFT C1

Here, and are energies from the DFT calculations, while μSn and μPt are the chemical potentials of the Sn and Pt elements, respectively. The Pt species is regarded as in equilibrium with Pt bulk, which undergoes the same amount of compression as in the Pt/ NTO interface. Thus, μPt is calculated using eq 2. DFT DFT μPt = (E NTO − Pt(4L) − E NTO − Pt(3L))/10

(2)

EDFT NTO−Pt(nL) is the DFT energy of the Pt/NTO interface including n Pt layers, each of which contains 10 Pt atoms. In this way, the same amount of strain is ensured in the calculation. The Sn element in the Pt/NTO interface is assumed to be in equilibrium with SnO2 bulk in the electrochemical environment, i.e., Sn + 2H2O ⇌SnO2 + 4H+ + 4e−. Thus, μSn is expressed in eq 3. DFT μSn = ESnO + 4(μH+ + μe− ) − 2μH O(l) 2 2

(3) 1,55

Using the standard hydrogen potential as the reference, the sum of the chemical potentials of a proton and an electron is calculated using eq 4. μ H + + μ e− =

1 0 μ + kBT ln a H+ − eU 2 H2(g)

(4)

μH0 2(g)

Here, is the chemical potential of a hydrogen molecule at the standard conditions (pH2 = 1 bar and T = 298.15 K) in the gas phase, aH+ is the activity of the protons, and U is the electrode potential relative to the standard hydrogen potential. From the DFT calculations, the μH0 2(g) term is expressed in eq 5, where ZPE represents the zero-point energy and S0 is the standard entropy taken from thermodynamic tables.56 μH0 (g) = E HDFT + ZPE H2(g) − TSH0 2(g) 2 2

(5)

Under the vapor−liquid equilibrium conditions, the chemical potential of water in the liquid phase μH2O(l) is equal to that of water in the gas phase μH2O(g) at the temperature T and the corresponding vapor pressure peq:g ⇌l (0.032 bar at 298.15 K and 0.474 bar at 353.15K). μH2O(l) can thus be calculated using eq 6.

Figure 6. Gibbs free energy diagram of the processes from the Nb segregation through Pt adlayers to the formation of the Nb2O5 bulk. The labels Ci represent different configurations. Three conditions, i.e., the unbiased condition (U = 0 V, T = 25 °C, pH = 1), start/stop conditions of PEMFCs (U = 1.2 V, T = 80 °C, pH = 1), and operating conditions (U = 0.8 V, T = 80 °C, pH = 1) are taken into account, with the corresponding results shown in blue, red, and green, respectively. The Sn (Nb) atoms are represented by gray (azure) spheres.

μH O(l) = μH O(g) = E HDFT + ZPE H2O(g) 2O(g) 2

2

− TSH2O(g)(eq:g ⇌ l)

(6)

The free energy change from C4 to C5 is calculated in a similar way and is expressed in eq 7. ΔGC4 → C5 = ECDFT + μ Nb − μPt − ECDFT 5 4

Among the six configurations, a few of them contain the same number and type of atoms, like C0 and C1, as well as C2, C3, and C4. Thus, the free energy changes from C0 to C1 and from C2 to C4 can be simply approximated as a variation of the DFT energies, because volume and entropy variations are usually negligible in such processes in solids. In the trans-

(7)

Here, the Nb element is regarded as in equilibrium with Nb2O5 bulk in the same electrochemical environment, i.e., 5 1 Nb + 2 H 2O ⇌ 2 Nb2 O5 + 5H+ + 5e−. μNb is thereby expressed in eq 8 and calculated in the same way as μSn. D


Article

Chemistry of Materials μ Nb =

1 DFT 5 ENb2O5 + 5(μ H + + μe− ) − μ H O(l) 2 2 2

(8)

The values of the ZPE and the entropy terms in the above calculations are listed in Table 1. Configurational entropy from locations of Nb atoms in different layers is below 0.1 eV at 353.15 K,54 which is too small and thus not included in the Gibbs free energy diagram. Table 1. ZPE and TS Terms of H2 and H2O (Unit: eV) in the Thermodynamic Analysisa ZPE TS (298.15 K) TS (353.15 K) a

H2(g)

H2O(g)(eq:g⇌l)

0.27 0.40 0.50

0.56 0.67 0.73

Figure 7. Gibbs free energy diagram of the processes from the Sn segregation through Pt adlayers to the formation of the SnO2 bulk. Results under the unbiased condition (U = 0 V, T = 25 °C, pH = 1), start/stop conditions of PEMFCs (U = 1.2 V, T = 80 °C, pH = 1), and operating conditions (U = 0.8 V, T = 80 °C, pH = 1) are shown in blue, red, and green, respectively. The Sn (Nb) atoms are represented by gray (azure) spheres.

