Photostimulated near-infrared persistentluminescence as a new optical read-outfrom Cr31-doped LiGa5O8Feng Liu1,2, Wuzhao Yan1,2, Yen-Jun Chuang1, Zipeng Zhen3, Jin Xie3 & Zhengwei Pan1,2

1College of Engineering, University of Georgia, Athens, GA 30602, USA, 2Department of Physics and Astronomy, University ofGeorgia, Athens, GA 30602, USA, 3Department of Chemistry, University of Georgia, Athens, GA 30602, USA.

In conventional photostimulable storage phosphors, the optical information written by x-ray or ultravioletirradiation is usually read out as a visible photostimulated luminescence (PSL) signal under the stimulationof a low-energy light with appropriate wavelength. Unlike the transient PSL, here we report a new opticalread-out form, photostimulated persistent luminescence (PSPL) in the near-infrared (NIR), from aCr31-doped LiGa5O8 NIR persistent phosphor exhibiting a super-long NIR persistent luminescence of morethan 1,000 h. An intense PSPL signal peaking at 716 nm can be repeatedly obtained in a period of more than1,000 h when an ultraviolet-light (250–360 nm) pre-irradiated LiGa5O8:Cr31 phosphor is repeatedlystimulated with a visible light or a NIR light. The LiGa5O8:Cr31 phosphor has promising applications inoptical information storage, night-vision surveillance, and in vivo bio-imaging.

When a storage phosphor is exposed to high-energy radiation, such as x-ray or ultraviolet (UV) light, partof the excitation energy is stored in the phosphor by capturing charge carriers (electrons or holes) intraps which are generally lattice defects or impurities1–3. The stored energy can then be liberated by

thermal, optical or other physical stimulations, resulting in stimulated emissions from the emitting centers in thephosphors. Storage phosphors, especially the thermally stimulated storage phosphors whose energy is releasedupon heating (including the thermal energy available at room temperature) based on thermoluminescence4–8 andphotostimulable storage phosphors whose energy is released by low-energy light (photon) illumination based onphotostimulated luminescence (PSL; also called as optically stimulated luminescence, OSL)1,2,9–12, have found aplethora of important applications, e.g., as self-sustained night-vision luminescent materials and in the fields ofdosimetry and x-ray imaging.

The photostimulable storage phosphors exhibiting PSL phenomenon can act as excellent imaging plates foroptical information write-in and read-out. The optical information write-in and read-out processes requireoptically illuminating the phosphor plate twice. The first exposure by an x-ray beam or a UV light ‘‘writes’’ alatent image in the form of trapped electrons on the phosphor plate. The number of trapped electrons isproportional to the amount of radiation absorbed locally and the latent image is optically readable within acertain time frame (usually within 8 h for practical use) after the exposure. The second illumination with a light ofappropriate wavelength, typically with an intense light with low energy (e.g., red He-Ne lasers), ‘‘reads’’ the imagein the form of higher energy visible PSL signal (e.g., violet-blue emission for BaFBr:Eu21 phosphor10,11). Most ofthe trapped electrons are extracted during the read-out process, and the residual electrons remaining on thephosphor plate can be erased through illumination with bright fluorescent light so that the plate can be used again.Therefore, the photostimulable storage phosphors which exhibit the PSL phenomenon can be used as erasableand rewritable optical memory media for many advanced optical storage applications. For example, the best-known and the most commercially successful photostimulable storage phosphor so far is BaFBr:Eu21, which isbeing widely used as the imaging plates in computed radiography10,11.

Here we report a new optical read-out form, photostimulated persistent luminescence (PSPL) in the near-infrared (NIR), from a Cr31-doped LiGa5O8 NIR persistent phosphor13. An Intense PSPL signal with an emissionpeaking at 716 nm can be repeatedly obtained in a period of more than 1,000 h when a UV light (250–360 nm)pre-irradiated LiGa5O8:Cr31 phosphor is repeatedly stimulated by a light between 380–1,000 nm.

ResultsPhotoluminescence and super-long persistent luminescence in the NIR. Cr31-doped LiGa5O8 was studied inthe 1970s but neither NIR persistent luminescence nor NIR PSPL was reported14. Figure 1a shows the normalized

SUBJECT AREAS:CONDENSED-MATTER

PHYSICS

OPTICAL MATERIALS ANDSTRUCTURES

OPTICAL PHYSICS

NANOSCALE MATERIALS

Received6 March 2013

Accepted8 March 2013

Published27 March 2013

Correspondence andrequests for materials

should be addressed toZ.W.P. (panz@uga.

edu)

SCIENTIFIC REPORTS | 3 : 1554 | DOI: 10.1038/srep01554 1

photoluminescence emission and excitation spectra of a LiGa5O8:Cr31 phosphor disc at room temperature. Under excitation at400 nm, the material exhibits a narrow-band emission peaking at716 nm. This NIR emission is characteristic of Cr31 ions and can beattributed to the spin-forbidden 2E R 4A2 transition. The associatedbroad background emission ranging from ,650 nm to ,850 nmoriginates mostly from the phonon sidebands of the 2E R 4A2

transition14. The photoluminescence excitation spectrum monitoredat 716 nm emission covers a very broad spectral region (from ,250to ,660 nm) and consists of three main excitation bands originatingfrom the inner transitions of Cr31, including the 300 nm bandoriginating from the 4A2 R 4T1(4P) transition, the 415 nm bandoriginating from the 4A2 R 4T1(4F) transition, and the 605 nmband originating from the 4A2 R 4T2(4F) transition.

Besides the intense NIR photoluminescence, the excitation of UVlight can also induce very-long-lasting NIR persistent luminescencein LiGa5O8:Cr31 phosphor with a persistence time .1,000 h (herethe persistence time is defined as the duration for which an eye can

see with the aid of a night-vision goggle in a dark room8; seeSupplementary Information). Figure 1b shows the persistent lumin-escence decay curve of a LiGa5O8:Cr31 phosphor disc monitored at716 nm after irradiation by 300 nm UV light for 20 min. Therecording lasted for 120 h. As can be seen, the persistent lumin-escence intensity drops quickly in the first several hours and thendecays slowly. After 120 h of persistent emission, the persistentluminescence intensity is still significantly high, indicating that theNIR persistent luminescence should last much longer than 120 h.The persistent luminescence intensities of the LiGa5O8:Cr31 phos-phor disc at 10 min and 1 h after ceasing the excitation were esti-mated to be ,4.7 mW m22 and ,1.2 mW m22, respectively(Supplementary Information and Supplementary Fig. S1). The upperinset of Fig. 1b shows a persistent luminescence emission spectrumrecorded at 1 h after the stoppage of the 300 nm UV light irradiation.The profile of the persistent luminescence emission spectrum isalmost identical to that of the photoluminescence emission spectrum(Fig. 1a), indicating that the NIR persistent luminescence originates

Figure 1 | Photoluminescence and persistent luminescence of LiGa5O8:Cr31 phosphor discs at room temperature. (a) Normalized excitation and

emission spectra for photoluminescence. The emission spectrum is acquired under 400 nm light excitation and the excitation spectrum is obtained by

monitoring 716 nm emission. (b) NIR persistent luminescence decay curve monitored at 716 nm after irradiation by 300 nm light for 20 min. The upper

inset shows the persistent luminescence emission spectrum recorded at 1 h after the stoppage of the irradiation. The bottom inset is the persistent

luminescence excitation spectrum obtained by plotting the persistent luminescence intensity (I10s) monitored at 716 nm as a function of the excitation

wavelengths over the 250–600 nm spectral range. The disc was irradiated for 5 min at each measured wavelength using a xenon arc lamp. (c–j) NIR

images of four phosphor discs taken at different persistent luminescence times (10 min to 1,080 h) after irradiated by a 254 nm lamp for 10 s to 5 min.

The imaging parameters are: (c–f) manual/ISO 400/10 s, (g–i) manual/ISO 800/30 s, and (j) manual/ISO 1600/30 s.

www.nature.com/scientificreports

SCIENTIFIC REPORTS | 3 : 1554 | DOI: 10.1038/srep01554 2

from the Cr31 emitting centers. However, unlike the NIR photolu-minescence which can be effectively excited by a wide range of wave-lengths (,250–660 nm) (Fig. 1a), the NIR persistent luminescencein LiGa5O8:Cr31 cannot be induced by low-energy visible light irra-diation. To evaluate the effectiveness of different excitation wave-lengths (energies) for persistent luminescence, we measured thepersistent luminescence decay curves monitored at 716 nm afterthe excitation of monochromatic light with different wavelengthsbetween 250–600 nm in 10 nm steps (Supplementary Fig. S2),recorded the persistent luminescence intensity at 10 s after the stop-page of each excitation (I10s), and plotted the intensity I10s as afunction of the excitation wavelengths, as the ball curve shown inthe bottom inset of Fig. 1b. It is clear that the NIR persistent lumin-escence can be effectively achieved by UV irradiation between 250–360 nm (250 nm is the low limit of the xenon emission in theFluoroLog-3 spectrofluorometer), and the effectiveness increases asthe excitation moves to shorter wavelengths.

