Investigation of the luminescent properties of Ce³⁺ doped and Ce³⁺/Mn²⁺ co-doped CaAl₂Si₂O₈ phosphors

Abstract

A series of Ce³⁺ doped and Ce³⁺/Mn²⁺ co-doped CaAl₂Si₂O₈ phosphors have been synthesized in a reducing atmosphere (Ar/H₂) via a solid-state route and their photoluminescent properties have been investigated at room temperature. For the CaAl₂Si₂O₈:Ce³⁺ phosphor, a broad emission band at 430 nm is observed in addition to characteristic bands of Ce³⁺ peaking at 334 nm and 356 nm. The extra band is assigned to the presence of cationic vacancies naturally created to counterbalance the substitution of Ce³⁺ ions for Ca²⁺ alkaline-earth species. Its intensity can be significantly reduced when Ca²⁺ cations are replaced by the reaction i.e. when the amount of alkaline earth vacancies induced by the presence of Ce³⁺ cations is minimized. Moreover, the co-doping Ce³⁺/Mn²⁺ goes along with the appearance of energy transfers that make possible fine tuning of the CIE(x, y) parameters from blue to warm white and even yellow.

---

1. Introduction

White light-emitting diodes (LEDs) have been extensively investigated because they are promising candidates as light sources for everyday life. In particular, they offer a huge variety of benefits such as low power consumption, high efficiency, long lifetime, high reliability and eco-friendliness.^1–5 To achieve white inorganic LEDs, three approaches are commonly considered: (1) the assembly of three LEDS emitting in the red, green and blue, respectively, (2) the use of a UV-LED chip coated with a blend of UV-excitable RGB phosphors, and (3) the combination of a yellow phosphor (namely, Y₃Al₂Si₃O₁₂:Ce³⁺) deposited at the surface of a blue chip (e.g. InGaN). Unfortunately, such devices still suffer from a poor colour rendering index (CRI of about 70%) that limits the full deployment of such an innovative technology. Moreover, the last technical solution, often favoured so far for cost reasons, exhibits a bright blue-halo effect that is currently subject of many controversies due to its potential risk for the retina of young children and elderly persons.⁶ Moreover, there still exist many problems to overcome for such devices to capture a share of the lighting market, such as the low stability of phosphor colour temperature with time and temperature, the untimely deterioration of phosphors, their uncorrelated lifetimes when mixed together in a blend, the difficulty in generating warm white lights, etc.⁷ In that context, an option to provide access to white UV-LEDs with high efficiency and reproducibility could consist in synthesising a single-phase phosphor with a broad emission band mimicking the solar spectrum and easily stimulated by a UV radiation.^8–13 This would alleviate the complicated optimization of adjusting the composition of phosphor blends and controlling the absorption of the blue emission, but it would require a good adaptability and flexibility of the host lattice to accommodate many substitutions to finely tune optical characteristics (i.e. emission spectrum, quantum yield, etc.).

In oxide-based luminescent materials, Ce³⁺ and Eu²⁺ ions are well-known to efficiently absorb UV rays, and to supply broad emission bands in the visible range.^14,15 Nevertheless, to achieve white light, co-doping seems to be inevitable even if the aforementioned activators may occupy distinguishable crystal sites leading to different emission wavelengths. In that context, the use of co-dopants can be of particular interest to adapt the emission spectra to specific requirements, for instance to enrich the emission of red and yellow. Namely, numerous possible energy transfers between luminescent centres can generate warm white light with enhanced colour rendering index. In that framework, Mn²⁺-doped materials are reported to exhibit emission bands ranging from about 500 to 700 nm depending on the strength of the crystal field experienced by the activator. Namely, a green emission band is observed for a weak crystal field in tetrahedral sites, whereas red luminescence is associated with a strong crystal field in octahedral sites. An intermediate crystal field will give rise to yellow or orange fluorescences, all of which are related to the ⁴T₁ → ⁶A₁ transition.^14,16,17 However, as the d–d electric dipole absorptions are forbidden by spin and parity, Mn²⁺ cations are very difficult to pump. A universal solution to figure out this situation consists in associating Mn²⁺ cations with sensitizers such as Eu²⁺ or Ce³⁺, i.e. materials with a high absorption cross-section.^8–13 This concept has already been applied to several matrices to provide potential phosphors for white LEDs.

Eu²⁺-doped alumino-silicate hosts with the CaAl₂Si₂O₈ composition are of great interest for luminescent materials because they exhibit efficient luminescence in the visible spectrum coupled with a high chemical stability. Moreover, they are also quite easy to synthesize.^9,18–21 In that context, Yang et al.⁹ investigated the energy transfer between Eu²⁺ and Mn²⁺ in CaAl₂Si₂O₈ and claim that this material could be regarded as a potential phosphor for white light UV-LED. If such an assertion holds under the stringent quality requirements for such a specific application, it is clear that the CIE chromatic parameters can be tuned continuously from (0.17, 0.11) to (0.33, 0.31) going from Ca_0.99Eu_0.01Al₂Si₂O₈ to Ca_0.74Eu_0.01Mn_0.25Al₂Si₂O₈ compositions under excitation at 354 nm. This prompted us to seek new single phase phosphors with adaptable chromatic characteristics. Besides, it should be recalled that Eu doped CaAl₂Si₂O₈ materials have also been reported as long persistent luminescent materials by Clabau et al.¹⁸ Consequently, in addition to producing white light, these materials might be used in the distant future for lighting displays supplied with electrical current in intermittent ways that could save energy. So far, no investigation has been carried out on energy transfers from Ce³⁺ to Mn²⁺ in this series of compounds. Thus we embarked on the elucidation of the optical properties of CaAl₂Si₂O₈ host lattices doped and co-doped with Ce³⁺ and Mn²⁺ cations.