The ZPE values are taken from ref 55.

In Figure 6, we consider three conditions of PEMFCs, i.e., the unbiased condition (blue), start/stop conditions (red), and operating conditions (green). It can be seen that the most stable configuration is C3 under the unbiased condition, where the Nb dopant alloys with the Pt layers. Nevertheless, since the Nb dopants do not prefer to locate at the outermost layer of the SnO2 substrate, this transport is hindered by a thermodynamic barrier of 0.45 eV as well as a potential kinetic barrier. We note in passing that the higher doping concentration (20.8 at. %) in this case leads to a larger Nb thermodynamic barrier compared to that in Figure 3 (4.2 at. %). The corresponding kinetic barrier is expected to be even higher, which further limits the transport process. Upon increasing the electrode potential, we find a larger thermodynamic driving force for the Nb oxidation. However, because the barrier of the Nb transport into the Pt layers exhibits a significant increase under these conditions, such a process is expected to be kinetically limited by the Pt adsorbates. In other words, the Pt catalysts can help to keep the Nb dopants inside the SnO2 host. This result is consistent with the experiments, which find an enhancement of the electronic conductivity upon the attachment of Pt on NTO.38 As a note, if the Pt layers do not completely cover the NTO substrate, the Nb dopants may leave the NTO system, leading to the Nb depletion and a consequent accumulation of oxides. In this sense, covering of the NTO substrate by Pt, given that it can be experimentally realized, may be used to improve the performance of the electrodes in PEMFCs. In a similar way, we investigate the alloying of Sn with the Pt layers and its oxidation to SnO2. The results are schematically shown in Figure 7. Under the unbiased condition, the Sn atoms prefer to penetrate into and alloy with the Pt layers, e.g., forming Pt3Sn, which could affect the activity of the Pt catalysts toward the ORR. Under operational conditions, the Sn atoms are less likely to alloy with Pt but tend to be oxidized on the catalyst surfaces. The thermodynamic barrier of the Sn transport under a potential of 0.8 V, interestingly, is much lower than the value of Nb, indicating a higher possibility of oxide formation in the case of Sn than that of the Nb dopants. As a result, SnO2 oxides may cover the Pt catalysts and lead to a decrease of the activity as well as the electronic conductivity of the electrodes. Charge Transport Properties across Pt/NTO Interfaces. The device model for the electronic transport studies has been shown in Figure 1. Since the Nb atoms prefer to aggregate in the SnO2 host, two adjacent Nb dopants are placed

in the left electrode, corresponding to a doping concentration of 8.3 at. %, and remaining unchanged in the calculations. We first investigate the dependence of the electric current on the location of Nb dopants. A single Nb atom is placed sequentially from the first to the fourth layer of the SnO2 substrate, with the calculated I−V curves shown in Figure 8a. It can be seen that the current is similar for different Nb locations, especially at a low voltage. It means that the transport properties are not particularly sensitive to a specific position of the Nb dopant. The results are in sharp contrast with the Sbdoping, where the current can vary ∼500% depending on the doping sites.33 We then increase the number of Nb dopants and investigate the effect of concentration on the transport properties. The Nb atoms are increased one by one, with the maximum concentration corresponding to five dopants out of 24 doping sites (i.e., 20.8 at. %) in the subsurface regions. The calculated I−V curves are shown in Figure 8b. Here, all the configurations correspond to the most stable ones at the respective concentrations. We find that the electric current does not increase monotonically, but instead it shows an irregular trend. Under a positive bias, the interface that contains two Nb dopants has the lowest conductance (blue line), while the electric current becomes much larger and is almost the same in the cases of three (red line) and five (green line) Nb dopants. The results show that the dependence on Nb concentrations is much more complicated compared with the Sb-doping,33 but generally, a relatively higher conductance is observed for odd numbers of Nb-dopant atoms compared to the even numbers, which would tend to support the formation of Nb4+−Nb4+ pairs that have been found not to contribute to the electronic conductivity.44 Thus, depending on the experimental preparation and the Nb distribution in NTO samples, the electric conductivity could change a lot. This finding is consistent with the variety of the reported values in different experiments.29,31,34−37,39 Codoping at Pt/SnO2 Interface. At the Pt/SnO2 interface, the preferred doping site for Nb and Sb atoms varies. The Nb dopants tend to localize in the subsurface regions of the SnO2 substrate, whereas for the Sb dopants, the first layer is preferred.33 This indicates that the two elements do not compete with each other in terms of the preferred doping E


