Water as probe molecule for midgap states in nanocrystalline strontium titanate by conventional and synchronous luminescence spectroscopy under ambient conditions

Water as probe molecule for midgap states in nanocrystalline strontium titanate by conventional and synchronous luminescence spectroscopy under ambient conditions

Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 174 (2017) 54–61 Contents lists available at ScienceDirect Spectrochimica Acta P...

2MB Sizes 0 Downloads 17 Views

Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 174 (2017) 54–61

Contents lists available at ScienceDirect

Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy journal homepage: www.elsevier.com/locate/saa

Water as probe molecule for midgap states in nanocrystalline strontium titanate by conventional and synchronous luminescence spectroscopy under ambient conditions Sean Taylor, Alexander Samokhvalov ⁎ Chemistry Department, Rutgers University, 315 Penn St., Camden, NJ 08102, USA

a r t i c l e

i n f o

Article history: Received 5 March 2016 Received in revised form 16 October 2016 Accepted 12 November 2016 Available online 13 November 2016 Keywords: Strontium titanate Luminescence Synchronous Midgap state Water adsorption

a b s t r a c t Alkaline earth metal titanates are broad bandgap semiconductors with applications in electronic devices, as catalysts, photocatalysts, sorbents, and sensors. Strontium titanate SrTiO3 is of interest in electronic devices, sensors, in the photocatalytic hydrogen generation, as catalyst and sorbent. Both photocatalysis and operation of electronic devices rely upon the pathways of relaxation of excited charge in the semiconductor, including relaxation through the midgap states. We report characterization of nanocrystalline SrTiO3 at room temperature by “conventional” vs. synchronous luminescence spectroscopy and complementary methods. We determined energies of radiative transitions in the visible range through the two midgap states in the nanocrystalline SrTiO3. Further, adsorption and desorption of vapor of water as “probe molecule” for midgap states in the nanocrystalline SrTiO3 was studied, for the first time, by luminescence spectroscopy under ambient conditions. Emission of visible light from the nanocrystalline SrTiO3 is significantly increased upon desorption of water and decreased (quenched) upon adsorption of water vapor, due to interactions with the surface midgap states. © 2016 Elsevier B.V. All rights reserved.

1. Introduction Strontium titanate SrTiO3 is a broad bandgap semiconductor which has been studied as a photocatalyst [1], catalyst [2], sorbent [3], material in electronic devices [4] and gas sensors [5]. Above 106 K including the room temperature, SrTiO3 is present in distorted cubic perovskite structure [6]. For the single crystals of SrTiO3, the energy is 3.75 eV for a direct bandgap and 3.25 eV for an indirect bandgap [7]. Electronic midgap states play an important role in modulation of optical absorption in semiconductors [8] and in charge transport, in particular as charge-trapping centers which affect photoconductivity [9] and photocatalytic reaction rates [10]. The photoluminescence (PL) spectroscopy is well suited for characterization of electronic states in powdered semiconductors [11] including metal oxides [12], zeolites [13,14], etc. The “conventional” PL emission spectrum is obtained by recording the emission wavelength λemiss under photoexcitation at a constant λexc. The PL excitation spectrum is collected by changing λexc while the emission is recorded at a constant λemiss. The narrower emission spectra can be obtained by using synchronous luminescence spectroscopy [15] in which the λexc and λemiss are changed simultaneously at a constant difference Δλ = λemiss − λexc. Applications of synchronous fluorescence spectroscopy ⁎ Corresponding author at: Department of Chemistry, Rutgers University, Camden, NJ 08102, USA. E-mail address: [email protected] (A. Samokhvalov).

http://dx.doi.org/10.1016/j.saa.2016.11.011 1386-1425/© 2016 Elsevier B.V. All rights reserved.

have recently been reviewed [16,17]. Synchronous luminescence spectroscopy has not been reported in characterization of nanocrystalline strontium titanate, to our knowledge. Adsorption based applications frequently involve moisture in ambient air, so this is important to study mechanisms of adsorption of water in the vapor phase. The water molecule has a high energy of stretching vibration at ca. 3500 cm−1, and water has been studied as a “probe molecule” in the PL quenching experiments with chemical compounds of several classes. These include inorganic ions in solution [18], metal-organic frameworks [19], semiconductor nanoparticles with lanthanide metal dopants [20–22] as well as without dopant such as CdS [23] and PbS [24]. Recently, we used the PL spectroscopy to study emission from monomers and excimers of molecules of aromatic hydrocarbons [25] and aromatic sulfur compounds [26]. We also applied the in-situ PL spectroscopy to learn about the direction of transfer of photoexcited charge in photocatalytic colloids of metal-doped titanium dioxide [27] and the ex-situ PL spectroscopy to study radiative transitions in the nanocrystalline binary nitrogen and metal codoped titanium dioxide [28]. Recently, we compared the “conventional” and synchronous luminescence spectra of the nanocrystalline calcium titanate with orthorhombic lattice measured at 77 K inside a liquid nitrogen Dewar [29]. In this work, we study water in the vapor phase as the universal spectroscopic “probe molecule” interacting with electronic midgap states in the nanocrystalline SrTiO3 with cubic lattice under ambient conditions. Reversible adsorption/desorption of water causes a reversible quenching/

S. Taylor, A. Samokhvalov / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 174 (2017) 54–61

55

increase of the photoluminescence in the visible range originating from the midgap states in SrTiO3. We characterize midgap states in the nanocrystalline SrTiO3 by synchronous luminescence spectroscopy at a convenient temperature of 25 °C versus “conventional” PL emission spectroscopy, and by the complementary methods. In addition, we determine the pathways of relaxation of excited charge in the nanocrystalline cubic SrTiO3 through the midgap electronic states. 2. Materials and Methods 2.1. Materials Nanopowder of SrTiO3 (99.95% purity, b100 nm nominal particle size, cubic phase) has been obtained from the U.S. Research Nanomaterials. Sulfuric acid and hydrogen peroxide were from Fisher. Fig. 1. XRD pattern of nanocrystalline SrTiO3.