2. Experimental section

2.1. Raw materials and synthesis process

Conventional high temperature solid state reactions were performed to synthesize Ce³⁺/Mn²⁺ co-activated CaAl₂Si₂O₈ compounds (hereafter abbreviated CASO). Oxide and carbonate materials, i.e. CaCO₃ (99.997%, Alfa Aesar), SiO₂ (99.99% Chempur), γ-Al₂O₃ (99.99%, Alfa Aesar), Ce₂(CO₃)₃ (99.5%, Rhodia), and MnO (99.99%, Alfa Aesar) were used as starting materials. Precursors were weighed in stoichiometric amounts and ball milled with a Fritsch Pulverisette 7 for two hours in ethanol. After drying at 100 °C, mixtures were annealed at 1350 °C for 50 h in argon/hydrogen (95/5 vol%) atmosphere.

2.2. Structure characterizations

The phase purity and crystal structure of the powder were examined by powder X-ray diffraction (XRD) profiles which were measured with a Bruker AXS D8 advanced automatic diffractometer with Cu K-L₃ radiation (germanium monochromator) operated at 40 kV and 40 mA. Furthermore, the structure refinements based on Gebert's crystal investigations²² are carried out with the Jana2006 Beta version software.²³ The instrumental function was expressed in terms of the geometry of the diffractometer with the relevant parameters which reported in Table 1. The XRD patterns were collected in the 10–90° 2θ range and all the doped samples diffraction peaks are in good agreement with those reported in JCPDS files 89-1462 (CASO) as exemplified in Fig. 1 for CASO:9%Ce³⁺, 6%Mn²⁺. No characteristic peaks of the dopants (Ce³⁺/Mn²⁺) were observed after a full pattern matching analysis, which means a solid-state solution is produced in all the samples to be researched. Note here, the experimental spectrum is not fitted well with the calculate spectrum. Three reasons can be put forward to explain, (i) the sample is the powder diffraction and CASO structure is triclinic which has lower symmetry, these factors easily induce overlap of the diffraction peaks and, (ii) due to the distort of the bond angle and length induced by the different radii between the Eu³⁺ and Ca²⁺ ions and, (iii) in order to make the charge balance, there would exists lots of holes and many Eu³⁺ ions may go into the interval of the lattice also caused the bad refinement induced by the different valence between Ce³⁺ and Ca²⁺ ions. Meanwhile, the lattice parameters for Ca_0.91Mn_0.09Al₂Si₂O₈, Ca_0.865Ce_0.09Al₂Si₂O₈ and Ca_0.775Ce_0.09Mn_0.09Al₂Si₂O₈ powder were also calculated based on the experimental XRD profiles with cell refinement software (Jana2006) by using a Rietveld procedure with the fundamental parameter approach in the I1̄ space group (Table 2).

Table 1 Instrumental data used for Rietveld refinements of CASO and its Ce³⁺, Mn²⁺-doped derivatives

Primary and second radius	217.5 mm
Receiving slit length	16 mm
Glancing angle	13.65°
Source and sample length	12 mm
Primary soller slit aperture	2.5°
Reception slit divergence angle	0.2°
Receiving slit width	0.1 mm
Peak-shape function	Lorentzian

---

Fig. 1 Observed, calculated and difference X-ray diffraction pattern of an Ce³⁺ (0.09) and Mn²⁺ (0.06)-co-doped CASO phosphor in the [10–45] 2θ range (inset is given the total pattern).

Table 2 Crystallographic details for Ca_0.925Eu_0.05Al₂Si₂O₈ and CASO HL

Formula	Ca0.91Mn0.09Al2Si2O8	Ca0.865Ce0.09Al2Si2O8	Ca0.775Ce0.09Mn0.09Al2Si2O8
a According to Angel et al., the anorthite form of CASO may crystallize in the I1̄ space group with similar cell parameters than those reported in Table 2. Then the multiplicity of the Wyckoff sites occupied by calcium cations is doubled compared to P1̄ SG but the occupancy rate is about 50%. So far, based on X-ray patterns collected in this study, we couldn't distinguish between the two models and we systematically privileged the I1̄ SG.b Rwp = weighted profile residual factor.c GOF = goodness-of-fit on F^2.
Formula weight (g mol^−1)	2236.36	2286.26	2373.11
Cryst. syst.	Anorthic(triclinic)	Anorthic(triclinic)	Anorthic(triclinic)
Space group^a	I1̄	I1̄	I1̄
a (Å)	8.1721(3)	8.1917(3)	8.1778(3)
b (Å)	12.8683(4)	12.8763(5)	12.8695(4)
c (Å)	14.1643(5)	14.1741(5)	14.1628(5)
α (deg.)	93.348(3)	93.069(3)	93.312(3)
β (deg.)	115.769(2)	115.688(2)	115.716(2)
γ (deg.)	91.156(2)	91.212(3)	91.135(2)
V (Å^3)	1337.39(10)	1343.74(10)	1339.01(10)
Z	8	8	8
Dcalcd (g cm^−1)	2.7629(2)	2.7493(3)	2.7590(2)
Rp (%)	10.79	11.10	12.58
Rwp (%)^b	13.56	14.50	16.08
GOF^c	1.50	1.33	1.59

---

In Fig. 2 are reported the evolution of refined lattice volume for CASO matrixes with different dopant concentrations (Ce³⁺ and Mn²⁺ separately). For the CASO matrix, the slight difference of ionic radii²⁴ between Ce³⁺ and Ca²⁺ (r (Ce³⁺) = 1.01 Å and r (Ca²⁺) = 1.00 Å when CN = 6, r (Ce³⁺) = 1.07 Å and r (Ca²⁺) = 1.06 Å when CN = 7) leads to a continuous increase of the lattice volume. At the opposite, a shrink of the cell volume is observed when Ca²⁺ cations are replaced by Mn²⁺ ones (r (Mn²⁺) = 0.67 Å and 0.90 Å when CN = 6 and 7, respectively). All these results indicate that in all cases, a solid solution is formed because the evolution of the cell parameters regularly evolves with dopant ions concentration.