The very-long-lasting NIR persistent luminescence of theLiGa5O8:Cr31 phosphor discs was also visually evaluated using anight-vision monocular in a dark room. Figure 1c–j shows thechanges of NIR emission ‘‘brightness’’ with a decay time up to1,008 h for four LiGa5O8:Cr31 discs after exposure to a 4 W254 nm UV lamp for 10 s, 30 s, 1 min and 5 min. Figure 1c–j clearlyshows that the LiGa5O8:Cr31 phosphor discs can be effectively acti-vated by the 254 nm UV lamp and seconds to minutes of UV irra-diation can result in 1,008 h of persistent NIR emission. It should benoted that, as the PSPL phenomenon that will be discussed below, theenergy stored in the material during excitation is not completelyreleased even after 1,008 h room-temperature decay. Thus, the actualpersistent luminescence decay time of the LiGa5O8:Cr31 phosphorshould be longer than 1,008 h.

Photochromism. In irradiating the LiGa5O8:Cr31 phosphor discsusing a 254 nm UV lamp in room light environment, we observedthat the body color of the LiGa5O8:Cr31 sample changed fromgreenish to reddish, as the digital pictures shown in Figs. 2a and2b. This UV irradiation induced coloration phenomenon is long-lived at room temperature, lasting for more than one month (thesamples need to be stored in the dark); however, it can be quicklybleached by external stimulations, such as through illumination withbright fluorescent light or through heating at around 400uC. Thiscoloration/bleaching process can be repeatedly carried out withoutleaving any permanent changes to the optical performance of theLiGa5O8:Cr31 samples. The change of body color due to UVirradiation is qualitatively reflected by measuring the diffusereflectance absorption on the samples with and without UVirradiation, as the spectra shown in Fig. 2c. Compared with thenon-irradiated sample (curve 3), the spectra of the UV-irradiatedsample (curves 1 and 2) contain an additional green absorptionband peaking at ,500 nm, as the dash-line curve and dot-dash-line curve shown in Fig. 2c. The presence of the additional greenabsorption band, together with the strong absorption of the materialto blue light, make the UV-irradiated LiGa5O8:Cr31 discs appear tobe reddish in room light condition.

Since the repetitive coloration/bleaching processes do not causepermanent change to the samples, the coloration phenomenon canbe attributed to the formation of photochromic centers15,16, whichmay result from the trapping of photogenerated electrons by thedeep-level lattice defects (such as oxygen vacancies adjacent to thechromium ions) in LiGa5O8:Cr31. Also, since the fading of colorationis accompanied by the decrease of persistent luminescence intensityat room temperature, the deep-level lattice defects used to formthe photochromic centers may also act as deep electron traps res-ponsible for the long persistent luminescence. To further understandthe properties of electron traps and photochromic centers inLiGa5O8:Cr31, we conducted thermoluminescence measurements

on phosphor discs undergoing different delay times, as shown inFig. 3a. The dot-dash-line curve in Fig. 3a shows the thermolumines-cence curve acquired immediately (delay time, 10 s) after the stop-page of 300 nm light irradiation (for 20 min). The curve consistsof two broad bands with maxima at 150uC and 220uC, which corre-spond to the shallow and deep traps, respectively. When the delaytime increases to 120 h (the solid-line curve in Fig. 3a), a majorityof the shallow-trap band disappears and the deep-trap band stillexists, indicating that it is the deep traps that are responsible forthe super-long persistent luminescence at room temperature. Sincethe photo-stimulation method has been frequently used to study thephotochromic centers in storage phosphors2, we then conductedphotostimulated thermoluminescence measurements on the de-cayed LiGa5O8:Cr31 samples. The dash-line curve in Fig. 3a showsthe thermoluminecence curve of a 120 h-decayed LiGa5O8:Cr31 discafter being exposed to 400 nm illumination for 100 s. Comparedwith the 120 h-decayed sample without photostimulation, the400 nm-light stimulation significantly changes the thermolumines-cence curve profile, i.e., the deep-trap band intensity decreases whilethe shallow-trap band reappears. This means that after the 400 nm-light photostimulation, some of the electrons in the deep traps arephoto-released and the emptied shallow traps are refilled. (Note thatthe phenomenon of electron transfer from deep traps to shallowtraps under external stimulation was also recently observed in somemechanoluminescent materials17,18).

Figure 2 | UV-irradiation-induced coloration and changes in diffusereflectance absorption of LiGa5O8:Cr31 phosphor plates. (a) Digital

image of a 15 3 15 mm2 LiGa5O8:Cr31 phosphor plate with its center

covered by an 8 3 8 mm2 black paper in room light environment. Scale

bar, 5 mm. (b) The same plate as the one in (a) after exposed to a 254 nm

UV lamp for 5 min. The paper was removed after the irradiation.

(c) Diffuse reflectance absorption spectra acquired on LiGa5O8:Cr31 plates

with and without UV irradiation. Curve 1 and curve 2 were recorded on a

300 nm-light-irradiated plate (for 20 min) with delay times of 10 s and

120 h, respectively. Curve 3 was acquired on a bleached plate (without UV

pre-irradiation). The dash-line curve is the difference between curve 1 and

curve 3. The dot-dash-line curve is the difference between curve 2 and

curve 3.

www.nature.com/scientificreports

SCIENTIFIC REPORTS | 3 : 1554 | DOI: 10.1038/srep01554 3

Photostimulated persistent luminescence (PSPL). The photosti-mulation induced electron trap redistribution, especially the refillof the shallow traps, suggests that photostimuation can affect thepersistent luminescence behaviors of the UV pre-irradiatedLiGa5O8:Cr31 phosphor. To verify this assumption, we illuminateda 120 h-decayed LiGa5O8:Cr31 disc with 400 nm light for 100 s andmeasured its persistent luminescence decay curve (monitored at716 nm), as the brown curve shown in Fig. 3b. Figure 3b clearlyshows that the 400 nm light illumination increases the persistentluminescence intensity and thus a new PSPL phenomenon occurs.For comparison, we also conducted the same measurement on acompletely bleached (i.e., without UV pre-irradiation) LiGa5O8:Cr31 disc, as the grey curve shown in Fig. 3b. No persistentluminescence was observed in the bleached sample after the400 nm light illumination, which is consistent with the persistentluminescence excitation spectrum shown in the bottom inset ofFig. 1b. The inset of Fig. 3b is the PSPL emission spectrum of the120 h-decayed disc, which was recorded at 10 s after the stoppage ofthe stimulation. The profile of the PSPL emission spectrum is almostidentical to that of the photoluminescence emission spectrum(Fig. 1a) and the persistent luminescence emission spectrum(upper inset of Fig. 1b).

