Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 97 (2012) 825–829

Contents lists available at SciVerse ScienceDirect

Spectrochimica Acta Part A: Molecular andBiomolecular Spectroscopy

journal homepage: www.elsevier .com/locate /saa

Infrared to visible upconversion in Ho3+/Yb3+ co-doped Y2O3 phosphor: Effectof laser input power and external temperature

Monika Rai a, K. Mishra a, S.K. Singh b, R.K. Verma a, S.B. Rai a,⇑a Laser and Spectroscopy Laboratory, Department of Physics, Banaras Hindu University, Varanasi 221005, Indiab Department of Physics, Changwon National University, Changwon 641-773, Republic of Korea

h i g h l i g h t s

" Synthesis and opticalcharacterization of Ho3+/Yb3+ co-doped Y2O3 phosphors.

" Effect of the laser input power andexternal temperature on the UCemission.

" Lifetime analysis of the thermallycoupled levels.

1386-1425/$ - see front matter � 2012 Elsevier B.V. Ahttp://dx.doi.org/10.1016/j.saa.2012.07.071

⇑ Corresponding author. Tel.: +91 542 230 7308; faE-mail address: sbrai49@yahoo.co.in (S.B. Rai).

g r a p h i c a l a b s t r a c t

a r t i c l e i n f o

Article history:Received 9 May 2012Received in revised form 14 July 2012Accepted 17 July 2012Available online 25 July 2012

Keywords:PhosphorRare-earthSol–gel chemistryOptical properties

a b s t r a c t

In the present paper, Ho3+ doped and Ho3+/Yb3+ co-doped Y2O3 phosphors have been synthesized usingsolution combustion technique and characterized for its structure and upconversion (UC) fluorescenceas a function of Yb3+ concentration. Effect of a variation in laser input power and external temperatureon the UC emission intensity has been studied to explore the UC mechanism and temperature dependentbehavior of the phosphor, respectively. On excitation with near infrared (NIR) light (976 nm), the phos-phor emits strong green emission along with relatively weak emission bands in red and blue regions at553, 670 and 497 nm due to 5S2 ?

5I8, 5F5 ?5I8 and 5F3 ?

5I8, respectively. The emission shows adecrease in intensity with an increase in external temperature, however contrary to the normal behaviorof Ho3+, no significant change in the FIR (fluorescence intensity ratio) of 5F4 ?

5I8 and 5S2 ?5I8 transitions

is noted in the present host. This peculiar behavior of the sample with external temperature has beenexplained by temperature dependent lifetime study of the thermally coupled levels.

� 2012 Elsevier B.V. All rights reserved.

Introduction

Photonic materials have attracted much interest of researchersin recent decades. Especially the inorganic materials bearing theproperty of upconversion (UC) show wide applications. UC de-scribes a nonlinear optical process in which low energy photonsare used to generate high-energy photons, usually infrared (IR)radiation is used to generate visible and ultraviolet (UV) light [1].

ll rights reserved.

x: +91 542 236 9889.

Electronic transitions in such RE doped materials involve 4f orbi-tals shielded with 5s, 5p orbitals (intraconfigurational f–f transi-tions). Due to this shielding, they exhibit unique optical featuressuch as long-lived excited states (in a range of 10�3–10�6 s) andnearly line like (FWHM � 10 nm) emission [2,3]. In addition to this,spectral emission in these cases may cover entire UV–IR regionwhich makes them promising for many potential applications suchas for lighting, display devices, sensors, energy harvesting, bio-imaging, and security purposes [4–10].

The efficiency of UC fluorescence in such materials depends onvarious parameters including the host matrix, particle shape and

826 M. Rai et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 97 (2012) 825–829

size, synthesis technique, excitation mechanisms, etc. A host withlow phonon energy is preferred for UC process as it minimizesthe non-radiative losses. This is important because UC is very sen-sitive to quenching by high energy vibrations according to the en-ergy gap law [11]. Yttrium oxide (Y2O3) seems to be an ideal matrixas it is a chemically very stable, high band gap (5.6 eV), transparenthost with a low phonon energy (�430–550 cm�1) etc. [12]. Besidethis, the ionic radius of Y3+ is almost similar to the other trivalentRE ions (RE3+) and so by Shannon effective ionic radius theory [13],it can be easily and effectively doped with RE3+ ions. The RE3+ ionsattach at a suitable site in the matrix and interact with the field ofhost lattice giving rise to a characteristic strong, sharp and Starksplitted visible emission bands of RE3+ ions.