2.2. Specimen Characterization 3. Results and Discussion XRD data were obtained using Rigaku SmartLab diffractometer system with Cu K-alpha line at 0.15418 nm. The Raman spectra were collected at room temperature with XploRA confocal microscope from Horiba Scientific which is equipped with lasers at 532 nm, 638 nm and 785 nm, and has a cut-off filter at 50 cm−1. The UV–Visible diffuse reflectance spectra, UV–Vis DRS were measured at room temperature with Cary 5000 spectrometer equipped with Praying Mantis attachment from Harrick Scientific. As white reference, finely grinded BaSO4 of 99.998% purity from Alfa Aesar has been used. 2.3. Drying and Hydration of SrTiO3 The as-obtained SrTiO3 nanopowder denoted asisSrTiO3 was dried at 105 °C in the oven overnight, yielding dried material drSrTiO3. The drSrTiO3 was hydrated using the procedure recently reported by us [19], by being placed inside the desiccator with liquid water and kept in contact with water vapor at relative humidity RH ~ 100% at 25 °C overnight, resulting in hydrated material hydSrTiO3. 2.4. Measurements by the Photoluminescence (PL) Spectroscopy All spectra were recorded using Fluorolog spectrometer FL3-22 from Horiba Scientific. This instrument is equipped with dual monochromator gratings on the excitation and emission optical pathways. In order to minimize artifacts due to primary and secondary re-absorption of light in solids [30], all spectra were collected in the Front Face (FF) geometry using FL-1001 accessory from Horiba Scientific. In addition, to avoid the effects of fluctuations in the intensity of the excitation light source on the spectra, the signal from the sample (S1) has been divided by the reference signal (R1) generated by the excitation beam before reaching the sample, and the ratio S1/R1 has been utilized in all cases. The 0.5 cm3 quartz cuvette was cleaned with Piranha solution (sulfuric acid and hydrogen peroxide), rinsed with distilled water, and dried. All PL measurements were conducted at 25 °C, with the sample placed into a freshly cleaned and dried 0.5 cm3 quartz cuvette, closed with polytetrafluoroethylene (PTFE) stopper, and sealed with Parafilm tape to exclude ambient moisture. The temperature of the sample in the cavity of the spectrometer was maintained at 25 °C by water circulation thermostat model A25 from Thermo Scientific. The PL emission spectra were collected at the excitation and emission optical slits at 2 nm. The excitation wavelength λexc has been varied from 250 nm (an extra bandgap excitation in SrTiO3) to 460 nm (the sub bandgap excitation) in steps of 10 nm. In synchronous luminescence spectroscopy, the Δλ parameter has been changed from Δλ = 10 nm to Δλ = 120 nm. Numeric fitting of the spectra was performed with Microcal Origin 2015 program.

3.1. Structural Characterization of Nanocrystalline SrTiO3 Fig. 1 shows an XRD pattern of our nanocrystalline SrTiO3; the peaks correspond to the perovskite phase, they can be indexed to the cubic space group, and the pattern matches the PDF card # 01-089-4934 (tausonite). The cubic lattice of our nanocrystalline SrTiO3 is also consistent with the JCPDS card # 05-0634 of cubic perovskite in previous reports [31,32]. No splitting of the (100), (110), and (200) reflections was observed which indicates an absence of a tetragonal distortion of cubic lattice. We analyzed the strongest (110) diffraction peak of our SrTiO3 at 2θ = 32.20 deg. by the Scherrer's equation [33], D = k λ / β cos(θ), where k is a constant (the shape factor with numeric value of 0.9), λ is an X-ray wavelength, β is the full-width at half-maximum (FWHM) of the diffraction peak of interest (in radian), and θ is the Bragg angle. This analysis yields the average nanocrystalline size of our nanocrystalline SrTiO3 at 36 nm. Fig. S1 shows the Raman spectrum of nanocrystalline SrTiO3 at room temperature at λexc = 532 nm; the Raman spectrum at λexc = 638 nm was similar (data not shown). The Raman spectrum in Fig. S1 is consistent with published spectrum [32] of nanocrystalline cubic strontium titanate. For cubic SrTiO3 of space group Pm3m, the phonons are represented by the symmetry 3F1u + F2u, and neither represents the first order Raman-active mode, since the center of symmetry results in the zero polarizability of the lattice [32,34]. Assignments of the Raman peaks of our nanocrystalline SrTiO3 (Fig. S1) are provided in Table 1. Recent reports showed that vibrational modes can be modified due to electrostatic forces, oxygen vacancies and external factors, so that the F1u mode is divided into a doubly degenerate E and a nondegenerate A1 modes, while the F2u mode is divided into E and B1 modes [32]. The long-range electrostatic forces separate the degenerate mode E and the nondegenerate mode A1 into the transverse optical (TO) and longitudinal optical (LO) modes, which are observed in the Raman spectra of strontium titanate at room temperature [34].

Table 1 Raman peaks of nanocrystalline cubic SrTiO3 with assignments. The Raman shift, cm−1 v1 v2 v3 v4 v5

= = = = =

185 275 550 731 805

The Raman shift in Ref. [32], cm−1

Vibrational mode in Ref. [32]

179 270 544 727 801

TO2 (transverse optical) TO3 (transverse optical) TO4 (transverse optical) TO (transverse optical) LO4 (longitudinal optical)