Article


Figure 8. Calculated I−V curves of the Pt/NTO interface (a) with one Nb dopant located in the different SnO2 layers and (b) with different Nb concentrations in the subsurface regions of SnO2.

position, especially at low concentrations. Meanwhile, both Nb and Sb atoms could potentially contribute an extra electron as the charge carrier. Thus, it is interesting to investigate whether the transport properties of the Pt/SnO2 interface could be further improved by codoping with Nb to a system which has already been Sb-doped. We substitute one Nb atom and one Sb atom into the SnO2 host, to study the situation of a low doping concentration. We place the Nb dopant in the second layer of the SnO2 substrate, i.e., its most favorable doping site, and adjust the position of the Sb atom. Similarly, we also put the Sb dopant in the first layer, and change the Nb location accordingly. In such a way, seven codoped configurations are constructed, as listed in Table 2. We find that the most stable configuration corresponds to the one where the two dopants locate at their respective favorable doping sites. Table 2. Relative Energies of the Seven Constructed Codoped Configurations at the Pt/SnO2 Interfacea configuration

energy (eV)

Sb(first)−Nb(first) Sb(first)−Nb(second) Sb(first)−Nb(third) Sb(first)−Nb(fourth) Sb(second)−Nb(second) Sb(third)−Nb(second) Sb(fourth)−Nb(second)

0.30 0.00 0.07 0.09 0.70 0.56 0.54

a

The labels in the parentheses represent the layers of location of the two dopants.

The calculated I−V curves are shown in Figure 9. Two different types of left electrodes, ATO (4.2 at. % Sb) and NTO (8.3 at. % Nb), are used in the device model, in order to simulate an Sb-dominant and a Nb-dominant situation, respectively. With the ATO electrode (Figure 9a), the codoped interface shows improved transport properties compared with the single-dopant case. Remarkably, the codoped interface has a larger conductance than the Sb-doped interface containing two Sb atoms, although the latter potentially provides the same number of charge carriers. This result indicates that a small amount of Nb may improve the transport properties of a Pt/ ATO interface under a low dopant concentration. From Figure 9b, we find that, upon full substitution of the first layer by Sb, further addition of a Nb dopant does not increase the conductivity, indicating that a saturation point may also exist in the concentration−conductance relation as in the Sb-doping case.33 The results of the NTO electrode are presented in Figure 9c. It is found that the conductance of the codoped

Figure 9. Comparison of the calculated I−V curves between the codoped and the single-dopant systems. An ATO electrode, doped by one Sb atom, is used for parts a and b, and an NTO electrode, doped by two Nb atoms, is used for part c.

interface is only slightly larger than that of the Pt/NTO interface (1 Nb atom), meaning that, at a Nb-dominant interface, addition of additional Sb atoms does not improve the conductivity of the system. Catalytic Properties of the Pt Adlayers toward the Oxygen Reduction Reaction. Pt is an excellent catalyst toward the ORR at the cathode of PEMFCs. Its catalytic activity can be affected by the supporting materials like the NTO substrates studied in this work. Here, we investigate the effects of the concentration and location of the Nb dopants on the catalytic properties. The adsorption energy of an OH group is used as the descriptor for the activity, from which the optimal F



■

CONCLUSIONS In summary, through a combination of DFT and NEGF calculations, we investigate the Nb segregation at the Pt/NTO interfaces under different electrochemical conditions. The corresponding effects on the electronic transport and the catalytic properties are also studied. We find that the Nb dopants tend to aggregate in the subsurface layers of the NTO host, whereas under the operating condition of PEMFCs, their transport across the interfaces is limited due to a high thermodynamic barrier. In contrast, the similar transport of Sn is more facile, indicating the possible formations of Sn−Pt alloys and tin oxides. The electronic conductivity of the Pt/ NTO interface is not particularly sensitive to the distance of the Nb dopants relative to the interface, but depends explicitly on the Nb concentration and configuration, e.g., Nb−Nb pairing. Moreover, the NTO substrate can improve the catalytic activity of the Pt adlayers toward the ORR via a ligand effect. This work sheds further light on future applications of Pt/NTO-based materials as possible cathodes for PEMFC applications.