Among different RE ions, the large absorption cross-section ofYb3+ and efficient energy transfer from the resonance levels ofYb3+ to Ho3+ make the co-dopant (Ho3+ and Yb3+) a suitable couplefor efficient luminescence. Moreover, the 5F4 ? 5I8 and 5S2 ? 5I8

transitions of Ho3+ having a separation of �667 cm�1 are thermallycoupled levels and usually show a temperature dependent varia-tion in emission intensity [14,15]. The temperature dependentvariation usually arises due to the temperature dependence ofnon-radiative rates and lifetimes of the energy levels of interest.Therefore, by measuring the intensity of fluorescence or lifetimeof the particular levels, the temperature could in principle be in-ferred. In the present paper, Ho3+ doped and Ho3+/Yb3+ co-dopedY2O3 phosphors have been synthesized using solution combustionroute and characterized for its structure and UC fluorescence. Ef-fect of laser input power and external temperature on the UC emis-sion has also been studied to explore the UC mechanism andtemperature dependent behavior of the sample, respectively.

Fig. 1. X-ray diffraction (XRD) patterns of Ho3+/Yb3+:Y2O3 phosphor (as-synthesizedand annealed at 1200 �C/2 h) along with JCPD card. Inset shows the Lorentzianfitting for the (222) reflection, centered at 2h � 29.06�.

Experimental

Sample preparation

Analytical reagent (AR) grade yttrium oxide (Y2O3, 99.99%,Himedia), holmium oxide (Ho2O3, 99.9%, Loba Chemie), nitric acid(Merck, 99.9%) and urea (Loba Chemie) were used for the synthesisof phosphor using solution combustion method [16]. The composi-tion was selected as follows,

ð100� x� yÞY2O3 þ x Ho2O3 þ y Yb2O3

where, x = 0.6 and y = 0, 0.5, 1.0, 1.5, 2.0 mol%.These oxides were dissolved in concentrated HNO3 and a small

amount of urea was then added as an organic fuel. This whole mix-ture was then heated on a magnetic stirrer at �333 K to removethe excess water, nitrates and other volatile organic impurities.As the water content in the solution was reduced, the solutionchanged into a transparent gel. The gel thus obtained was kept ina platinum crucible and burnt in a closed furnace at 723 K in airatmosphere. This resulted in a white foamy voluminous structure,which is then grinded to obtain a fine powder. The sample thus ob-tained is referred to as ‘‘as-synthesized’’. The as-synthesized phos-phor powders were post annealed at 1473 K for 2 h to removevolatile impurities (namely OH, CO2, NOx, etc.), if any, left duringthe synthesis process [16].

Characterization

To study the crystal structure, X-ray diffraction (XRD) patternswere recorded using 18 kW Cu rotating anode based high resolu-tion Rigaku X-ray powder diffractometer (XRD) fitted with acurved crystal monochromator in the diffracted beam. Data wereobtained from 2h = 10�–80� at a scanning speed of 3�/min. Surfacemorphology was characterized using scanning electron microscopy

(SEM) on Quanta-200 model operated at 20 kV. A near infrareddiode laser emitting at 976 nm used to excite the sample and theUC fluorescence were recorded using a iHR320 Jobin Yvon spec-trometer equipped with R928 photon counting photo multipliertube. The laser beam is focused on the sample with spot size ofabout 0.4 mm in diameter using collimating optics.

For the UC fluorescence measurements at different tempera-tures, sample was prepared in the form of pellets (12 mm in diam-eter, 0.5 mm in thickness). The pellet was placed on an aluminumplate kept in a closed heater and the temperature of the samplewas read with the help of a thermocouple. Band pass filters wereused to record the spectrum for a particular region to avoid back-ground emission. The decay time measurements of 5F4 ? 5I8 and5S2 ? 5I8 transitions of Ho3+ was carried out at different tempera-tures using 976 nm wavelength of laser. Laser beam was choppedwith a mechanical chopper and data were acquired using an oscil-loscope (analog digital scope-HM1507) supported with a softwareSP107. The decay time was determined using non-linear leastsquares fit method.