56

S. Taylor, A. Samokhvalov / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 174 (2017) 54–61

3.2. Characterization of Nanocrystalline SrTiO3 by Diffuse Reflectance Spectroscopy (DRS) Reflectance R(λ) spectra are useful in characterization of nanocrystalline semiconductors [35]. Fig. 2 shows the reflectance spectrum R(λ) of our nanocrystalline SrTiO3. In Fig. 2, the low reflectance (high absorbance) at ca. 240–330 nm may correspond to a strong absorption of photons due to an excitation across a direct optical bandgap Edirect. Then, a progressively increasing reflectance at ca. λ N 320 nm in Fig. 2 corresponds to a decreasing absorbance, when the incident photon energy becomes equal and then smaller than the bandgap. The arrows in Fig. 2 indicate the two inflection points in the reflectance spectrum which may indicate transitions with the energies Edirect and Eindirect of nanocrystalline SrTiO3; thus, it is necessary to determine the bandgap(s) of our strontium titanate. The Kubelka-Munk (KM) function F(R) with R(λ) being reflectance spectrum is commonly utilized in optical characterization of nanocrystalline semiconductors [36,37]. The definition of the KM function is F(R) = (1 − R)2/2R = k / s, where k is the Kubelka-Munk absorption coefficient and s is the scattering constant [38]. For semiconductors in the powder form, optical bandgaps are obtained through the Tauc plots (α ∗ E)1/n = E − Eg, where E is the photon energy, Eg is the bandgap, α is an absorption coefficient, and n is the Tauc constant [39]. The Tauc plot can be obtained from F(R) function [37,40] in which case the Tauc formula becomes (F(R) ∗ E)1/n = E − Eg. This is accepted [40] that for a direct transition the value of the Tauc constant is n = 1/2 and for an indirect transition n = 2. For single crystals of SrTiO3, published value of a direct bandgap is Edirect = 3.75 eV, and Eindirect = 3.25 eV for an indirect bandgap [7]. Optical bandgaps of nanocrystalline SrTiO3 were reported between 3.2 eV [41] and 3.77 eV [32]. We determined Edirect and Eindirect of our nanocrystalline SrTiO3 by the Tauc method, Fig. 3. Analysis of Fig. 3A and B results in Edirect = 3.85 eV and Eindirect = 3.30 eV which are consistent with published values [7] for “bulk” cubic strontium titanate. It has been reported that the quantum size effects in strontium titanate are rather weak; when the crystalline size of SrTiO3 has been varied between 1 μm (the “bulk”) and 6.5 nm, the proton reduction potential [1] has changed just by 0.05 V. This is consistent with our data in Fig. 3 where the obtained values of Edirect and Eindirect of our nanocrystalline strontium titanate are close to published data [7] for SrTiO3 single crystals. 3.3. Characterization of Nanocrystalline SrTiO3 by the Excitation Wavelength-Dependent PL Emission Spectroscopy In spectroscopic characterization of chemical compounds, the socalled three-dimensional (3D) PL spectra are often recorded which are

Fig. 2. Optical reflectance spectrum of nanocrystalline SrTiO3.

Fig. 3. The Tauc plots for nanocrystalline SrTiO3. A) Direct bandgap. B) Indirect bandgap.

termed “total luminescence spectra” [42]. In this variety of the PL spectroscopy, the excitation wavelength λexc is systematically varied, and the respective PL spectra are recorded for each λexc resulting in the 3D emission “profiles” [42]. When the set of the 2D PL spectra is used for data presentation instead, this is termed the excitation wavelength dependent PL emission spectroscopy [19]. The nanocrystalline SrTiO3 in our studies has Edirect = 3.85 eV (322 nm) and Eindirect = 3.30 eV (375 nm), Fig. 3. When the energy of the photoexcitation of our nanocrystalline SrTiO3 was close to or higher than the Edirect = 3.85 eV (322 nm), very weak emission was observed at 25 °C (data not shown). On the other hand, when the energy of the photoexcitation was E ≤ 3.5 eV i.e. close to the Eindirect, the much higher intensity emission spectra were observed, consistently with the phonon-assisted excitation at room temperature. Fig. 4A and B show the excitation wavelength dependent PL emission spectra with the λexc varied at 10 nm increments; an empty quartz cuvette (a sample container) has shown a negligibly small emission. In Fig. 4, the Rayleigh lines at λexc are partially shown to aid in spectral assignments. In Fig. 4A, at the excitation energy close to the Eindirect, one can see an increasing PL signal when the λexc increases, and at λexc = 400 nm the spectrum is clearly resolved into the two shoulders. In Fig. 4, spectral shoulder at 465 nm (2.67 eV) is denoted “blue light”; this emission is consistent with the PL spectrum of SrTiO3 at room temperature [32] with the maximum at ca. 460 nm. In Fig. 4, the shoulder at ca. 490 nm (2.53 eV) is denoted “green light”; the similar peak was reported in the PL spectrum of nanocrystalline SrTiO3 but not assigned [43]. To our knowledge, the PL spectra of SrTiO3 at the subbandgap excitation have not been reported. Fig. 4B shows the PL spectra of our nanocrystalline SrTiO3 at the subbandgap excitation with λexc ≥ 410 nm with a strong emission in the visible range within ca. 450–650 nm. The PL emission can be conveniently represented by “contour maps” [44] in which the emission intensity is plotted with the λexc as an X axis and the λemiss as the Y axis. For example, the fluorescence contour maps have been used in spectroscopic identification of organic compounds [44]. In the luminescence contour maps [45,46], each chemical

S. Taylor, A. Samokhvalov / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 174 (2017) 54–61

57

to the absorption transition involving the electronic midgap states below the CBM in the nanocrystalline strontium titanate. 3.4. Characterization of Nanocrystalline SrTiO3 by the PL Excitation Spectroscopy

Fig. 4. The excitation wavelength dependent PL emission spectra of nanocrystalline SrTiO3. A) At the excitation energy close to Eindirect. B) At the subbandgap excitation. (For interpretation of the references to color in this figure, the reader is referred to the web version of this article.)

compound is visualized as a set of concentric elliptic lines. To our knowledge, the PL contour maps were not reported for strontium titanate. Fig. 5 shows the PL emission contour map of our nanocrystalline SrTiO3 at 25 °C. The set of high intensity contours starting from the left bottom corner (λexc = 360 nm and λemiss = 360 nm) and directed at the angle of 45 degrees is due to the Rayleigh line (the excitation), and it is labeled “Rayleigh”. The remaining contours in Fig. 5 are due to the photoluminescence from our nanocrystalline SrTiO3 and their shape is consistent with contours due to individual light emitting compounds, e.g. [45,46]. One can see in Fig. 5 that the strongest emission is achieved under the subbandgap photoexcitation at ca. 430 nm, which can be attributed

Fig. 5. The PL contour map of nanocrystalline SrTiO3.

