† Corresponding author. E-mail:
Project supported by the National Natural Science Foundation of China (Grant Nos. 51172236, 51502292, 51272254, 51102239, 61205173, and 61405206).
The Judd–Ofelt theoretic transition intensity parameters
Phosphors activated by rare-earth ions have been used widely in many fields such as displays, illumination, lasers, and so on. It is important to study the transition intensity of rare-earth in solids and aqueous solution by theory and experiments. Judd and Ofelt independently gave a quantitative description of the transition intensity in 1962.[1,2] The original idea in Judd’s and Ofelt’s papers is that the wavefunctions of ng and nd configuration with opposite parities are mixed into the 4f configuration by odd crystal field and the parity-forbidden rule of 4f–4f transition of electric dipole is broken down, resulting in the sharp lines of 4f–4f observed experimentally. Ab initio calculation has shown that the contribution from the static electric interaction is not dominant, but overlap, covalence, and polarization are also very important for the rare-earth ions in a crystal field. Hence, a more general phenomenal equation, which is very similar to that obtained by Judd and Ofelt but represents more physical effects such as overlap, covalence, and polarization, was developed by Reid et al. and Newman et al.[3–6]
Transition intensity of rare-earth ions based on the Judd–Ofelt work has been accepted and used widely to analyze the optical transition of rare-earth ions in solids, aqueous solutions, glass, and so on, which are generally applied to their absorption spectra. So far, it has been very rare to use emission spectra of rare-earth ions to fit transition intensity parameters. As pointed out by M. F. Reid, emission spectra can be particularly troublesome, since absolute measurements are difficult. It is possible to carry out fitting using relative intensities within crystal field splitting multiplets, but is not as satisfactory as using absolute intensities.[7] It will be shown in this paper that it is not feasible to determine intensity parameters reliably if no other constraints are incorporated into the fitting. Generally, for absorption measurement of solids, it is necessary to grow large, high-quality single crystal, requiring much time and money. In most cases, it is not so easy to grow single crystal whose size and quality are desirable for absorption measurement. Due to this limitation, the emission spectra of rare-earth ions in polycrystalline have rarely been investigated by transition intensity theory. While it is possible in some cases to calibrate the emission data by assuming that calculated magnetic dipole intensities are accurate,[5] in the great bulk of cases, the magnetic dipole transition is not so strong or is discerned difficultly, and no intensity calibration is viable.
On the other hand, integral quantities such as line strength and oscillator strength of the absorption spectrum are usually used to fit intensity parameters. The number of intensity parameters
In order to overcome these difficulties of fitting the transition intensity of rare-earth ions to emission or absorption spectra, we propose a “full profile method.” As a practical example, the intensity parameters and Huang–Rhys factors were fitted to the emission spectrum of Yb:GdTaO4, a promising ultra-short pulse laser medium, at 8 K. Also, some problems related to Yb3+ in solids are discussed. The results of fitting to the emission spectrum of Yb3+:GdTaO4 indicate that the full-profile method is a viable way to determine transition intensity parameters through a powder crystalline emission spectrum, and the difficulty that the number of the intensity integral quantities of the spectrum is smaller than that of the intensity parameters and other relative parameters is overcome.
Considering an emission band from the transition between the initial energy levels EI with an initial population number NI and final energy levels EF, whose initial and final wavefunctions are |Iiα⟩ (α = 1,…, gi) and |Ffβ⟩ (β = 1,…, gf), where gi and gf are the degeneration of energy levels EI and EF, respectively, the electronic and magnetic dipole transition line strength
The non-polarized luminescence line strength for powder polycrystalline should be averaged as
According to Judd–Ofelt parameterization, the matrix element of electric dipole transition is
The spontaneous emission transition probability AIF can be calculated with the values
Using the spontaneous transition probability
The values of σ and s2σ cannot be discerned experimentally, indicating that the uncertainty of the instrumental constant results in the uncertainty of
Equation (
It is obvious that σI depends not only on the instrumental constant σ but also on the population NI in the initial state I. It is a great problem for spectroscopy to know the excited state population NI, which further indicates that it is difficult to obtain the absolute fluorescence intensity
Assuming that the transition line profile function from EI to EF is ϕIF(λ, λIF,wIF) against λ with the full width at half maximum (FWHM) wIF and peak center λIF, ϕIF(λ, λIF,wIF) is normalized by
The total intensity of a spectrum from all transitions can be expressed as
If the integral intensity of a spectral band from EI0 to EF0 is called
If a spectrum is only from one initial state, 𝓡IF = 1, equation (
Now every data point
According to Eqs. (
That is,
On the other hand, if all
To a great extent, the reliability of determining the parameters depends on the degree of consistency between
1)
2) To what degree does the calculated
In order to solve for variable
The recorded spectral intensity is a series of discrete wavelength and intensity values (λ1, I(λ1)), (λ2, I(λ2)), …, (λN, I(λN)), so equation (
The minimum of the normal equation of Eq. (
The
The (5 at.%) Yb3+:GdTaO4 was prepared by a high temperature solid state reaction method. The raw materials were Yb2O3, Gd2O3, and Ta2O5 with purity 99.999%. They were weighted according to the constitute ratio of Yb0.05Gd0.95TaO4, mixed and ground thoroughly in an agate mortar by an agate pestle. The mixtures were calcined at 1500 degree Celsius for 72 h. After the samples were cooled down to room temperature, they were thoroughly re-ground as polycrystalline powder and used as photoluminescence measurement samples. The x-ray diffraction indicates that it was M-Yb3+:GdTaO4 with space group I12/a1 (No.15),[9] and its structure was refined by GSAS using the structure data of M-GdTaO4 as the initial parameters. The obtained structure data of M-Yb3+:GdTaO4 was used to calculate the structural data needed by superpostion mode.