Results and discussion

X-ray diffraction (XRD) measurements

X-ray diffraction (XRD) data were analyzed for the identifica-tion of phase and crystallite size. Fig. 1 shows the diffraction pat-terns of the samples at two different temperatures (as-synthesized and annealed at 1473 K for 2 h). All the diffractionpeaks match well to the cubic bixbyte Y2O3 structure (JCPDS no.43-1036) which belongs to the Ia-3 (206) space group. No impuritypeak is seen in the structure which confirms the synthesis of phasepure material. Both the diffraction patterns show sharp diffractionpeaks confirming the well crystalline nature of the sample. Theintensity of the diffraction peaks in the case of annealed samplesshows an appreciable increase which depicts the improvement incrystallinity of the sample.

Further, it is also marked that as the sample is annealed at high-er temperatures the FWHM (full width at half maximum) of thediffraction peaks decreases and peaks become sharper. This sug-gest for an increase in the crystallite size of the annealed powder.To confirm this, crystallite size of both samples was calculatedusing the Scherrer formula [17],

D ¼ k� 0:89b� cos h

M. Rai et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 97 (2012) 825–829 827

where, D is crystallize size, k is the wavelength of incident X-ray[CuKa (1.54056 ÅA

0

)], b is the FWHM and h is the diffraction anglefor (h k l) plane. For the crystallite size calculation, three mostintense peaks (�29.06�, 48.44� and 57.44�) were selected. TheFWHM of these peaks were taken by their Lorentzian peak fitting.Inset to the Fig. 1 shows the goodness of fit for the most intensepeak centered at 2h � 29.06�. Before the calculation of particlesize, instrumental correction in the measured FWHM (which isestimated by recording the XRD pattern for standard silica sample)was made to get the correct FWHM. The average crystallite size forthe as-synthesized and the annealed samples calculated throughthe aforementioned procedure comes out to be �17 and �86 nm,respectively.

Scanning electron microscopic (SEM) analysis

Supplementary Fig. S1 (see Supplementary information) showsthe SEM image of the Ho3+/Yb3+ co-doped Y2O3 phosphor (annealedat 1473 K for 2 h). The image shows the characteristic surfacemorphology of the combustion product. Particles are highlyagglomerated and form a connected network type of structure withsome vacant spaces among them, which is expected due to theevolution of different gases during combustion of the gel. Further,it reveals inhomogeneous distribution of the particles (a variationin size and shape of particles). The size of some of the particles liewell below 100 nm and may be regarded as nano-particles whichis in well agreement to the crystallite size calculation using XRD.However, most of the particles are of sub-micron size due toagglomeration of many crystallites and lack of separate grainboundary. The shapes of the particles are irregular and seem to bepolygonal.

Upconversion measurement and effect of laser input power

Fig. 2 shows the UC spectra of Ho3+ in phosphor with and with-out Yb3+ ions. It is found that the phosphor doped with Ho3+ singlygives a weak UC emission. Since, Ho3+ has no level resonant with976 nm excitation, the nearest lying levels (namely 5I5, 5I6) couldbe populated only through phonon assisted ground state absorp-tion (GSA) process. The host matrix has a phonon frequency ofthe order of 500 cm�1 so one or two phonons would be sufficientto bridge the non-resonant energy gap, which opens the possibilityof population of 5I6 level by GSA. Since, the 5I6 level has a compar-atively long lifetime (of the order of ms) [18], the ions in this state

Fig. 2. Room temperature upconversion (UC) spectra of Ho3+ and Ho3+/Yb3+:Y2O3

phosphors (annealed at 1473 K/2 h) on excitation with 976 nm radiation. Insetshows the effect of variation in Yb3+ concentration on UC emission intensity.

may re-absorb photons (excited state absorption, ESA) and popu-late upper lying 5S2 and 5F4 levels. However, this is a less probableand weak process and hence the emission intensity is found to beweak.

