1 Introduction
Upconversion (UC) luminescence is the process of achieving higher energy visible emission photons by continuously absorbing at least two low-energy photons [
1]. Rare-earth ions doped UC luminescent materials have attracted the attention of many researchers, including potential applications in biological imaging, solid-state lasers, fluorescent anti-counterfeiting and temperature sensors [
1–
5]. Of particular interest is the optical temperature sensors based on fluorescence intensity ratio (FIR), which enable non-contact, high-precision and fast temperature measurement [
6–
8]. The FIR technology is based on the intensity ratio of emission bands, corresponding to the two thermally coupled energy levels of rare earth ions, to realize the detection of temperature-dependent UC luminescence intensity [
8]. At present, numerous Yb
3+ and Er
3+ ions codoped temperature sensing luminescent materials based on FIR technology have been effectively synthesized, including NaGd(MoO
4)
2 [
9], La
4GeO
8 [
10], Y
2O
3 [
11], Cs
2NaGdCl
6 [
12], NaYF
4 [
13], among others. Unfortunately, we found that these researchers were not very comprehensive in their discourse on the mechanism of UC luminescence, and even some of the literature neglected to describe the energy back-transfer (EBT) mechanism.
Numerous scholars have extensively investigated the UC luminescent materials codoped with Yb
3+ and Er
3+ ions, with a discussion on the UC luminescence mechanism [
14–
16]. Among them, some literatures also mentioned the EBT mechanism. It is noteworthy that EBT is intricately linked to excitation power, but the excitation power used in the existing optical temperature measurement experiments is not consistent. For example, Chen et al. [
10] focused on the optical temperature measurement of Y
2GeO
5:Er
3+, Yb
3+ phosphors, revealing a peak absolute sensitivity (
SA) of 1.85 × 10
3 K
−1 at 473 K (excitation power: 150 mW). Similarly, Liao et al. [
17] explored the optical temperature measurement of Gd
2TiO
5: 2 at% Yb
3+/2 at% Er
3+ phosphors, identifying a maximum
SA of 40.76 × 10
4 K
−1 at 565 K (excitation power density: 3 W/cm
2). Additionally, Li et al. [
18] observed that KBaY(MoO
4)
3:Yb
3+, Er
3+ phosphors exhibited a peak
SA of 0.01206 K
−1 at 420 K (excitation power density: 4.6 W/cm
2). EBT represents a crucial mechanism state affecting the layout of Yb
3+/Er
3+ ion intermediate states [
3,
19]. Regrettably, current literature often only mentions the existence of this mechanism, yet fails to provide an in-depth analysis encompassing aspects such as sample conditions, doped Yb
3+ ion concentrations, and the correlation between excitation power and EBT.
Consequently, to comprehensively investigate the process of EBT in the UC luminescence mechanism, various samples were synthesized with differing states (phosphors and crystals) and different doping conditions to conduct experiments. The experiment aimed to enhance the understanding of the EBT mechanism, and establish a pathway for the optimum excitation power point for optical temperature measurements. The characteristics of the EBT process between the phosphor (polycrystalline) and the crystal (monocrystalline), that is, the local order and long-range order systems, were analyzed and compared. The impact of Yb3+ ion concentration on the EBT process was scrutinized under various conditions including Er3+ ions doped Yb3+ ions self-activation and Yb3+/Er3+ ions codoping configurations.
2 Experimental section
The NaYb(MoO
4)
2: 5 at% Er
3+ phosphor was synthesized by high temperature solid state method, and the crystals of NaYb(MoO
4)
2: 5 at% Er
3+ and NaBi(MoO
4)
2: 10 at% Yb
3+, 1 at% Er
3+ were successfully grown by Czochralski method using a JGD-400 top-mounted single crystal growth furnace (produced by CETC26 th) [
20]. The UC emission spectra of the samples were measured using an Omni-
λ5028i and a charge coupler (Andor DU401 BVF). The excitation source for the UC emission was a 980 nm semiconductor laser with a maximum output power of 10 W (BWT Beijing Ltd.).
The phase structural analysis was performed using a Bruker D8 Advance X-ray diffractometer equipped with Cu-Kα radiation (λ = 1.5406 Å). The diffraction angle (2θ) ranged from 10° to 80° with a step size of 0.02°. The diffuse reflectance UV–Vis–IR absorption spectra were measured on a UV-3600 spectrophotometer.