The PL excitation spectra are useful in finding the absorption transitions which result in the strongest emission peaks. Fig. 6 shows the PL excitation spectra of our nanocrystalline SrTiO3 for emission maxima at 460 nm and 500 nm measured at 25 °C. The PL excitation spectra of empty quartz cuvettes have shown the negligibly small photoluminescence. In Fig. 6A and B, the emission at 460 nm and 500 nm has a high intensity at ca. λexc = 380 nm (the maximum in Fig. 6A and a shoulder in Fig. 6B). The transition at λexc = 380 nm corresponds to the photoexcitation at the energy close to Eindirect in our nanocrystalline SrTiO3. On the other hand, the emission at 500 nm is efficiently induced by the subbandgap photoexcitation at ca. 430 nm, see the maximum in Fig. 6B. The strong emission at 500 nm (“green light”) at the subbandgap excitation with λexc = 430 nm is consistent with the maximum in the PL contour map in Fig. 5. Our experimental data obtained using powder XRD (Fig. 1), the Raman spectroscopy (Fig. S1), optical diffuse reflectance spectroscopy (Fig. 2), the Tauc plots (Fig. 3), the wavelength-dependent PL emission spectroscopy (Fig. 4), the PL contour map (Fig. 5), and the PL excitation spectroscopy (Fig. 6) have been used to construct the energy diagram for excitation and relaxation of charge in the nanocrystalline cubic SrTiO3, see Fig. 7. The potential of the CBM in SrTiO3 versus the normal hydrogen electrode (NHE) was taken in Ref [47]. In Fig. 7, solid lines show radiative transitions (absorption and emission) and dashed lines indicate non radiative transitions which cannot be studied by the PL spectroscopy. Fig. 7a shows an extra bandgap excitation followed by: 1) a radiative recombination of an electron in the CBM of SrTiO3 with the hole trapped on the low-energy midgap state (“blue light”), and 2) a non-radiative relaxation of an electron from the CBM to the electron trapping high energy midgap state, followed by a radiative recombination with the

Fig. 6. The PL excitation spectra of nanocrystalline SrTiO3 at 25 °C. A) Emission at 460 nm. B) Emission at 500 nm.

58

S. Taylor, A. Samokhvalov / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 174 (2017) 54–61

Fig. 7. Energy diagram for the photoexcitation induced transitions in nanocrystalline SrTiO3. A) Indirect extrabandgap excitation. B) Subbandgap excitation. (For interpretation of the references to color in this figure, the reader is referred to the web version of this article.)

holes. This energy diagram is consistent with generic model of midgap states in strontium titanate [32]. Origin of midgap states in strontium titanate which cause emission in visible range has been investigated by experiment for both single crystals [48] and nanoparticles [49]. In both cases, emission of visible light was observed as more than one PL band and assigned to self-trapped excitons localized on surface states (oxygen vacancies) [48]. In addition, ground and excited states of strongly trapped excitons localized on oxygen vacancies in strontium titanate have been studied by quantum chemical computations [50] and emission in visible range was predicted. We speculate that midgap states in Fig. 7 are due to strongly trapped excitons localized on oxygen vacancies on the surface of the nanocrystals of strontium titanate. To our knowledge, the energy diagram of SrTiO3 under the subbandgap photoexcitation was not reported. In Fig. 7b, the subbandgap excitation results in emission at ca. 500 nm (“green light”). The excitation at ca. 430 nm in Fig. 7B is consistent with the maximum on the PL contour maps in Fig. 5. This relatively high intensity transition is due to the abundant filled electronic states (the initial state for this transition) at the valence band maximum (VBM) in strontium titanate. In Fig. 7, the midgap states are shown as the multiplets of the respective electron and hole charge-trapping states. The individual states of different energy within these multiplets (shown as horizontal lines in Fig. 7) reflect structural heterogeneity of lattice sites in the nanocrystals. To our knowledge, the spectral widths of luminescence bands of the nanocrystalline SrTiO3 were not reported. 3.5. Synchronous Luminescence Spectroscopy of Nanocrystalline SrTiO3 The PL emission shoulder at 2.8 eV (“blue light”) due to the transition between the CBM and the midgap hole state and the shoulder at 2.4 eV (“green light”) [48–50] due to the transition between the exciton state and the midgap hole state in SrTiO3 were observed in “conventional” PL emission spectra, Fig. 4. However, the better resolution of these overlapping PL peaks is desirable. Synchronous luminescence spectroscopy has been found useful to achieve significant narrowing of emission spectra of many organic compounds [16]. To our knowledge, synchronous luminescence spectra were not reported for nanocrystalline SrTiO3. In order to better resolve broad emission bands at ca. 460 nm (“blue light”) and 500 nm (“green light”), we performed synchronous luminescence spectroscopic measurements at 25 °C. In synchronous

Fig. 8. Synchronous luminescence spectra of nanocrystalline SrTiO3 at variable Δλ. (For interpretation of the references to color in this figure, the reader is referred to the web version of this article.)

luminescence spectroscopy [15], a major variable is the Δλ parameter, Δλ = λemiss − λexc. The Δλ = 0 nm would correspond to the resonance light scattering (RLS) experiment [51]. Fig. 8 shows synchronous luminescence spectra of our nanocrystalline SrTiO3 at Δλ = 20–130 nm. One can see in Fig. 8 that synchronous luminescence spectra show the two peaks overlapping to a certain degree, while positions of spectral maxima depend on the value of the Δλ parameter. Position of spectral maxima and the peak widths in synchronous fluorescence spectra depend on the Δλ which is well known for organic molecules [15]. In Fig. 8, the strongest emission from our nanocrystalline SrTiO3 has been obtained at the Δλ = 80 nm which is considered, for this compound, an optimal value. Spectral positions of the peaks at ca. 460 nm (“blue light”) and at ca. 500 nm (“green light”) in synchronous luminescence spectrum at Δλ = 80 nm (Fig. 8B) are consistent with spectral shoulders in “conventional” PL emission spectrum, Fig. 4. 3.6. Optical Transitions through Midgap States in Nanocrystalline SrTiO3: Synchronous Luminescence vs. Conventional PL Spectroscopy This is of interest to compare the capabilities of “conventional” PL emission spectroscopy and synchronous luminescence spectroscopy at 25 °C in resolving the two major emission peaks of nanocrystalline SrTiO3. To determine the energies of these transitions and the respective spectral full widths at the half maxima (the FWHMs), plotting and numeric fitting of the spectra in the photon energy, rather in the wavelength domain is preferred. Fig. 9a shows “conventional” PL emission spectrum of nanocrystalline SrTiO3 at λexc = 400 nm with the two shoulders of interest as plotted in the photon energy domain. This spectrum is compared with synchronous luminescence spectrum of SrTiO3 (Fig. 9b) at an optimal value Δλ = 80 nm. Both emission spectra in Fig. 9 have been fitted with the multi-Gaussian function, and an adjusted fitting parameter R2 = 0.99 indicates a good quality of fitting; using the Lorentian model resulted in poor fitting quality. In Fig. 9, the synchronous luminescence spectrum is significantly narrower and better resolved than