On the other hand, Ho3+/Yb3+:Y2O3 phosphor material emitsstrong UC emission and intense luminescence peaks are observedat 497, 538, 553, 670, and 757 nm which have been assigned toarise due to electronic transitions 5F3 ? 5I8, 5F4 ?

5I8, 5S2 ? 5I8,5F5 ?

5I8 and 5S2 ?5I7 of Ho3+, respectively. Effect of the Yb3+ dop-

ing on the emission intensity of green emission (553 nm) is shownin the inset of Fig. 2. It is obvious that at 2 mol% Yb3+ concentra-tions the emission intensity is maximum (about 48 times). Authorsdid not try a higher concentration of Yb3+ as it may change the lat-tice parameter of the host material appreciably [19]. Thus, Yb3+

ions act as sensitizer for the emission of Ho3+ and the emissionintensity is enhanced appreciably. The complete UC mechanism in-volved in the process is shown schematically in Fig. 3. The red andgreen transitions have been shown to involve two photon pro-cesses while a three photon process has been shown for bluetransition.

In order to support the UC mechanism shown in the energy le-vel diagram in Fig. 3, number of photons involved for a particulartransition in the UC process has been verified through laser inputpower dependence study. Fig. 4 shows the variation of emissionintensity in green band with laser input power while inset to thefigure shows lnI (intensity of UC emission) vs lnP (applied laser in-put power) plot. It is well known that the slope of these curves (n)gives the number of photons involved in the excitation process ofdifferent UC bands. The plot shows a sharp increase in intensityin the beginning and subsequently saturation like behavior at highinput laser powers. The transitions 5F4 ?

5I8, 5S2 ? 5I8 and5F5 ?

5I8 in Ho3+ are well known to arise due to two photon pro-cesses (n � 2). For the present phosphor the slope value is 2.41and 2.17 for 5F4, 5S2 ? 5I8 and 5F5 ? 5I8 transitions, respectivelywhich also indicate the two photon process. A value of the slopemore than two however indicates that beside of a direct populationin 5F4, 5S2 and 5F5 levels through ETU (energy transfer UC from Yb3+

to Ho3+ through well known co-operative process) the 5F4, 5S2 and5F5 levels have a feed back of population from upper levels also (asis also shown in the energy level diagram through non-radiativerelaxation). Further, it has been noted that the slope of the curvedecreases at higher pump powers. This decrease in slope value isdue to the fact that UC rate depends on pump power. The workby Singh et al. [20] from our group also reports a similar behavior

Fig. 3. Schematic energy level diagram showing the mechanism involved in variousupconversion (UC) transitions.

Fig. 5. Upconversion (UC) spectra of Ho3+/Yb3+:Y2O3 in green region as a function ofexternal temperature. Inset to the figure shows the variation of fluorescenceintensity ratio (FIR) of 5F4 ?

5I8 and 5S2 ? 5I8 transitions. (For interpretation of thereferences to colour in this figure legend, the reader is referred to the web version ofthis article.)

Fig. 4. Effect of laser input power on the intensity of different transitions in Ho3+/Yb3+:Y2O3 phosphor under 976 nm excitations. Inset to figure shows the log I(emission intensity) vs logP (applied laser input power) plot. The slope of the plotgives the involvements of number of photons in the UC process.

828 M. Rai et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 97 (2012) 825–829

for Er3+, Yb3+ in Gd2O3 matrix and explains it in detail by using rateequations that increasing pump power increases the UC rate andhence the slope of the lnI–lnP plot decreases from 2 to 1 for twophoton process.

Supplementary Fig. S2 (see Supplementary information) showsCIE (international commission on illumination) chromaticity dia-gram which is used to study the perception of color in terms ofmathematically defined color spaces. Most modern contemporarylighting sources bear specifications which refer to color in termsof the 1931 CIE chromatic color coordinates. The green emissionof the Ho3+/Yb3+:Y2O3 phosphor has CIE chromaticity coordinates(0.31, 0.67) as shown in Supplementary Fig. S2 (Supplementaryinformation). The coordinate present at the very boundary sup-ports the high color purity in the sample.

Fig. 6. Luminescence decay curves for 5F4 ?5I8 and 5S2 ?

5I8 transitions atdifferent temperatures showing a decrease in lifetime of 5F4 and 5S2 states withtemperature.

