YANG Ting,JIANG Hongyi,DONG Ye,GAO Shuai,LIU Shuning

(School of Materials Science and Engineering, Wuhan University of Technology, Wuhan 430070, China)

Abstract: The low-valence cations Na+ and Sr2+ were selected as the co-dopants to increase the vacancies concentration in the Y2.982Ce0.018Al2Ga3O12 phosphor.The successful incorporation of Na+ and Sr2+ was confirmed by the X-Ray Diffraction (XRD) results.All the samples show 5d-4f green persistent luminescence of Ce3+ after 450 nm excitation.The decay curves demonstrate that the persistent luminescence is effectively enhanced with Na+ and Sr2+ doping.The thermoluminescence glow curves also show not only does the trap concentration increase,but also the distribution of trap depths is broadened.In addition,the air-and H2/Ar-annealing treatments were conducted on every as-made sample.The experimental results prove that the increased traps after the Na+/Sr2+ doping are mainly attributed to the oxygen vacancies,and the traps have a continuous and broad distribution of trap depths.We hope this work could give new inspiration for designing a high-performance persistent phosphor.

Key words: persistent luminescence;YAGG;Na+;Sr2+;oxygen vacancy;charge compensation;solidstate reaction

1 Introduction

Persistent luminescence (PersL) refers to a phenomenon that a phosphor shows continuous luminescence for a long duration from several minutes to hours after ceasing optical excitation.Since Matsuzawaet al[1]reported the pioneering work about the SrAl2O4:Eu2+,Dy3+phosphor in 1996 and proposed a mechanism to explain the persistent luminescence,numerous persistent phosphors have been discovered in various host compounds with different luminescent centers,and several different PersL mechanisms have been reported.However,there is still no well-developed theory that could explain all the behavior of these persistent phosphors,and detailed studies are still required to supplement the mechanism.Usually,the PersL phenomenon originates from the delayed recombination of the excited carriers that are stored by the trapping centers.Hence,selecting proper traps are necessary to design the best-performing persistent material.The carrier traps in phosphor usually include deliberately introduced dopants (eg,Dy3+,Cr3+,and Yb3+)and intrinsic defects (eg,antisite defects,vacancies).Among them,the intrinsic defects have received less attention because they are not as easy to control as the introduced dopants.However,studies have shown that they have a great impact on the PersL performance,the most typical being oxygen vacancy (VO)[2,3].

Aitasaloet al[4]considered theVOas an important carrier trap in their electron model.Clabauet al[5]reported that the RE3+co-doping would not change the chemical nature of the trap in the SrAl2O4:Eu2+phosphor,and they ascribed the trap to theVOdefect in the SrAl2O4:Eu2+,RE3+.Holsaet alalso considered that theVOplays an important role in the outstanding PersL of Sr2MgSi2O7:Eu2+,Dy3+[6],and they identified theVOtrap in CaAl2O4:Eu2+,R3+by electron paramagnetic resonance (EPR) measurements[7].We reported the important effect ofVOon the PersL of Y3Al2Ga3O12:Ce3+,Yb3+(YAGG:Ce,Yb)[8]and manipulated theVOconcentration through annealing at various temperatures in different atmospheres[9].In these works,we considered that the reduction from Yb3+to Yb2+generates an electron at the original site,which is favorable for the formation of the positive chargedVOunder the charge compensation effect,resulting in the enhancement of PersL.The defect equilibrium equation of theVOformation is expressed as

Regarding the charge compensation effect,Aitasaloet alalso reported that the aliovalent substitution leads to the spontaneous creation of defects for charge compensation[4].On these bases,a reasonable conjecture is generated that introducing the cation (R+,R2+)whose valence is lower than the trivalent yttrium Y3+into the YAGG system should be conducive to generating large numbers ofVOand stabilizing theVOtraps,due to their compensation effect on theVO.Thus,the low-valence state cation could enhance the PersL performance,even though the cations themselves cannot act as the proper traps in phosphors.