catalyst should bind the OH group about 0.1 eV weaker than Pt(111).57−60 The OH group is placed at the top site of the Pt layers, and its binding energy is calculated using eq 9, with reference to H2O and H2 as in ref 1. 1 DFT DFT Ead(OH) = E Pt/NTO E H (g) − OH + 2 2 DFT − (E Pt/NTO + E HDFT ) 2O(g)

(9)

In Table 3, we list the relative OH binding energies on a few Pt/NTO interfaces compared to the value on an isolated and Table 3. Relative Binding Energy (Unit: eV) of an OH Group on the Different Pt Pt/NTO Systemsa 3 layers of Pt no Nb atom 1 Nb atom (s5) 1 Nb atom (2s5) 1 Nb atom (3s4) 1 Nb atom (4s1) 2 Nb atoms 3 Nb atoms 4 Nb atoms (a) 4 Nb atoms (b) 5 Nb atoms Pt slab (strain) Pt slab (unstrained)

Article

0.11 0.09 0.11 0.11 0.10 0.13 0.11 0.12 0.12 0.16 0.07 0.00

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Phone: 0045-46775818. ORCID

Qiang Fu: 0000-0002-6682-8527 Present Address †

Institut für Physik and IRIS Adlershof, Humboldt-Universität zu Berlin, Zum Großen Windkanal 6, 12489 Berlin, Germany.

a

The values on isolated Pt(111) slabs without a NTO substrate are also listed for comparison.

Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We gratefully acknowledge financial support from the European Union’s Seventh Framework Programme (FP7/2013-2017) for the Fuel Cells and Hydrogen Joint Technology Initiative under Grant Agreement 325327SMARTCAT. Q.F. thanks Umberto Martinez, Marcin Dulak, and Vladimir Tripkovic for helpful discussions. J.M.G.L. acknowledges support from the Spanish Ministry of Economy and Competitiveness under Projects FIS2012-30996 and FIS2013-46159-C3-1-P, and from the Villum Foundation’s Young Investigator Programme (fourth round, project: In silico design of efficient materials for next generation batteries. Grant: 10096).

unstrained three-layer Pt(111) slab. With this approach, the functional dependence on the adsorption energy is largely canceled. It is important since the PBE functional leads to an overbinding of the OH group compared to, for example, the RPBE functional.61 According to Table 3, the OH binding on the Pt/NTO systems is weakened by 0.09−0.16 eV, compared to that on a pristine unstrained Pt(111) slab. This value is very close to the requirement for the optimal catalyst toward the ORR. In a comparison with the Ead(OH) result on an isolated Pt slab under the same amount of compression, we find that the strain effect, coming from the construction of the Pt/NTO interface model, accounts for 0.07 eV. The rest of the binding energy change is therefore attributed to the NTO substrate. In fact, since NTO is an n-doped semiconductor, the dopant induced ligand effect could also contribute to a weakened OH binding. For example, the NTO substrate may donate electrons to the Pt layers, which results in a further decrease of Ead(OH). Focusing solely on the dopant induced ligand effect, the binding energy of OH is weakened by an additional 0.03−0.09 eV, depending on the specific concentration and configuration of the Nb dopants. Thus, the strain and the ligand effects in this case are equally important. It is worth noting that an improvement of the Pt activity by the NTO substrate has also been experimentally reported.38 This could lead to enhanced catalytic activity of the Pt/NTO system, and a detailed analysis on the effects of, e.g., Pt layer thickness and possible surface reconstructions would be interesting for further studies.