Effect of external temperature on UC emission

The UC emission has also been recorded at different tempera-tures to monitor the effect of external temperature on the lumines-cence intensity. In fact, 5F4 and 5S2 levels of Ho3+ are very close toeach other and thermally coupled so the focus of the study hasbeen to observe the behavior of the emission from these two levelsto explore the possibility of temperature sensitivity. The emissionspectra recorded at different temperatures upto 500 K is shown inFig. 5. Two important results could be noticed from the spectra.First, the overall decrease in the intensity of emission is attainedwith an increase in sample temperature, which can be describedin terms of lattice vibration of the host material by the followingexpression,

NðtÞ ¼ Nð0Þ expð�t=sÞ

where, N(t) is the population in the emitting level at temperature tand s is its lifetime [21]. The inverse of s is equal to the sum of theradiative emission rate (R) and the non-radiative relaxation rate(NR). Since, NR increases with temperature this causes a decreasein s and hence R is further reduced. This increase in NR modifiesthe fluorescence intensity ratio. Thus as temperature increases fluo-rescence intensity decreases.

The second important and unusual result is observed for thefluorescence intensity ratio (FIR) of 5F4 ?

5I8 and 5S2 ? 5I8 transi-tions arising from the thermally coupled levels 5F4, and 5I8. Ther-mally coupled levels show uneven variation in intensity of

emission with an increase in temperature and so the FIR from themusually shows a linear increase. This linear behavior is used as astandard curve and can be used for the measurement of tempera-ture in terms of fluorescence [22]. However, for the present case,the FIR of 5F4 ? 5I8 and 5S2 ? 5I8 transitions arising from thethermally coupled levels remains almost constant (see inset to

Table 1Lifetime of 5F4 and 5S2 levels at three different temperatures.

Sample temperature (K) Decay time (ls)

5F4 ?5I8

5S2 ?5I8

300 1056 ± 30 1120 ± 39378 688 ± 11 729 ± 14473 518 ± 12 645 ± 14

M. Rai et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 97 (2012) 825–829 829

the Fig. 5). There is only a minor change of 0.04 for an increase intemperature from 300 to 550 K, which is almost negligible. It hasalready been reported by many authors that FIR based temperaturesensing behavior depends upon several factors including the sepa-ration between two thermally coupled levels and phonon fre-quency of the host lattice, etc [23]. For Ho3+, the separationbetween the two thermally coupled levels is �667 cm�1 and it var-ies slightly from host to host [24]. Y2O3 with a low phonon fre-quency �500 cm�1 is an efficient host for UC emission but theorder of phonon frequency of the host and separation betweentwo thermally coupled levels is almost same. Thus it is expectedthat for the present host, electrons maintain a thermal equilibriumbetween both the levels even at higher temperatures. Thus the FIRdoes not change appreciably and maintains almost a constantvalue.

This surmise has been further supported with the decay behav-ior of the transition 5F4 ? 5I8 and 5S2 ? 5I8 arising from the ther-mally coupled levels 5F4 and 5S2 with a variation in temperature.Lifetimes of both the levels decrease with an increase in tempera-ture, but in different ratio. The decay patterns of both the transi-tions with a variation in temperature are shown in Fig. 6. Table 1shows the quantitative values of decay time of 5F4 ?

5I8 and5S2 ? 5I8 transitions at three different temperatures. From thetable it is evident that the decrease in lifetime of 5F4 level is morerapid as compared to 5S2 level. This causes the population of 5F4 le-vel to decay faster with temperature. Therefore, as soon as the ionsin 5S2 level are promoted to 5F4 level due to thermal energy, theyquickly decay non-radiatively to 5S2 level again and thus FIR re-mains almost constant even with an increase in temperature.

Conclusions

Ho3+/Yb3+ co-doped Y2O3 phosphors were synthesized success-fully by combustion technique. The phosphor emits strong UCemission with most intense transition in green region along withrelatively weak emissions in red and blue regions. UC emissionfor green and red transitions shows saturation behavior at higherinput laser power due a change in pumping rates. In addition tothis, material shows sensitivity towards temperature and it is ob-

served that the overall fluorescence is reduced with an increasein sample temperature (external heating), however contrary tothe usual process in Ho3+ emission, the two thermally coupled lev-els (5F4 and 5S2) of Ho3+ do not show any change in intensity ratioof 5F4 ?