The NIR PSPL phenomenon in LiGa5O8:Cr31 is analogous to thevisible PSL in conventional photostimulable storage phosphors, indi-cating that the LiGa5O8:Cr31 phosphor has the potential to be used asa new type of erasable and rewritable optical memory media foroptical information write-in and read-out. For the measurement inFig. 3b, the write-in and read-out sources are 300 nm UV light and400 nm visible light, respectively. To precisely determine the range ofwrite-in and read-out energies needed for the NIR PSPL inLiGa5O8:Cr31, we plotted PSPL write-in and read-out spectra bymeasuring PSPL decays over a wide range of wavelengths between250–660 nm. In plotting the PSPL write-in spectrum, we irradiated ableached LiGa5O8:Cr31 disc using monochromatic UV light between

250–380 nm in 10 nm steps and recorded the persistent lumin-escence decay curves with and without a 400 nm light stimulation,as shown by the three sets of representative curves in Fig. 4a. Wedefined the PSPL intensity as the difference between the persistentluminescence intensities of each set of decays with and withoutphotostimulation at the time of 10 s after the stoppage of the stimu-lation, as indicated by the vertical double arrowheads in Fig. 4a. ThePSPL write-in spectrum was then obtained by plotting the PSPLintensities as a function of the write-in wavelengths (250–380 nm),as the ball curve shown on the left panel of Fig. 4c. The write-in spectrum reveals that in order to induce NIR PSPL inLiGa5O8:Cr31, the write-in wavelength should be shorter than360 nm and within the measured wavelengths of 250–360 nm theshorter the excitation wavelength the more effective the write-inprocess is. Moreover, by comparing Fig. 4c with the bottom insetof Fig. 1b, it can be found that the PSPL write-in spectrum is identicalin shape to the persistent luminescence excitation spectrum eventhough they were obtained by different methods and their physicalmeanings are different (The PSPL write-in spectrum shows theenergy required to fill the deep traps, i.e., to form the photochromiccenters, while the persistent luminescence excitation spectrumreveals the energy needed to photoionize the localized electrons fromCr31 to the conduction band8). The coincidence of these two spectraclearly indicates that the filling of the deep traps (i.e., the formation ofthe photochromic centers) accompanies the photoionization of Cr31

in LiGa5O8.In acquiring the PSPL read-out spectrum, the write-in wavelength

was fixed at 300 nm, while the read-out wavelengths were tunedbetween 380–660 nm in 10 nm steps. Using the same method asabove, we acquired PSPL decay curves and plotted the PSPL read-out spectrum, as shown in Fig. 4b and on the right panel of Fig. 4c,respectively. The read-out spectrum reveals that the PSPL phenom-enon can be induced by the entire visible spectrum and that theshorter the stimulation wavelength the more effective the read-out

Figure 3 | Thermoluminescence spectra and PSPL decay curves of LiGa5O8:Cr31 phosphor discs. (a) Thermoluminescence curves monitored at

716 nm emission over 20–280uC. The samples were pre-irradiated by 300 nm UV light for 20 min. The dot-dash-line curve and solid-line curve were

acquired at delay times of 10 s and 120 h, respectively. The dash-line curve was acquired on a 120 h-decayed disc after stimulation by 400 nm light for

100 s. (b) PSPL decay curves monitored at 716 nm. The brown curve was acquired on a 120 h-decayed disc (pre-irradiated by 300 nm light for 20 min),

while the grey curve was recorded on a bleached disc (without UV pre-irradiation). The inset is the PSPL emission spectrum of the 120 h-decayed disc,

which was recorded at 10 s after the stoppage of the stimulation. The wavelength of the stimulation light is 400 nm.

www.nature.com/scientificreports

SCIENTIFIC REPORTS | 3 : 1554 | DOI: 10.1038/srep01554 4

process is. Moreover, our read-out experiments further showed thatthe energy stored in the LiGa5O8:Cr31 phosphor can also be read outby NIR light stimulation, such as a 980 nm NIR laser, as the PSPLdecay curve shown in Supplementary Fig. S3. Thus, our read-outexperiments reveal that unlike the visible PSL phenomenon in con-ventional photostimulable phosphors in which the read-out wave-length is longer than the emission wavelength, the stimulationwavelength for the NIR PSPL in LiGa5O8:Cr31 phosphor can beeither shorter or longer than the emission wavelength.

It is worth noting from Fig. 4c that there is no overlap between thePSPL write-in and PSPL read-out spectra in the LiGa5O8:Cr31 phos-phor. This is important because while the visible light is reading theoptical information written by the UV light, it does not write newinformation into the material. In addition, the very broad write-inand read-out spectra allow one to easily find suitable optical sources,such as inexpensive laser diodes and light emitting diodes (LEDs), forthe optical information write-in and read-out processes.

The write-in and read-out capabilities of LiGa5O8:Cr31 phosphorbased on NIR PSPL phenomenon was also visually demonstrated in adeliberate imaging experiment using a night vision monocular. Theimaging experiment was carried out on the square LiGa5O8:Cr31

ceramic plate in Fig. 2a and lasted for 1,008 h. In the write-in process,the 15 3 15 mm plate with its center covered by an 8 3 8 mm squarepaper was exposed to a 254 nm UV lamp for 20 min. Due to

shielding by the square paper, the center of the plate remained un-activated; thus, the shape of the paper was written on the plate (seeFig. 2b). Under the night vision monocular, the UV-irradiatedLiGa5O8:Cr31 plate appeared to be a bright square ring, as shownin the NIR image in Fig. 5a. The stored image faded slowly in the darkat room temperature and after 120 h of decay the image can still beclearly seen (Fig. 5a–d). In the decay period of 120–720 h, the righthalf of the plate was stimulated by a YAG:Ce-based white LED for20 s at every 120 h (the left half was covered by a piece of paperduring the stimulation; the effectiveness of the white LED for thePSPL is given in Supplementary Fig. S4). As expected, only the UVpre-irradiated region on the right half of the plate exhibited enhancedNIR persistent luminescence after each stimulation (Figs. 5e and 5f,and Supplementary Fig. S5a; the PSPL images were taken at 10 s afterceasing the white LED stimulation) and the enhancement lasted formore than 5 h (Fig. 5e1–e3 and Fig. 5f1–f2). In the decay period of744–1,008 h, the entire plate was exposed to the white LED for 20 s atevery 24 h. As a result, the entire UV pre-irradiated region on theplate emitted enhanced NIR persistent luminescence after eachstimulation, as shown in the NIR images in Fig. 5g–k (see moreimages in Supplementary Fig. S5b). Since the left half of the platereceived LED stimulations 6 times fewer than the right half (at 744 h,the left half experienced its first stimulation, while the right halfreceived its seventh), the PSPL in the left half is slightly brighter than

Figure 4 | Optical write-in and read-out spectra for PSPL in LiGa5O8:Cr31. (a) Persistent luminescence decay curves monitored at 716 nm with (colored

solid-line curves) and without (grey dash-line curves) photostimulation. The write-in wavelengths (lwi) are 320 nm, 340 nm and 360 nm. The read-out

wavelength (lro) is 400 nm. (b) Persistent luminescence decay curves monitored at 716 nm with (colored solid-line curves) and without (grey dash-line

curve) photostimulation. The write-in wavelength (lwi) is 300 nm. The read-out wavelengths (lro) vary between 380 nm and 660 nm. For the

measurements in (a) and (b), the read-out process (i.e., stimulation) starts at 5 min after ceasing the write-in process (i.e., UV pre-irradiation). The

vertical double arrowheads in (a) and (b) represent the PSPL intensities. (c), PSPL write-in spectrum (the ball curve in left panel) and read-out spectrum

(the triangle curve in right panel) plotted according to the data from (a) and (b), respectively.

www.nature.com/scientificreports

SCIENTIFIC REPORTS | 3 : 1554 | DOI: 10.1038/srep01554 5

that in the right half (see also Fig. 5g1–g3). Remarkably, Figure 5kreveals that after 1,008 h of decay and 18 and 12 white LED stimula-tions on the right half and left half of the plate, respectively, the storedimage on the LiGa5O8:Cr31 plate can still be clearly read out. (Ourprolonged imaging experiments on other samples further showedthat the stored information can be clearly read out after even muchlonger decay, such as 2,000 h, as shown in Supplementary Fig. S6.).

Figure 5e–k also reveals that the PSPL intensity decreases gradu-ally as the read-out times increase due to the photo-release of elec-trons from the deep traps after each stimulation (the natural decayalso contributes to the decrease of PSPL intensity). The decrease ofthe PSPL intensity can be quantitatively evaluated as pixel intensityby converting each pixel to grayscale using ImageJ software19.Figure 5i–o shows the pixel intensities acquired from Fig. 5e,Fig. 5f, Fig. 5g, and Fig. 5k, respectively. The dependence of the pixel

intensities on the read-out times in Fig. 5i–o is consistent with thevisual observation.

Besides the ceramics, we also fabricated LiGa5O8:Cr31 phosphor inthe form of micro-powders (particle sizes: 2–10 mm) at a lower tem-perature (1,250uC) by the solid-state reaction method. The powdersexhibit similar persistent luminescence and PSPL performances to theceramics. One advantage of the powder-form phosphor is that thepowders can be easily incorporated into paints (for example, transpar-ent acrylic polyurethane vanish) to form NIR persistent luminescentpaints which exhibit very good persistent luminescence and PSPLproperties, as shown in the NIR images in Supplementary Fig. S7.

