Blue Photoluminescence and Cyan-Colored Afterglow of Undoped SrSO4 Nanoplates.

Undoped SrSO4 nanoplates were synthesized via the composite hydroxide-mediated approach. The products were characterized by means of X-ray diffractometry, scanning electron microscopy, X-ray energy-dispersive spectroscopy, X-ray photoelectron spectroscopy, photoluminescence (PL) spectroscopy, electron spin resonance technique, afterglow spectroscopy, and thermoluminescence dosimetry. The steady-state PL spectrum of undoped SrSO4 nanoplates can be deconvoluted into two distinct Gaussian bands centered at 2.97 eV (417.2 nm) and 2.56 eV (484.4 nm), respectively. The nature of the defect emissions is confirmed through the emission-wavelength-dependent PL decays as well as the excitation-wavelength-dependent PL decays. A cyan-colored afterglow from undoped SrSO4 nanoplates can be observed with naked eyes in the dark, and the afterglow spectrum of the undoped SrSO4 nanoplates exhibits a peak at about 492 nm (2.52 eV). The duration of the afterglow is measured to be 16 s. The thermoluminescence glow curve of the undoped SrSO4 nanoplates shows a peak at about 40.1 °C. The trapping parameters are determined with the peak shape method, the calculated value of the trap depth is 0.918 eV, and the frequency factor is 1.2 × 1014 s-1. Using density functional calculations, the band structures and densities of states of oxygen-deficient SrSO4 and strontium-deficient SrSO4 are presented. The mechanisms of the cyan-colored afterglow are discussed for undoped SrSO4, and the oxygen vacancies in SrSO4 are proposed to be the luminescence center of the afterglow.

Introduction

Afterglow materials are materials that continue to emit light after the removal of the excitation source.[1] The new era of afterglow materials came when long-lasting afterglow of Eu2+- and Dy3+-codoped SrAl2O4 was reported in 1996.[2] Since then, intensive research activities have been focused on rare-earth doped alkaline-earth aluminates (MAl2O4, M = Mg, Ca, Sr, and Ba) due to their advantages of long afterglow duration, high light intensity, and non-toxicity.[3−8] However, these afterglow materials exhibit poor resistance to moisture and water due to the hydrolysis of aluminates. Apparently, it is of significance to develop water-resistant afterglow materials in order to meet the increasing demand on water-resistant afterglow materials for outdoor applications.

An essential prerequisite of water-resistant afterglow materials is that host materials are poorly soluble in water. With the solubility of 1 part in 8800, strontium sulfate (SrSO4) is poorly soluble in water. The low aqueous solubility and low toxicity make SrSO4 an ideal candidate for water-resistant afterglow materials provided that SrSO4 can display afterglow. We have noticed that some undoped inorganic materials can give off afterglows. For instance, a blue afterglow peaking at 480–490 nm was reported for undoped HfO2,[9,10] and a blue afterglow peaking at 498 nm was reported for undoped inverse spinel Mg2SnO4.[11] Additionally, a blue afterglow peaking at about 440 nm has been recently reported for undoped CaAl2O4 nanocrystals.[12,13] These reports indicate that an alternative strategy toward designing water-resistant afterglow materials is through the examination of water-insoluble undoped inorganic materials. However, the afterglow characteristics of undoped SrSO4 are not reported yet. In this paper, the afterglow properties of undoped SrSO4 nanoplates are presented. The mechanisms of the cyan-colored afterglow are discussed for undoped SrSO4. Our results demonstrate that the oxygen vacancies in the oxide are responsible for the observed afterglow in undoped SrSO4. This work aims at shedding light on the origin of the blue photoluminescence (PL) and providing a comprehensive understanding on the mechanisms of cyan-colored afterglow from water-insoluble undoped SrSO4.

Results and Discussion

Phase and Morphology of Undoped SrSO4

Figure depicts the XRD pattern of the undoped SrSO4 nanoplates. The open circles (in black) in Figure represent the raw XRD data of the undoped SrSO4 nanoplates. As can be seen in Figure , the XRD profile of the undoped SrSO4 nanoplates exhibits distinct peaks at 20.984, 23.579, 25.932, 27.039, 28.063, 30.043, 32.765, 33.484, 34.714, 37.817, 39.984, 40.835, 42.173, 44.346, 45.329, 46.610, and 49.013°. For comparison, the standard diffraction data of SrSO4, which are available in a data file provided by the Joint Committee on Powder Diffraction Standards (JCPDS) as JCPDS no. 05-0593, are displayed at the bottom by the vertical bars (in pink). According to JCPDS no. 05-0593, these peaks can be assigned to the reflections from the (011), (111), (002), (210), (102), (211), (112), (020), (301), (212), (220), (103), (221), (113), (401), (410), and (321) crystallographic planes of SrSO4.[14,15] Some of the diffraction peaks in Figure are marked with Miller indices. Obviously, the XRD data in Figure confirm the formation of orthorhombic SrSO4via the composite hydroxide-mediated approach.

Figure 1

XRD curve and its Rietveld analysis of the undoped SrSO4 nanoplates. Open circles: raw data; solid green curve: Rietveld diffractogram. The standard diffraction data of SrSO4 (JCPDS no. 05-0593) are displayed at the bottom for comparison.

Being widely used in the field of powder XRD, Rietveld analysis is an effective method to determine the crystal structures and lattice parameters for a large diversity of crystals. Rietveld analysis was performed through the wide pattern fitting of the XRD pattern in Figure . The program FULLPROF Suite 2014 was used to perform the Rietveld refinement. The solid green curve in Figure shows the Rietveld diffractogram of the undoped SrSO4 nanoplates. The refined unit cell parameters for the undoped SrSO4 nanoplates are a = 0.8361 nm, b = 0.5351 nm, and c = 0.6871 nm with the unit cell volume of 0.3072 nm3. Apparently, the lattice parameters derived from our SrSO4 nanoplates are very close to those of SrSO4 in the standard file (ICSD 85808) whose a = 0.8359 nm, b = 0.5351 nm, and c = 0.6869 nm.