The photoluminescence spectra of (5 at.%) Yb:GdTaO4 at 8 K were measured with Fluorolog-3-Tau made by JOBIN YVON. Its excitation and emission spectra are shown in Figs.
2F5/2: 10779 cm−1, 10460 cm−1, 10285 cm−1;
2F7/2: 0 cm−1, 283 cm−1, 417 cm−1, 607 cm−1.
Some obvious vibronic peaks can be observed in the photoluminescence spectrum of Yb3+:GdTaO4 at low temperature and the lattice relaxation during optical emission cannot be omitted. Single frequency approximation[10] was adopted to describe the lattice relaxation during emission. If pIF phonons with frequency ΔIF and Huang–Rhys factor SIF are released during de-excitation of an electron, the matrix element Miα,fβ in Eqs. (
The wavefunctions, |Iiα⟩ and |Ffβ⟩ mentioned previously, calculated by the superposition model parameters in Ref. [12], were used to calculate spectral transition intensity. The initial parameters
Another quantity to be evaluated is
According to Krupke’s approximation,[13]
In Krupke’s table,[11] ⟨4f|r|5d⟩, ⟨4f|r3|5d⟩, ⟨4f|r5|5d⟩, ⟨4f|r4|4f⟩, ⟨4f|r6|4f⟩, ⟨4f|r8|4f⟩, Δ(5d) and Δ(n′g) are listed for Pr3+, Nd3+, Eu3+, Tb3+, Er3+ and Tm3+. In oxide compounds, Δ (5d) can be estimated by the work of Dorenbos.[14] Goldner et al.[15] give estimations of Δ(5d) and Δ(5g) for Yb3+, which are 21700 cm−1, 51340–72340 cm−1, respectively. The radial integrals and energy denominators Δ(5d) and Δ(5g) for Tm3+ and Eu3+ are also listed in Krupke’s work.[16]
Based on these works, the values of ⟨4f|rt|5d⟩ and ⟨4f|rt + 1|4f⟩ used to estimate the values of Ξ(t,k) are shown in Table
In the site symmetry C2, there are twenty-five non-vanishing parameters
The energy levels of Yb3+:GdTaO4 are shown in Fig.
The lattice relaxation shifted the emission wavelength by 5.3 cm−1, so the energy level position of 10285 cm−1 was changed and fixed as 10280 cm−1 during the fitting procedure in order to keep the calculated transition wavelength consistent with the corresponding experimental line peaks. The vibronic peaks 1009.2 nm, 1020.0 nm, and 1038.6 nm were accounted for, their frequency center shifts are 88 cm−1, 59 cm−1, and 45 cm−1 relative to the electronic state transition peaks 999.8 nm, 1013.4 nm, and 1033.3 nm, and were approximated by single-frequency phonons here, whose Huang–Rhys factors labeled by S1, S2, and S3 were also fitted. On the other hand, the FWHM of the lines of 1000.3 nm, 1014 nm, and 1033 nm peaks with Lorentz profiles, which were labeled as w1, w2, and w3, respectively, were also included in the fitting. In sum, the fitting variable
A detailed Levenberg–Marquardt iteration algorithm with Eq. (
The refractive index of Yb3+:GdTaO4, which is necessary for calculating transition probabilities, is calculated by the equation[18]
After 88 iterations, the relative residual R is 11.75%, and the experimental and fitting curves are shown in Fig.
The transition probabilities and line strengths obtained from the intensity fitting are shown in Table
The emission cross sections obtained by this equation are also shown in Fig.
Although the transition intensity parameters
When Yb3+:GdTaO4 is used as a laser medium, it is expected that laser operation at the zero-phonon line 972 nm is not realizable due to strong reabsorption. If the laser operates as follows: Yb3+ is pumped to the excited level E6, rapidly relaxed to E5, and then transitions to the background state E1, a three-level system forms, and a very high laser threshold is expected, so the optical transition of E5 → E1 is nonsense for laser operation. It can be seen that the emission cross sections in the range of 1000–1019 nm from E5 → E2, E3 are high, and no great reabsorption occurs at low temperatures, the prospect of realizing laser operation in this channel is promising. When the working temperature of laser material increases, the thermal population of E2 increases, and the reabsorption of 1000 nm increases, which result in the case that the strongest line is 1014 nm, rather than 1000 nm at room temperature, which is shown as Fig.
At the same time, an increase in host temperature also results in broadening the FWHM of the emission spectrum, and a continuous emission band from 1001 nm to 1035 nm can be observed at room temperature (shown in Fig.
Full profile fitting to the emission spectrum of rare-earth ions in solids to obtain transition intensity parameters was proposed, in which the contribution of a transition to the emission spectrum is expressed as the product of transition probability, line profile function, instrument measurement constant and transition center frequency or wavelength, and the whole emission spectrum is the sum of the transitions. Then the emission spectrum is expressed as a function with independent variables intensity parameters
1 | |
2 | |
3 | |
4 | |
5 | |
6 | |
7 | |
8 | |
9 | |
10 | |
11 | |
12 | |
13 | |
14 | |
15 | |
16 | |
17 | |
18 | |
19 | |
20 |