DiscussionThe above results on PSPL phenomenon in LiGa5O8:Cr31 at roomtemperature indicates that the PSPL write-in process fills the deep

Figure 5 | NIR images for PSPL in LiGa5O8:Cr31 phosphor plate. The plate is the same as the one in Fig. 2a. Before imaging the plate was exposed to a

254 nm UV lamp for 20 min. (a–d) Natural decay to 120 h at room temperature. (e–f) From 120 to 720 h, the right half of the plate was stimulated by a

white LED for 20 s at every 120 h. (e1–e3) and (f1–f2) show the natural decay after each stimulation. (g–k) From 744 to 1,008 h, the entire plate was

stimulated by a white LED for 20 s at every 24 h. (g1–g3) show the natural decay after the stimulation. All the PSPL images were taken at 10 s after ceasing

the white LED stimulation. The numbers on the top of the bright square ring in (e–k) are the times of stimulation (i.e., read-out). The imaging parameter

is manual/ISO 400/10 s. (i–o) Pixel intensity acquired along the white dot-dash lines in (e), (f), (g), and (k), respectively, by converting each pixel to

grayscale using ImageJ software. The number on the top of each peak is the times of stimulation.

www.nature.com/scientificreports

SCIENTIFIC REPORTS | 3 : 1554 | DOI: 10.1038/srep01554 6

traps with electrons (i.e., forms the photochromic centers), while thePSPL read-out process releases the captured electrons from the filleddeep traps (i.e., the photochromic centers) to the conduction band,followed by the refill of the emptied shallow traps. To gain insightinto the interaction between the traps and the Cr31 emitting centers,we conducted extended thermoluminescence measurements startingat liquid nitrogen temperature (2196uC, i.e., 77 K) and decay mea-surement at 77 K on LiGa5O8:Cr31 discs, as shown in Fig. 6a andFig. 6b, respectively. Figure 6a shows two thermoluminescencecurves recorded over 2196–280uC (i.e., 77–553 K) after irradiatingthe samples with 300 nm UV light for 20 min at 77 K (dash-linecurve) and at room temperature (solid-line curve). (For the case ofroom-temperature irradiation, the sample was quickly transferred toliquid nitrogen environment for measurement after the stoppage ofthe irradiation.). In high-temperature region (from room temper-ature to 280uC), the two curves exhibits the same profile, which isidentical in shape to the thermoluminescence curve measured over20–280uC (the dash-line curve in Fig. 3a). In low-temperature region(from 77 K to room temperature), however, the 77 K pre-irradiatedsample shows an additional thermoluminescence band, while theroom-temperature pre-irradiated sample does not. This means thatthere exist low-temperature trap levels in LiGa5O8:Cr31, but theselow-temperature traps are inactive (i.e., emptied) during room tem-perature irradiation due to the thermal energy available at roomtemperature. This trap information indicates that the low-temper-ature traps will not contribute to the persistent luminescence at 77 Kfor the samples irradiated at room temperature.

Figure 6b shows a decay curve measured at 77 K for a LiGa5O8:Cr31 disc pre-irradiated at room temperature. Before immersing thesample into liquid nitrogen for the measurement, the sample wasirradiated by 300 nm light at room temperature for 20 min. Theinterval between the stoppage of the room-temperature irradiation

and the starting of the low-temperature measurement is 120 s. Thisdecay curve was also plotted as a function of reciprocal persistentluminescence intensity versus time, as the linear curve shown in theinset of Fig. 6b. The presence of low-temperature persistent lumin-escence without the contribution from the low-temperature trapsand the linear dependence of the reciprocal persistent luminescenceintensity versus time are characteristics of a quantum tunneling pro-cess20,21, meaning that at 77 K the electrons captured in the deep trapscan directly recombine with the nearby ionized Cr31 ions viaquantum tunneling, instead of going through the conduction band8.This tunneling process, which is temperature independent, proceedsat a slow rate, which on the one hand leads to the super-long-persistent luminescence at room temperature, and on the other handmaintains the electron population in the deep traps for a long timefor the PSPL process.

Based on the above results and discussions, we propose a mech-anism to account for the super-long NIR persistent luminescenceand the new NIR PSPL phenomenon at room temperature inLiGa5O8:Cr31, as schematically shown in Fig. 6c. To simplify thedescription, we assign the shallow traps and the filled deep traps(i.e., the photochromic centers) in the PSPL process as TRAP-1and TRAP-2, respectively. Under UV (250–360 nm) excitation, theground-state electrons of Cr31 ions are photoionized to the conduc-tion band (process 1). The conduction electrons are subsequentlycaptured by TRAP-1 (process 2) and TRAP-2 (process 3). In theinitial stage of the persistent luminescence process, the electronscaptured in TRAP-1 escape thermally via the conduction band (pro-cess 4) and recombine with the ionized Cr31 ions, which dominatesthe initial intense persistent NIR emission. Several hours later, therelease of electrons through the conduction band can be neglecteddue to the depletion of TRAP-1. The NIR persistent luminescencesubsequently originates mainly from TRAP-2 via quantum tunneling

Figure 6 | Low-temperature thermoluminescence and persistent luminescence measurements and NIR PSPL mechanisms in LiGa5O8:Cr31.(a) Thermoluminescence curves recorded by monitoring at 716 nm emission over 2196–280uC (77–553 K) on a LiGa5O8:Cr31 disc. The dash-line curve

and solid-line curve were recorded after 300 nm UV light irradiations for 20 min at 77 K and at room temperature, respectively. (b) Persistent

luminescence intensity (I) monitored at 716 nm as a function of time (t) at 77 K for a LiGa5O8:Cr31 disc pre-irradiated at room temperature. Inset shows

the same data plotted as I21 versus t. (c) Schematic representation of the NIR persistent luminescence and photostimulated NIR persistent luminescence

mechanisms. The straight-line arrows and curved-line arrows represent optical transition and electron transfer processes (see text for further detail),

respectively.

www.nature.com/scientificreports

SCIENTIFIC REPORTS | 3 : 1554 | DOI: 10.1038/srep01554 7

(process 5), giving weak but very-long NIR persistent luminescence.Under the stimulation of a visible or NIR light for a short time atroom temperature, some of the electrons in TRAP-2 can be photo-released to the conduction band (process 6) and some of which refillthe depleted TRAP-1 (process 7), resulting in enhanced persistentluminescence (i.e., the PSPL phenomenon). Since only some of theelectrons in TRAP-2 can be photo-released after photostimulation(owing to short-time, moderate-intensity light stimulation, e.g., a85-lumen white LED for 20 s in the present study), the PSPL phe-nomenon can thus occur many times, as the images shown in Fig 5.

The above results show that we have synthesized a novelLiGa5O8:Cr31 NIR phosphor that can function as both a self-sus-tained NIR persistent material exhibiting a super-long (.1,000 h)NIR persistent luminescence and a new type of photostimulablestorage material exhibiting repeatedly photostimulated NIR persist-ent luminescence. This NIR phosphor can be used as a promisingstorage medium for long-time optical information storage and read-out. The material is also expected to be used as invisible (to nakedeyes) markers or signage in defense and security8 and as opticalprobes in medical diagnosis22. Especially, the LiGa5O8:Cr31 phos-phor in the form of nanoparticles is very suitable for in vivo bio-imaging because it can avoid some critical inherent problems (e.g.,autofluorescence, poor signal-to-noise ratio, and shallow penetrationdepth) encountered in conventional optical imaging23–25 owing to thefollowing three key characteristics. Firstly, the material emits intense,long-lasting NIR persistent light after the removal of the excitation.This allows the nanoparticles to be tracked in vivo without externalexcitation, permitting the complete removal of the autofluorescenceand hence the background noise originating from in situ excitation.The resulting imaging is bestowed with a significantly improvedsignal-to-noise ratio, allowing for detection in rather deep-tissueswith high sensitivity. Secondly, the emission (peaking at 716 nm)is in the tissue transparency window (i.e., in 700–900 nm range)25, inwhich light attenuation is largely due to scattering rather thanabsorption. This can further increase the detection depth. Thirdly,the UV pre-irradiated samples can be repeatedly stimulated by long-wavelength light (e.g., white LED flashlight), which allows forlongitudinal (days or even weeks) monitoring/tracking of theLiGa5O8:Cr31 labeled cells in vivo.