Figure depicts the SEM micrographs of undoped SrSO4. As shown in Figure , we can see clearly that SrSO4 nanoplates are formed via the composite hydroxide-mediated approach, and the typical thickness of the undoped SrSO4 nanoplates is around 80 nm. However, the other two lateral dimensions of the nanoplates vary in the range of 200–800 nm. For example, the length and width of the nanoplate marked in Figure are about 543 and 200 nm, respectively. As documented in the literature, the composite hydroxide-mediated approach is a powerful methodology for creating a diversity of nanostructures.[16,17] For instance, Eu2+-doped SrSO4 nanostructures in the morphology of nanoneedle, nanorod, octagonal disk, and hexagonal disks were synthesized via the hydrothermal approach,[15] but SrSO4 in the morphology of nanoplates was not reported yet. Thus, our undoped SrSO4 nanoplates are unique in their morphology.

Figure 2

SEM micrograph of the undoped SrSO4 nanoplates.

Figure depicts the EDX spectrum of the undoped SrSO4 nanoplates. As can be seen in Figure , the three X-ray emission peaks in the EDX spectrum are located at 0.53, 1.81, and 2.31 keV, which can be attributed to the characteristic X-ray emissions of O(Kα1), Sr(Lα1,2), and S(Kα1,2), respectively. In addition to those peaks, the X-ray emissions of Au(Mα1) and Au(Lα1) can also be identified at 2.122 and 9.713 keV, respectively. As described previously, the Au element in the specimen was introduced in the process of Au sputtering for the convenience of SEM characterization.[3,18] As expected, the EDX spectrum of undoped SrSO4 confirms that elements Sr, S, and O are present in our synthesized compound.

Figure 3

EDX spectrum of the undoped SrSO4 nanoplates.

XPS Spectrum of Undoped SrSO4

Figure illustrates the XPS survey scan (a), XPS spectra of S 2p (b), Sr 3d (c), and O 1s (d) in SrSO4 nanoplates. Obviously, the photoelectron lines of S 2p, O 1s, and Sr 3d can be identified in Figure a, which confirms the presence of elements S, O, and Sr in our target compound SrSO4 nanoplates. The peak of S 2p3/2 in Figure b is located at 168.98 eV. According to the report by Vasquez, the photoelectron line of S 2p3/2 in SrSO4 is located at 168.75 eV.[19] Thus, our recorded photoelectron line of S 2p3/2 is very close to that documented in the literature. The binding energy peak of Sr 3d5/2 can be derived from Figure c. As shown in Figure c, the binding energy peak of Sr 3d5/2 is located at 133.68 eV. When compared to the peak of Sr 3d5/2 (133.85 eV) in the SrSO4 nanofilm, the peak of Sr 3d5/2 in our undoped SrSO4 nanoplates is shifted 0.17 eV toward the lower binding energy.[20] Interestingly, in Figure c, we also recorded a photoelectron line at 135.38 eV, which can be tentatively attributed to the binding energy peak of Sr 3d3/2. Finally, the XPS spectral profile of O 1s is displayed in Figure d. As can be seen in Figure d, the XPS spectral profile of O 1s is located at 531.88 eV, which is very close to the reported values of 531.8 and 531.9 eV for SrSO4 by Vasquez.[19,20] Obviously, it is hard to deconvolute this high-resolution XPS spectrum of O 1s into two components. Generally speaking, the high-resolution XPS spectrum of O 1s can be deconvoluted into two components if the population density of oxygen vacancy in SrSO4 is high enough since the chemical environment of Sr–O bond in perfect SrSO4 is different from that in oxygen-deficient SrSO4. Thus, the data in Figure d suggest that the population density of oxygen vacancy in the SrSO4 matrix is not sufficiently high.

Figure 4

XPS survey scan (a), XPS spectra of S 2p (b), Sr 3d (c), and O 1s (d) in SrSO4 nanoplates.

Steady-State PL Spectrum of Undoped SrSO4

Figure illustrates the PL spectrum of the undoped SrSO4 nanoplates in the energy scale. The hollow circles in Figure represent the raw PL data. Apparently, this broad PL spectrum can be deconvoluted into two Gaussian bands. The first Gaussian band, as shown by the blue curve in Figure , is peaked at 2.97 eV (417.5 nm). The second Gaussian band, as shown by the green curve in Figure , is centered at 2.56 eV (484.4 nm). For clarity, the former PL band is denoted as PL band A, while the latter PL band is denoted as PL band B. The area ratio of the two PL bands A to B is 1.42:1. By using the techniques reported earlier,[21] the Commission Internationale de l’Eclairage (CIE) chromaticity coordinates of the undoped SrSO4 are calculated to be (0.167, 0.197), and the PL color of the undoped SrSO4 nanoplates is identified as blue.

Figure 5

PL spectrum of the undoped SrSO4 nanoplates in energy scale and its Gaussian band deconvolution.

SrSO4, which is a typical insulator with a band gap of around 7.6 eV,[22] is often utilized as the host material of rare-earth dopants to develop efficient luminescent materials. Due to the large band gap of SrSO4, the photon energy of our ultraviolet laser (3.82 eV) is not high enough to pump carriers from its valence band to conduction band. As a result, the two PL sub-bands in Figure cannot originate from the band-to-band recombination.[14] Thus, the two PL bands are very likely connected with intrinsic defects in SrSO4. Examples of the intrinsic defects in SrSO4 include oxygen and strontium vacancies, oxygen and strontium interstitials, and oxygen and strontium antisites. After having considered the high formation energies of antisites and interstitials, only the oxygen vacancies and the strontium vacancies are the most probable intrinsic defects in SrSO4. The oxygen vacancy is generally known to be one of the important intrinsic defects in oxides, and the oxygen vacancy-related emissions are documented for a large diversity of host materials such as CaAl2O4,[12,13] SrAl2O4,[4−6] BaAl2O4,[8] SrSO4,[14] ZnWO4,[23,24] and ZnMoO4.[25] Consequently, the two PL bands A and B in Figure can be attributed to the oxygen and strontium vacancies in SrSO4.

