Reversible GaInP2 Surface Passivation by Water Adsorption: A Model

Article pubs.acs.org/JPCC

Reversible GaInP2 Surface Passivation by Water Adsorption: A Model System for Ambient-Dependent Photoluminescence James L. Young,†,‡ Henning Döscher,†,§ John A. Turner,† and Todd G. Deutsch*,† †

National Renewable Energy Laboratory, Golden, Colorado 80401, United States Materials Science and Engineering Program, University of Colorado, Boulder, Colorado 80309, United States § Philipps-Universität Marburg, D-35032 Marburg, Germany ‡

ABSTRACT: The photoluminescence (PL) intensity of semiconductors can be modulated by their ambient. We show GaInP2 to respond reversibly to water vapor, irreversibly to oxygen, and with a time dependence to air. We characterize the reversible PL response to water vapor in a set of steady-state measurements that reveal a systematic dependence on pressure. We derive a model for this behavior based on Langmuir adsorption and Shockley− Read−Hall recombination principles to describe how partial pressure controls luminescence. The expression for the GaInP2/ water vapor system shows excellent agreement to measurements. Combined, the PL monitoring technique and model demonstrate a quantitative approach for probing gas/surface defect interactions of semiconductor processing steps, luminescence-based sensors, and photocatalytic surfaces.

■

INTRODUCTION

significant implications for photocatalysis, photovoltage, and photocorrosion.11−13

■

Understanding semiconductor surface defects and controlling their populations is critical to the development and performance of solid-state electronic devices. Some mitigation strategies include gas-phase processing steps to passivate defect states manifested as dangling bonds or surface oxides that cause Fermi-level pinning and/or nonradiative recombination losses.1−3 Chemical sensors and chemical field-effect transistors (ChemFETs) show luminescence that is modulated by a reversibly adsorbing species.4,5 Reshchikov6 showed that GaN photoluminescence (PL) responds reversibly and described a model where adsorbed oxygen induces surface states to reduce radiative recombination. Liu et al.5 fabricated ZnO oxygen sensors whose PL decreases with increasing oxygen partial pressure and attributed it to dissociative adsorption on surface defects. Kocha et al.7 observed a PL increase after etching GaInP2 in sulfuric acid but did not consider any effect of performing the measurement in air. Furthermore, none of these earlier reports suggested a quantitative model for the dependence of PL intensity on ambient pressure. We show that the PL enhancement of GaInP2 only persists in air (or other humid ambient) and depends systematically on the partial pressure of water. The model derived here not only is important to solid-state electronic devices but also is critically important to water-splitting devices where GaInP2 operates in contact with an aqueous electrolyte.8,9 Such systems demonstrate record 12−14% solar-to-hydrogen conversion efficiencies and are capable of the maximum practical 25%.10 The interaction of water and oxygen with photoelectrodes has © 2016 American Chemical Society

METHODS We perform PL measurements of GaInP2 in both air and controlled gas-phase ambient defined by the vacuum and control systems of an atomic layer deposition (ALD) reactor14 modified for in situ PL monitoring. An attachment houses a vertically oriented sample stage positioned ∼1 cm behind a ConFlat-mounted 1 in. diameter quartz window while pneumatic and leak valves introduce and maintain water vapor, oxygen, or air pressure measured by a Baratron capacitance manometer. A collimated laser diode (ThorLabs) with center wavelength λ = 532 nm adjusted to 0.25 mW provides illumination that an objective lens outside of the window focuses on the sample while collecting PL. An Ocean Optics USB2000+ spectrometer monitors the PL signal using an integration time of more than 1 s as necessary for sufficient signal at room temperature. With a ∼100 μm spot size, the illumination intensity of 3 W/cm2 is considerably higher than necessary for flat-band conditions15 where surface-charge effects are compensated.16 For example, flat-band potential measurements on GaAsPN17 saturate with only 0.6 W/cm2. Because surface charge is compensated, changes in PL intensity result from changes in the surface defect population. A dichroic mirror (ThorLabs) with 50% transmission/reflection at λ = 567 Received: December 21, 2015 Revised: February 8, 2016 Published: February 9, 2016 4418