To demonstrate the potential of LiGa5O8:Cr31 in in vivo bio-imaging, we prepared LiGa5O8:Cr31 nanoparticles with diametersof 50–150 nm by a sol-gel method followed by high-temperaturecalcination (Supplementary Fig. S8). The LiGa5O8:Cr31 nanoparti-cles were coated with polyethylenimine (PEI), which were then usedto label 4T1 murine breast cancer cells. The PEI-LiGa5O8:Cr31 nano-particles labeled 4T1 cells were illuminated with a 254 nm UV lamp,and subcutaneously injected into the back of a nude mouse. Theimaging experiment was performed on an IVIS Lumina II imagingsystem in the bioluminescence mode for a period of 10 days, asshown in Fig. 7 (the methods used for nanoparticle synthesis,PEI coating, 4T1 cell labeling, and imaging are described inSupplementary Information). The NIR persistent luminescencefrom the LiGa5O8:Cr31 nanoparticles can be clearly imaged even4 h after the injection (Fig. 7a). As expected, the NIR persistentluminescence can be repeatedly rejuvenated in vivo by stimulationwith a white LED flashlight (Figs. 7a1–e1), and the resulting PSPLsignals can last for more than 5 min (Figs. 7a2–e2). Such repeatedNIR PSPL signals permit a wide tracking window and the signalsfrom the LiGa5O8:Cr31 nanoparticles can be clearly detected even 10days after the injection (Fig. 7e1 and 7e2). Similar performance wasalso observed in a phantom study with PEI-LiGa5O8:Cr31 nanopar-ticles labeled 4T1 cells, as shown in Supplementary Fig. S9.

MethodsMaterials synthesis. LiGa5O8:Cr31 phosphors in the forms of solid discs and micro-scale powders were synthesized by a solid-state reaction method. Stoichiometricamounts of Li2CO3, Ga2O3 and Cr2O3 powders were ground to form a homogeneous

fine powder (the Cr content in the LiGa5O8:Cr31 phosphor demonstrated here is 1atom%). The mixed powder was then prefired at 800uC in air for 2 h. The prefiredmaterial was again ground to fine powder suitable for sintering. Part of the prefiredpowder was pressed into discs with diameters of 15 mm and 40 mm using a 16-tonhydraulic press. The discs were then sintered at 1,300uC in air for 2 h to form a solidceramic, and the powder samples were sintered at 1,250uC in air for 2 h. The 15-mm-diameter discs were used as-is. Some 40-mm-diameter discs were cut into 15 3

15 mm square plates for deliberate imaging experiments.LiGa5O8:Cr31 phosphor in the forms of nanoparticles was synthesized by a sol-gel

method followed by calcination at high temperature. The detailed fabrication pro-cedure is described in Supplementary Information.

Characterization methods. The spectral properties (excitation and emission spectra,decay curves, persistent luminescence emission and excitation spectra, and PSPLwrite-in and read-out spectra) of the LiGa5O8:Cr31 discs were measured using a

Figure 7 | Images of PEI-LiGa5O8:Cr31 nanoparticles labeled 4T1 cellsinjected into a nude mouse for a 10-day tracking using an IVIS imagingsystem. The PEI-LiGa5O8:Cr31 nanoparticles labeled 4T1 cells (,2.5 3

107 cells) were illuminated by a 4-W 254 nm UV lamp for 15 min, and

then subcutaneously injected into the back of a nude mouse. (a) Image

taken at 4 h after cell injection. To get the PSPL signal, the mouse was

exposed to a white LED flashlight (Olight SR51, 900 lumens) for 15 s. (a1)

and (a2), Images taken at 10 s and 5 min after the stimulation, respectively.

The signals were attributed to PSPL. (b–e) The mouse was exposed daily to

the LED flashlight for 15 s, and images were taken at 10 s and 5 min after

the stimulation. All images were acquired in the bioluminescence mode

with an exposure time of 2 min. The images were processed using

Living Image software at binning of 4 and smooth of 5 3 5. The color scale

bar represents the luminescence intensity in the unit of radiance,

p/sec/cm2/sr.

www.nature.com/scientificreports

SCIENTIFIC REPORTS | 3 : 1554 | DOI: 10.1038/srep01554 8

Horiba FluoroLog-3 spectrofluorometer equipped with a 450 W xenon arc lamp anda R928P photomultiplier tube (250–850 nm). All spectra were corrected for theoptical system responses. Appropriate optical filters were used to avoid stray light inall spectral measurements. The themoluminescence curves were recorded using ahomemade measurement setup (temperature range, 2196–280uC; heating rate,4uC/s). An ITT PVS-14 Generation III night vision monocular was used to monitorthe persistent luminescence ‘‘brightness’’ and to take NIR images through a Pentaxdigital SLR camera connected to the front of the monocular. The monitoring andimaging experiments were conducted in a dark room. Before all the spectralmeasurements and imaging, the samples were heat-treated in a muffle oven at 400uCfor 20 min to completely empty the electron traps.

Four light sources were used to activate the LiGa5O8:Cr31 discs/powders forspectral measurements and imaging: a 450 W xenon arc lamp (in a FluoroLog-3spectrofluorometer), a 4 W 254 nm UV lamp, a 400 mW 980 nm infrared laser, anda YAG:Ce-based white LED flashlight (85 lumens).

The in vivo bio-imaging experiments were performed on an IVIS Lumina IIimaging system in the bioluminescence mode. A 900-lumen white LED flashlight(Olight SR5) was used as the stimulation source (see Supplementary Information fordetail). Animal studies were performed according to a protocol approved by theInstitutional Animal Care and Use Committee (IACUC) of University of Georgia.

1. Yukihara, E. G. & McKeever, W. S. Optically Stimulated Luminescence:Fundamental and Applications (John Wiley & Sons Ltd., 2011).

2. Schweizer, S. Physics and current understanding of x-ray storage phosphors. Phys.Stat. Sol. (a) 187, 335–393 (2001).

3. Brito, H. F. et al. Persistent luminescence mechanism: human imagination atwork. Opt. Mater. Express 2, 371–381 (2012).

4. Holsa, J. Persistent luminescence beats the afterglow: 400 years of persistentluminescence. Electrochem. Soc. Interface 18, 42–45 (2009).

5. van den Eeckhout, K., Smet, P. F. & Poelman, D. Persistent luminescence inEu21-doped compounds: a review. Materials 3, 2536–2566 (2010).

6. Matsuzawa, T., Aoki, Y., Takeuchi, N. & Murayama, Y. A new longphosphorescent phosphor with high brightness, SrAl2O4:Eu21,Dy31.J. Electrochem. Soc. 143, 2670–2673 (1996).

7. Yamamoto, H. & Matsuzawa, T. Mechanism of long phosphorescence ofSrAl2O4:Eu21,Dy31 and CaAl2O4:Eu21,Nd31. J. Lumin. 72, 287–289 (1997).

8. Pan, Z. W., Lu, Y. Y. & Liu, F. Sunlight-activated long-persistent luminescence inthe near-infrared from Cr31-doped zinc gallogermanates. Nat. Mater. 11, 58–63(2012).

9. Meijerink, A. & Blasse, G. Photostimulated luminescence and thermallystimulated luminescence of some new x-ray storage phosphors. J. Phys. D: Appl.Phys. 24, 626–632 (1991).

10. von Seggern, H. Photostimulable x-ray storage phosphors: a review of presentunderstanding. Brazilian J. Phys. 29, 254–268 (1999).

11. Riesen, H. & Liu, Z. Q. Optical storage phosphors and materials for ionizingradiation. Current Topics in Ionizing Radiation Research (Ed. Nenoi, M.) (InTech,2012) pp. 625–648.

12. Nanto, H. et al. Advanced optical storage phosphor materials for erasable andrewritable optical memory utilizing photostimulated luminescence. Proc. SPIE3802, 258–265 (1999).

13. Pan, Z. W. & Yan, W. Z. Near infrared doped alkali metal, gallium oxide, alkalimetal gallate phosphors. University of Georgia Research Foundation, PCT/USpatent application WO 2011/035294 A2 (2011).

14. Szymczak, H., Wardzynska, M. & Mylnikova, I. E. Optical spectrum of Cr31 in thespinel LiGa5O8. J. Phys. C: Solid State Phys. 8, 3937–3943 (1975).

15. Alig, R. C. Theory of photochromic centers in CaF2. Phys. Rev. B 3, 536–545(1971).

16. Staebler, D. L. & Schnatterly, S. E. Optical studies of a photochrominc color centerin rare-earth-doped CaF2. Phys. Rev. B 3, 516–526 (1971).

17. Chandra, V. K. & Chandra, B. P. Dynamics of the mechanoluminescence inducedby elastic deformation of persistent luminescent crystals. J. Lumin. 132, 858–869(2012).