Afterglow Spectrum and Afterglow Decay Profile of Undoped SrSO4

The afterglow spectrum of undoped SrSO4 nanocrystals is shown in Figure a. As can be seen in Figure a, the peak position of this afterglow spectrum is located at 492 nm (2.52 eV). This afterglow spectrum spans across a wide spectral range from 375 to 625 nm, but the profile of this afterglow spectrum is relatively narrow when compared to the PL spectrum of the undoped SrSO4 nanoplates in Figure . Obviously, one part of this afterglow spectrum falls into the blue spectral regime, while the other part of this spectrum falls into the green spectral regime. Generally speaking, green and blue mix to produce cyan. According to the method described in our previous work,[21,23] the CIE chromaticity coordinates of the afterglow are derived to be (0.124, 0.327), and the calculated value of its color temperature is 24659 K. As a result, the color of the afterglow can be depicted as cyan. The inset in Figure a depicts the afterglow photo of the undoped SrSO4 nanoplates. Indeed, the color of the afterglow is cyan. The afterglow decay curve of undoped SrSO4 is presented in Figure b. As shown in Figure b, this decay curve exhibits that the cyan-colored afterglow of undoped SrSO4 can last for about 16 s. Although it is many times shorter than those of rare-earth doped SrAl2O4 and BaAl2O4,[4−6,8] the duration of the cyan-colored afterglow of undoped SrSO4 is in the same order of those of undoped CaAl2O4, undoped HfO2, and undoped Mg(Mg,Sn)O4.[9−13]

Figure 6

(a) Afterglow spectrum of the undoped SrSO4 nanoplates. (b) Afterglow decay profile of the undoped SrSO4 nanoplates. Inset in (a): afterglow photo of the undoped SrSO4 nanoplates. Inset in (b): afterglow photos of the undoped SrSO4 nanoplates at 0, 4, 8, and 12 s after the extinction of excitation.

Although SrSO4 is an excellent host material of rare-earth dopants to develop efficient phosphors, neither rare-earth-doped SrSO4 nor undoped SrSO4 was reported to give off afterglow. Therefore, the key issue is why undoped SrSO4 can exhibit the cyan-colored afterglow. A comparison of Figure a with Figure reveals that the peak energy of the afterglow spectrum (2.52 eV) in Figure is very close to that of PL band B in Figure (2.56 eV). This fact suggests that the blue afterglow of undoped SrSO4 is likely correlated with the intrinsic defects in SrSO4. Actually, oxygen vacancies are proposed as the luminescence center of the afterglow in undoped CaAl2O4,[12,13] undoped HfO2,[9,10] and undoped inverse spinel Mg2SnO4.[11] Moreover, Zhou et al. recently reported blue afterglow from undoped boron oxide.[26] Thus, the knowledge on defect energy levels of intrinsic defects, particularly the oxygen and strontium vacancies, in SrSO4 can help in understanding the origin of cyan afterglow of SrSO4.

Both the PL and afterglow of these SrSO4 nanoplates appear to be fairly stable after being stored in moisture for 2 years. Interestingly, both the afterglow intensity and the lifetime of SrSO4 nanoplates are found to be chemically stable when the phosphor is in contact with acidic solutions (hydrochloric acid) and basic solutions (sodium hydroxide). Such advantages over aluminate-based afterglow materials can attract considerable attention due to the outstanding chemical stability of SrSO4. Furthermore, we checked the thermal stability of the afterglow of undoped SrSO4 by annealing the phosphors at 200, 300, 400, 500, 600, and 700 °C. The duration of each annealing was 2 h. The afterglow photos of undoped SrSO4 after annealing at different temperatures are given in Figure S1. Both the afterglow intensity and the afterglow duration of undoped SrSO4 are found to increase monotonically with the annealing temperature up to 400 °C. After having reached their apex, both the afterglow intensity and the afterglow duration turn their heads down as the annealing temperature increases from 500 to 700 °C. In particular, the afterglow of undoped SrSO4 becomes quenched after annealing at 700 °C for 2 h. Additionally, we measured the PL spectra of undoped SrSO4 nanoplates at 200, 300, 400, and 500 °C on a heating console. As shown in Figure S2, the integrated PL intensity of undoped SrSO4 nanoplates is quite stable as the temperature of the heating console varies from 200 to 400 °C. As a contrast, the integrated PL intensity of undoped SrSO4 nanoplates is severely degraded when the temperature of the heating console is increased to 500 °C.

Electronic Structures of Strontium-Deficient SrSO4 and Oxygen-Deficient SrSO4

To get better insights into this problem, accurate calculations on the band structures and densities of states are required for defect-bearing SrSO4. Figure represents the density functional theory calculated band structures and densities of states of oxygen-deficient SrSO4 (i.e., SrSO4−δ, where δ = 0.0625). In the density functional calculations, the exchange–correlation functional was treated within the GGA + U scheme by the Perdew–Burke–Ernzerhof potential. The U parameter was selected as U2p = 4 eV for oxygen. As shown in Figure , the calculated band gap of the oxygen-deficient SrSO4 is 7.23 eV. It is found that the oxygen vacancy can introduce two defect energy levels in the band gap of SrSO4, one of which is located at EV + 2.38 eV, while the other is located at Ev + 0.30 eV. These defect energy levels can be clearly identified in the density of states, as shown in Figure b. Apparently, our calculated band gap value of SrSO4 is underestimated when compared to the experimental band gap value of 7.6 eV for SrSO4.[22] The density functional calculations are often known to underestimate the band gap value of materials. For example, Zhai et al. reported that SrSO4 is an insulator with an indirect band gap of 6.14 eV using the GGA approach in the density functional calculations.[14] Hu et al. reported that the indirect band gap of SrSO4 is 6.0 eV using the plane-wave pseudopotential GGA approach.[27] Using the scissors operator to overcome this underestimation, the band gap of oxygen-deficient SrSO4 is adjusted to be 7.6 eV; meanwhile, the oxygen vacancy-introduced defect energy levels are adjusted to EV + 2.50 eV and EV + 0.31 eV.