DOI: 10.1021/acs.jpcc.5b12498 J. Phys. Chem. C 2016, 120, 4418−4422

Article

The Journal of Physical Chemistry C

Figure 1. (a) Band-to-band PL spectrum of GaInP2 as measured in air ambient before (black) and after (green) etching in concentrated sulfuric acid. (b) The PL intensity response of as-received (black) and etched (gray) GaInP2 during air-to-vacuum transition as well as the PL intensity of etched GaInP2 continuously monitored in air (green). (c) Starting from vacuum, PL is monitored while introducing air (green), oxygen (red), and water vapor (blue). The PL continuously monitored in vacuum (black) is stable. (d) Response to water vapor (blue) is reversible and that to oxygen (red) is irreversible upon evacuation.

experiments were conducted to expose GaInP2 to vacuum, oxygen, water vapor, and air. Figure 1c shows that etched GaInP2 PL is stable in vacuum (black), increases in the presence of water vapor (blue), decreases in oxygen (red), and first increases for 30 s and then decays in air (green). Because air contains oxygen and water vapor, we attribute the initial PL increase to adsorbed water and the subsequent decay to oxygen. We exposed another etched GaInP2 sample and followed with evacuation (Figure 1d) to show that the PL response to water vapor (blue) is reversible and that to oxygen (red) is irreversible. Longer evacuation times (∼20 min) are necessary for complete reversibility, as will be shown. Presumably, the quenched PL results from reaction with oxygen under illumination to form a surface oxide that introduces recombination sites. Kocha et al.7 characterized its composition with XPS. By avoiding the irreversible effect of oxygen exposure, we characterize the pressure-dependent PL response of GaInP2 to water vapor necessary to validate our model. We expand on Reshchikov’s6 description of oxygen-induced traps on GaN to derive a quantitative model based on principles of Langmuir adsorption18 and Shockley−Read−Hall (SRH) carrier recombination.19,20 Both concepts illustrated in Figure 2 apply to luminescent semiconductors reversibly passivated by a gas-phase adsorbate in general. Reversibility is necessary to achieve a modulated response from gas sensors and ChemFETs and is desirable for catalytic surfaces such as GaInP2 photoelectrodes, where interactions of intermediate strength are optimal, as suggested by Sabatier’s principle. Irreversible passivation is desirable in other solid-state devices for which PL monitoring and associated kinetic models

nm separates reflected illumination from the PL signal, which is centered and constant at λ = 661 nm (Figure 1a), allowing the peak height to serve as the parameter for PL intensity when monitoring it in changing ambient. The 2 μm thick Ga0.51In0.49P (GaInP2) epilayers that are Zn-doped p-type to 1 × 1017 cm−3 are grown on degenerately Zn-doped (001) GaAs substrates miscut 2° toward (110) by ambient-pressure organometallic vapor-phase epitaxy (AP-OMVPE). The doping density is confirmed with Mott−Schottky measurement and analysis. Samples are etched for 2 min in concentrated sulfuric acid with a sequence of solvent rinses (deionized water, acetone, methanol, deionized water) and blown dry with a nitrogen gun before and after etching.

■

RESULTS AND DISCUSSION We first characterize the PL response of GaInP2 to ambient. Figure 1a shows that the etching treatment enhances PL relative to unetched samples when PL measurements are performed in air. The enhancement slowly decays by ∼20% over 5 min in air (Figure 1b, green trace). However, evacuating air after 1 min causes a rapid drop in the PL intensity (Figure 1b, gray trace) that approaches that of the unetched sample. Notably, the PL intensity of the unetched sample was unperturbed by vacuum (Figure 1b, black trace), suggesting that the observed PL enhancement results from interaction of the surface with air (oxygen and/or water). The PL intensity remains stable for 2 days in air when taking measurements intermittently (similar to Kocha et al.7) but decays during continuous measurement, indicating higher GaInP2/air reactivity under illumination. To probe the effect of the individual components of air on the PL intensity, 4419

DOI: 10.1021/acs.jpcc.5b12498 J. Phys. Chem. C 2016, 120, 4418−4422

Article

The Journal of Physical Chemistry C

G = UB,R +

1 US,NR d

(2)

The recombination rate at the surface, US,NR, is proportional to recombination velocity S and excess carrier concentration n′.22 With sufficient illumination intensity for the flat-band condition,15,17 n′ is uniform to give UB,R = G −

1 n′S d

(3)

The PL signal PLdet originates from UB,R within a volume V but is not measured directly in practice. A fraction C1 of UB,R will escape the sample, of which a fraction C2 is detected depending on the collection efficiency and system throughput. Thus, we express the detected luminescence as Figure 2. Semiconductor with conduction and valence band energies (ECB, EVB) and surface defect (*) population exposed to gas A with partial pressure PA adsorbing reversibly as A*. Recombination is radiative (1), yielding PL, or nonradiative (2), mediated by a surface defect. PA modulates A* and * populations and, thus, the frequency of (1) and (2), making PL intensity a function f(PA).