To verify this conjecture,Na+and Sr2+ions are selected as the low-valence co-dopants and incorporated into the YAGG:Ce system individually.a series of Y3-0.018-xCe0.018RxAl2Ga3O12(whereRis the Na+or Sr2+,andxranges from 0 to 0.014) were successfully synthesized by the high-temperature solid-state reaction method.Subsequently,all the as-synthesized phosphors were annealed in oxidizing and reducing atmospheres to observe the concentration change inVOand discuss the correlation between theVOand the traps.The photoluminescence (PL),PersL,and thermoluminescence(TL) spectra of all the samples are recorded.EPR spectroscopy is a useful method for studying materials with unpaired electrons,and therefore it was performed to measure theVOconcentration.It has been detailedly discussed about the promoting effects of the Na+and Sr2+ions on the formation ofVOconcentration and their enhancement on the PersL performance of YAGG:Ce phosphor.Our conjecture is also verified by the results.

2 Experimental

2.1 Material preparation

Synthesis of YAGG:Ce,(Na/Sr) phosphors.The phosphors Y3-0.018-xCe0.018NaxAl2Ga3O12and Y3-0.018-xCe0.018SrxAl2Ga3O12(x=0,0.002,0.006,0.01,and 0.014) were synthesized by solid-state reactions.The chemicals of Y2O3(4N),Al2O3(4N),Ga2O3(4N),CeO2(4N),Na2CO3(4N),and SrCO3(4N) were used as the raw materials.All of them were purchased from Shanghai Macklin Biochemical Co.,Ltd.5wt% H3BO3(AR,Sinopharm Chemical Reagent Co.,Ltd) was added as a fluxing agent to improve the crystallinity.The powder with stoichiometric compositions was mixed thoroughly for 12 h by a ball milling system with ethanol.After the mixing,the obtained slurry was dried at 70 ℃ for 6 h.Then,the dried mixture was sintered at 1 600 ℃ for 2 h under a CO2atmosphere.After cooling to room temperature,the obtained bulk was crushed and ground into fine yellowish powder.Two equal amounts of each as-made sample were weighed out,waiting for the following post-treatment.

Thermal treatment.The two as-made (CO2-sintered) samples with an equal amount were further annealed at 1 000 ℃ for 2 h in 10% H2/Ar and air atmosphere,respectively.

2.2 Characterization method

The crystalline phases of each sample were identified using an X-ray diffractometer (D/MAX-RB,RIGAKU,Japan) with a non-monochromated Cu Kα X-Ray,and the scanning speed was 0.5 degree/min.Photoluminescence excitation (PLE),PL,and PersL spectra were measured by a fluorescence spectrometer(F-380,Gangdong,China).The PersL spectra were excited by 450 nm for 300 seconds and recorded after 1 min as the excitation ceased.The PersL decay curves(mcd/m2) were detected by a long afterglow material optical test system (CY-1000B,EVERFINE,China),and all the samples were irradiated by a xenon lamp(1000 ± 5 lx) for 5 min,and then the measurements commenced after 1 s as soon as the excitation stopped.The PersL duration time was counted until the intensity reaches 0.32 mcd/m2.The TL curves were measured with a TL instrument (TOSL-3DS,Guangzhou Radiation Technology,China).Before each test,the sample is first heated to 300 ℃ to thoroughly bleach all the carriers,and then it is charged by 254 nm from a UV lamp for 1min.Subsequently,the TL measurement operated and the temperature increased at a linear heating rate of 5 ℃/s up to 300 ℃.The EPR spectra of the air-and H2/Ar-samples were measured using an EPR spectrometer(A300,Bruker,Germany) operating in the X-band near 9.48 GHz at a temperature of 77 K.

3 Results and discussion

3.1 Structural characterization and photoluminescence spectra

X-ray diffraction (XRD) measurements were performed to identify the crystalline phases of each asmade sample.The results are presented in Fig.1.The Bragg angles and the relative intensities of the diffraction peaks are well indexed by the standard cubic (Ia3d)Y3Al2Ga3O12phase (ICSD# 280107) data.The patterns do not show obvious characteristic diffraction peaks of impurity phases suggesting that both the Na+and the Sr2+are successfully incorporated into and homogeneously distributed in their respective lattices.