18. Botterman, J. et al. Mechanoluminescence in BaSi2O2N2:Eu. Acta Mater. 60,5494–5500 (2012).

19. Schneider, C. A., Rasband, W. S. & Eliceiri, K. W. NIH image to Image J: 25 yearsof image analysis. Nat. Methods 9, 671–675 (2012).

20. Delbecq, C. J., Toyozawa, Y. & Yuster, P. H. Tunneling recombination of trappedelectrons and holes in KCl:AgCl and KCl:TlCl. Phys. Rev. B 9, 4497–4505 (1974).

21. Avouris, P. & Morgan, T. N. A tunneling model for the decay of luminescence ininorganic phosphor: the case of Zn2SiO4:Mn. J. Chem. Phys. 74, 4347–4355(1981).

22. de Chermont, Q. L. M. et al. Nanoprobes with near-infrared persistentluminescence for in vivo imaging. Proc. Natl. Acad. Sci. USA 104, 9266–9271(2007).

23. Michalet, X. et al. Quantum dots for live cells, in vivo imaging, and diagnostics.Science 307, 538–544 (2005).

24. Ntziachristos, V., Bremer, C. & Weissleder, R. Fluorescence imaging with near-infrared light: new technological advances that enable in vivo molecular imaging.Eur. Radiol. 13, 195–208 (2003).

25. Frangioni, J. V. In vivo near-infrared fluorescence imaging. Curr. Opin. Chem.Biol. 7, 626–634 (2003).

AcknowledgementsZ.P. acknowledges financial support from the National Science Foundation (CAREERDMR-0955908), the American Chemical Society Petroleum Research Fund (PRF50265-DN10), and the US Office of Naval Research (N00014-07-1-0060). J.X.acknowledges the support by an NCI/NIH R00 grant (5R00CA153772). We thankJohn D. Budai for reading the manuscript. We thank Rick Tarleton for allowing us to use theIVIS imaging system.

Author contributionsZ.P. and F.L. conceived and designed the materials synthesis and characterizationexperiments and wrote the paper. F.L. and W.Y. carried out material synthesis and spectralmeasurements. F.L. and Z.P. carried out the NIR imaging. F.L. investigated the persistentluminescence mechanism. Y.-J.C. synthesized nanoparticles and analyzed the pixelintensity. J.X. designed the bio-imaging experiments. Y.-J.C., Z.Z. and J.X. carried outbio-imaging experiments. All the authors discussed the results.

Additional informationSupplementary information accompanies this paper at http://www.nature.com/scientificreports

Competing financial interests: The authors declare no competing financial interests.

License: This work is licensed under a Creative CommonsAttribution-NonCommercial-NoDerivs 3.0 Unported License. To view a copy of thislicense, visit http://creativecommons.org/licenses/by-nc-nd/3.0/

How to cite this article: Liu, F. et al. Photostimulated near-infrared persistent luminescenceas a new optical read-out from Cr31-doped LiGa5O8. Sci. Rep. 3, 1554; DOI:10.1038/srep01554 (2013).

www.nature.com/scientificreports

SCIENTIFIC REPORTS | 3 : 1554 | DOI: 10.1038/srep01554 9

1

Supplementary Information

Photostimulated near-infrared persistent luminescence as a new

optical read-out from Cr3+-doped LiGa5O8

Feng Liu, Wuzhao Yan, Yen-Jun Chuang, Zipeng Zhen, Jin Xie, Zhengwei Pan

1. Definition of NIR persistence time for Cr3+

-doped persistent phopshors

The high-energy side of Cr3+

emission, generally at around 700 nm, is the boundary between

deep red and near-infrared (NIR). However, the human’s eyes are insensitive to this wavelength,

especially at low light levels (such as the persistent luminescence described in present article),

due to the Purkinje effect. So, we refer the persistent luminescence from Cr3+

ions as NIR.

For visible persistent luminescence, the persistence time is defined as the time for the

afterglow brightness to decay to 0.32 mcd/m2, which is approximately 100 times the limit of

human perception with dark-adapted eyes. For NIR persistent luminescence, however, this

definition is no longer valid because the NIR signal is invisible to unaided eyes. The persistence

time for NIR persistent phosphors should then be determined by the sensitivity of infrared

detection systems such as nigh vision goggles, infrared cameras, or infrared detectors. In present

article, we define the persistence time of our NIR persistent phosphors as the duration it takes for

an eye can see with the aid of a night vision monocular (e.g., ITT PVS-14 Generation III) in a

dark room.

2. Estimation of NIR persistent luminescence intensity of LiGa5O8:Cr3+

phosphor

We used a comparison method to estimate the persistent luminescence intensities of

LiGa5O8:Cr3+

phosphor in absolute unit, e.g., mW/m2 (refer to the Supplementary Information in

ref. S1 for detailed method). The known LumiNova Green phosphor (a Eu2+

, Dy3+

-codoped

strontium aluminate, called LumiNova-G hereafter) manufactured by Nemoto & Co. Ltd. and

our recently reported Cr3+

-doped Zn3Ga2Ge2O10 (called ZGGO:Cr hereafter) NIR persistent

phosphor[S1]

were used as the benchmark materials. In its official documents[S2,S3]

, the

manufacturer provided the persistent luminescence luminance of LumiNova-G at 10 min and 1 h

after ceasing the excitation as ~300 mcd/m2 and ~30 mcd/m

2, respectively. These luminance

values in unit of mcd/m2 can be converted to intensity values in unit of mW/m

2. The conversion

revealed that the persistent luminescence intensities of LumiNova-G at 10 min and 1 h after

ceasing the excitation are ~7.7 mW/m2 and ~0.77 mW/m

2, respectively

[S1]. The estimated

persistent luminescence intensities of ZGGO:Cr NIR persistent phosphors at 10 min and 1 h after

ceasing the excitation are ~12.3 mW/m2 and ~1.38 mW/m

2, respectively

[S1]. We then measured

the persistent luminescence emission spectra for LiGa5O8:Cr3+

, ZGGO:Cr, and LumiNova-G at

10 min and 1 h after ceasing the excitation and compared the integrated intensities of the three

emission spectra to obtain the integrated intensities of LiGa5O8:Cr3+

. For a fair comparison, the

following conditions and cautions were used in our measurements.

(1) The persistent luminescence measurements were carried out at room temperature on a

FluoroLog-3 spectrofluorometer equipped with a standard R928P photomultiplier tube.

The instrument parameters kept the same for the three materials. All the measurements

2

were corrected for the detector response (corrected emission range of the

spectrofluorometer is 290850 nm).

(2) The LumiNova-G powder was pressed into a disk with a diameter of 15 mm and

thickness of ~2 mm, the same as the ZGGO:Cr and LiGa5O8:Cr3+

phosphor discs.

(3) Before the measurements, all the samples were left in dark for 7 days to empty the traps

(Note: the traps in ZGGO:Cr and LiGa5O8:Cr3+

can be emptied by heating to ~400 C.

But heating LumiNova-G can cause degradation of its persistent luminescence

performance).

(4) During excitation, the LumiNova-G and ZGGO:Cr discs were exposed to a 4-W 365 nm

UV lamp for 5 min, while the LiGa5O8:Cr3+

disc was exposed to a 4-W 254 nm UV lamp

for 5 min.

(5) In the persistent luminescence measurements, the emitting areas for the three discs were

fixed at 7 mm ×14 mm.

Figure S1a shows the measured persistent luminescence emission spectra for LiGa5O8:Cr3+

,

ZGGO:Cr, and LumiNova-G at 10 min and 1 h after ceasing the excitation. By comparing the

integrated intensities of LiGa5O8:Cr3+

with those of the known LumiNova-G and ZGGO:Cr

phosphors, we estimated that the persistent luminescence intensities of LiGa5O8:Cr3+

phosphor

disc at 10 min and 1 h after ceasing the excitation are ~4.7 mW/m2 and ~1.2 mW/m

2,

respectively. The results show that at 10 min after ceasing the excitation, the persistent

luminescence intensity of LiGa5O8:Cr3+

phosphor (~4.7 mW/m2) is lower than those of

LumiNova-G phosphors (~7.7 mW/m2) and ZGGO:Cr phosphor (~12.3 mW/m

2); however, after

1 h persistent luminescence, the emission intensity of LiGa5O8:Cr3+

phosphor (~1.2 mW/m2) is

higher than that of LumiNova-G phosphors (~0.77 mW/m2) and is comparable to that of

ZGGO:Cr phosphor (~1.38 mW/m2).