PLdet = C1C2VUB,R

Both C1 and C2 may be characterized individually, but here, we let C0 = C1C2 while substituting n′ = Gτ, where τ is the bulk carrier lifetime. ⎡ τ ⎤ PLdet = C0GV ⎢1 − S ⎥ ⎣ d ⎦

could be developed, but here, we limit our scope to reversible adsorption. Langmuir18 described how steady-state adsorbate populations depend on ambient pressure, while SRH related19,20 carrier recombination rates to defect populations. The following derivation links these principles when an adsorbate passivates or induces surface defects. Fixed-intensity illumination generates a steady-state carrier population within a volume V of the semiconductor defined by its illumination spot area A and optical absorption depth d. The generation rate G [cm−3 s−1] is equal to the recombination rate U [cm−3 s−1], which has bulk UB [cm−3 s−1] and surface US [cm−2 s−1] components. Each component has radiative (subscript R) or nonradiative (subscript NR) contributions according to G = U = UB +

(4)

or

S=

PLdet ⎤ d⎡ ⎢1 − ⎥ τ⎣ C0GV ⎦ (5)

The parameter S is equal to vthσnNt, where vth, σn, and Nt are carrier thermal velocity, trap capture cross section, and number of surface traps, respectively. When the presence of an adsorbate activates or deactivates a trap, S changes proportionally with Nt, which has a maximum value Nsites. A fraction θactive of Nsites is active such that Nactive = Nsitesθactive, making S a function of θactive. The maximum value of S occurring when θactive = 1 is a constant Smax = vthσnNsites; therefore, we write S = vthσnNt = vthσnNsitesθactive = Smaxθactive

(6)

Next, we consider two cases where the presence of an adsorbate either activates or passivates surface traps. Reschikov6 and Liu et al.5 report that oxygen decreases luminescence, whereas we showed here that water vapor passivates GaInP2. Thus, two cases are possible where θactive is equal to the fraction of occupied sites θoccupied or the fraction of unoccupied sites 1 − θoccupied Case (A): trap active when occupied, decreasing PLdet (θactive = θoccupied) or

1 1 US = (UB,R + UB,NR ) + (US,R + US,NR ) d d (1)

We neglect US,R because US,NR is dominant at surfaces and UB,NR because it is small for highly luminescent semiconductors such as GaInP221 (although UB,NR may be retained as constant for others) to give

Figure 3. (a) The PLdet measurements are plotted versus the water-vapor partial pressure P. The inset shows a measurement sequence for P = 36 mTorr, where the PLdet value is the average over 1 min after stabilization (black square) plotted in the main figure (black square). (b) PLdet is plotted versus P with a linear-log plot inset to compare best fits of the eq 9 model using either a molecular (eq 7, red line) or dissociative (eq 8, black line) adsorption isotherm for fAD(P). The fitting parameters are C = 614, S*max = 0.62, and α = 10.7 for molecular adsorption and C = 722, S*max = 0.80, and α = 3.1 for dissociative adsorption. 4420