■

REFERENCES

(1) Nørskov, J. K.; Rossmeisl, J.; Logadottir, A.; Lindqvist, L.; Kitchin, J. R.; Bligaard, T.; Jónsson, H. Origin of the Overpotential for Oxygen Reduction at a Fuel-Cell Cathode. J. Phys. Chem. B 2004, 108, 17886−17892. (2) Escudero-Escribano, M.; Verdaguer-Casadevall, A.; Malacrida, P.; Grønbjerg, U.; Knudsen, B. P.; Jepsen, A. K.; Rossmeisl, J.; Stephens, I. E. L.; Chorkendorff, I. Pt5Gd as a Highly Active and Stable Catalyst for Oxygen Electroreduction. J. Am. Chem. Soc. 2012, 134, 16476−16479. (3) Greeley, J.; Stephens, I. E. L.; Bondarenko, A. S.; Johansson, T. P.; Hansen, H. A.; Jaramillo, T. F.; Rossmeisl, J.; Chorkendorff, I.; Nørskov, J. K. Alloys of Platinum and Early Transition Metals as Oxygen Reduction Electrocatalysts. Nat. Chem. 2009, 1, 552−556. (4) Jennings, P. C.; Lysgaard, S.; Hansen, H. A.; Vegge, T. Decoupling Strain and Ligand Effects in Ternary Nanoparticles for Improved ORR Electrocatalysis. Phys. Chem. Chem. Phys. 2016, 18, 24737−24745. (5) Kinoshita, K. Carbon: Electrochemical and Physicochemical Properties, 1st ed.; Wiley-Interscience, 1988. G


Article

Chemistry of Materials (6) Pourbaix, M. Atlas of Electrochemical Equilibria in Aqueous Solutions, 2nd ed.; National Association of Corrosion, 1974. (7) Maass, S.; Finsterwalder, F.; Frank, G.; Hartmann, R.; Merten, C. Carbon Support Oxidation in PEM Fuel Cell Cathodes. J. Power Sources 2008, 176, 444−451. (8) Avasarala, B.; Moore, R.; Haldar, P. Surface Oxidation of Carbon Supports due to Potential Cycling under PEM Fuel Cell Conditions. Electrochim. Acta 2010, 55, 4765−4771. (9) Liu, Z. Y.; Zhang, J. L.; Yu, P. T.; Zhang, J. X.; Makharia, R.; More, K. L.; Stach, E. A. Transmission Electron Microscopy Observation of Corrosion Behaviors of Platinized Carbon Blacks under Thermal and Electrochemical Conditions. J. Electrochem. Soc. 2010, 157, B906−B913. (10) Smith, M. C.; Gilbert, J. A.; Mawdsley, J. R.; Seifert, S.; Myers, D. J. In Situ Small-Angle X-ray Scattering Observation of Pt Catalyst Particle Growth during Potential Cycling. J. Am. Chem. Soc. 2008, 130, 8112−8113. (11) Schulenburg, H.; Schwanitz, B.; Linse, N.; Scherer, G.; Wokaun, A.; Krbanjevic, J.; Grothausmann, R.; Manke, I. 3D Imaging of Catalyst Support Corrosion in Polymer Electrolyte Fuel Cells. J. Phys. Chem. C 2011, 115, 14236−14243. (12) Mayrhofer, K. J.; Meier, J. C.; Ashton, S. J.; Wiberg, G. K.; Kraus, F.; Hanzlik, M.; Arenz, M. Fuel Cell Catalyst Degradation on the Nanoscale. Electrochem. Commun. 