S. Taylor, A. Samokhvalov / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 174 (2017) 54–61

59

“nominal” value of 2. On the other hand, in “conventional” PL spectrum in Fig. 9a the ratio FWHM (“green light”)/FWHM (“blue light”) = 0.59 eV/0.06 eV = 9.8. Such a high ratio is much further from the expected “nominal” ratio = 2 than the ratio = 3.4 observed in synchronous luminescence spectra in Fig. 9b as discussed above. Therefore, synchronous luminescence spectra are better suited than “conventional” PL emission spectra to learn about the origin of radiative transitions involving midgap states in strontium titanate. 3.7. Water in Vapor as Spectroscopic “Probe Molecule”: Adsorption and Desorption on Nanocrystalline SrTiO3

Fig. 9. Emission spectra of nanocrystalline SrTiO3 with Gaussian fitting. a) “Conventional” PL emission spectrum at λexc = 400 nm. b) Synchronous luminescence spectrum at Δλ = 80 nm.

“conventional” PL emission spectrum. This is consistent with narrowing [15] of synchronous luminescence spectra of organic compounds compared to “conventional” PL spectra. In addition, the excitation (the Rayleigh) line appears at 3.1 eV (400 nm) in Fig. 9a which prevents a reliable curve fitting in the high-energy part of this “conventional” PL spectrum. This “Rayleigh artifact” is not present in synchronous luminescence spectrum in Fig. 9b which significantly aids in numeric curve fitting. In “conventional” PL spectrum (Fig. 9a), the fitted peaks have the following parameters: 1) the peak at 2.53 eV (“green light”) with the FWHM = 0.59 eV and 2) the peak at 2.69 eV (“blue light”) with the FWHM = 0.06 eV. In synchronous luminescence spectrum (Fig. 9b), the fitted peaks have the parameters: 1) the peak at 2.47 eV (“green light”) with the FWHM = 0.47 eV and 2) the peak at 2.69 eV (“blue light”) with the FWHM = 0.14 eV. Thus, in both spectra in Fig. 9, the two peaks corresponding to “blue light” and “green light” are identified, consistently with the energy diagram in Fig. 7. The FWHMs of the peaks are expected to correlate with their spectral assignments. Namely, emission peak due to “blue light” originating in the transition from the CBM to the low-energy midgap multiplet of states is expected to be narrower than emission peak due to “green light” originating in the transition from the high energy multiplet to the low-energy multiplet of midgap states, see Fig. 7. In synchronous luminescence spectrum (Fig. 9b), the scheme of radiative transitions in Fig. 7 is consistent with the relative FWHMs of the peaks due to “blue light” (2.69 eV) and “green light” (2.47 eV). Namely, the FWHM = 0.47 eV due to “green light” (2.47 eV) and the FWHM = 0.14 eV due to “blue light” (2.69 eV) yield the ratio FWHM (“green light”)/FWHM (“blue light”) = 0.47 eV/0.14 eV = 3.4. Assuming the same width, in the energy domain, for the high-energy and low-energy multiplets of midgap states in Fig. 7, the ratio FWHM (“green light”)/FWHM (“blue light”) is expected to be equal to 2. However, the high-energy and low-energy multiplets of the midgap states are due to the different kind of states, therefore their spectral widths are not likely to be exactly the same. This explains that an experimentally obtained ratio FWHM (“green light”)/FWHM (“blue light”) = 3.4 deviates from its expected

It would be interesting to determine by experiment the location of the midgap states in strontium titanate using luminescence spectroscopy and the suitable “probe” molecule. The water molecule with the high energy of its stretching vibration at ca. 3500 cm−1 has been studied as a “probe molecule” in the PL quenching experiments. For instance, the high intensity luminescence of CdS nanoparticles prepared in reversed micelles and vacuum-dried has been quenched by adding small amounts of liquid water [23]. Evaporation of the solvent and reconstituting the dried micellar solution with isooctane resulted in a total recovery of the luminescence, so luminescence quenching was due to adsorption of water on CdS [23]. Interactions of water molecules with (001) surface of cubic SrTiO3 studied by the DFT simulations revealed up to two monolayers of adsorbed water [52], with H-bonds between the hydroxyl hydrogens and surface oxygens and intermolecular interactions of adsorbed water molecules, but not bulk diffusion. Heats of water vapor sorption on cubic SrTiO3 at 30–45 nm in size were reported and surface adsorption was concluded to occur [53]. Our nanocrystalline SrTiO3 is ca. 36 nm in size as found by XRD (Fig. 1). To our knowledge, quenching of the PL emission in the nanocrystalline SrTiO3

Fig. 10. The PL emission spectra of SrTiO3. a) The asisSrTiO3 before drying. b) The drSrTiO3 after drying. c) The hydSrTiO3 after hydration.

60

S. Taylor, A. Samokhvalov / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 174 (2017) 54–61

Table 2 Change in mass of nanocrystalline SrTiO3 after drying and hydration, and integrated PL spectra at the extrabandgap (λexc = 380 nm) and subbandgap (λexc = 430 nm) excitation. Material

Change in mass, g

asisSrTiO3 – drSrTiO3 −0.0105 hydSrTiO3 +0.0108

Integrated PL signal at λexc = 380 nm, a.u.

Integrated PL signal at λexc = 430 nm, a.u.