DOI: 10.1021/acs.jpcc.5b12498 J. Phys. Chem. C 2016, 120, 4418−4422

Article

The Journal of Physical Chemistry C

This technique and model uniquely demonstrate systematic modulation of S manifest by PLdet. Thus, lifetime measurements that inherently convolute τ and S as an effective lifetime27 may be deconvoluted when S is varied independently. Langmuir adsorption and SRH recombination principles may also be applied to semiconductors where UB,NR is constant rather than negligible. An alternate derivation may account for UB,NR as another term subtracted from G in eq 3, and a model similar to eq 9 would result where its fit to PLdet = f(P) measurements can help evaluate the nature of adsorption while quantifying α. Additionally, characterizing the measurement system constants C0 and G would allow UB,NR to be quantified. Oxide systems such as ZnO/O nanorods studied by Liu et al.5 may have significant UB,NR, but such a nanoscale system with a high ratio of surface area to volume may still have negligible UB,NR. Their data suggest that the ZnO PL intensity is linear over much of the oxygen partial pressure range, while two data points at low pressure do not fit. Our model gives a physical explanation: such behavior is characteristic of adsorption isotherms that appear linear at higher pressure but have a stronger dependence at low pressure (see Figure 3b). The conceptual model for GaN/O suggested by Reshchikov6 includes recombination pathways that contribute to UB,NR and competing channels in a near-surface region that may require additional recombination-rate terms in eq 2. Regardless, PLdet = f(P) fit with adsorption isotherms can help evaluate the nature of adsorption and quantify α. PL measurements are relatively simple among techniques that are surface-sensitive, requiring only an excitation source, spectrometer, minimal optics, and line of sight to the sample. However, PL is not inherently quantitative unless combined with a model. We connected adsorption models to the surface defect population to describe why and how S is modulated by ambient pressure. Varying S quantitatively could also provide a way to delineate surface and near-surface (or bulk) recombination effects to help explain phenomena such as GaN yellow-band luminescence as discussed by Izpura.28 The simplicity of PL monitoring and the value of a quantitative model could be especially useful when applied in concert with techniques that characterize adsorbed species such as infrared spectroscopy,29 ambient-pressure X-ray photoelectron spectroscopy,25,30 or reflection anisotropy spectroscopy.31,32

Case (B): trap inactive when occupied, increasing PLdet (θactive = 1 − θoccupied). Finally, we consider the nature of the adsorption reaction in order to select an appropriate adsorption isotherm function θoccupied = fAD(P), where P is the partial pressure. Conversely, one could compare the quality of fit to data for several different fAD’s to determine its nature. Figure 2 illustrates a reversible process where A adsorbs to passivate one surface trap * as the adsorbed molecule A*, which is described by the molecular isotherm fAD,M θoccupied = fAD,M (P) =

αP 1 + αP

(7)

where α is the adsorption reaction equilibrium constant.23 A dangling bond (unpaired electron) is a common type of surface trap that is passivated by a bonding reaction. This implies that a bond in A breaks (dissociates) and forms two species that may both adsorb such that the dissociative isotherm fAD,D depends on P1/223 θoccupied = fAD,D (P) =

αP1/2 1 + αP1/2

(8)

5

Liu et al. suggested that competition between oxygen and water for surface sites could help explain a temperature dependence in their results. Here, a competitive adsorption isotherm23 should be applied. A number of other theoretical and empirical isotherms, such as those described by Keller and Staudt,24 may be considered for other systems. Combining eqs * = τ Smax , we arrive at 5 and 6 and constants C = C0GV and Smax d the simplest form of the equation describing the GaInP2/water vapor system * (1 − f (P))] PLdet = C[1 − Smax AD

(9)

The measured dependence of GaInP2 PLdet on water vapor pressure P is used to validate eq 9. We measure PLdet at fixed illumination for a set of water-vapor partial pressures P in Figure 3a. The inset shows that exposing GaInP2 to water vapor (P = 36 mTorr, ∼15 min) results in a stabilized PLdet = 380 counts, and PLdet returns to its original level after evacuating for 15−20 min. We then fit the PLdet versus P data set to eq 9 in Figure 3b using either fAD(P) for molecular (eq 7) or dissociative (eq 8) adsorption. Because α > 1, adsorption is favorable, which has been predicted by Wood et al.11 for InP and observed by Zhang et al.25 for GaP. This is illustrated in Figure 3 by the strong dependence of PLdet at low P, where half of the surface traps are populated (θoccupied = 0.5) at only 0.2 Torr. The better fit to eq 9 for the dissociative isotherm would imply that both −H and −OH passivate surface traps because dissociation resulting in one passivating species would behave as molecular adsorption. On the basis of modeling of water adsorption on InP surfaces11 and XPS studies of H2O/GaP,25 we expect that water dissociates on GaInP2 with −OH on Ga and In sites and −H on P sites. * = τ Smax = 0.80. Combined From the fit, we also obtain Smax d with other measurements or literature values, for example, d = 200 nm based on the optical absorption depth of the λ = 532 nm illumination and τ = 100 ns,26 we obtain Smax = 160 cm/s. Because Smax = σsvthNsites and vth depend only on temperature, further measurements or modeling could be made to obtain σs or Nsites.