2008, 10, 1144−1147. (13) Mayrhofer, K. J.; Ashton, S. J.; Meier, J. C.; Wiberg, G. K.; Hanzlik, M.; Arenz, M. Non-destructive Transmission Electron Microscopy Study of Catalyst Degradation under Electrochemical Treatment. J. Power Sources 2008, 185, 734−739. (14) Shao, Y.; Liu, J.; Wang, Y.; Lin, Y. Novel Catalyst Support Materials for PEM Fuel Cells: Current Status and Future Prospects. J. Mater. Chem. 2009, 19, 46−59. (15) Sharma, S.; Pollet, B. G. Support Materials for PEMFC and DMFC Electrocatalysts - A Review. J. Power Sources 2012, 208, 96− 119. (16) Okanishi, T.; Matsui, T.; Takeguchi, T.; Kikuchi, R.; Eguchi, K. Chemical Interaction between Pt and SnO2 and Influence on Adsorptive Properties of Carbon Monoxide. Appl. Catal., A 2006, 298, 181−187. (17) Saha, M. S.; Li, R.; Sun, X. Composite of Pt-Ru Supported SnO2 Nanowires Grown on Carbon Paper for Electrocatalytic Oxidation of Methanol. Electrochem. Commun. 2007, 9, 2229−2234. (18) Masao, A.; Noda, S.; Takasaki, F.; Ito, K.; Sasaki, K. CarbonFree Pt Electrocatalysts Supported on SnO2 for Polymer Electrolyte Fuel Cells. Electrochem. Solid-State Lett. 2009, 12, B119−B122. (19) Zhang, P.; Huang, S.-Y.; Popov, B. N. Mesoporous Tin Oxide as an Oxidation-Resistant Catalyst Support for Proton Exchange Membrane Fuel Cells. J. Electrochem. Soc. 2010, 157, B1163−B1172. (20) Dou, M.; Hou, M.; Liang, D.; Lu, W.; Shao, Z.; Yi, B. SnO2 Nanocluster Supported Pt Catalyst with High Stability for Proton Exchange Membrane Fuel Cells. Electrochim. Acta 2013, 92, 468−473. (21) Fabbri, E.; Rabis, A.; Kötz, R.; Schmidt, T. J. Pt Nanoparticles Supported on Sb-doped SnO2 Porous Structures: Developments and Issues. Phys. Chem. Chem. Phys. 2014, 16, 13672−13681. (22) Fröhlich, D.; Kenklies, R.; Helbig, R. Band-Gap Assignment in SnO2 by Two-Photon Spectroscopy. Phys. Rev. Lett. 1978, 41, 1750− 1751. (23) Reimann, K.; Steube, M. Experimental Determination of the Electronic Band Structure of SnO2. Solid State Commun. 1998, 105, 649−652. (24) Szczuko, D.; Werner, J.; Behr, G.; Oswald, S.; Wetzig, K. Surface-related Investigations to Characterize Different Preparation Techniques of Sb-doped SnO2 Powders. Surf. Interface Anal. 2001, 31, 484−491. (25) Szczuko, D.; Werner, J.; Oswald, S.; Behr, G.; Wetzig, K. XPS Investigations of Surface Segregation of Doping Elements in SnO2. Appl. Surf. Sci. 2001, 179, 301−306. (26) Dobler, D.; Oswald, S.; Werner, J.; Arabczyk, W.; Behr, G.; Wetzig, K. X-ray Photoelectron Spectroscopy Investigation of