We also acquired the persistent luminescence decay curves of LiGa5O8:Cr3+

, ZGGO:Cr, and

LumiNova-G phosphors at their respective emission maxima (in unit of CPS) with a time period

up to 100 h, as shown in Fig. S1b. Basing on these decay curves, we can compare the persistent

luminescence intensities (in unit of CPS) of LiGa5O8:Cr3+

, ZGGO:Cr, and LumiNova-G

phosphors at different instants. The results on some important instants are given below.

(1) The measured intensities of LiGa5O8:Cr3+

are about 2.8 times and 6.7 times of those of

LumiNova-G at 10 min and 1 h after the stoppage of the excitation, respectively.

(2) The measured intensity of LiGa5O8:Cr3+

at 100 h is higher than that of LumiNova-G at 24

h after the excitation.

(3) At any instant from 10 min to 100 h after the excitation, the measured intensity of

LiGa5O8:Cr3+

is higher than that of ZGGO:Cr.

3. Synthesis of LiGa5O8:Cr3+

nanoparticles

The LiGa5O8:Cr3+

nanoparticles were synthesized by a sol-gel method followed by

calcinations at high temperature. The synthesis process is briefly described as follows. A solution

was prepared by dissolving stoichiometric amount of lithium nitrate (LiNO3, 99%), gallium

nitrate [Ga(NO3)3, 99.999%], and chromium nitrate [Cr(NO3)3, 98.5%,] into methanol. The

solution was then stirred for several hours on a magentic stirrer. During the stirring process,

3

chelating agent, acetylacetone, was added into the solution to form chelate complexes. Aqueous

ammonium solution (28 vol.%) was also added to stabilize the complexes and to adjust the pH of

the sol. When a homogeneous sol was formed, the solution was heated to 80100 C for several

hours to form dry gel. The dry gel was then calcinated in a muffle furnace at 10001100 C for

35 hours to form LiGa5O8:Cr3+

nanoparticles exhibiting NIR persistent luminescence and

photostimulated persistent luminescence (PSPL) properties. Finally, by centrifugation and

filtration, LiGa5O8:Cr3+

nanoparticles with diameters in the range of 50150 nm were obtained

for subsequent bio-imaging experiments (see Supplementary Fig. S8).

Note that the LiGa5O8:Cr3+

nanoparticles show lower persistent luminescence and PSPL

intensities (Supplementary Fig. S10) and a shorter overall persistence time than the discs and

micro-powders owing to the lower sample density and lower sintering temperature (a higher

sintering temperature can generally enhance the persistent luminescence of the materials[S4]

).

4. Coating LiGa5O8:Cr3+

nanoparticles with polyethylenimine

Before loading the LiGa5O8:Cr3+

nanoparticles into mouse breast cancer 4T1 cells, the

nanoparticles were coated with polyethylenimine (PEI, MW: ~25000, Aldrich) because PEI is a

polycation molecule known for facilitating cell uptake of nanoparticles via endocytosis[S5,S6]

. In

the process of coating, LiGa5O8:Cr3+

nanoparticles were dispersed in 12.5 wt% PEI(aq) and then

stirred for two days. The raw PEI coated LiGa5O8:Cr3+

nanoparticles (PEI-LiGa5O8:Cr3+

) were

washed with deionized water for three times and collected by centrifugation. The purified PEI-

LiGa5O8:Cr3+

nanoparticles were dispersed in 1× phosphate buffered saline (PBS, pH 7.4, 0.067

M, HyClone, Thermo Scientific), and sterilized by UV light in a laminar flow hood for 72 hours.

5. PEI-LiGa5O8:Cr3+

nanoparticles labeled 4T1 cells

A total of ~1×106 murine breast cancer 4T1 cells were seeded and cultured in RPMI-1640

medium (Cellgro, Mediatech, Inc.) with 10% fetal bovine serum (FBS, Cellgro, Mediatech, Inc)

and 100 μg/ml sterilized PEI-LiGa5O8:Cr3+

at 37 °C in a humidified atmosphere containing 5%

CO2 for two days in a 25 cm2 cell culture flask. The adhesive 4T1 cells were then harvested by

the treatment of trypsin-EDTA solution (Cellgro, Mediatech, Inc). Part of these alive PEI-

LiGa5O8:Cr3+

labeled 4T1 cells was stored in 1× PBS buffer (pH 7.4) at 4 C for the subsequent

experiments of subcutaneous injection into athymic nude mice, and part was fixed with 70%

ethanol aqueous solution (200 proof, Decon Labs, Inc.) at 4 C for 12 h and then stored at 4 C

for subsequent phantom study. Blank samples of fixed, un-labeled 4T1 cells were also prepared

by the same procedure, except for no loading of PEI-LiGa5O8:Cr3+

nanoparticles.

6. Imaging and tracking of PEI-LiGa5O8:Cr3+

nanoparticles labeled 4T1 cells

The PEI-LiGa5O8:Cr3+

labeled 4T1 cells were tracked in vivo and in vitro for a period of up

to 10 days using an IVIS Lumina II imaging system in the bio-luminescence mode. For in vivo

tracking, the PEI-LiGa5O8:Cr3+

labeled 4T1 cells (~2.5×107

cells) were illuminated by a 4 W 254

nm UV lamp for 15 min, and then subcutaneously injected into the back of a nude mouse. For in

vitro tracking, the fixed PEI-LiGa5O8:Cr3+

labeled 4T1 cells (~2×106 cells in 1.5 mL ethanol

aqueous solution) and the un-labeled 4T1 cells (blank sample, ~2×106 cells in 1.5 mL ethanol

aqueous solution) were irradiated by a 254 nm UV lamp for 15 min in a 24-well cell plate. To get

4

the PSPL signal, the mouse/plate was daily exposed to a white LED flashlight (Olight SR51, 900

lumens) for 15 s with the flashlight being placed at 6 cm above the mouse/plate, and images were

then taken at 10 s and 5 min after the stimulation. The exposure time is 2 min. The experiments

on the mouse and plate lasted for 10 days (Fig. 7 in main text) and 8 days (Supplementary Fig.

S9), respectively. The images were processed by Living Image (Version 4.3.1 SP1, PerkinElmer)

at binning of 4 and smooth of 5×5.

References:

[S1] Pan, Z. W., Lu. Y. Y. & Liu. F. Sunlight-activated long-persistent luminescence in the near-

infrared from Cr3+

-doped zinc gallogermanates. Nat. Mater. 11, 5863 (2012).

[S2] http://www.use.com.tw/material/e5.pdf

[S3] http://www.umccorp.com/print/LuminovE%20(04-16-2009).pdf

[S4] Trojan-Piegza, J., Niittykoski, J., Hölsä, J. & Zych, E. Thermoluminescence and kinetics of

persistent luminescence of vacuum-sintered Tb3+

-doped and Tb3+

, Ca2+

-codoped Lu2O3

materials. Chem. Mater. 20, 22522261 (2008).

[S5] Canton, I. & Battaglia, G. Endocytosis at the nanoscale. Chem. Soc. Rev. 41, 27182739

(2012).

[S6] Cheng, Z. L., Zaki, A. A., Hui, J. Z., Muzykantov, V. R. & Tsourkas, A. Multifunctional

nanoparticles: cost versus benefit of adding targeting and imaging capabilities. Science 338,

903910 (2012).

Figure S1. Comparison of persistent luminescence performances of LiGa5O8:Cr3+ NIR

persistent phosphor, Cr3+-doped Zn3Ga2Ge2O10 (ZGGO:Cr) NIR persistent phosphor,

and LumiNova green (LumiNova-G) persistent phosphor. a, Persistent luminescence

spectra of LiGa5O8:Cr3+, ZGGO:Cr, and LumiNova-G at 10 min and 1 h after ceasing the

UV irradiations. b, Persistent luminescence decay curves of LiGa5O8:Cr3+, ZGGO:Cr, and

LumiNova-G phosphors. The monitoring wavelengths for LiGa5O8:Cr3+, ZGGO:Cr, and

LumiNova-G are 716 nm, 713 nm, and 520 nm, respectively.