Figure 7

Density functional theory calculated band structures (a) and density of states (b) of oxygen-deficient SrSO4 (i.e., SrSO4−δ, where δ = 0.0625). The exchange–correlation functional was treated within the GGA + U scheme by the Perdew–Burke–Ernzerhof potential. The U parameter was selected as U2p = 4 eV for oxygen.

Figure represents the density functional theory calculated band structures and densities of states of strontium-deficient SrSO4 (i.e., Sr1−δSO4, where δ = 0.0625). As shown in Figure , the calculated band gap of strontium-deficient SrSO4 is 7.72 eV, which is slightly larger than the reported band gap value of 7.6 eV for SrSO4.[22,27] Obviously, the defect energy levels of strontium-deficient SrSO4 are very close to the top of the valence band of SrSO4. Detailed analysis reveals that the peak of these defect energy levels is located at EV + 0.177 eV. After the scissors operation to overcome the band gap overestimation problem, the band gap of strontium-deficient SrSO4 is adjusted to 7.6 eV; meanwhile, the peak of defect energy levels of strontium vacancy is adjusted to EV + 0.175 eV.

Figure 8

Density functional theory calculated band structures (a) and density of states (b) of Sr-deficient SrSO4 (i.e., Sr1−δSO4, where δ = 0.0625). The exchange–correlation functional was treated within the GGA + U scheme by the Perdew–Burke–Ernzerhof potential. The U parameter was selected as U2p = 4 eV for oxygen.

The main motivation of the calculated band structures and densities of states of oxygen-deficient SrSO4 and strontium-deficient SrSO4 is to highlight the possible roles of oxygen and strontium vacancies in the PL and afterglow of undoped SrSO4. The results in Figures and 8 show that both the oxygen vacancy and the strontium vacancy can generate defect energy levels in the band gap of SrSO4. On the one hand, the oxygen vacancy introduces one deep energy level (EV + 2.50 eV) and one shallow energy level (EV + 0.31 eV) in the band gap of SrSO4. Since they are positively charged, they can work as electron traps. On the other hand, the strontium vacancy can generate shallow defect energy levels in the band gap of SrSO4 (EV + 0.175 eV). These strontium vacancies can work as hole traps because they are negatively charged. Apart from acting as carrier traps, these vacancies might be involved in the radiative recombination processes of undoped SrSO4 nanoplates. Therefore, both the oxygen vacancies and the strontium vacancies might play important roles in the PL and afterglow of undoped SrSO4. As discussed in the following section, comparison of the experimental data with the calculated electronic structures can unveil that the oxygen vacancy acts not only as the luminescence center of PL but also as a trap center and a luminescence center of afterglow.

Possible PL and Afterglow Mechanisms of Undoped SrSO4

The possible PL and afterglow mechanisms of undoped SrSO4 are illustrated in Figure . The band gap of SrSO4 is assumed to be 7.6 eV.[22] As shown in Figure , one of the defect energy levels of oxygen vacancy is located at EV + 2.50 eV, while the other defect energy level of oxygen vacancy, which is located at EV + 0.31 eV, is not shown in Figure for the purpose of clarity. Similarly, the defect energy level of strontium vacancy is located at EV + 0.175 eV. Upon the excitation of the ultraviolet photons with a wavelength of 325 nm (3.82 eV), the SrSO4 lattice can absorb some excitation energy due to the presence of various kinds of intrinsic defects in practical SrSO4 (process ①). After non-radiative relaxation, the hot electrons can be captured either by oxygen vacancies in the SrSO4 lattice (process ②) or by electron traps present in SrSO4 (process ③). In the case of our undoped SrSO4, the electron traps are the oxygen vacancies themselves because, qualitatively speaking, the oxygen vacancies in SrSO4 can work simultaneously as electron traps, the luminescence center of PL, and the luminescence center of afterglow.[12,13] After the capture of electrons at oxygen vacancies, the electrons can recombine radiatively with holes in the valence band of SrSO4, resulting in a PL band peaking at about 2.50 eV, provided that the defect energy levels of oxygen vacancy are accurately located at EV + 2.50 eV (process ④). Such an emission band corresponds to the PL band peaking at about 496 nm, which is denoted here as PL band A* for the convenience of discussion. Besides the recombination with holes in the valence band, the electrons captured at the oxygen vacancy apparently have a certain possibility to recombine radiatively with holes trapped at strontium vacancies, yielding another PL band peaking at about 2.33 eV, provided that the defect energy level of strontium vacancy is accurately located at EV + 0.175 eV (process ⑤). Such a PL band corresponds to a PL band peaking at about 523 nm, which is denoted as PL band B*.

Figure 9

Schematic illustration of the PL and afterglow mechanisms of undoped SrSO4. Process ①: absorption of the excitation energy by various kinds of intrinsic defects in the SrSO4 lattice. Process ②: non-radiative relaxation of hot electrons to oxygen vacancies in the SrSO4 lattice. Process ③: non-radiative relaxation of hot electrons to electron traps (oxygen vacancies) in the SrSO4 lattice. Process ④: radiative recombination of electrons captured at the oxygen vacancy with holes in the valence band of SrSO4, which results in a broad PL band. Process ⑤: radiative recombination of electrons captured at oxygen vacancies with holes captured at strontium vacancies, yielding another PL band. Process ⑥: thermal release of electrons from the traps in SrSO4, and the subsequent radiative recombination via process ⑤ results in the cyan-colored afterglow.

When comparing the PL bands A* and B* in Figure with the two PL bands A and B in Figure , we can see that the peak energies of the PL bands A* (2.50 eV) and B* (2.33 eV) in Figure are smaller than those of the two PL bands A (2.97 eV) and B (2.56 eV) in Figure . Theoretically speaking, the peak energies of the PL bands A* and B* in Figure should be 2.97 and 2.56 eV, respectively, if our density functional theory calculations could give correct values of the defect energy levels for SrSO4. Unfortunately, the density functional theory calculated values of the defect energy levels are underestimated. Such discrepancies between the calculated peak energies and the experimental peak energies of the two PL bands depend on the density functional theory itself. After having considered the not well-described band gaps of semiconductors and insulators in semilocal approximations to density functional theory, it is hard to reliably determine the defect energy levels within the band gap of SrSO4.[28] That is the reason why the predicted peak energies of the two PL bands A* and B* in Figure are 15.8 and 7.4% smaller than the actual peak energies of the two PL bands A and B in Figure .