Segregation Processes at Sb and In Doped SnO2. Chem. Phys. 2003, 286, 375−383. (27) Saadeddin, I.; Hilal, H. S.; Pecquenard, B.; Marcus, J.; Mansouri, A.; Labrugere, C.; Subramanian, M. A.; Campet, G. Simultaneous Doping of Zn and Sb in SnO2 Ceramics: Enhancement of Electrical Conductivity. Solid State Sci. 2006, 8, 7−13. (28) Lee, K.-S.; Park, I.-S.; Cho, Y.-H.; Jung, D.-S.; Jung, N.; Park, H.Y.; Sung, Y.-E. Electrocatalytic Activity and Stability of Pt Supported on Sb-doped SnO2 Nanoparticles for Direct Alcohol Fuel Cells. J. Catal. 2008, 258, 143−152. (29) Wang, Y.; Brezesinski, T.; Antonietti, M.; Smarsly, B. Ordered Mesoporous Sb-, Nb-, and Ta-Doped SnO2 Thin Films with Adjustable Doping Levels and High Electrical Conductivity. ACS Nano 2009, 3, 1373−1378. (30) Stroppa, D. G.; Montoro, L. A.; Beltrán, A.; Conti, T. G.; da Silva, R. O.; Andrés, J.; Leite, E. R.; Ramirez, A. J. Dopant Segregation Analysis on Sb:SnO2 Nanocrystals. Chem. - Eur. J. 2011, 17, 11515− 11519. (31) Takasaki, F.; Matsuie, S.; Takabatake, Y.; Noda, Z.; Hayashi, A.; Shiratori, Y.; Ito, K.; Sasaki, K. Carbon-Free Pt Electrocatalysts Supported on SnO2 for Polymer Electrolyte Fuel Cells: Electrocatalytic Activity and Durability. J. Electrochem. Soc. 2011, 158, B1270− B1275. (32) Alsac, A. A.; Yildiz, A.; Serin, T.; Serin, N. Improved Conductivity of Sb-Doped SnO2 Thin Films. J. Appl. Phys. 2013, 113, 063701. (33) Fu, Q.; Colmenares Rausseo, L. C.; Martinez, U.; Dahl, P. I.; García Lastra, J. M.; Vullum, P. E.; Svenum, I.-H.; Vegge, T. Effect of Sb Segregation on Conductance and Catalytic Activity at Pt/Sb-doped SnO2 Interface: a Synergetic Computational and Experimental Study. ACS Appl. Mater. Interfaces 2015, 7, 27782−27795. (34) Gokulakrishnan, V.; Parthiban, S.; Jeganathan, K.; Ramamurthi, K. Investigations on the Structural, Optical and Electrical Properties of Nb-doped SnO2 Thin Films. J. Mater. Sci. 2011, 46, 5553−5558. (35) Stefik, M.; Cornuz, M.; Mathews, N.; Hisatomi, T.; Mhaisalkar, S.; Grätzel, M. Transparent, Conducting Nb:SnO2 for Host-Guest Photoelectrochemistry. Nano Lett. 2012, 12, 5431−5435. (36) Cavaliere, S.; Subianto, S.; Savych, I.; Tillard, M.; Jones, D. J.; Rozièr e, J. Dopant-Driven Nanostructured Loose-Tube SnO 2 Architectures: Alternative Electrocatalyst Supports for Proton Exchange Membrane Fuel Cells. J. Phys. Chem. C 2013, 117, 18298−18307. (37) Tsukatsune, T.; Takabatake, Y.; Noda, Z.; Daio, T.; Zaitsu, A.; Lyth, S. M.; Hayashi, A.; Sasaki, K. Platinum-Decorated Tin Oxide and Niobium-Doped Tin Oxide PEFC Electrocatalysts: Oxygen Reduction Reaction Activity. J. Electrochem. Soc. 2014, 161, F1208−F1213. (38) Senoo, Y.; Kakinuma, K.; Uchida, M.; Uchida, H.; Deki, S.; Watanabe, M. Improvements in Electrical and Electrochemical Properties of Nb-doped SnO2−δ Supports for Fuel Cell Cathodes due to Aggregation and Pt Loading. RSC Adv. 2014, 4, 32180−32188. (39) Hu, Y.; Yella, A.; Guldin, S.; Schreier, M.; Stellacci, F.; Grätzel, M.; Stefik, M. High-Surface-Area Porous Platinum Electrodes for Enhanced Charge Transfer. Adv. Energy Mater. 2014, 4, 1400510. (40) Kakinuma, K.; Kim, I.-T.; Senoo, Y.; Yano, H.; Watanabe, M.; Uchida, M. Electrochemical Oxidation of Hydrolyzed Poly Oxymethylene-dimethyl Ether by PtRu Catalysts on Nb-Doped SnO2−δ Supports for Direct Oxidation Fuel Cells. ACS Appl. Mater. Interfaces 2014, 6, 22138−22145. (41) Savych, I.; Subianto, S.; Nabil, Y.; Cavaliere, S.; Jones, D.; Rozière, J. Negligible Degradation upon in situ Voltage Cycling of a PEMFC Using an Electrospun Niobium-Doped Tin Oxide Supported Pt Cathode. Phys. Chem. Chem. Phys. 2015, 17, 16970−16976. (42) Oakton, E.; Tillier, J.; Siddiqi, G.; Mickovic, Z.; Sereda, O.; Fedorov, A.; Copéret, C. Structural Differences between Sb- and Nbdoped Tin Oxides and Consequences for Electrical Conductivity. New J. Chem. 2016, 40, 2655. (43) Lü, X.; Yang, W.; Quan, Z.; Lin, T.; Bai, L.; Wang, L.; Huang, F.; Zhao, Y. Enhanced Electron Transport in Nb-Doped TiO2 NanoH