Fig. 2 The variations in the lattice volume with Ce³⁺ and Mn²⁺ concentrations in CASO.

2.3. Optical measurements

The photoluminescence (PL) and PL excitation (PLE) spectra were recorded at room temperature on a Fluorolog-3 spectrofluorometer (Jobin Yvon Instruments) equipped with a xenon light as excitation source. Emission and excitation spectra were corrected for the contribution of the monochromator and the wavelength dependence of the lamp output. Chromaticity coordinates in the CIE 1931 color space were calculated by using the OptPropMatlab toolbox.²⁵ The room PL decays were acquired with a streak camera Hamamatsu C7700 coupled to an imaging spectrograph. The excitation line was obtained with a Spectra Physics Hurrican X laser system (800 nm, 100 fs, 1k Hz). Samples were excited at λ_ex of 267 nm by the frequency-tripled Ti sapphire laser line.

3. Results and discussion

From structure point of view, CASO has both layered and framework structure. In the layered structure, the Ca²⁺ ions are located in the interlayers of the double tetrahedral layer and are expected to exchange readily through the interlayers. According to the Wyckoff, CASO structure contains four crystallographically different Ca sites at the same Wyckoff position, 4i (Fig. 3). Due to the charge balance, all the four sites are not full occupied and induced the disorder of bond angle and length. We can divide the four Ca²⁺ sites into two groups in accord to their different number of folders. Namely, one type of Ca²⁺ ion occupies an octahedral site with six oxygen atoms and the average Ca–O bond distance is 2.485 Å. Other Ca²⁺ ions occupy three kinds of polyhedral sites with seven coordinated oxygen atoms and their average bond distances are 2.508, 2.531, and 2.562 Å, respectively (Fig. 4). Therefore, it is expected that the Ce³⁺ ions doped in this host will exhibit two kinds of luminescence properties according to their different energy level. Al and Si atoms both occupy tetrahedral sites with four coordinated oxygen atoms, and the average bond distances for Al–O and Si–O are 1.735 and 1.611 Å, respectively. On the basis of steric considerations, it is clear that Ce³⁺ cations will occupy exclusively alkaline-earth crystallographic sites like Eu²⁺ cations in Eu²⁺ MASO derivatives (M = Ca, Sr, Ba). In the same way, the ionic radii of Mn²⁺ cation in the different polyhedra tend to suppose their presence in the alkaline-earth sites. This assumption was a posteriori verified by Yang and co-workers⁹ and their investigations of optical properties of Mn²⁺-doped CASO phases.

Fig. 3 CASO structure represented in cell along the c axis.

Fig. 4 Calcium environments and bond length in CASO (Wyck.: 4i).

3.1. Optical properties of C_1−3x/2ASO:xCe³⁺

The PLE and PL spectra of CASO:2%Ce³⁺ are displayed in Fig. 5. PLE spectra are composed of two broad bands peaking at 250 and 294 nm for CASO (λ_em = 356 nm). Without doubt, these absorption bands are associated with 4f–5d intra-atomic electronic transitions of Ce³⁺, the d-block being split as usual by the ligand field. Since the optical gap of the CASO host lattice is estimated around 4 eV (310 nm),¹⁸ it can be speculated that the 4f-block of Ce³⁺ lies at the top of the valence band. Formally, on the basis of optical and XPS investigations carried out on CaAl₂Si₂O₈:Eu,¹⁸ the Eu-d block is expected to lie just below the conduction band and the 4f⁷ block of Eu²⁺ cations. Thus, as the position in energy of d-orbitals is not expected to significantly change within the lanthanide series²⁶ this leads us to conclude on the aforementioned positioning of the 4f⁷ block of Ce³⁺.

Fig. 5 PLE and PL spectra for phosphors with composition of CASO:2%Ce³⁺.

Emission spectra were monitored at both excitation wavelengths, i.e. 294 and 250 nm. No significant evolution in the shape of PL was noticed vs. excitation wavelengths. The PL spectra of the CASO:Ce samples exhibit an intense band at 356 nm with a shoulder at 334 nm and a much broader band around 430 nm. The two emission bands placed at low wavelength can be assigned to Ce³⁺ cations occupying the Ca²⁺ sites. Formally, in sake of clarity, PL spectra of CASO:9%Ce sample has been deconvoluted by Gaussian functions taking into account the ∼0.25 eV energy separation between the 5d → ²F_5/2 and 5d → ²F_7/2 transitions due to spin-orbit coupling.¹⁴ This clearly leads for CASO:Ce compound to distinguish the contributions of six-fold and seven-folded coordinated Ce³⁺ cations (Fig. 6), the former giving rise to emissions at 352 and 379 nm, the latter to emissions at 327 and 355 nm. This assignment, based only on the widespread postulate that the higher the coordination of Ce³⁺ cations, the lower the crystal field splitting and the shorter the emission wavelength, must be moderated by the existence of counter-examples.^27,28

Fig. 6 Emission spectra of CASO:9%Ce³⁺ with decomposition in Gaussian curves under λ_ex = 294 nm.