500 600 700 8000.0

5.0x105

1.0x106

1.5x106

2.0x106

LumiNova-G

ZGGO:Cr

1 h

LiGa5O

8:Cr

Inte

nsity (

arb

. u

nits)

Wavelength (nm)

10 min

10 min

1 h

1 10 100 1000

102

103

104

105

106

107

ZGGO:Cr

LumiNova-G

LiGa5O

8:Cr

background

Decay Time (minutes)

CP

S

10

0 h2

4 h

1 h

10

min

a

b

Supplementary Figures

5

0 10 20 30 40 50 60

10-2

10-1

100

101

380 nm

360 nm

340 nm

Inte

nsity (

arb

. units)

Time (s)

ex

=

320 nm

I10s

Figure S2. Room temperature persistent luminescence decay curves

of LiGa5O8:Cr3+ phosphor discs irradiated by monochromatic light

between 250600 nm for 5 min. The monitoring wavelength is 716 nm.

The effectiveness of excitation decreases when the excitation wavelength

is increased from 250 nm to 600 nm. The persistent luminescence

intensity at time of 10 s after the stoppage of the irradiation (I10s) was

used to plot the persistent luminescence intensity as a function of

excitation wavelengths shown in the bottom inset of Fig. 1b.

6

0 100 200 300 400 500 600

10-1

100

101

Inte

nsity (

arb

. u

nits)

Decay Time (s)

24 h delay

980 nm laser stimulation

Figure S3. A 980 nm laser-stimulated persistent

luminescence in a LiGa5O8:Cr3+ phosphor disc. Before

stimulation, the disc was pre-irradiated by 300 nm light for

20 min and decayed 24 h at room temperature. The

photostimulated persistent luminescence decay was

monitored at 716 nm emission. The stimulation source is

a 400 mW 980 nm laser pointer.

7

Figure S4. Effectiveness of white LED for the photostimulated

persistent luminescence (PSPL) in LiGa5O8:Cr3+. a, Thermoluminescence

curves monitored at 716 nm over 20280 C. The samples were pre-

irradiated by 300 nm UV light for 20 min. The solid-line curve was acquired

at a delay time of 120 h. The dash-line curve was acquired on a 120 h-

decayed disc after stimulation by a YAG:Ce-based white LED for 100 s. b,

The solid-line curve is the emission spectrum of the white LED used in the

present study. The dots curve is the PSPL read-out spectrum which is the

same as the one in the right panel of Fig. 4c.

20 60 100 140 180 220 260

0.00

0.15

0.30

0.45

white LED stimulation

Therm

olu

min

escence Inte

nsity (

arb

. units)

Temperature (°C)

120 h delay

350 400 450 500 550 600 650

WLE

D E

mis

sio

n In

tensity

(arb

. units)

Wavelength (nm)

PS

PL Inte

nsity (

arb

. units)

a

b

8

a

b 768 h 792 h 816 h 840 h 864 h

888 h 912 h 936 h 960 h 984 h 1008 h

744 h

120 h 240 h 360 h 480 h 600 h 720 h

Figure S5. NIR images for photostimulated persistent luminescence (PSPL) in

LiGa5O8:Cr3+ phosphor plate. The plate is the same as the one in Fig. 2a. Before

imaging the plate was exposed to a 254 nm UV lamp for 5 min. a, From 120 to 720 h,

the right half of the plate was stimulated by a YAG:Ce based white LED (85 lumens)

for 20 s at every 120 h. b, From 744 to 1,008 h, the entire plate was stimulated by the

white LED for 20 s at every 24 h. All the PSPL images were taken at 10 s after

ceasing the white LED. The imaging parameter is manual/ISO 400/10 s. The images

taken at 120 h, 720 h, 744 h, 816 h, 888 h, 960 h, and 1,008 h are the same as the

ones shown through Fig. 5e to Fig. 5k, respectively.

9

Figure S6. NIR images of four LiGa5O8:Cr3+ phosphor discs taken at

different persistent luminescence times (10 min to 2,000 h) after irradiation

with a 254 nm lamp for 10 s to 5 min. The images in a and b are the same as

the ones in Fig. 1c and Fig. 1j, respectively. At decay times of 1,008 h and 2,000

h, the discs were stimulated by a YAG:Ce based white LED (85 lumens) for 20 s

and exhibited enhanced NIR persistent luminescence, as shown in c and d,

respectively (The PSPL images were taken at 10 s after ceasing the white LED

stimulation). The image in d clearly shows that after 2,000 h of decay, the

information stored in the LiGa5O8:Cr3+ phosphor discs can be clearly read out by

white LED stimulation, even after a short 10 s of 254 nm light irradiation (the left

most disc). The imaging parameters are: a, c and d, manual/ISO 400/10 s, and

b, manual/ISO 1600/30 s.

10 s 30 s 1 min 5 min

10 min

White LED stimulation

White LED stimulation

1008 h

2000 h

a

b

c

d

1008 h

10

Figure S7. NIR images of persistent luminescence and photostimulated persistent

luminescence (PSPL) of the University of Georgia logo. The logo was drawn by

using a NIR paint made by mixing 40 wt.% LiGa5O8:Cr3+ powder with acrylic

polyurethane vanish. Before imaging the logo was exposed to a 254 nm UV lamp for 5

min. a, NIR images taken at 10 s to 24 h after the stoppage of the UV irradiation. b,

From 24 to 144 h, the left half of the logo was stimulated by a YAG:Ce based white LED

(85 lumen) for 20 s at every 24 h. From 168 to 240 h, the entire logo was stimulated by

the white LED for 20 s at every 24 h. For each stimulation, the PSPL images were taken

at 10 s, 5 min, 30 min, and 24 h after ceasing the white LED. The imaging parameter is

manual/ISO 400/15 s.

24 h

48 h

96 h

10 s 1 h 5 h 24 h

144 h

168 h

192 h

10 s 5 min 30 min 24 h

240 h

Pers

iste

nt

Lu

min

escen

ce

P

ho

tosti

mu

late

d P

ers

iste

nt

Lu

min

escen

ce (

PS

PL

)

a

b

11

Figure S8. Transmission electron microscopy image

of LiGa5O8:Cr3+ nanoparticles. The diameters of the

nanoparticles are in the range of 50150 nm.

200 nm

12

0.5

1.0

1.5

x105

LGO:Cr3+ labeled 4T1 Cells 4T1 Cells

(a) Before PSPL (b) PSPL, 10 s (c) Before PSPL (d) PSPL, 10 s

24 h

72 h

120 h

192 h

Figure S9. Phantom images of 4T1 cells labeled and un-labeled with PEI-LiGa5O8:Cr3+

nanoparticles for a 8-day (192 h) tracking using an IVIS imaging system. a,b, 4T1 cells

labeled with PEI-LiGa5O8:Cr3+ nanoparticles (2x106 cells in 1.5 mL ethanol aqueous solution)

in a 24 well plate. c,d, Un-labeled 4T1 cells (2x106 cells in 1.5 mL ethanol aqueous solution)

in a 24 well plate. The un-labeled 4T1 cells were used for comparison purpose. Both the

labeled and un-labeled 4T1 cells were irradiated by a 254 nm UV lamp for 15 min. The two

types of cells (Columns a and c) were imaged daily using an IVIS Lumina II imaging system

in the bioluminescence model, followed by stimulation with a white LED flashlight (Olight

SR51, 900 Lumens) for 15 s with the flashlight being placed at 6 cm above the plate. Images

were then taken at 10 s after the stimulation (Columns b and d). The detected signals were

attributed to PSPL. The signals from the LiGa5O8:Cr3+ nanoparticles can be clearly detected

even 8 days after the UV irradiation. All images were acquired in the bioluminescence mode

with an exposure time of 2 min. The color scale bar represents the luminescence intensity in

the unit of radiance, p/sec/cm2/sr.

13

Figure S10. Comparisons of persistent luminescence and PSPL properties of

LiGa5O8:Cr3+ discs and LiGa5O8:Cr3+ nanoparticles. a, Persistent luminescence

decay curves. Both samples were irradiated by 300 nm xenon light for 5 min. The

monitored wavelength is 716 nm. b, PSPL decay curves. Before measurements, both

samples were irradiated by 300 nm xenon light for 5 min. The monitored wavelength is

716 nm. The stimulation light is 400 nm xenon light. The results show that the

nanoparticles have lower persistent luminescence and PSPL intensities than the discs.

63 66 69 720 10 20 30 40 50 60

Inte

nsity (

arb

. u

nits)

Decay Time (min)

40

0 n

m lig

ht

stim

ula

tio

n

phosphor disc

nanoparticles

Inte

nsity (

arb

. u

nits)

Decay Time (min)

Background

phosphor disc

nanoparticles

a b

14