Article

Chemistry of Materials particles via Pressure-Induced Phase Transitions. J. Am. Chem. Soc. 2014, 136, 419−426. (44) Sahasrabudhe, G.; Krizan, J.; Bergman, S. L.; Cava, R. J.; Schwartz, J. Million-fold Increase of the Conductivity in TiO2 Rutile through 3% Niobium Incorporation. Chem. Mater. 2016, 28, 3630− 3633. (45) Tripković, V.; Abild-Pedersen, F.; Studt, F.; Cerri, I.; Nagami, T.; Bligaard, T.; Rossmeisl, J. Metal Oxide-Supported Platinum Overlayers as Proton-Exchange Membrane Fuel Cell Cathodes. ChemCatChem 2012, 4, 228−235. (46) Kresse, G.; Furthmüller, J. Efficiency of ab-initio Total Energy Calculations for Metals and Semiconductors Using a Plane-Wave Basis Set. Comput. Mater. Sci. 1996, 6, 15−50. (47) Kresse, G.; Furthmüller, J. Efficient Iterative Schemes for ab initio Total-Energy Calculations Using a Plane-Wave Basis Set. Phys. Rev. B: Condens. Matter Mater. Phys. 1996, 54, 11169−11186. (48) Perdew, J. P.; Burke, K.; Ernzerhof, M. Generalized Gradient Approximation Made Simple. Phys. Rev. Lett. 1996, 77, 3865−3868. (49) Chadi, D. J.; Cohen, M. L. Special Points in the Brillouin Zone. Phys. Rev. B 1973, 8, 5747−5753. (50) Atomistix ToolKit version 2014.0; QuantumWise A/S (www. quantumwise.com). (51) Brandbyge, M.; Mozos, J.-L.; Ordejón, P.; Taylor, J.; Stokbro, K. Density-Functional Method for Nonequilibrium Electron Transport. Phys. Rev. B: Condens. Matter Mater. Phys. 2002, 65, 165401. (52) Soler, J. M.; Artacho, E.; Gale, J. D.; García, A.; Junquera, J.; Ordejón, P.; Sánchez-Portal, D. The SIESTA Method for ab initio Order- N Materials Simulation. J. Phys.: Condens. Matter 2002, 14, 2745−2779. (53) Wendt, S.; Sprunger, P. T.; Lira, E.; Madsen, G. K. H.; Li, Z.; Hansen, J. Ø.; Matthiesen, J.; Blekinge-Rasmussen, A.; Laegsgaard, E.; Hammer, B.; Besenbacher, F. The Role of Interstitial Sites in the Ti3d Defect State in the Band Gap of Titania. Science 2008, 320, 1755− 1759. (54) Xing, W.; Toyoura, K.; Norby, T. Effect of Cation Disorder on the Solubility and Result of Doping in Oxides. Int. J. Hydrogen Energy 2012, 37, 8062−8065. (55) Man, I. C.; Su, H.-Y.; Calle-Vallejo, F.; Hansen, H. A.; Martínez, J. I.; Inoglu, N. G.; Kitchin, J.; Jaramillo, T. F.; Nørskov, J. K.; Rossmeisl, J. Universality in Oxygen Evolution Electrocatalysis on Oxide Surfaces. ChemCatChem 2011, 3, 1159−1165. (56) NIST-JANAF Thermochemical Tables; (http://kinetics.nist.gov/ janaf/). (57) Stephens, I. E. L.; Bondarenko, A. S.; Perez-Alonso, F. J.; CalleVallejo, F.; Bech, L.; Johansson, T. P.; Jepsen, A. K.; Frydendal, R.; Knudsen, B. P.; Rossmeisl, J.; Chorkendorff, I. Tuning the Activity of Pt(111) for Oxygen Electroreduction by Subsurface Alloying. J. Am. Chem. Soc. 2011, 133, 5485−5491. (58) Viswanathan, V.; Hansen, H. A.; Rossmeisl, J.; Nørskov, J. K. Universality in Oxygen Reduction Electrocatalysis on Metal Surfaces. ACS Catal. 2012, 2, 1654−1660. (59) Hansen, H. A.; Viswanathan, V.; Nørskov, J. K. Unifying Kinetic and Thermodynamic Analysis of 2 e− and 4 e− Reduction of Oxygen on Metal Surfaces. J. Phys. Chem. C 2014, 118, 6706−6718. (60) Tripkovic, V.; Hansen, H. A.; Rossmeisl, J.; Vegge, T. First Principles Investigation of the Activity of Thin Film Pt, Pd and Au Surface Alloys for Oxygen Reduction. Phys. Chem. Chem. Phys. 2015, 17, 11647−11657. (61) Hammer, B.; Hansen, L.; Nørskov, J. Improved Adsorption Energetics within Density-functional Theory Using Revised PerdewBurke-Ernzerhof Functionals. Phys. Rev. B: Condens. Matter Mater. Phys. 1999, 59, 7413−7421.

I


Pt Interfaces: Dopant

Recommend Documents