If the assignment of Ce³⁺ emission bands to the luminescence spectrum of Ce doped CASO compound is elucidated, there still remains an unexpected broad peak (hereafter called “defect” peak) at 430 nm for CASO. To shed light on the origin of this unassigned peak, PLE spectra were monitored at these two wavelengths (Fig. 7). Their examination clearly shows that they differ from the PLE of Ce³⁺. At very first sight, based on previous investigations carried out on Eu activated CASO phosphor, emissions at 430 nm might be attributed to 4f⁶5d¹ → 4f⁷(⁸S_7/2) transitions of Eu²⁺ possibly present as undesired traces since this activator gives rise to luminescence bands peaking around 440 nm for CASO host lattices.¹⁸ To address this issue, samples were prepared in new alumina crucibles with fresh precursors. They systematically led to materials with the same optical characteristics, i.e. a broad “defect” emission band in addition to those of Ce³⁺. The determination of the decay times for this band in CASO was then initiated and revealed lifetimes of about 22 ns for CASO:Ce samples which turns to be very different to that commonly expected for Eu²⁺ (ref. 29) and that collected for CASO:Eu materials by Yang et al. (i.e. 60–70 μs).⁹ Consequently, the assignment of these unforeseen bands to the presence of Eu²⁺ luminescent centers can be unambiguously eliminated. This is in good agreement with the observation that room temperature excitation spectra of CASO:Ce³⁺ at 430 nm differ from those of CASO:Eu²⁺ at 440.

Fig. 7 PLE spectra for CASO:2%Ce³⁺ monitored at 356, 430 nm.

To explain the presence of these broad bands in the absence of Eu²⁺ cations, one hypothesis can be proposed. Namely, the substitution of Ce³⁺ ions for Ca²⁺ alkaline-earth specie must be combined with the creation of vacancies according to the formal equation 3Ca²⁺ → 2Ce³⁺ + V_Ca to respect the charge balance. Thus, “defect” emissions would be correlated with the presence of these alkaline-earth vacancies (probably distributed in close proximity of for electrostatic reasons). This would explain why the intensity of such bands increases with cerium concentration as observed experimentally (Fig. 8).

Fig. 8 Evolution PL of CASO:Ce³⁺ and overlap between Ce³⁺ and ‘defects’ (insert).

To test this hypothesis, investigations were undertaken on CASO:La³⁺, CASO:La³⁺, K⁺ and CASO:Ce³⁺, K⁺ materials (prepared with K₂CO₃ (99.99% Alfa Aesar) and La₂O₃ (99.9% Alfa Chempur) precursors), as well as CASO host lattices themselves. Fig. 9 displays the PL of CASO, CASO:5%Ce, CASO:5%La, CASO:5%Ce, 5%K and CASO:5%La, 5%K monitored at 294 nm. First, undoped CaAl₂Si₂O₈ host lattices do not luminesce at all. Second, CASO:La³⁺ materials exhibit only the defect emissions, i.e. similar spectra than those of CASO:Ce³⁺ missing the 5d → 4f contributions. Third, the intensities of the defect bands strongly decrease when K⁺ ions are co-inserted in the host lattices with lanthanide cations in equimolar concentration. Namely, the alkali metal plays the role of charge compensator according to the formal reaction that diminishes the amount of alkaline earth vacancies naturally created under La³⁺ substitution. On the basis of the examination of emission spectra of CASO:Ce and coupled with decay time measurements, are clearly indicative that the emission bands at 430 nm in Ce doped CASO are related to alkaline earth vacancies and not to the presence of undesired Eu²⁺ cations. CASO:Ce, K, it is worth noting that the intensities of the 5d → ²F_5/2 and 5d → ²F_7/2 Ce contributions are not affected by the insertion of K⁺ and that the “defect” bands cannot be totally annihilated by the insertion of K⁺ cations suggesting that V_Ca vacancies still exist. Thus, we can conclude that Ce³⁺ (La³⁺) cations are unfortunately systematically introduced in larger amount than K⁺ cations. Regardless, the three aforemention observations, coupled with decay time measurements, are clearly indicative that the emission bands at 430 nm in Ce doped CASO is related to alkaline earth vacancies and not to the presence of undesired Eu²⁺ cations.

Fig. 9 PL spectra monitored at 294 nm for CASO HL and Ce³⁺, La³⁺, Ce³⁺/K⁺ and La³⁺/K⁺ co-doped materials.

The dependence of the PL of Ca_1−3x/2ASO:xCe³⁺ materials (x lower than 16%) are depicted in Fig. 8. The emissions associated with Ce³⁺ lines increase up to quenching concentrations of about 9% for CASO. Beyond these limits, the intensity of 5d → ²F_5/2 and 5d → ²F_7/2 lines decreases whereas the one of the “defect” bands continues to slightly increase and to tend towards a plateau. This trend maybe explained from the fact that (i) the higher the Ce³⁺ concentration the higher the Ca²⁺ vacancies, (ii) the existence of possible energy transfers from Ce³⁺ to V_Ca as suggested by the spectral overlap between the Ce³⁺ PL spectra and the “defect” PLE spectra in CASO:Ce³⁺ materials (the inset in Fig. 8), and (iii) a quenching phenomenon specific to defects themselves.