The recorded two PL bands A and B in Figure indicate that oxygen vacancies in SrSO4 work as the luminescence center of PL upon the excitation of the ultraviolet laser. As illustrated in Figure , a portion of electrons can also be captured by the traps (oxygen vacancies) in the SrSO4 lattice (process ③). These trapped electrons contribute nothing to the PL of SrSO4 if they are not released from the traps. Once the ultraviolet excitation of the laser is stopped, processes ①–④ are ceased, but the electrons trapped at the positively charged traps (oxygen vacancies) begin their work. Under thermal activation, these trapped electrons can be released from the electron traps (process ⑥). Afterglow can be resulted via the recombination path ⑤ once the released electrons are captured by the luminescence center of PL (i.e., oxygen vacancies). Hence, according to Figure , cyan-colored afterglow with weak intensity can be expected for undoped SrSO4. Obviously, this afterglow mechanism involves the gradual release of electrons from electron traps (i.e., oxygen vacancies), followed by electron migration to strontium vacancies. Therefore, the luminescence center of the afterglow of undoped SrSO4 is the combination of oxygen vacancy and strontium vacancy. If so, the peak energy of the afterglow from undoped SrSO4 is consequently determined by the energy difference between the defect energy levels of oxygen vacancy and strontium vacancy in the band gap of SrSO4. If the proposed afterglow mechanism in Figure is reasonable, the peak energy of the afterglow of undoped SrSO4 should be equal to the peak energy of PL band B (2.56 eV) in Figure . In our case, the peak energy of the afterglow of undoped SrSO4 (2.52 eV) is nearly equal to the peak energy of the PL band B in Figure (2.56 eV). Although it would be better to exactly determine the absolute positions of these defect states in the band gap of SrSO4via the density functional calculations or via any reliable experimental techniques, it is currently not possible for us to do so due to a lot of limitations.

ESR Spectrum of Undoped SrSO4

Electron spin resonance (ESR) spectroscopy is helpful in studying chemical species with unpaired electrons. Figure illustrates the ESR spectrum of undoped SrSO4 measured at room temperature. The microwave frequency was 9.856 GHz. From its crossover point, the center field of this resonance is determined to be 3515.5 G. Thus, the ESR spectrum in Figure shows a signal at about 3515.5 G. This signal corresponds to a gyromagnetic g value of about 2.0031, which is very close to the g value of free electron (2.0023). The data in Figure indicate that the undoped SrSO4 contains unpaired electrons, which probably originate from electrons trapped in positively charged vacancies such as oxygen vacancies. Actually, such positively charged oxygen vacancies are present in a variety of inorganic materials such as CaAl2O4.[12]

Figure 10

ESR spectrum of undoped SrSO4 nanoplates measured at room temperature. Sweep width is 100 G. Microwave frequency: 9.856 GHz; microwave power: 20 mW; modulation frequency of the receiver: 100 Hz; and modulation amplitude of the receiver: 2 G.

Thermoluminescence Glow Curve of Undoped SrSO4

Thermoluminescence is an important research tool to study the energy levels of impurity in solids.[29]Figure represents the thermoluminescence glow curve of undoped SrSO4. As shown in Figure , a broad thermoluminescence band peaking at about 40.1 °C (313.25 K) can be identified for undoped SrSO4. This thermoluminescence band indicates that certain traps are present in undoped SrSO4. In combination with the information derived from the PL and ESR measurements, these traps can be ascribed to the oxygen vacancies in the matrix of SrSO4. The depths of the electron traps (or hole traps) are critically important to the duration of afterglow. Generally speaking, shallow traps in afterglow materials can lead to an afterglow, which decays very quickly. As a contrast, afterglow materials with deep traps have long afterglow time. The trap depth can be calculated using the thermoluminescence glow curve.[30,31] Based on the shape of glow curve, we can calculate the trapping parameters with Chen’s peak shape method. The geometry factor of the thermoluminescence glow curve, μg, is defined by eq where Tm is the temperature at the maximum, T1 is the half width temperature at the low-temperature side of the peak, and T2 is the half width temperature at the high-temperature side of the peak. The readings of T1, Tm, and T2 in Figure are 298.95, 313.25, and 329.75 K, respectively. Thus, the calculated value of the geometry factor is 0.5375 for the thermoluminescence glow curve of undoped SrSO4 nanoplates. This value is very close to the value of μg = 0.52, which indicates the second-order kinetics of the thermoluminescence of undoped SrSO4 nanoplates. Since the thermoluminescence glow curve exhibits second-order kinetics, a considerable amount of retrapping of charge carriers takes place in undoped SrSO4 nanoplates. The trap depth in undoped SrSO4 can be estimated using the following equationwhere E is the trap depth and k is Boltzmann’s constant.[32,33] It is found that the trap depth is 0.918 eV for undoped SrSO4 nanoplates. The calculated value of the trap depth in undoped SrSO4 is indicative of the formation of traps in SrSO4 nanoplates. Obviously, these traps are too shallow to give off sufficiently long afterglow for undoped SrSO4 because a little amount of energy is needed to detrap the charge carriers from the trap centers. The corresponding frequency factor s (in the unit of s–1), which is the frequency of an electron escaping the trap, can be calculated using the second-order kinetics formula as given in eq

Figure 11

Thermoluminescence glow curve of undoped SrSO4. The temperature rising rate was 2 °C/min.

The calculated value of the frequency factor is 1.2 × 1014 s–1. In an attempt to escape from the potential well, the frequency factor represents the product of the number of times an electron hits the wall and the wall reflection coefficient when the trap is treated as a potential well. It is clear that the estimations of the trap depth and the frequency factor utilize the values of shape parameters (T1, T2, and Tm) along with geometry factor (μg). Lifetime is generally estimated using the equationwhere E is the trap depth, s is the frequency factor, and k is Boltzmann’s constant. The value of room-temperature lifetime of undoped SrSO4 nanoplates is estimated to be 17.9 s, which can be utilized to justify the physical basis of the output parameters.