Fig.1 XRD patterns of (a) Y2.982-xCe0.018NaxAl2Ga3O12 and (b) Y2.982-xCe0.018SrxAl2Ga3O12 with different doping concentrations (x=0.002,0.006,0.01,and 0.014)

Figs.2(a) and 2(b) depict the room-temperature PL spectra (excited by 450 nm) and the PLE spectra(monitoring at 510 nm) of Y2.982Ce0.018Al2Ga3O12with different concentrations of Na+and Sr2+.The PLE spectra present two broad bands at around 350 and 450 nm,which is ascribed to the transition of Ce3+ions from the ground 4f level to the second-lowest 5d level (5d2) and the lowest level (5d1),respectively.Under 450 nm excitation,every sample exhibits the same broad emission band ranging from 470 to 650 nm with a peak at around 510 nm,which is attributed to the 5d1-4f transition of Ce3+ions.Fig.2 demonstrates that the introduction of Na+and Sr2+shows little or no effect on the profile of the PLE and PL bands,suggesting that the luminescent center in YAGG has not changed,remaining as the Ce3+ions,and no other luminescent defect appears after incorporating the Na+and Sr2+.Interestingly,the positions of the emission peaks shift toward longer wavelengths in the presence of Na+or Sr2+,which is attributable primarily to the high sensitivity of the Ce3+emission wavelength to the crystallographic environment.It is known that the splitting of the Ce3+5d excitation state is determined by the local crystal field surrounding the Ce3+ions.Na+and Sr2+ions typically tend to substitute the Y3+ions in dodecahedral sites,which are also the location of the luminescent center Ce3+,because of their similar radii.However,both the ionic radius of Na+(1.18 Å) and Sr2+(1.26 Å) are larger than the Y3+(1.019 Å)[10]in 8-fold coordination.Their substitutions for Y3+induce the expansion of unit-cell and the lattice strain in YAGG,which increase the crystal field strength surrounding Ce3+ions[8,11].Thus,the crystal field splitting of the Ce3+increases,resulting in the following redshift in Fig.2.Finally,it should be mentioned that there are increases in both the emission intensities after the Na or the Sr doping,even though the increases are small,which implies that there is no detectable energy transfer between the Ce3+and the Na+/Sr2+[12],and the Na and Sr do not act as the traps in YAGG phosphors.The intensity increases are probably the result of the decrease in the Ce4+concentration.The low valence incorporation is usually accompanied by negative charges,which would inhibit the oxidation of Ce3+and increase the concentration of luminescent centers.This inference will be detailedly discussed in the following heating-treatment section.

Fig.2 PL (excited by 450 nm) and PLE (monitoring 510 nm) spectra of (a) Y2.982-xCe0.018NaxAl2Ga3O12 and (b) Y2.982-xCe0.018SrxAl2Ga3O12 samples with different doping concentrations (x=0-0.014).The insets in (a) and (b) show their corresponding intensities of the PL peaks

3.2 Persistent luminescence spectra

To characterize the effect of Na+and Sr2+on the persistent phosphor performance of the YAGG:Ce3+,the PersL decay measurements are carried out after artificial light irradiation for 5 min.The PersL decay curves of Y2.982Ce0.018Al2Ga3O12with different concentrations of Na+and Sr2+are recorded and presented in Figs.3(a) and 3(b).As expected,the low-valence cations incorporating can effectively lengthen the decay duration of the YAGG:Ce,as a result of the large generation of theVOtraps.Similar to the duration times,the initial PersL brightness,as shown in the insets,is remarkably enhanced with the Na and Sr doping.The sample when the Sr reachesx=0.01 is almost ten times the intensity of the singly-doped one.It indicates that the newly created traps are mainly shallow types.The PersL durations do not increase monotonically with the Na+or Sr2+concentration increasing.The decay time and initial brightness decrease when the substitution ratio exceeds 0.01,which is presumably due to concentration quenching.Energy can be transferred far from a luminescent center and finally dissipated by jumping to a quenching center in the form of heat.The transfer is ineffective at low concentrations due to the large average distance between the luminescent center and the quenching center.However,the average distance shortens as the Na+or Sr2+concentration increases,which makes it easier for the transfer and thus increases the probability of non-radiative transitions.

The Na and the Sr themselves cannot serve as the proper electron traps in YAGG:Ce,but as mentioned above,their substitutions in the YAGG lattice are generally accompanied by negative net charges.These negative charges need to be compensated by the point defects with opposite charges,i e,theVO,to maintain charge neutrality[2].Moreover,it has been reported that VOpossesses a continuous and broad distribution of trap depths in YAGG:Ce and has an important effect on the PersL[8,9],which agrees well with the increase in Fig.3.Therefore,the increases are probably attributed to the increases inVOconcentration.The PersL spectra are measured at 1 min after 450 nm light charging for 300 s by the above fluorescence spectrometer,and the results are shown in Figs.3(c) and 3(d).The variation trends of the PersL intensity with the Na+/Sr2+doping concentration are found to be similar to that of the decay durations.This similarity confirms the reliability of the two results as they are obtained from different measurements.Figs.3(c) and 3(d) show the band shape and wavelength of all the PersL curves are generally in good agreement with that of the above PL curves,exhibiting a single broad band at around 520 nm.It indicates that the PersL at 1 min originates from the 5d-4f transition of Ce3+.Namely,the Ce ions remain as the luminescent centers in the YAGG during the persistent luminescence process.