3.2. Optical properties of C_1−yASO:yMn²⁺

The PL and PLE spectra for purely Mn²⁺ activated CASO host lattices were also recorded (Fig. 10). They tend to be very similar to those reported in the literature.^9,21 In brief, excitation spectra consist for both host matrix of several bands in the UV and visible regions corresponding to the ⁴A₁(⁶S) to ⁴E(⁴D), ⁴T₂(⁴D), [⁴A₁(⁴G), ⁴E(⁴G)], ⁴T₂(⁴G), and ⁴T₁(⁴G) transition levels, respectively. Concerning the emission spectra, the main peak is located at 563 nm for CASO. Due to the photoluminescence property of Mn²⁺ ions and the low dopant concentration in the CASO matrix, it is difficult to observe the PL and PLE spectra for the C_0.995ASO:0.5%Mn²⁺ sample. However, excitation and emission can be observed for higher dopant rate. Mn²⁺ ion in solid materials is usually characterized by a green emission in tetrahedral coordination and by a yellow-red emission in octahedral coordination.^10,21 Therefore, the emission at 568 nm also suggests that Mn²⁺ cations are located at the alkaline-earth site instead of being at the Al or Si sites as already proposed by for CASO⁹ which was expected on the basis of steric considerations. As the d–d transitions of Mn²⁺ ion are difficult to pump, the transitions energy between the ground state and excited sub-levels states are weak. Thus, we can suppose that the energy differences between the Mn₁ (occupying the Ca₁ site) and the other sites Mn_2,3,4 (occupying the Ca_2,3,4 sites) are also weak and thus not easy to distinguish. Under this consideration, we cannot give proper assignments to the emission and excitation spectra of Mn²⁺. Furthermore, the large Stocks shift (0.9 eV) of Mn²⁺ centers in CASO matrix indicates that a strong electron–lattice interaction occurs inducing a bad resolution of emission band of Mn²⁺.

Fig. 10 PL and PLE spectra for purely Mn²⁺ activated CASO.

3.3. Overlap of the PL of CASO:Ce and PLE of CASO:Mn

The comparison of the PL spectra for CASO:Ce³⁺ and the PLE spectra for CASO:Mn²⁺ reveals a heavy spectral overlap (Fig. 11). Thus, energy transfers from Ce³⁺ to Mn²⁺ can be anticipated as already exemplified in many phosphors including Ca₈MgLa(PO₄)₇,³⁰ Ca₃Sc₂Si₃O₁₂,³¹ SrAl₂O₄,³² BaAl₂O₄,³³ Ba₂Ca(BO₃)₂,⁸ Ca₉Y(PO₄)₇ (ref. 10) or LiCaBO₃.³⁴ Nevertheless, in contrasts to these materials, transfers could also take place in the case of CASO from alkaline-earth-vacancy based defects to Mn²⁺ and from Ce³⁺ to Mn²⁺ via V_Ca according to the proposed Fig. 12.

Fig. 11 Spectral overlap between the photoluminescence spectrum of C_0.865ASO:9%Ce³⁺ (solid line) and photoluminescence excitation spectrum of C_0.91ASO:9%Mn²⁺.

Fig. 12 Possible energy transfers in CASO:Ce, Mn compounds.

Moreover, Fig. 13 gathers the PLE of CASO:9%Ce³⁺ monitored at 353 nm (Ce³⁺ emission), and CASO:9%Ce³⁺, 9%Mn²⁺ monitored at 353 nm and 568 nm (Mn²⁺ emission). It can be observed for CASO:Ce, Mn that photoluminescence excitation spectrum monitored by the Mn²⁺ emission (λ_em = 568 nm, green line) has a profile very similar to that of the excitation band monitored by the Ce³⁺ emission (λ_em = 353 nm, red line). This reveals that the Mn²⁺ ions can be essentially excited in the same energy range than Ce³⁺ ions. This observation, coupled with the strong decrease of the PLE going from CASO:9%Ce (black line) to CASO:9%Ce, 9%Mn (red line) confirms also that an ET occurs from to Mn²⁺. The photoluminescence excitation spectra of the Mn²⁺ ions in the Ce³⁺ and Mn²⁺ co-activated sample extend from 230 to 340 nm, indicating that this phosphor could be, in principle, used in a UV white-light LED.

Fig. 13 Photoluminescence excitation of CASO:9%Ce³⁺ (black line), CASO:9%Ce³⁺, 9%Mn²⁺ (red line) under λ_em = 353 nm and CASO:9%Ce³⁺, 9%Mn²⁺ (green line) under λ_em = 568 nm.

3.4. Optical properties of co-doped CASO:Ce, Mn

Emission spectra of Ce³⁺/Mn²⁺ co-doped CASO is reported in Fig. 14. Ce³⁺ concentration has been fixed at 9% for CASO, namely, at the quenching concentration to favour energy transfers. On PL spectra, it can be observed that the insertion of Mn²⁺ in CASO:Ce compounds triggers a decline of the intensity of the emission bands associated with both Ce³⁺ and “defects” concomitant with an enhancement of the Mn²⁺ emission. The decrease of the Ce³⁺ emission is regular and continuous with Mn²⁺ concentration, whereas the Mn²⁺ emission progressively increases to a Mn insertion rate of about 9% before decreasing due to concentration quenching.

Fig. 14 Photoluminescence spectra for C_0.865−yASO:9%Ce³⁺, yMn²⁺ phosphors versus the Mn²⁺ doping content (y), under λ_ex = 294 nm.