4.35 × 107 1.39 × 108 3.72 × 107

5.54 × 107 1.57 × 108 5.08 × 107

upon adsorption of water was not reported. We conducted the PL experiments with removal of adsorbed water by heating and the following hydration via adsorption of water vapor (see Experimental). Fig. 10 shows the excitation wavelength dependent PL emission spectra of our nanocrystalline SrTiO3 before drying (Fig. 10a), after drying (Fig. 10b), and after the subsequent hydration (Fig. 10c) in water vapor at room temperature. The λexc was varied from that corresponding to the excitation at Eindirect (λexc = 380 nm) to λexc = 430 nm corresponding to the transitions from/to midgap states (see Fig. 7). Upon drying, there was a factor ~ 3 increase of the PL signal, compare Fig. 10b and Fig. 10a. Upon the following hydration of SrTiO3, there was a factor ~ 3 decrease of the PL signal, compare Fig. 10c and Fig. 10b. Table 2 shows the change of mass of strontium titanate after drying and the following hydration, and the resultant changes in the integrated PL spectrum. For a quantitative analysis of efficiency of the PL quenching by adsorbed water, an excitation with the photon energies as in Fig. 7 was conducted, namely at λexc = 380 nm and λexc = 430 nm. At λexc = 380 nm, both “blue light” at ca. 460 nm and “green light” at ca. 500 nm were observed as spectral shoulders, while at λexc = 430 nm a “green light” at ca. 500 nm formed the maximum in Fig. 10. Integration of the PL spectra at λexc = 380 nm was conducted in the range λemiss = 400–650 nm; for λexc = 430 nm the integration range was λemiss = 450–700 nm. Upon hydration of drSrTiO3 towards hydSrTiO3, a slightly more water was adsorbed than it was desorbed in previous drying step of asisSrTiO3, Table 2. Consequently, the integrated PL spectrum at λexc = 380 nm for hydSrTiO3 at 3.72 × 107 a.u. was smaller than that for asisSrTiO3 at 4.35 × 107 a.u. This indicates that the higher amount of adsorbed water has caused the stronger PL quenching; the same was observed for the PL under the subbandgap excitation at λexc = 430 nm (Table 2). In summary, for both λexc = 380 nm and λexc = 430 nm, numeric value of the integrated PL spectrum increased by the factor ~ 3 upon desorption of water and subsequently decreased (quenched) upon adsorption of water. Quenching of the PL from nanocrystalline SrTiO3 after hydration was also observed in the synchronous luminescence spectra at Δλ = 80 nm, Fig. S2. The reversible increase and quenching of emission in the visible range upon desorption/adsorption of water vapor on the surface [53] of nanocrystalline SrTiO3 suggests that initial and final states for radiative transitions are, at least in part, associated with surface states, see Fig. 11. These excited states are, tentatively, the strongly trapped excitons localized on oxygen vacancies on the surface of strontium titanate. The large surface/bulk ratio in the nanocrystals of our SrTiO3 would allow a substantial number of such “surface excitons”. Water as polar adsorbate could interact with surface oxygen vacancies via donor-acceptor mechanism, likely by electron donation from oxygen atom of water molecule to the oxygen vacancy. This electron donation by adsorbed water can be interpreted either in terms of surface dipole [54] or in terms of increasing the Fermi level in the semiconductor nanoparticle. The increase in the Fermi level upon adsorption of water would result in population of some previously vacant electronic midgap states below the conduction band minimum (CBM), so that they could not be populated with photoexcited electrons. Hence, the luminescence from these midgap states in the nanocrystalline SrTiO3 would be quenched by the water molecule. This proposed mechanism is illustrated in Fig. 11 by the water molecule approaching the surface with its

Fig. 11. Proposed energy diagram for transitions in the nanocrystalline SrTiO3 with adsorbed water. A) Indirect extrabandgap excitation. B) Subbandgap excitation. (For interpretation of the references to color in this figure, the reader is referred to the web version of this article.)

oxygen atom. Consequentially, upon the removal of adsorbed water, the Fermi level in the nanocrystal of SrTiO3 would be lower which would deplete the higher energy midgap states of negative charge. Hence, these higher energy midgap states can become transiently populated by the photoexcited electrons, so that the luminescence increases. It is of interest to determine the relative magnitude of contribution from the surface states vs. bulk states to emission of “green” and “blue” light by the nanocrystalline SrTiO3 at room temperature. Experiments are in progress by the time-dependent PL spectroscopy to investigate this behavior in more detail. We anticipate that the room temperature synchronous luminescence spectroscopy of nanocrystalline strontium titanate to study optically emissive midgap states as proposed herein can lead to the new capabilities in characterization of electronic properties and defects in other metal oxides, in studies of charge relaxation in the semiconductors, in metal-semiconductor devices, photovoltaics etc. Further, reversible quenching of the photoluminescence in the nanocrystalline SrTiO3 at room temperature due to adsorption of water is of interest for emerging applications of this material as heterogeneous photocatalyst in the aqueous phase, sorbent, working element of humidity sensors, and beyond.

4. Conclusions We characterized nanocrystalline cubic strontium titanate at 25 °C using “conventional” PL emission, excitation and synchronous luminescence spectroscopy, the Raman, optical reflectance spectroscopy, and XRD. Upon photoexcitation, the following optical transitions take place: a) emission at ca. 430 nm from the midgap state below the conduction band minimum, b) “blue” emission at ca. 460 nm to the midgap state, and c) “green” emission at ca. 500 nm between the two midgap states. Synchronous luminescence spectra of nanocrystalline cubic strontium titanate measured at 25 °C are significantly narrower and better resolved than “conventional” PL emission spectra which aids in spectral assignments. Utilization of water in the vapor phase as spectroscopic “probe molecule” has revealed, for the first time, that quenching of emission of visible light by the nanocrystalline SrTiO3 upon water adsorption at room temperature proceeds via interactions with surface midgap states. Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.saa.2016.11.011.