3.3 Thermoluminescence curves

Trap parameters,such as trap depth and density,determine the final PersL performance of phosphors.TL measurement is an effective tool to obtain the trap parameters and reveal the characteristic of the trap.Thus,the TL spectra of the Y3-0.018-xCe0.018NaxAl2Ga3O12and Y3-0.018-xCe0.018SrxAl2Ga3O12with different doping concentrations are measured and shown in Fig.4.Before the excitation,all the samples were first heated to 300℃ to completely bleach all the carriers in traps.After the bleaching,they are exposed to UV illumination for 1 min,followed by the formal heating process at a 5 K/s heating rate.As shown in Fig.4,the curves of the singly-doped sample without the Na+/Sr2+present a strong main peak centered at 367 K ranging from 303 to 433 K,and note that there exists a much weak peak at 463 K which overlaps with the high-temperature tail of the main peak.Correspondingly,the co-doped curves have a similar profile and show a closely overlapping broad band ranging from 303 to 573 K.The variation trends of the TL intensity with the Na+and the Sr2+doping concentration are in accordance with that of the decay durations.Generally,the Na+and the Sr2+have the same effect on the TL performance,and both of them can effectively enhance the overall intensity of the TL curve.The singly-doped sample already has a clear TL peak,which is probably the intrinsic trapVO.The further increase in the concentrations and depths with the Na/Sr doping is obviously on the basis of this peak.There is an obvious increase in the high-temperature region.Not only does the previous weak peak at 463 K gain a significant enhancement,but new deeper traps emerge from the region with higher temperatures ranging from 500 to our testing limit 573 K.Furthermore,as shown in Fig.4(b),the traps of the Sr-doped samples seem to even have a distribution in the region higher than 573 K.The increase in the medium-high-temperature region corresponds to the prolongation of the PersL duration,while the increase in low temperature corresponds to the enhancement of the initial brightness.The main peak at 367 K seems to shift towards low temperature,but in fact,it is because the new trap appears in the region where the temperature is lower than 367 K,and the newly formed peak gains a drastic increase in TL intensity and covers the previous main peak.In general,the incorporation of the Na+and the Sr2+can lead to the generation of new traps in YAGG:Ce and effectively enhance its TL intensity.Furthermore,we will demonstrate in the next section that the new traps originate from oxygen vacancy defects and can be eliminated in the presence of Na+or Sr2+.

Fig.4 Thermoluminescence glow curves of (a) Y2.982-xCe0.018NaxAl2Ga3O12 and (b) Y2.982-xCe0.018SrxAl2Ga3O12 samples with different doping concentrations (x=0-0.014).All the curves were measured at a heating rate of 5 K/s after UV charging for 1 min

3.4 Thermal treatment

To ascertain the property of the trap in the YAGG:Ce,Na/Sr and investigate the effect ofVOon the optical properties of YAGG:Ce,we select the Y2.972Ce0.018Na0.01Al2Ga3O12and Y2.972Ce0.018Sr0.01Al2Ga3O12with the best performance in PersL and TL and conduct the annealing treatments under air and 10% H2/90% Ar atmospheres for them.The singly-doped sample Y2.982Ce0.018Al2Ga3O12is also annealed as a control.The characterization ofVOconcentration and optical properties have been carried out on each sample,and the results of Y2.972Ce0.018Al2Ga3O12,Y2.982Ce0.018Na0.01Al2Ga3O12,and Y2.972Ce0.018Sr0.01Al2Ga3O12are presented in Figs.5-8.