To quantify the optical characteristics of phosphors and to judge the impact of Mn²⁺ co-doping on the generated color, the CIE(x, y) parameters of prepared materials have to be determined. These values are reported below as shown in Tables 3 and 4 for CASO:Ce, CASO:Ce, Mn, respectively. All data were collected with an excitation wavelength of 294 nm.

Table 3 CIE chromaticity coordinates for phosphors C_1−3x/2ASO:xCe³⁺ (λ_ex = 294 nm)

C1−3x/2ASO:xCe^3+	CIE(x, y)
x = 0.005	(0.1976, 0.1564)
x = 0.02	(0.2162, 0.1791)
x = 0.05	(0.2148, 0.1876)
x = 0.09 (point 1 in Fig. 15)	(0.2270, 0.2212)
x = 0.15	(0.2022, 0.1868)

---

Table 4 CIE chromaticity coordinates for phosphors C_0.865−yASO:0.09Ce³⁺, yMn²⁺ (0.01 < y < 0.15)

C0.865−yASO:0.09Ce^3+, yMn^2+	CIE(x, y)
x = 0.01	(0.2968, 0.3142)
x = 0.015	(0.3176, 0.3242)
x = 0.02	(0.3416, 0.3342)
x = 0.03	(0.3577, 0.4016)
x = 0.06	(0.3982, 0.4535)
x = 0.09	(0.4272, 0.4794)
x = 0.12	(0.2022, 0.1868)
x = 0.15	(0.4551, 0.4797)

---

For the C_0.865−yASO:9%Ce³⁺, yMn²⁺ (y = 0 to 15%) series, a regular shift from blue for y = 0% to white (y = 1, 1.5 and 2%) and to yellow (y > 3%) is clearly evidenced (Fig. 15 and Tables 3 and 4). This corresponds to a change from (0.2270, 0.2212) to (0.4551, 0.4797) of the chromatic coordinates for Mn²⁺ concentrations increasing from 0 to 15%. This clearly evidences the role of ET from Ce (and defect) towards Mn²⁺ cations in these materials and the importance to control the Mn concentration to generate a white color. The white color would be obtained for a Ca_0.845Ce_0.090Mn_0.02Al₂Si₂O₈ composition, this composition being probably not optimal.

Fig. 15 Evolution of CIE chromaticity coordinates for C_0.865−yASO:0.09Ce³⁺, yMn²⁺ excited at 294 nm. The numbers from 1–9 correspond to: 1, y = 0; 2, y = 0.01; 3, y = 0.015; 4, y = 0.02; 5, y = 0.03; 6, y = 0.6; 7, y = 0.09; 8, y = 0.12; 9, y = 0.15.

3.5. Mechanism of the ET from Ce³⁺ to Mn²⁺

At present, we will examine the efficiency (η_T) of the ET from to Mn²⁺. Generally, the ET efficiency from a sensitizer to activator can be expressed by the following eqn (1):^10,19,33where η_T is the ET efficiency and I₀ and I_S are the luminescence intensity of a sensitizer in the absence and presence of an activator, respectively. In our case, I_S0 is the intrinsic emission intensity of the sensitizer (Ce³⁺) and I_S is the emission intensity of the sensitizer (Ce³⁺) in the presence of the activator (Mn²⁺) for identical cerium concentrations. Fig. 16 shows the results of ET efficiency from Ce³⁺ to Mn²⁺ calculated by the previous eqn (1) for the C_0.865−yASO:9%Ce³⁺, yMn²⁺ (y = 1 to 15%) samples. As shown in Fig. 16, the ET efficiency increases with increasing Mn²⁺ concentration. However, the slope of the emission intensity gradually decreases when Mn²⁺ concentration increases. This indicates that the quantity of energy that can transfer from Ce³⁺ to Mn²⁺ is gradually restricted for high Mn²⁺ concentration due to: (1) the fixed concentration of Ce³⁺ and (2) the concentration quenching of Mn²⁺ to a lesser extent.

Fig. 16 Dependence of the energy transfer efficiency η_T on the Mn²⁺ content (y) for the C_0.865−yASO:9%Ce³⁺, yMn²⁺ samples under λ_ex = 294 nm.

Normally, two types of resonant ET mechanism exist. One corresponds to exchange interactions and the other to multipolar interactions.^10,14,19 It is well-known that if ET is based on exchange interaction, the critical distance between the sensitizer and activator should be shorter than about 5 Å. We previously saw that the critical distance can be calculated from the quenching concentration. From Fig. 16, the critical concentration can be determined to be about 0.161 (= 0.09 (Ce³⁺ concentration) + 0.071 (Mn²⁺ concentration)) from the total concentration of Ce³⁺ and Mn²⁺ for which the ET efficiency is 0.5. Therefore, the critical distance R_C (Ce³⁺–Mn²⁺) was calculated to be ∼15.21 Å by using the eqn (2),^10,11,13,14 where V is the unit-cell volume of the sample, N is the number of formula units per unit cell and X_C is the critical concentration. This value is too large to lead to exchange interaction via orbital overlaps. Consequently, the electric multipolar interaction must be considered for ET between Ce³⁺ and Mn²⁺ ions in the CASO matrix.