Doping a host material with rare-earth ions can enhance the trap depth and hence increase the peak temperature of the thermoluminescence glow curve.[34] For example, Atone et al. reported that the most prominent glow peak appears at 140 °C for their Tb-doped SrSO4 phosphors after exposure to the irradiation of gamma rays;[35] Khadijeh et al. reported that the dominant glow peak is located at 217 °C for Dy- and Tb-codoped SrSO4 after exposure to the irradiation of gamma ray.[36] The comparison of these thermoluminescence glow curves reveals that doping SrSO4 with rare-earth ions is one effective method to enhance the trap depths in SrSO4. Additionally, Ambast and Sharma reported that the trap depth in CaWO4 varied by doping with Dy3+ alone or by codoping with Dy3+ and K+.[37] The results indicate that trap depth varies after doping or codoping with specific dopants. Consequently, it is expected that doping SrSO4 nanoplates can adjust the trap depth to a suitable value so that sufficiently long afterglow can be achieved. In addition to doping, both thermal annealing and high energy electron irradiation can control the population density of intrinsic defects in SrSO4.[7,34,38] In the case of oxygen vacancy as the luminescence center of PL, the relative value of the population density of the oxygen vacancy can be measured through the PL intensity measurement, and the absolute value of the population density of the oxygen vacancy can be derived via the ESR technique.

Emission-Wavelength Dependence of Time-Resolved PL for Undoped SrSO4 Nanoplates

As mentioned in Figure , the differences in the steady-state PL spectrum and the afterglow spectrum are explained in terms of the presence of multiple luminescence paths in undoped SrSO4. It is known that investigation of time-resolved PL can offer important information on the luminescence paths in a diversity of materials.[5,8,23,24,39] In the light of the mechanisms proposed in Figure , there should be three relaxation paths, at least, for the excited carriers in undoped SrSO4, and the PL lifetime of the carriers will depend on the emission wavelength. In order to study the recombination dynamics in undoped SrSO4 nanoplates, we measured the time-resolved PL spectra at different detection wavelengths when the excitation wavelength was fixed at 375 nm. Figure represents the time-resolved PL spectra of undoped SrSO4 nanoplates at detection wavelengths of 418 nm (a) and 485 nm (b). The two detection wavelengths are close to the peaks of the PL bands A and B in Figure . Indeed, the PL decay profile of undoped SrSO4 is heavily dependent on the emission wavelength, as can be seen in Figure . Detailed analysis shows that each PL decay curve in Figure can be fitted with one quadruple exponential functionwhere I(t) refers to the PL intensity at time t, A0 is the baseline, A is the pre-exponential factor of the ith decay component, and τ is the decay time constant of the ith decay component (i = 1–4).[23,39] The fitting parameters τ and I (i = 1–4) of the time-resolved PL spectra in Figure are listed in Table . The parameter χ2 in Table represents the goodness of fit, and the average lifetime, τavg, is calculated by using the formula

Figure 12

Time-resolved PL spectra of undoped SrSO4 at different detection wavelengths: (a) λem = 418 nm and (b) λem = 485 nm. The excitation wavelength is fixed at 375 nm.

Table 1

Fitting Parameters of the Time-Resolved PL Spectra of Undoped SrSO4 Nanoplates with Different Detection Wavelengths When the Excitation Wavelength Is Fixed at 375 nm

			I1	I2	I3	I4
excitation wavelength (nm)	detection wavelength (nm)	A0	τ1 (ns)	τ2 (ns)	τ3 (ns)	τ4 (ns)	τave (ns)	χ2
375	418	7.63	47139.8	32818.5	8114.1	401.2	3.52	1.153
			0.51	1.88	5.59	15.97
375	485	14.43	40532.4	40611.6	14262.1	1106.4	4.40	1.247
			0.46	1.77	5.49	15.07

These fitting parameters bear important information on the kinetics of carrier recombination. The most prominent feature of the data in Table is that there are four decay components for each time-resolved PL spectrum in Figure . For example, the four decay time constants are τ1 = 0.51 ns, τ2 = 1.88 ns, τ3 = 5.59 ns, and τ4 = 15.97 ns for SrSO4 when the detection wavelength is 418 nm. It is noted that τ1 is at the limit of the measurement capability of the instrument, and therefore, it merely represents the order of the short decay time constant.[5,23,25,39] Consequently, the time-resolved PL spectra in Figure suggest that there are three independent radiative relaxation paths that contribute to the blue emissions of the undoped SrSO4. As discussed in Figure , the three different relaxation paths in undoped SrSO4 nanoplates are correlated to processes ④ and ⑤ and the combined process ⑥ and ⑤. The second prominent feature of the data in Table is the increase in the average lifetime with increasing detection wavelength. As listed in Table , the weighed average lifetimes are calculated to be 3.52 and 4.40 ns for the detection wavelengths of 418 and 485 nm, respectively. Obviously, the average lifetime of the visible emission at 485 nm (4.40 ns) is much longer than that of the emission at 418 nm (3.52 ns). Similar emission-wavelength-dependent lifetime was reported for undoped CaAl2O4 nanocrystals.[12] As documented in the literature, this is the characteristic of emissions involving deep trap states.[40,41] Therefore, the emission-wavelength-dependent PL decay suggests that the emissions involved deep trap states in SrSO4.