3 Results and discussion
Figure 1a displays the X-ray diffraction (XRD) patterns for the relevant phosphor and single crystals. All patterns correspond to the standard reference compounds NaYb(MoO4)2 (PDF#57–0839) and NaBi(MoO4)2 (PDF#51–1508), demonstrating the pure phase state of the prepared samples and that the crystal structure of the main lattice is not significantly altered by the addition of dopant ions. The photograph of the as-prepared samples is given in Fig. 1b.
Figure 1c shows the measured diffuse reflection spectra of Er
3+ doped NaYb(MoO
4)
2 samples and Er
3+/Yb
3+ co-doped NaBi(MoO
4)
2 single crystal in the range of 200–1000 nm. In the Er
3+ doped NaYb(MoO
4)
2 samples, nine absorption bands with considerable intensity were displayed, corresponding to the transition from the
4I
15/2 ground state of Er
3+ to different excited states. These absorption peaks are located at 379, 407, 451, 489, 523, 545, 655, 802, and 970 nm, corresponding to the transitions of the excited states
4G
11/2,
2H
9/2,
4F
5/2,
4F
7/2,
2H
11/2,
4S
3/2,
4F
9/2,
4I
9/2, and
4I
11/2, respectively [
21]. Among them, the absorption band appears wider near 970 nm, which may be due to the superposition of two absorption peaks, attributed to the
4I
15/2 →
4I
11/2 transition of Er
3+ and the
2F
7/2 →
2F
5/2 transition of high concentration of Yb
3+ ions in the host. In addition, an absorption band corresponding to the transition from the ground state
4I
15/2 to the excited state
2H
11/2 of Er
3+ ions was observed in the Er
3+/Yb
3+ co-doped NaBi(MoO
4)
2 single crystal, and its absorption peak was at 523 nm.
The optical band gap energy (
Eg) of the samples is estimated from the diffuse reflectance spectra using the Tauc relation given in [
22]
where
α is the absorption constant,
h is Planck's constant,
ν is the frequency of the incident photon,
A is the proportionality constant,
Eg is the material band gap, the index
n represents the transition properties that occur in the sample, and their values are
n = 1/2 and 2 for direct and indirect transitions, respectively. Here, we consider the direct band gap, so the value of
n remains 1/2. The Tauc plots of the prepared samples are shown in Fig. 1d. The band gap of Er
3+/Yb
3+ co-doped NaBi(MoO
4)
2 single crystal is observed to be 3.20 eV, while the band gap of Er
3+ doped NaYb(MoO
4)
2 phosphor is calculated to be 3.68 eV. This is attributed to the larger ionic radius of Bi
3+ ions, leading to a smaller band gap in the former [
23]. In addition, the band gap of Er
3+ doped NaYb(MoO
4)
2 single crystal is slightly larger than that of the phosphor. This is due to the segregation coefficient of Er
3+ within the NaYb(MoO
4)
2 host, resulting in incomplete substitution of Yb
3+ by doped Er
3+ during crystal growth.
To investigate the correlation between the sensitizer Yb3+ ions and the UC luminescence of the materials, the UC spectra of the three synthesized samples were measured under 980 nm laser excitation, as illustrated in Fig. 2a. The samples exhibited two prominent green emission bands and a weak red emission band within the visible light band. Among them, the green emission bands are a result of the transitions 2H11/2 → 4I15/2 (centered at 530 nm) and 4S3/2 → 4I15/2 (centered at 554 nm) involving the Er3+ ions. The weak red emission band originates from the transition 4F9/2 → 4I15/2 (centered at 660 nm) of the Er3+ ions. It is worth noting that the concentration of Yb3+ ions in the matrix does not exert a significant influence on the emission peak position of the UC emission spectra.
The contour plots for different power densities of three different samples were examined to learn more about the UC luminescence mechanism. The corresponding contour plots obtained after normalizing the data were shown in Fig. 2b–d. Over a broad power density range, the concentration of Yb3+ ions and the pump power density were examined in relation to the energy back transfer (EBT) process of Yb3+-Er3+ ions. A comparison of the three spectra reveals that the spectrum of the NaBi(MoO4)2 crystal remains constant throughout the power density range, dominated by the green emission peak. Conversely, the spectra of the self-activated NaYb(MoO4)2 phosphor and NaYb(MoO4)2 crystal were distorted at higher power density, potentially attributed to the emergence of the EBT process between Yb3+-Er3+ ions under high-power density conditions.