On the basis of the Dexter's ET formula of multipolar interaction and Reisfeld's approximation, the following eqn (3) can be obtained:^9–11,21

where I_S0 and I_S have the same meaning as above. C is the concentration of Mn²⁺. α equal to 6, 8, 10 for dipole–dipole, dipole–quadrupole, quadrupole–quadrupole interactions, respectively. Plots of I_S0/I_S based on the above equation are shown in Fig. 17. As one can see, linear behavior is only observed when α = 6, which implies that ET from Ce³⁺ to Mn²⁺ occurred via a dipole–dipole mechanism. This is quite surprising because a dipole–quadripole interactions was expected however this was already observed in the literature for several series of materials.³⁵

Fig. 17 Dependence of I_S0/I_S of Ce³⁺ versus α = 6, 8 and 10 for the samples C_0.865−yASO:9%Ce³⁺, yMn²⁺.

To confirm these results based on intensity measurements only, the decay curves of the C_0.865−yASO:9%Ce³⁺, yMn²⁺ (y = 0 to 15%) samples were also collected. These ones are depicted in Fig. 18 while decay times are reported in Table 5. For the first three samples, the decay curve was best fitted with a monoexponential function whereas for others samples the decays curves were fitted with a biexponential function.

Fig. 18 Decay curve obtained under excitation λ_ex = 267 nm and emission λ_em = 360 nm (centered at the maximum emission of Ce³⁺ in the CASO matrix). Dotted lines represent the experimental value and solid lines represent the fitting value.

Table 5 The results of the decay times for the samples C_0.865−yASO:9%Ce³⁺, yMn²⁺ (y = 0 to 15%) (λ_em = 360 nm)

Phosphors	A1	A2	τ1 (ns)	τ2 (ns)	〈τ〉 (ns)
Ce9Mn0	—	—	23.524	—	23.524
Ce9Mn1	—	—	22 665	—	22.665
Ce9Mn3	—	—	20 280	—	20.280
Ce9Mn6	8.1	91.9	0.682	19 654	18.117
Ce9Mn9	7.0	93.0	3.067	17.963	16.920
Ce9Mn12	12.8	87.2	0.663	17.268	15.143
Ce9Mn15	17.5	82.5	4.697	17.803	12.073

---

From Table 5 and Fig. 18, one can also notice that the meantime value 〈τ〉 decreases when the concentration of Mn increases. Thus we can further confirm that the ET from Ce³⁺ to Mn²⁺ occurs in the CASO matrix. Furthermore, we can also conclude that the ET is a resonant non-radiative transfer because the decay times of the Ce³⁺ decreases with the Mn²⁺ ion concentration. The plots of τ_S0/τ_S that are calculated from the previous eqn (3) are reported in Fig. 19. This plot also proves that the dipole–dipole interactions are at work (α = 6) even if dipole–quadripole interactions cannot be now totally rebutted (α = 8).

Fig. 19 Dependence of τ_S0/τ_S of Ce³⁺ on α = 6, 8 and 10 for the samples C_0.865−yASO:9%Ce³⁺, yMn²⁺.

As mentioned above, an ET also occurs between the defects emission and the Mn²⁺ ions in the CASO matrix. The ET efficiency from to Mn²⁺ ions in the samples C_0.865−yASO:9%Ce³⁺, yMn²⁺ (y = 1 to 15%) phosphors were also calculated and the results are presented in Fig. 20. As one can see, the ET efficiency increases gradually with the increase of Mn²⁺ concentration. The rate of increase of the emission intensity dramatically decreases with increasing Mn²⁺ concentration and the efficiency in the transfer is higher than in the Ce³⁺–Mn²⁺ when compared to the ET from Ce³⁺ to Mn²⁺ shown in Fig. 16. These phenomena in the maybe due to a better overlap between the emission of the (∼420 nm) with the excitation of the Mn²⁺ (∼403 nm) compared to the emission of Ce³⁺ at 356 nm.

Fig. 20 Dependence of the energy transfer efficiency η_T from to Mn²⁺ (on the Mn²⁺ content (y)).

The critical concentration for energy transfer from to Mn²⁺ ions was also calculated and estimated to be around 17 Å for a concentration of ∼0.045 and a quenching concentration for Mn²⁺ of 2.43%. In the same way as previously described, the mechanism of multipolar interactions can be expected. Fig. 21 was used to determine the intensities of the emissions from the defects for samples C_0.865−yASO:9%Ce³⁺, yMn²⁺. For those samples, the Ce³⁺ concentration is fixed and consequently, the content of defects also . As one can see in Fig. 21, a linear behavior is only observed when α is equal to 6, meaning that a dipole–dipole interactions also occurs between vacancies and Mn²⁺ ions.

Fig. 21 Dependence of I_S0/I_S of defects emission . versus α = 6, 8 and 10 (fixed Ce³⁺ (thus, ), change the concentration of Mn²⁺).

4. Conclusions

In summary, we have synthesized and investigated the luminescent properties of CaAl₂Si₂O₈ phosphors (co)-activated with Ce³⁺ and Mn²⁺. The results have shown that the substitution of the alkaline-earth ions by Ce³⁺ (and La³⁺) implies the systematic creation of cationic vacancies even when alkali metals are introduced as counter-cations to prevent their formation. These vacancies strongly impact the photoluminescence, leading to broad emission bands at 430 nm in CaAl₂Si₂O₈ host lattice. Their intensities are naturally reinforced by enhancing their concentration when Ce doping increases (at least up to 16%Ce) but also by energy transfer mechanism from Ce³⁺ cations. Furthermore, Ce doped CASO compounds give rise to a cold blueish white luminescence. Changes in the spectral characteristics can be initiated by the use of Mn²⁺ cations as co-dopants. Then, due to energy transfers towards Mn²⁺ centres, emission shifts continuously from blue-white, to white, and even yellow-white vs. Ce and Mn concentrations.