Excitation-Wavelength Dependence of Time-Resolved PL for Undoped SrSO4 Nanoplates

It is proved that the luminescence lifetime is also dependent on the excitation wavelength if multiple luminescence centers are present.[42] Thus, we investigated the charge carrier dynamics of the relaxation processes in undoped SrSO4 nanoplates by measuring the time-resolved PL at different excitation wavelengths when the detection wavelength is fixed at 418 nm. Figure represents the time-resolved PL spectra of undoped SrSO4 nanoplates at excitation wavelengths of 255, 320, and 375 nm. The detection wavelength is fixed at 418 nm. It is found that the PL decay curves a and b can be reasonably fitted with one triple exponential function while the PL decay curve c can only be fitted with one quadruple exponential function. Table lists the fitting parameters of each time-resolved PL spectrum in Figure . Apparently, the most notable difference in the PL decays rests on the number of exponential components. As discussed above, the number of radiative recombination paths is 2 when the excitation wavelengths are 255 and 320 nm, respectively. As a sharp contrast, the number of radiative recombination paths is 3 when the excitation wavelength is 375 nm. The second prominent feature of the data in Table is the decrease in the average lifetime with increasing excitation wavelength. As can be seen in Table , the average lifetimes of the carriers in undoped SrSO4 nanoplates are 35.89, 3.80, and 3.52 ns for the excitation wavelengths of 255, 320, and 375 nm, respectively. It is clear that the average lifetime of the PL is longer at shorter excitation wavelengths but shorter at longer excitation wavelengths. These data indicate that the time-resolved PL spectrum of undoped SrSO4 nanoplates is excitation-wavelength-dependent.

Figure 13

Time-resolved PL spectra of undoped SrSO4 at different excitation wavelengths: (a) λex = 255 nm, (b) λex = 320 nm, and (c) λex = 375 nm. The detection wavelength is fixed at 418 nm.

Table 2

Fitting Parameters of the Time-Resolved PL Spectra of Undoped SrSO4 Nanoplates with Different Excitation Wavelengths When the Detection Wavelength Is Fixed at 418 nm

			I1	I2	I3	I4
excitation wavelength (nm)	detection wavelength (nm)	A0	τ1 (ns)	τ2 (ns)	τ3 (ns)	τ4 (ns)	⟨τ⟩ (ns)	χ2
255	418	99.116	727.60	221.97	17.64	none	35.89	1.072
			1.08	5.60	81.24	none
320	418	8.758	956.40	167.69	4.51	none	3.80	1.036
			1.25	5.35	21.17	none
375	418	7.63	47139.8	32818.5	8114.1	401.2	3.52	1.153
			0.51	1.88	5.59	15.97

The average lifetimes in Table contain important information on the energy spacing of the defects involved in the blue emissions from undoped SrSO4 nanoplates. Since the PL band A peaking at 418 nm is associated with the defect emissions in undoped SrSO4 nanoplates, the nature of the PL decay of these defect emissions is generally determined by the energy spacing between the defects associated with the blue emissions. In other words, defects that are close in energy space have a faster decay, while those with larger spacing in energy space have a slower decay.[42] Here, we suppose that the energy level of oxygen vacancy of SrSO4 is located at 2.97 eV above the valence band, as shown in Figure . Upon the excitation of 255 nm photons (4.86 eV), only those defects at EV + 4.86 eV can absorb the excitation energy of the incoming photons. In this case, the absorption species is 1.89 eV above the luminescence center (oxygen vacancies) in energy space. Similarly, the absorption species is only 0.34 eV above the luminescence center (oxygen vacancies) in energy space upon the excitation of 375 nm photons (3.31 eV). Due to the larger energy difference in the former case, it takes much longer time for the excited electron to complete its non-radiative relaxation process ② than the latter case. Consequently, the blue emission at 418 nm decays very fast upon the excitation of 375 nm, while this PL decays much more slowly upon the excitation of 255 nm.

Summary

SrSO4 nanoplates are derived via the composite hydroxide-mediated approach at 240 °C. These phosphors are characterized in detail by means of XRD, SEM, EDX, XPS, steady-state PL, time-resolved PL, and afterglow spectroscopies and thermoluminescence dosimetry. The steady-state PL spectrum of undoped SrSO4 nanoplates can be deconvoluted into two Gaussian bands centered at 2.97 eV (417.2 nm) and 2.56 eV (484.4 nm), respectively. The cyan-colored afterglow peaking at about 492 nm (2.52 eV) is recorded in the undoped SrSO4 material upon ultraviolet laser irradiation (325 nm, 13 mW). The duration of the afterglow is measured to be 16 s. The thermoluminescence glow curve of undoped SrSO4 nanoplates exhibits a peak at about 40.1 °C, and the trap depth is estimated to be about 0.918 eV using the peak shape method. In order to shed light on the origins of the blue PL and the cyan-colored afterglow, density functional theory calculations are performed to derive the band structures and densities of states for oxygen-deficient SrSO4 and strontium-deficient SrSO4. It turns out that both the oxygen vacancies and the strontium vacancies in the lattice of SrSO4 play important roles in the cyan-colored afterglow of undoped SrSO4 nanoplates. The emission wavelength- and the excitation wavelength-dependent PL decays are studied to investigate the nature of defect emissions in SrSO4 nanoplates. Our results demonstrate that (i) the blue emission at 418 nm decays much faster (3.52 ns) than the blue emission at 485 nm does (4.40 ns) when the excitation wavelength is fixed at 375 nm and (ii) the blue emission decays slower (35.89 ns) at shorter excitation wavelength (255 nm) but faster (3.52 ns) at longer excitation wavelength (375 nm) when the emission wavelength is fixed at 418 nm. These PL decay characteristics are indicative of the nature of defect emissions of undoped SrSO4 nanoplates. These data have demonstrated the triple roles played by oxygen vacancies in undoped SrSO4: (i) they work as the luminescence center of PL to be responsible for the blue emissions; (ii) they work as the luminescence center of afterglow to account for the cyan-colored afterglow; and (iii) they act as the carrier traps to be responsible for the accumulation of charge carriers at room temperature. Our findings can not only provide a comprehensive understanding on the PL and afterglow of undoped SrSO4 but also present an alternative strategy toward designing new afterglow materials through the examination of undoped inorganic materials.