■

CONCLUSIONS We have clarified that the enhancement and subsequent decay of GaInP2 PL after etching in concentrated sulfuric acid7 depend on the reversible reaction with water vapor and irreversible reaction with oxygen. We derived a relationship between PL intensity and the partial pressure of adsorbates that reversibly passivates surface defects. A linear combination of terms describes the dependence of PL intensity PLdet on the fractional surface coverage θoccupied, which we related to ambient pressure using adsorption isotherms. The model was validated with the water vapor/GaInP2 reversible adsorption system that is important to high-performance solar water splitting.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest. 4421

DOI: 10.1021/acs.jpcc.5b12498 J. Phys. Chem. C 2016, 120, 4418−4422

Article

The Journal of Physical Chemistry C

■

(19) Shockley, W.; Read, W. T. Statistics of the recombination of holes and electrons. Phys. Rev. 1952, 87, 835−842. (20) Hall, R. N. Electron-hole recombination in germanium. Phys. Rev. 1952, 87, 387. (21) Geisz, J. F.; Steiner, M. A.; García, I.; Kurtz, S. R.; Friedman, D. J. Enhanced external radiative efficiency for 20.8% efficient singlejunction GaInP solar cells. Appl. Phys. Lett. 2013, 103, 041118. (22) Muller, R. S.; Kamins, T. I. Device electronics for integrated circuits, 3rd ed.; John Wiley & Sons: New York, 2003. (23) Masel, R. I. Principles of adsorption and reaction on solid surfaces; Wiley: New York, 1951. (24) Keller, J. U.; Staudt, R. Gas Adsorption Equilibria: Experimental Methods and Adsorptive Isotherms; Springer US: Boston, MA, 2005; Vol. 33. (25) Zhang, X.; Ptasinska, S. Distinct and dramatic water dissociation on GaP(111) tracked by near-ambient pressure X-ray photoelectron spectroscopy. Phys. Chem. Chem. Phys. 2015, 17, 3909−3918. (26) King, R. R.; Fetxer, C. M.; Colter, P. C.; Edmondson, K. M.; Law, D. C.; Stavrides, A. P.; Yoon, H.; Kinsey, G. S.; Cotal, H. L.; Ermer, J. H.; et al. Lattice-matched and metamorphic GaInP/GaInAs/ Ge concentrator solar cells. Photovoltaic Energy Conversion, 2003. Proceedings of 3rd World Conference on 2003, 622−625. (27) Schröder, D. K. Carrier lifetimes in silicon. IEEE Trans. Electron Devices 1997, 44, 160−170. (28) Izpura, J. I. Surface-assisted luminescence: The PL yellow band and the EL of n-GaN devices. Adv. Condens. Matter Phys. 2013, 2013, 1. (29) Dobrich, A.; Kleinschmidt, P.; Döscher, H.; Hannappel, T. Quantitative investigation of hydrogen bonds on Si(100) surfaces prepared by vapor phase epitaxy. J. Vac. Sci. Technol. B Microelectron. Nanom. Struct. 2011, 29, 04D114. (30) Stoerzinger, K. A.; Hong, W. T.; Azimi, G.; Giordano, L.; Lee, Y.; Crumlin, E. J.; Biegalski, M. D.; Bluhm, H.; Varanasi, K. K.; ShaoHorn, Y. Reactivity of perovskites with water: Role of hydroxylation in wetting and implications for oxygen electrocatalysis. J. Phys. Chem. C 2015, 119, 18504−18512. (31) Döscher, H.; Dobrich, A.; Brückner, S.; Kleinschmidt, P.; Hannappel, T. Si(100) surfaces in a hydrogen-based process ambient. Appl. Phys. Lett. 2010, 97, 151905. (32) Brückner, S.; Döscher, H.; Kleinschmidt, P.; Hannappel, T. In situ investigation of hydrogen interacting with Si(100). Appl. Phys. Lett. 2011, 98, 211909.

ACKNOWLEDGMENTS J.L.Y. acknowledges support from the National Science Foundation Graduate Research Fellowship Program Award No. DGE 1144083 and acknowledges Steven George for advice and use of the ALD reactor modified for PL measurements. H.D. appreciates financial support by an EU Marie Curie fellowship (IOF No. 300971). This work was supported by the U.S. Department of Energy (DOE) Fuel Cell Technology Office under Contract No. DE-AC36-08GO28308 with NREL.