Acknowledgements

The author acknowledges S. Jobic, P. Deniard, H. Brault and F. Massuyeau for the technical assistance and fruitful discussion in IMN, Nantes University, CNRS, FRANCE. This work has been supported by the Chinese Scholarship Council (CSC), no. 2009615018.

Notes and references

1. C. You, Opt. Eng., 2005, 44, 111307.
2. T.-S. Chan, R.-S. Liu and I. Baginskiy, Chem. Mater., 2008, 20, 1215–1217.
3. C. Feldman, T. Justle, C. R. Ronda and P. J. Schmidt, Adv. Funct. Mater., 2003, 13, 511.
4. H. A. Hoppe, Angew. Chem., Int. Ed. Engl., 2009, 48, 3572–3582.
5. R.-J. Xie, N. Hirosaki, T. Suehiro, F.-F. Xu and M. Mitomo, Chem. Mater., 2006, 18, 5578–5583.
6. N. Cheung and T. Y. Wong, Surv. Ophthalmol., 2007, 52, 180–195.
7. D. A. Steigerwald, J. C. Bhat, D. Collins, R. M. Fletcher, M. O. Holcomb, M. J. Ludowise, P. S. Martin and S. L. Rudaz, IEEE J. Sel. Top. Quantum Electron., 2002, 8, 310–320.
8. C. Guo, L. Luan, Y. Xu, F. Gao and L. Liang, J. Electrochem. Soc., 2008, 155, J310–J314.
9. W.-J. Yang, L. Luo, T.-M. Chen and N.-S. Wang, Chem. Mater., 2005, 17, 3883–3888.
10. C.-H. Huang, T.-W. Kuo and T.-M. Chen, ACS Appl. Mater. Interfaces, 2010, 2, 1395–1399.
11. N. Guo, Y. Huang, H. You, M. Yang, Y. Song, K. Liu and Y. Zheng, Inorg. Chem., 2010, 49, 10907–10913.
12. S. Ye, F. Xiao, Y. X. Pan, Y. Y. Ma and Q. Y. Zhang, Mater. Sci. Eng., R, 2010, 71, 1–34.
13. G. Li, D. Geng, M. Shang, C. Peng, Z. Cheng and J. Lin, J. Mater. Chem., 2011, 21, 13334–13344.
14. G. Bless and B. C. Grabmaire, Luminescent Materials, Springer Verlag, Berlin, 1994.
15. C. Ronda, Luminescence: From Theory to Applications, Willey-VCH Verlag GmbH & Co., Weinheim, 2008.
16. S. C. Gedam, S. J. Dhoble and S. V. Moharil, J. Lumin., 2007, 124, 120–126.
17. S. H. M. Poort, A. Meijerink and G. Blasse, Solid State Commun., 1997, 103, 537–540.
18. F. Clabau, A. Garcia, P. Bonville, D. Gonbeau, T. Le Mercier, P. Deniard and S. Jobic, J. Solid State Chem., 2008, 181, 1456–1461.
19. J. K. Park, J. M. Kim, E. S. Oh and C. H. Kim, Electrochem. Solid-State Lett., 2005, 8, H6–H8.
20. C. Zhang, J. Yang, C. Lin, C. Li and J. Lin, J. Solid State Chem., 2009, 182, 1673–1678.
21. S. Ye, Z.-S. Liu, X.-T. Wang, J.-G. Wang, L.-X. Wang and X.-P. Jing, J. Lumin., 2009, 129, 50–54.
22. W. Z. Gebert, Kristallografiya, 1972, 135, 437.
23. V. D. petricek and L. Palatinus, The Crystallographic Computing System JANA, 2006 Beta, Academy of Sciences, Praha, Czeck Republic, 2006.
24. R. D. Shannan, Acta Crystallogr., Sect. A: Cryst. Phys., Diffr., Theor. Gen. Crystallogr., 1976, 32, 751.
25. J. Wagberg, Matlab Toolbox for Calculation of Color Related Optical Properties-version 2.1, More Research and DPC Digital Printing Center, 2007.
26. P. Dorenbos, J. Lumin., 2003, 105, 117–119.
27. G. Denis, P. Deniard, E. Gautron, F. Clabau, A. Garcia and S. Jobic, Inorg. Chem., 2008, 47, 4226–4235.
28. G. Denis, X. Rocquefelte, P. Deniard, M.-H. Whangbo and S. Jobic, J. Mater. Chem., 2009, 19, 9170–9175.
29. C. Fouassier, Luminescence in Encyclopedia of Inorganic Chemistry, New York, J. Wiley, 2nd edn, 2005.
30. Y. N. Xue, F. Xiao and Q. Y. Zhang, Spectrochim. Acta, Part A, 2011, 78, 1445–1448.
31. Y. Liu, X. Zhang, Z. Hao, Y. Luo, X. Wang, L. Ma and J. Zhang, J. Lumin., 2013, 133, 21–24.
32. X. Xu, Y. Wang, X. Yu, Y. Li and Y. Gong, J. Am. Ceram. Soc., 2011, 94, 160.
33. N. Suriyamurthy and B. S. Panigrahi, J. Lumin., 2007, 127, 483–488.
34. C. Guo, J. Yu, X. Ding, M. Li, Z. Ren and J. Bai, J. Electrochem. Soc., 2011, 158, J42–J46.
35. K. H. Kwon, W. B. Im, H. S. Jang, H. S. Yoo and D. Y. Jeon, Inorg. Chem., 2009, 48, 11525–11532.

---