S. Taylor, A. Samokhvalov / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 174 (2017) 54–61

Acknowledgments A.S. thanks Dr. Dmitri Barbash from Drexel University Core Facilities for providing access to the XRD instrument. S. T. thanks Rutgers University Camden for his Dean's Graduate Research Grant. References [1] T.K. Townsend, N.D. Browning, F.E. Osterloh, Nanoscale strontium titanate photocatalysts for overall water splitting, ACS Nano 6 (2012) 7420–7426. [2] K. Hbaieb, Exploring strontium titanate as a reforming catalyst for dodecane, Appl. Nanosci. (2015) 1–8. [3] D. Zhang, Adsorption and preconcentration capabilities of nano-barium-strontium titanate powder coated by dithizone for trace bismuth, Appl. Mech. Mater. 71–78 (2011) 3504–3507. [4] D. Marre, E. Bellingeri, I. Pellecchi, L. Pellegrino, A. Tumino, A.S. Siri, Electronic devices based on semiconducting strontium titanate, J. Phys. IV 11 (2001) 331–335. [5] J. Gerblinger, H. Meixner, Fast oxygen sensors based on sputtered strontium titanate, Sensors Actuators B Chem. 4 (1991) 99–102. [6] A. Okazaki, M. Kawaminami, Lattice constant of strontium titanate at low temperatures, Mater. Res. Bull. 8 (1973) 545–550. [7] K. van Benthem, C. Elsässer, R.H. French, Bulk electronic structure of SrTiO3: experiment and theory, J. Appl. Phys. 90 (2001) 6156–6164. [8] H. Yaghoubi, Z. Li, Y. Chen, H.T. Ngo, V.R. Bhethanabotla, B. Joseph, S. Ma, R. Schlaf, A. Takshi, Toward a visible light-driven photocatalyst: the effect of midgap-states-induced energy gap of undoped TiO2 nanoparticles, ACS Catal. 5 (2015) 327–335. [9] P. Nagpal, V.I. Klimov, Role of mid-gap states in charge transport and photoconductivity in semiconductor nanocrystal films, Nat. Commun. 2 (2011) 486. [10] J. Liu, S. Wen, X. Zou, F. Zuo, G.J.O. Beran, P. Feng, Visible-light-responsive copper(II) borate photocatalysts with intrinsic midgap states for water splitting, J. Mater. Chem. A 1 (2013) 1553–1556. [11] T.D. Shen, I. Shmagin, C.C. Koch, R.M. Kolbas, Y. Fahmy, L. Bergman, R.J. Nemanich, M.T. McClure, Z. Sitar, M.X. Quan, Photoluminescence from mechanically milled Si and SiO2 powders, Phys. Rev. B Condens. Matter 55 (1997) 7615–7623. [12] M. Kwiatkowski, I. Bezverkhyy, M. Skompska, ZnO nanorods covered with a TiO2 layer: simple sol-gel preparation, and optical, photocatalytic and photoelectrochemical properties, J. Mater. Chem. A 3 (2015) 12748–12760. [13] Z.H. Bai, M. Fujii, K. Imakita, S. Hayashi, Strong white photoluminescence from annealed zeolites, J. Lumin. 145 (2014) 288–291. [14] D. Saha, S.G. Deng, Adsorption equilibrium, kinetics, and enthalpy of N2O on zeolite 4A and 13X, J. Chem. Eng. Data 55 (2010) 3312–3317. [15] T. Vo-Dinh, Multicomponent analysis by synchronous luminescence spectrometry, Anal. Chem. 50 (1978) 396–401. [16] A. Andrade-Eiroa, G. de-Armas, J.-M. Estela, V. Cerda, Critical approach to synchronous spectrofluorimetry. I, Trends Anal. Chem. 29 (2010) 885–901. [17] A. Andrade-Eiroa, G. de-Armas, J.-M. Estela, V. Cerda, Critical approach to synchronous spectrofluorimetry. II, Trends Anal. Chem. 29 (2010) 902–927. [18] M. Moriyasu, Y. Yokoyama, S. Ikeda, Quenching of uranyl luminescence by water molecule, J. Inorg. Nucl. Chem. 39 (1977) 2211–2214. [19] J. Dai, M.L. McKee, A. Samokhvalov, Fluorescence of A100 MOF and adsorption of water, indole, and naphthalene on A100 by the spectroscopic, kinetic, and DFT studies, J. Phys. Chem. C 119 (2015) 2491–2502. [20] F. Wang, J. Wang, X. Liu, Direct evidence of a surface quenching effect on size-dependent luminescence of upconversion nanoparticles, Angew. Chem. Int. Ed. 49 (2010) 7456–7460. [21] J.-C. Boyer, M.-P. Manseau, J.I. Murray, F.C.J.M. van Veggel, Surface modification of upconverting NaYF4 nanoparticles with PEG–phosphate ligands for NIR (800 nm) biolabeling within the biological window, Langmuir 26 (2010) 1157–1164. [22] R. Arppe, I. Hyppänen, N. Perälä, R. Peltomaa, M. Kaiser, C. Würth, S. Christ, U. ReschGenger, M. Schäferling, T. Soukka, Quenching of the upconversion luminescence of NaYF4:Yb3+,Er3+ and NaYF4:Yb3+,Tm3+ nanophosphors by water: the role of the sensitizer Yb3+ in non-radiative relaxation, Nanoscale 7 (2015) 11746–11757. [23] M. Tata, S. Banerjee, V.T. John, Y. Waguespack, G.L. McPherson, Fluorescence quenching of CdS nanocrystallites in AOT water-in-oil microemulsions, Colloids Surf. A Physicochem. Eng. Asp. 127 (1997) 39–46. [24] S.W. Buckner, R.L. Konold, P.A. Jelliss, Luminescence quenching in PbS nanoparticles, Chem. Phys. Lett. 394 (2004) 400–404. [25] J. Dai, M. McKee, A. Samokhvalov, Adsorption of naphthalene and indole on F300 MOF in liquid phase by the complementary spectroscopic, kinetic and DFT studies, J. Porous. Mater. 1-19 (2014). [26] M. Demir, M. McKee, A. Samokhvalov, Interactions of thiophenes with C300 Basolite MOF in solution by the temperature-programmed adsorption and desorption, spectroscopy and simulations, Adsorption (2014) 1–14.