■

REFERENCES

(1) Houssa, M.; Nelis, D.; Hellin, D.; Pourtois, G.; Conard, T.; Paredis, K.; Vanormelingen, K.; Vantomme, A.; Van Bael, M. K.; Mullens, J.; et al. H2S exposure of a (100)Ge surface: Evidences for a (2 × 1) electrically passivated surface. Appl. Phys. Lett. 2007, 90, 222105. (2) Oktyabrsky, S.; Ye, P. D. Fundamentals of III-V semiconductor MOSFETs; Springer: New York, 2010 (3) Houssa, M.; Chagarov, E.; Kummel, A. Surface defects and passivation of Ge and III−V interfaces. MRS Bull. 2009, 34, 504−513. (4) Janata, J. Thirty years of CHEMFETs - A personal view. Electroanalysis 2004, 16, 1831−1835. (5) Liu, X.; Sun, Y.; Yu, M.; Yin, Y.; Yang, B.; Cao, W.; Ashfold, M. N. R. Incident fluence dependent morphologies, photoluminescence and optical oxygen sensing properties of ZnO nanorods grown by pulsed laser deposition. J. Mater. Chem. C 2015, 3, 2557−2562. (6) Reshchikov, M. A. Strong suppression of the yellow luminescence in C-doped GaN in air ambient. Appl. Phys. Lett. 2006, 89, 232106/1− 232106/4. (7) Kocha, S. S.; Peterson, M. W.; Nelson, A. J.; Rosenwaks, Y.; Arent, D. J.; Turner, J. A. Investigation of chemical wet-etch surface modification of Ga0.5In0.5P using photoluminescence, X-ray photoelectron-spectroscopy, capacitance measurements, and photocurrentvoltage curves. J. Phys. Chem. 1995, 99, 744−749. (8) Khaselev, O.; Turner, J. A. A monolithic photovoltaicphotoelectrochemical device for hydrogen production via water splitting. Science 1998, 280 (5362), 425−427. (9) May, M. M.; Lewerenz, H.-J.; Lackner, D.; Dimroth, F.; Hannappel, T. Efficient direct solar-to-hydrogen conversion by in situ interface transformation of a tandem structure. Nat. Commun. 2015, 6, 8286. (10) Döscher, H.; Geisz, J. F.; Deutsch, T. G.; Turner, J. A. Sunlight absorption in water − efficiency and design implications for photoelectrochemical devices. Energy Environ. Sci. 2014, 7, 2951− 2956. (11) Wood, B. C.; Ogitsu, T.; Schwegler, E. Ab initio modeling of water−semiconductor interfaces for photocatalytic water splitting: role of surface oxygen and hydroxyl. J. Photonics Energy 2011, 1, 016002. (12) Lewerenz, H.; Aspnes, B.; Miller, B.; Malm, L.; Heller, A. Semiconductor interface characterization in photoelectrochemical solar cells. J. Am. Chem. Soc. 1982, 104, 3325−3329. (13) Heller, A. Hydrogen-evolving solar cells. Science 1984, 223, 1141−1148. (14) Elam, J. W.; Groner, M. D.; George, S. M. Viscous flow reactor with quartz crystal microbalance for thin film growth by atomic layer deposition. Rev. Sci. Instrum. 2002, 73, 2981. (15) Turner, J. A. Energetics of the semicontuctor-electrolyte interface. J. Chem. Educ. 1983, 60, 327−329. (16) Otaredian, T. The influence of the surface and oxide charge on the surface recombination process. Solid-State Electron. 1993, 36, 905. (17) Chen, Z.; Jaramillo, T. F.; Deutsch, T. G.; Kleiman-Shwarsctein, A.; Forman, A. J.; Gaillard, N.; Garland, R.; Takanabe, K.; Heske, C.; Sunkara, M.; et al. Accelerating materials development for photoelectrochemical hydrogen production: Standards for methods, definitions, and reporting protocols. J. Mater. Res. 2010, 25, 3−16. (18) Langmuir, I. The adsorption of gases on plane surfaces of glass, mica and platinum. J. Am. Chem. Soc. 1918, 40, 1361−1403. 4422

DOI: 10.1021/acs.jpcc.5b12498 J. Phys. Chem. C 2016, 120, 4418−4422