3.4.1 EPR analysis

Characterizing theVOconcentration in materials is crucial to clarify the relationship betweenVOand PersL.EPR spectroscopy is a useful method for studying materials that have unpaired electrons,which is in coincidence with the property of oxygen vacant sites.Hence,EPR spectroscopy is used to measure theVOconcentration qualitatively.Only the curves between the magnetic field values of 3 480 G and 3 520 G are chosen to show because that is where theVOsignal is most likely to appear.The results are shown in Fig.5.The vertical axis represents the signal intensity of unpaired electrons.There is only one peak in Fig.5,and all the shapes and locations of the peaks are identical to each other,except for the differences in intensity.Their location corresponds to a central value of the magnetic field values between the peaks and valleys,which is 3 499.853 G.This value is used to calculate thegfactor ofVOusing the following equation[13]:

wherehis the Planck constant,vis the microwave frequency in X-Band and its value is 9.802 GHz,βis the Bohr magneton,andHdenotes the magnetic field.The g factor value at 3499.853 G is calculated to be 2.003 by Eq.(1),which is identified as the electron trapped in theVOaccording to the Refs.[14,15].The intensity of this signal is proportional to the concentration ofVO.Fig.5 show that the signal intensity atg=2.003 experience a significant decrease after the air annealing,while a slight but obvious increase after the H2/Ar annealing.This variation is in line with the expectation because theVOcan be removed from the system under an oxygen-rich environment and introduced under oxygen-deficient environments.Under the air atmosphere,the atmospheric oxygen atoms at the surface of phosphor particles diffuse deeply inside the grains at high temperatures,leading to a fast outward migration of oxygen vacancies and a drastic decrease in theVOconcentration.Note that every sample with the Na/Sr doping has a higher intensity than the corresponding singly-doped one,which strongly confirms our inference that the presence of Na/Sr in the YAGG lattice can greatly contribute to theVOgeneration.The formation ofVOwould disturb the lattice,resulting in the lattice strain.Moreover,the oxygen vacancies are often accompanied by positive charges (or) relative to the regular lattice,which brings the local charge imbalance.Thus,a high concentration ofVOis inhibited by the requirement of structural integrity and charge neutrality and is difficult to reach under normal conditions without codoping.When YAGG material is doped with Na+and Sr2+,they will substitute the Y3+in the lattice due to the similar radii.The low valence cations at the original trivalent site are generally accompanied by negative net charges,which could compensate theVOwith opposite charges to maintain charge neutrality in the system.Therefore,the limit of the maximum concentration is relaxed and a higher concentration of theVOin YAGG:Ce is achieved.In addition,the stabilizing effect of Na+and Sr2+onVOcan be observed in the comparison among the air-annealed samples.After the same air annealing,theVOconcentration of the Na and the Sr sample are both higher than the singly-doped one.This means that some vacancies that were supposed to be eliminated under this condition are stabilized by the Na and the Sr.

3.4.2 PL analysis

Fig.6 shows the PLE spectra monitoring 510 nm and the PL spectra excited by 450 nm for the YAGG:Ce,YAGG:Ce,Na,and YAGG:Ce,Sr annealed in air and H2/Ar atmospheres.All of them are recorded at room temperature.It can be observed that all the PLE and PL curves of the air and H2/Ar samples have the same profile as the original samples.It indicates that the luminescent centers are not changed by the air and H2/Ar annealing,which is still the Ce3+ions.However,the PLE and PL intensities are greatly changed.The intensities are less affected by the air annealing,especially in the singly-doped sample.When the original sample is oxidized,theVOconcentration significantly decreases,which means the trapping centers significantly decrease.This will also decrease the energy transfer efficiency from Ce3+toVO.Meanwhile,the elimination of vacancy defects reduces the strain in the YAGG lattice and makes a more ordered crystal structure.These changes should be conducive to the Ce3+luminescence,and therefore the PL intensity should increase.However,in fact,the PL intensities of the air-annealed samples have all decreased to some extent,which is due to the oxidation of the Ce3+ions.Cerium exhibits two valence states in host material: Ce3+and Ce4+,whose equilibrium is strongly influenced by the crystal structure of the host and by the preparation conditions[16].The Ce4+could serve as a quenching center for the Ce3+luminescence because the intense charge transfer absorption of Ce4+in the UV-visible spectral region,coupled with a ligand anion,generally recombine via nonradiative relaxation[17].The air annealing would lead to the oxidation of the Ce3+ions,which means a decrease in the luminescent centers.Therefore,the PL intensity of the air samples decreases.We also notice that compared with the singly-doped sample,the decreases in the Na and the Sr samples after the same air calcination are much larger.This is probably because the excessive negative charges caused by the aliovalent substitution will be compensated by the oxidation of the Ce3+ions after the oxygen vacancies with positive charges are removed.This will lead to a larger decrease than the singly-doped samples in the luminescent centers.Of the singly-doped sample after being heated in air,the increase in the Ce4+concentration is limited,as the singly-doped system does not need to compensate for extra negative charges.