In addition, to illustrate the phenomenon more intuitively, the luminescence images of the three samples were captured at different power densities, as shown in Fig. 3. The NaYb(MoO4)2 phosphor and crystal exhibit a yellow luminescence trend at higher power densities, and the phosphor has demonstrated yellow luminescence at lower power densities than the crystal. This may be due to the better thermal stability of the crystal compared to the phosphor, so that the power point at which its EBT occurs appears to be shifted back. The thermal conductivity of a material is intricately linked to its internal structure. In general, the long-range ordered structure will have better thermal conductivity. Thus, in contrast to phosphors exhibiting a locally ordered structure, crystals possessing a long-range ordered structure demonstrate better thermal stability, potentially accounting for the weaker EBT process within crystals.
Notably, the luminescence image of the Yb3+/Er3+ ions codoped NaBi(MoO4)2 crystal in Fig. 3c shows that the sample exhibits bright green emission throughout the power density range. This observation is consistent with the findings depicted in the contour plot of Fig. 2d, indicating the absence of an EBT process between Yb3+ and Er3+ ions. The results of the experiment revealed a direct correlation between the concentration of Yb3+ ions and the EBT process of Yb3+-Er3+ ions, with the excitation power density level also playing a role in this process. A higher concentration of Yb3+ ions was found to facilitate the EBT process between Yb3+ and Er3+ ions, leading to the suppression of green light emission and the enhancement of red light emission from Er3+ ions. Furthermore, it was observed that the phosphor initiated the EBT process at a lower power threshold compared to the crystal, which may be related to the poor thermal stability of the phosphor.
Figure 4a summarizes the qualitative relationship between pump power density and luminescence color for the three samples. The NaYb(MoO4)2 phosphor exhibits the earliest transition to yellow emission (lower power threshold), followed by the NaYb(MoO4)2 crystal, while the NaBi(MoO4)2 crystal retains green emission throughout. This trend aligns with the thermal stability and Yb3+ concentration differences between samples. The color shift is attributed to EBT-driven suppression of green emission and enhancement of red emission, as confirmed by spectral distortion analysis (Fig. 2b–d) and luminescence imaging (Fig. 3). Furthermore, a schematic diagram of the energy transfer relationship between Yb3+ and Er3+ ions is provided to enhance the understanding, as depicted in Fig. 4b.
The simplified energy level diagrams of Er3+ and Yb3+ ions and possible luminescence processes at low pump power densities are illustrated in Fig. 5. The large absorption cross-section of Yb3+ ions in the near-infrared region and the amazing energy level overlap between Yb3+ and Er3+ ions result in the energy transfer (ET) process of Yb3+ to Er3+ ions, which primarily contributes to the generation of visible green and red light emission in the current work. Under the irradiation of a 980 nm laser, a large number of Yb3+ ions absorb near infrared photons and are excited from the ground state of 2F7/2 to the excited energy level of 2F5/2. Subsequently, the excited Yb3+ ions excite the neighboring Er3+ ions from the 4I15/2 ground state to the 4I11/2 energy level through an effective energy transfer process (ET1: 2F5/2 + 4I15/2 → 2F7/2 + 4I11/2). Additionally, Er3+ ions can directly absorb near-infrared photons through the ground state absorption process (GSA) and undergo a transition from the 4I15/2 ground state energy level to the 4I11/2 energy level. The adjacent Yb3+ ions excite a portion of the Er3+ ions located at the 4I11/2 level to the higher 4F7/2 level through the secondary energy transfer process (ET3: 2F5/2 + 4I11/2 → 2F7/2 + 4F7/2). Similarly, Er3+ ions can absorb a near-infrared photon through the excited state absorption process (ESA1), achieving the transition from the 4I11/2 to the 4F7/2 level. The 4F7/2 energy level undergoes nonradiative relaxation (NR) filling into the 2H11/2 and 4S3/2 energy levels. Ultimately, the transitions 2H11/2 → 4I15/2 and 4S3/2 → 4I15/2 result in green UC emission centered at 530 nm and 554 nm. The electrons in the 2H11/2/4S3/2 (Er3+) state can be partially non-radiatively relaxed to the 4F9/2 state (Er3+), and the 4F9/2 level can be filled directly from the 4S3/2 level. In addition, the electrons in the 4I11/2 can be depopulate by ET3, ESA1 and nonradiative to 4I13/2. Obviously, the electrons of the 4I13/2 energy level can be further excited to the 4F9/2 energy level through the excited state absorption process ESA2: 4I13/2 (Er3+) + a photon → 4F9/2 (Er3+), or through the energy transfer process (ET2) of the adjacent Yb3+ ions. Eventually, the radiative transition 4F9/2 → 4I15/2 occurs, resulting in the generation of red UC emission.