5I8 and 5S2 ?5I8 transitions for this host.

Acknowledgments

Authors acknowledge the financial support from UniversityGrants Commission, New Delhi, India. Ms. K. Mishra and Mr. R.K.Verma are thankful to Council of Scientific and Industrial Research(CSIR), New Delhi for Senior Research Fellowship.

Appendix A. Supplementary data

Supplementary data associated with this article can be found, inthe online version, at http://dx.doi.org/10.1016/j.saa.2012.07.071.

References

[1] F. Auzel, Chem. Rev. 104 (2004) 139–174.[2] F. Wang, X.G. Liu, Chem. Soc. Rev. 38 (2009) 976–989.[3] S. Heer, K. Kompe, H.U. Gudell, M. Haase, Adv. Mater. 16 (2004) 2102–2105.[4] A. Rapaport, J. Milliez, M. Bass, A. Cassanho, H. Jenssen, J. Disp. Technol. 2

(2006) 307–311.[5] R. Walti, W. Luthy, H.P. Weber, S.Y.A. Rusanow, A.A. Yakovlev, A.I. Zagumenyi, I.

Shcherbakov, A.F. Umiskov, J. Quant. Spectrosc. Radiat. Transfer 54 (1995)671–681.

[6] K. Binnemans, Chem. Rev. 109 (2009) 4283–4374.[7] K. Mishra, N.K. Giri, S.B. Rai, Appl. Phys. B 103 (2011) 863–875.[8] V.K. Tikhomirov, L.F. Chibotaru, D. Saurel, P. Gredin, M. Mortier, V.V.

Moshchalkov, Nano Lett. 9 (2009) 721–724.[9] H.A. Hoppe, Angew. Chem. 48 (2009) 3572–3582.

[10] B.K. Gupta, D. Haranath, S. Saini, V.N. Singh, V. Shanker, Nanotechnology 21(2010) 55607–55615.

[11] I. Klink, G.A. Hebbink, L. Grave, F.C.J.M. van Veggel, D.N.R. Reinhoudt, L.H.Slooff, A. Polman, J.W. Hofstraat, J. Appl. Phys. 86 (1999) 1181–1186.

[12] F. Wang, Y. Han, C.S. Lim, Y. Lu, J. Wang, J. Xu, H. Chen, C. Zhang, M. Hong, X.Liu, Nature 463 (2010) 1061–1065.

[13] R.D. Shannon, Acta Crystallogr. Sect. A 32 (1976) 751–767.[14] A.K. Singh, Sens. Actuators, A 136 (2007) 173–177.[15] R.K. Verma, S.B. Rai, J. Quant. Spectrosc. Radiat. Transfer. 113 (2012) 1594–

1600.[16] S.K. Singh, K. Kumar, S.B. Rai, Appl. Phys. B 94 (2009) 165–173.[17] R.L. Snyder, J. Fiala, H.J. Bunge, Defect and microstructure analysis by

diffraction, Oxford University Press, USA, 1999.[18] X. Wang, S. Xiao, Y. Bu, X. Yang, J.W. Ding, Opt. Lett. 33 (2008) 2653–2655.[19] B. Allieri, L.E. Depero, A. Marino, L. Sangaletti, L. Caporaso, A. Speghini, M.

Bettinelli, Mater. Chem. Phys. 66 (2000) 164–171.[20] S.K. Singh, A.K. Singh, S.B. Rai, Nanotechnology 22 (2011) 275703. 10pp.[21] S.K. Singh, K. Kumar, S.B. Rai, Sens. Actuators, A 149 (2009) 16–20.[22] H. Kusama, O.J. Sovers, T. Yoshioka, Jpn. J. Appl. Phys. 15 (1976) 2349–2358.[23] S.A. Wade, J.C. Muscat, S.F. Collins, G.W. Baxter, Rev. Sci. Instrum. 70 (1999)

4279–4282.[24] V.K. Rai, Appl. Phys. B 88 (2007) 297–303.