Fig.6 PL (excited by 450 nm) and PLE (monitoring 510 nm) spectra of the (a) Y2.982Ce0.018Al2Ga3O12,(b) Y2.972Ce0.018Na0.01Al2Ga3O12,and (c)Y2.972Ce0.018Sr0.01Al2Ga3O12 annealed in H2/Ar and air atmospheres.The original,air-annealed,and H2/Ar-annealed samples are denoted in grey,red,and blue color,respectively

The changes after H2/Ar calcination are totally different from our expectations.Generally,the proportion of the Ce4+should decline after the reducing heating[16]and the PL should be enhanced,but it turns out to be an abnormal decrease.The main reason for the decrease is not the increase in theVOconcentration,though the excessive concentration of the vacancies might indeed lead to a decrease in emission intensity.Compared with the air samples,the original samples have higherVOconcentrations.If it is the vacancies that cause the luminescence reduction,the intensities of the original samples should be lower than that of the air sample.Furthermore,the intensity variations of the H2/Ar samples in Fig.6 and Fig.6 also illustrate that there is no correlation between the decrease in the PL intensities and theVOconcentration.The reason for the decrease is not yet clear,and future studies are still needed.

3.4.3 PEL analysis

To reveal the effect ofVOon the PersL property,the persistent decay measurements are conducted after 5 min artificial light irradiation.The decay curves of all the samples are recorded and presented in Fig.7.As expected,the PersL performance is seriously weakened by air annealing.After it,overall PersL intensities have a drastic decrease,including the initial brightness and the PersL duration time.First,the decrease is mainly attributed to the elimination of traps rather than the luminescence quenching.Even though the PL quenching caused by the oxidation of Ce3+might have some contributions to the PersL quenching,the decreases in the PL intensities in Fig.7 are not in accordance with the decrease here in magnitude.Furthermore,the TL spectra in the next section will demonstrate that the carrier traps indeed have a drastic decrease in the trap concentration after the air calcination,which is consistent with the decrease in Fig.7.Second,the eliminated traps are mainly ascribed to the oxygen vacancies,because there is no other defect in the YAGG system that can surely be removed by the oxidizing annealing and introduced under the oxygen-deficient atmosphere.Moreover,combining the results in Fig.5 with Fig.7,we found that the variations in theVOconcentration agree well with the changes in the PersL.Finally,it should be mentioned that after the air heating the luminance through the whole decay process has decreased,suggesting that theVOmay have a continuous distribution of trap depths in the YAGG phosphor.

Fig.7 PersL decay curves of the (a) Y2.982Ce0.018Al2Ga3O12,(b) Y2.972Ce0.018Na0.01Al2Ga3O12,and (c) Y2.972Ce0.018Sr0.01Al2Ga3O12 annealed in H2/Ar and air atmospheres after xenon lamp excitation for 5 min.The original,air-annealed,and H2/Ar-annealed samples are denoted in grey,red,and blue color,respectively.The insets show their corresponding initial brightness

Unlike the changes in the PL spectra,the PersL performance can be efficiently enhanced by the H2/Ar annealing,which of course is due to an increase in theVOconcentration.However,the increases are different.For the singly-doped sample,the H2/Ar annealing increases its overall luminance,including the initial brightness and duration time,while for the Na/Sr-doped samples,only the latter parts of the decay curves and the duration times have obvious increases.This is because the concentrations of the shallow traps responsible for the fast decay process are already high enough in the original (CO2) co-doped sample.As mentioned in the EPR section,the low-valence cations doping is beneficial to the existence of the VOunder the charge compensation effect.With the help of the oxygen-deficient atmosphere (CO2) and high temperature,the concentration of theVOrelative to the shallow traps could reach a very high level and approach the limit,as demonstrated in Fig.5.This also means that a further significant increase in these traps is very diffi-cult to achieve under annealing at the same temperature(1 000 ℃) and pressure (room pressure).Thus,we only observe the increase in the slow decay process.As for the original (CO2) singly-doped sample,without the low-valence cation acting as the stabilizer,the concentration of the overallVOis not high enough,so there is still room for enhancement.