Experimental Section

Synthesis of SrSO4 Nanoplates

SrSO4 nanoplates were synthesized via the composite hydroxide-mediated approach.[16,17] All chemicals used in this work were provided by Sinopharm Chemical Reagents Ltd (Shanghai, China). Analytical reagents NaOH (0.515 mol) and KOH (0.485 mol) were thoroughly mixed in a clean beaker to form the composite hydroxides in order to form eutectic melts at 165 °C and above. Analytical reagents Sr(NO3)2 and Na2SO4 were used as the starting materials to form SrSO4 nanoplates. Sr(NO3)2 (0.1 mol) and Na2SO4 (0.1 mol) were mixed into the composite hydroxides. The mixture was filled into a Teflon vessel (with a capacity of 100 mL). After covering with a Teflon lid, the Teflon vessel was put in a furnace for chemical reactions. The temperature in the furnace was set at 240 °C. After reacting for 24 h, the vessel was taken out of the furnace. The solid product in the vessel was dissolved in deionized water. The SrCO3 impurity in the solids was removed by washing with dilute hydrochloric acid. After repeated washing with deionized water to remove the composite hydroxides on the surface of the particles, the solids were filtered and dried in an oven overnight.

Phase, Morphology, Elemental Composition, and Chemical State of SrSO4 Nanoplates

The X-ray diffraction (XRD) profiles of undoped SrSO4 nanoplates were recorded on an X-ray diffractometer (D/max 2500 PC, Rigaku Corporation, Akishima, Japan) using Cu Kα radiation (λ = 0.15405 nm). A scanning electron microscope (model S-4800, Hitachi, Tokyo, Japan) was employed to analyze the morphology of the synthesized products. The scanning electron microscope was coupled with a silicon-drifted detector as the X-ray analyzer to record the energy-dispersive X-ray (EDX) spectrum of the synthesized products. X-ray photoelectron spectroscopy (XPS) analysis was carried out on an Escalab 250Xi spectrophotometer (Thermo Scientific, Waltham, MA, USA). The incident X-ray came from Al Kα radiation with an energy of 1486.6 eV.

Steady-State PL and Time-Resolved PL Spectra of Undoped SrSO4

A spectrophotometer (Tianjin Gangdong Ltd., Tianjin, China) was used to acquire the steady-state PL spectra of undoped SrSO4. The excitation source of the PL measurement was provided by a helium–cadmium laser (Kimmon Electric Co. Ltd., Tokyo, Japan). The wavelength of laser radiation was 325 nm, and the output power of the laser radiation was 13 mW. The time-resolved PL spectra of the SrSO4 nanoplates were obtained at room temperature on a picosecond fluorescence lifetime spectrometer (LifeSpec II, Edinburgh Instruments, Edinburgh, UK) utilizing a time-correlated single-photon counting method with a pulsed light source. Two picosecond pulsed light-emitting diodes and one pulsed diode laser were employed in this work to provide emission wavelengths of 255, 320, and 375 nm with an average power of 0.3 μW, 0.8 μW, and 0.1 mW, respectively. Their pulse widths at an operating repetition frequency of 5 MHz were 900, 700, and 90 ps, correspondingly. Details on the time-resolved PL analyses could be found elsewhere.[43,44] All the measurements were carried out at room temperature.

ESR Measurement of Undoped SrSO4

The X-band ESR spectrum was measured at room temperature using an X-band ESR spectrometer (Bruker E500) with a 100 kHz magnetic field modulator. The settings of the ESR spectrometer were as follows: the center field was 3515.5 G and the sweep width was 100 G. The resonance frequency of its cavity was 9.856 GHz. The modulation frequency of its receiver is 100 Hz, while the modulation amplitude of the receiver is 2 G.

Afterglow Spectrum and Thermoluminescence Glow Curve of Undoped SrSO4

The afterglow spectrum of undoped SrSO4 was recorded with the same PL spectrophotometer (Tianjin Gangdong Ltd., Tianjin, China) after the laser irradiation was blocked off. The thermoluminescence glow curves of undoped SrSO4 were measured on a thermoluminescence meter constructed according to the scheme given by Yamashita et al.[45] Prior to the thermoluminescence measurements, undoped SrSO4 nanoplates were exposed to the irradiation of ultraviolet light of 254 nm for 10 min. The thermoluminescence signals of undoped SrSO4 nanoplates were recorded when SrSO4 was heated from 10 to 200 °C at a rate of 2 °C/s.

Electronic Structure Calculations of SrSO4 Nanoplates

First-principles density functional calculations of the band structures and the densities of states of SrSO4 were performed using the density functional theory module of the Quantumwise Atomistix ToolKit 11.8 (Atomistix ToolKit 11.8 package, Copenhagen, Denmark). The exchange–correlation functional was treated within the generalized gradient approximation (GGA) + U scheme by the Perdew–Burke–Ernzerhof potential.[46] The U parameter was selected as U2p = 4 eV for oxygen. Orthorhombic SrSO4 belonged to the space group Pnma (62). The unit cell of SrSO4 consisted of 4 Sr atoms, 4 S atoms, and 16 O atoms. The initial structural data of SrSO4 were taken from the Inorganic Crystal Structure Database (ICSD) with the ICSD number of 85808. The lattice parameters of a = 0.8359 nm, b = 0.5351 nm, and c = 0.6869 nm were used in the present work. The considered electronic configurations were 3d104p65s2 for Sr, 2s22p4 for O, and 3s23p4 for S. A 2 × 2 × 1 supercell was constructed for the oxygen-deficient SrSO4. Such a supercell consisted of 16 Sr sites, 16 S sites, and 64 O sites. When one oxygen site was vacant, oxygen-deficient SrSO4 resulted. The resultant SrSO4 was denoted as SrSO4−δ in this work, where δ = 0.0625. Similarly, Sr-deficient SrSO4 resulted after the removal of one strontium site from the supercell. The resultant SrSO4 was denoted as Sr1−δSO4 in this work, where δ = 0.0625. Double-zeta single-polarized basis sets were chosen for each element. The electronic wave functions were expanded in plane waves up to a typical kinetic energy cutoff value of 125 Hartree. The Monkhorst–Pack scheme k-point grid sampling was set at 5 × 5 × 5 for the Brillouin zone. The Brillouin zone sampling and the kinetic energy cutoff were sufficient to guarantee an excellent convergence for the calculated band structures. Details on the density functional calculations were available elsewhere.[12−14,47]