To directly explain the EBT process in the sample, we monitored the change of the emission intensity of Yb3+ ions under different power density excitation of 5 at% Er3+:NaYb(MoO4)2 crystal. Figure 6a gives the contour plot of the sample at low power density. The spectrum shows that the emission intensity of Yb3+ gradually decreases with the increase of power density, and the emission process of Yb3+ becomes weaker. As described in the energy transfer process of Fig. 5, this realizes the positive energy transfer from Yb3+ to Er3+ ions. In the case of high power density, as shown in Fig. 6b, it can be found that the very wide emission band of Yb3+ around 1um centered at 1048 nm is gradually weakened. In contrast, it shows an enhanced signal around 1010 nm. This is most likely due to the energy transfer of Er3+ to Yb3+ ions, that is, the possible EBT process occurs, resulting in the enhancement of Yb3+ signal. Finally, the yellow luminescence shown in Fig. 3b is realized.
As shown in the contour plot of Fig. 5b, when the concentration of Yb
3+ ions in the sample reaches a significant level and the pump power is high enough, it cannot be explained by the traditional UC mechanism of Fig. 5. The influence of other mechanisms on the luminescence of the sample should be considered, as shown in Fig. 7. At sufficient concentrations of Yb
3+ ions, the red luminescence can be accounted for by the following two processes: the energy back transfer (EBT) process from Er
3+ to Yb
3+ (
4S
3/2 +
2F
7/2 →
4I
13/2 +
2F
5/2) [
5,
24,
25] and the cross-relaxation (CR) process between the Er
3+ ions (
4F
7/2 +
4I
11/2 →
4F
9/2 +
4F
9/2) [
26,
27]. The EBT process can layout
2F
5/2 energy levels and lowering the layout of the
4S
3/2 energy level. Since the energy level difference between
4F
7/2 and
4F
9/2 (5200 cm
−1) is comparable to the energy level difference between
4I
11/2 and
4F
9/2 (5100 cm
−1), the CR process diminishes the intensity of green light emission [
17,
28]. More importantly, the CR process becomes the dominant process only when the Er
3+ ions concentration is high. Therefore, the EBT process is more efficient and dominates among the two processes [
29], leading to the enhancement of red UC emission at high Yb
3+ ion concentrations. This EBT-dominated mechanism leads to suppression of green UC luminescence and enhancement of red UC luminescence at higher Yb
3+ ion concentrations, which is consistent with experimental observations. Notably, under high pump power excitation, it may be necessary to consider the energy exchange between two Er
3+ ions. The highly filled
4F
7/2 energy level reaches saturation, and any further energy contribution is transferred through the virtual energy level of the Yb–Yb cluster to the nearest
4F
7/2 energy level of the other Er
3+ ion, leading to a simultaneous transfer of energy through energy bridging between the two Er
3+ ions [
30].
4 Conclusions
Overall, rare-earth ions doped NaYb(MoO4)2 phosphor and crystal, along with NaBi(MoO4)2 crystal, were successfully synthesized. The three samples exhibited two prominent green emission bands and a weak red emission band in the visible band. The luminescence images clearly show that the Er3+ ions doped Yb3+ ions self-activated NaYb(MoO4)2 phosphor and crystal exhibit yellow luminescence at high power density levels, attributed to the efficient EBT process (4S3/2 + 2F7/2 → 4I13/2 + 2F5/2). In contrast, the Yb3+/Er3+ ions codoped NaBi(MoO4)2 crystal displays bright green emission in the whole power density range. The findings indicate that the high Yb3+ ion concentration promotes an enhanced EBT process between Yb3+ and Er3+ ions, leading to the suppression of green emission and enhancement of red emission from Er3+ ions. In addition, crystals with long-range ordered structure exhibit weaker EBT process than phosphors with local ordered structure, a phenomenon possibly ascribed to their superior thermal stability. The relative intensity change of Yb3+ emission in 5 at%Er3+:NaYb(MoO4)2 crystal was further monitored, and the EBT process in self-activated samples at high power density was visually displayed. Based on the experimental results and comprehensive understanding of EBT mechanism, we provide a reliable direction for determining the optimal excitation power for optical temperature measurement.