3.4.4 TL

Fig.8 shows the TL spectra of the YAGG:Ce,YAGG:Ce,Na,and YAGG:Ce,Sr annealed in air and H2/Ar atmospheres.First,it can be clearly observed that the air annealing could remarkably decrease the trap concentration and weaken the TL intensities.This decrease is consistent with the PersL quenching in Fig.8,confirming that the removal of the traps is the main reason for the luminance decrease in the PersL.Combining the fact that these traps can be easily eliminated through annealing in the oxidizing atmosphere with the decrease in theVOconcentration in Fig.6,we can conclude that the eliminated traps are theVOdefects.The overall TL intensity has decreased after the air annealing,confirming the discussion above that theVOhas a broad and continuous distribution of trap depths.Then,as discussed in Figs.4(a) and 4(b),the TL intensity of the YAGG:Ce has been significantly enhanced the Na/Sr doping leads to the increase in the existing traps and the generation of the new traps.Moreover,Fig.8 shows that after annealing in air,the TL intensity of the co-doped samples decreases to less than that of the singly-doped one,which means the traps originally introduced by the co-doping can be removed by the air annealing.This result confirms the inference mentioned in Fig.4 that the traps introduce by the Na/Sr doping arise from theVOdefects.Last,comparing the three red curves,we find after the air annealing under the same conditions the co-doped samples still have a greater TL intensity than the singly-doped one,and the distribution of trap depths is also wider.This indicates,as we illustrated in the EPR and the PersL sections,that some vacancies that should disappear under 1 000 ℃and room pressure become difficult to be removed after incorporating the Na and the Sr,which again confirms the stabilizing effect of the low-valence cations.

Fig.8 Thermoluminescence glow curves of the (a) Y2.982Ce0.018Al2Ga3O12,(b) Y2.972Ce0.018Na0.01Al2Ga3O12,and (c) Y2.972Ce0.018Sr0.01Al2Ga3O12 annealed in H2/Ar and air atmospheres.All the curves were measured at a heating rate of 5 K/s after UV charging for 1 min.The original,air-annealed,and H2/Ar-annealed samples are denoted in grey,red,and blue color,respectively

In contrast to the air annealing,the H2/Ar annealing gives rise to an increase in the concentration of theVOtraps,resulting in the TL intensity increase.The increases in the H2/Ar-annealed curves are mainly concentrated in the medium-high-temperature region,especially the high-temperature part ranging from 475 to 550 ℃,while the low-temperature part shows a limited increase.Namely,the increased traps are predominantly the deep traps,while the increase in the shallow traps is insignificant,which is in good agreement with the result obtained from the PersL decay curves.Then,though the TL intensity of the singly-doped sample increase after the H2/Ar annealing,it is still lower than the intensities of the H2/Ar-annealed co-doped samples,and the distribution of trap depths is also narrower.This is mainly because the negative net charges brought about by the low-valence cation doping compensate for the positive charges carried by theVO,raising the limit of theVOconcentration in the YAGG:Ce system.

4 Conclusions

The green persistent phosphors Y2.982-xCe0.018(Na/Sr)xAl2Ga3O12have been successfully synthesized.The low-valence cations Na+and Sr2+are chosen as the co-dopants to promote the generation ofVO.The XRD results have shown that these two cations are successfully incorporated into the YAGG lattice.The PL spectra prove that the luminescent center remains as the Ce3+ions with Na and Sr doping,while the PersL decay curves demonstrate the PersL performance has been effectively enhanced,and both the initial brightness and duration time have obvious increases after the doping.The TL spectra indicate that not only has the trap concentration increased,but also the distribution of trap depths has been broadened.In addition,each original sample has been annealed in air and H2/Ar atmospheres to ascertain the property of the trap in the YAGG:Ce,Na/Sr.The results prove that the increased traps after the Na+/Sr2+doping are attributable primarily to the oxygen vacancies,and the traps possess a continuous and broad distribution of trap depths.The effect ofVOon the optical properties of YAGG:Ce and the charge compensation effect have been discussed.Our work shed light on the promoting effect of Na+and Sr2+onVOand theVOrole in the PersL of YAGG:Ce.We believe it could be helpful in understanding the PersL mechanism in other systems and give new inspiration for designing a high-performance persistent phosphor.

Conflict of interest

All authors declare that there are no competing interests.