61

[27] M. Mehta, S. Taylor, A. Samokhvalov, Photochemical synthesis, characterization, photoinduced electron transfer, charging and discharging in copper-titania colloid, J. Photochem. Photobiol., A 317 (2016) 1–8. [28] S. Taylor, M. Mehta, A. Samokhvalov, Production of hydrogen by glycerol photoreforming using binary nitrogen-metal-promoted N-M-TiO2 photocatalysts, ChemPhysChem 15 (2014) 942–949. [29] A. Alzahrani, A. Samokhvalov, Conventional and cryo-synchronous luminescence spectra of orthorhombic calcium titanate, J. Lumin. 178 (2016) 430–436. [30] J.R. Lakowicz, Principles of Fluorescence Spectroscopy, Springer, New York, 2006. [31] F.P. Cai, Y.B. Tang, F.Y. Chen, Y. Yan, W.D. Shi, Enhanced visible-light-driven photocatalytic degradation of tetracycline by Cr3+ doping SrTiO3 cubic nanoparticles, RSC Adv. 5 (2015) 21290–21296. [32] A.E. Souza, G.T.A. Santos, B.C. Barra, W.D. Macedo, S.R. Teixeira, C.M. Santos, A. Senos, L. Amaral, E. Longo, Photoluminescence of SrTiO3: influence of particle size and morphology, Cryst. Growth Des. 12 (2012) 5671–5679. [33] W.F. Zhang, Y.L. He, M.S. Zhang, Z. Yin, Q. Chen, Raman scattering study on anatase TiO2 nanocrystals, J. Phys. D. Appl. Phys. 33 (2000) 912. [34] X. Wu, D. Wu, X. Liu, Negative pressure effects in SrTiO3 nanoparticles investigated by Raman spectroscopy, Solid State Commun. 145 (2008) 255–258. [35] H. Tang, H. Berger, P.E. Schmid, F. Levy, Optical properties of anatase (TiO2), Solid State Commun. 92 (1994) 267–271. [36] F.C. Jentoft, Ultraviolet-visible-near infrared spectroscopy in catalysis: theory, experiment, analysis, and application under reaction conditions, in: B.C. Gates, H. Knozinger (Eds.), Advances in Catalysis, Vol. 52, Elsevier Academic Press Inc, San Diego 2009, pp. 129–211. [37] H. Lin, C.P. Huang, W. Li, C. Ni, S.I. Shah, Y.H. Tseng, Size dependency of nanocrystalline TiO2 on its optical property and photocatalytic reactivity exemplified by 2chlorophenol, Appl. Catal., B 68 (2006) 1–11. [38] H.G. Hecht, The interpretation of diffuse reflectance spectra, J. Res. Natl. Bur. Stand. A Phys. Chem. 80A (1976) 567–583. [39] D.L. Wood, J. Tauc, Weak absorption tails in amorphous semiconductors, Phys. Rev. B 5 (1972) 3144–3151. [40] J.B.M. Goodall, S. Kellici, D. Illsley, R. Lines, J.C. Knowles, J.A. Darr, Optical and photocatalytic behaviours of nanoparticles in the Ti-Zn-O binary system, RSC Adv. 4 (2014) 31799–31809. [41] Q.I. Rahman, M. Ahmad, S.K. Misra, M. Lohani, Efficient degradation of Methylene Blue dye over highly reactive Cu doped strontium titanate (SrTiO3) nanoparticles photocatalyst under visible light, J. Nanosci. Nanotechnol. 12 (2012) 7181–7186. [42] Y. Talmi, D.C. Baker, J.R. Jadamec, W.A. Saner, Fluorescence spectrometry with optoelectronic image detectors, Anal. Chem. 50 (1978) 936A–952A. [43] P. Jayabal, V. Sasirekha, J. Mayandi, K. Jeganathan, V. Ramakrishnan, A facile hydrothermal synthesis of SrTiO3 for dye sensitized solar cell application, J. Alloys Compd. 586 (2014) 456–461. [44] P.J. Hargis, T.J. Sobering, G.C. Tisone, J.S. Wagner, S.A. Young, R.J. Radloff, Ultraviolet fluorescence detection and identification of protein, dna, and bacteria, Optical Instrumentation for Gas Emissions Monitoring and Atmospheric Measurements, Spie - Int Soc Optical Engineering, Bellingham 1995, pp. 147–153. [45] S.E. Cabaniss, Theory of variable-angle synchronous fluorescence spectra, Anal. Chem. 63 (1991) 1323–1327. [46] L.J. Soltzberg, S. Lor, N. Okey-Igwe, R. Newman, 3D fluorescence characterization of synthetic organic dyes, Am. J. Anal. Chem. 3 (2012) 622–631. [47] X. Yong, M.A.A. Schoonen, The absolute energy positions of conduction and valence bands of selected semiconducting minerals, Am. Mineral. 85 (2000) 543–556. [48] D. Kumar, R.C. Budhani, Defect-induced photoluminescence of strontium titanate and its modulation by electrostatic gating, Phys. Rev. B 92 (2015). [49] W.F. Zhang, Z. Yin, M.S. Zhang, Study of photoluminescence and electronic states in nanophase strontium titanate, Appl. Phys. A Mater. Sci. Process. 70 (2000) 93–96. [50] R. Viteri, L.M. Procel, F. Tipan, D.F. Ortiz, A. Stashans, This reference is available at http://spie.org/Publications/Proceedings/Paper/10.1117/12.515664. [51] A. Bhogale, N. Patel, P. Sarpotdar, J. Mariam, P.M. Dongre, A. Miotello, D.C. Kothari, Systematic investigation on the interaction of bovine serum albumin with ZnO nanoparticles using fluorescence spectroscopy, Colloids Surf., B 102 (2013) 257–264. [52] R.A. Evarestov, A.V. Bandura, V.E. Alexandrov, Adsorption of water on (001) surface of SrTiO3 and SrZrO3 cubic perovskites: hybrid HF-DFT LCAO calculations, Surf. Sci. 601 (2007) 1844–1856. [53] S.K. Sahu, P.S. Maram, A. Navrotsky, Thermodynamics of nanoscale calcium and strontium titanate perovskites, J. Am. Ceram. Soc. 96 (2013) 3670–3676. [54] K.S. Williams, J.L. Lenhart, J.W. Andzelm, S.V. Bandara, N.F. Baril, N.C. Henry, M.Z. Tidrow, First principles investigation of water adsorption and charge transfer on III–V(110) semiconductor surfaces, Surf. Sci. 622 (2014) 71–82.