Core–shell engineering is generally utilized to reduce surface quenching effects [29]. Shell passivation could promote the upconversion process by promoting the occurrence of energy hopping in higher-lying excited states [30]. Therefore, lifetime is affected by the shell layers, including the shell thickness and shell composition.

Excessive emitters would increase the probability of cross-relaxation, leading to concentration quenching. For a Yb³⁺-Er³⁺-Ho³⁺ tri-doped nanoparticle, higher content of Ho³⁺ could shorten the interionic distances of Yb³⁺-Er³⁺/Ho³⁺, Er³⁺-Ho³⁺ and Ho³⁺-Ho³⁺, which could enhance the non-radiative process and energy transfer (ET) sensitization. A decrease in the lifetime of Ho³⁺ (643 nm, 750 nm, 895 nm), Er³⁺(525 nm, 545 nm, 655 nm, 810 nm, 845 nm) and Yb³⁺ (1020 nm) was observed when increasing the concentration of Ho³⁺ [44]. Similarly, as the mole content of Tm³⁺ rose from 0.2% to 8% in NaYF₄: 20%Yb,x%Tm (40 nm), the lifetime of the blue emission was reduced from 662.4 µs to 25.6 µs (Figure 3a). Meanwhile, as the mole percentage of Yb³⁺ rose from 10% to 30% in NaYF₄: x%Yb,1%Tm, the lifetime of the same blue emission decreased from 206.7 µs to 120.2 µs [20]. The higher concentration of part of emitters would likely cause self-aggregation and cross relaxation, resulting in a shortened lifetime. When the Er³⁺ concentration was varying from 1% to 70% in the β-NaYbF₄@NaY_0.8−xEr_xGd_0.2F₄@ NaY_0.8Gd_0.2F₄, the lifetime of Er³⁺ at the level of ⁴S_3/2 (540 nm) decreased sequentially. The shorter lifetime was attributed to the noneffective passivation on the surface when dopant with a higher concentration of emitters. Due to the greater energy from Yb³⁺ (²F_5/2) being transferred to Er³⁺ (⁴S_3/2, ⁴S_9/2), the lifetime of Yb³⁺ was shortened. With the lack of cross-relaxation pathways for the ⁴S_9/2 level of Er³⁺, the lifetime rarely has significant changes with the Er³⁺ concentration variation. Interestingly, the lifetime of Er³⁺ at 654 nm in the β-NaYbF₄@NaY_0.8−xEr_xGd_0.2F₄ structure had a long lifetime due to more incredible energy transferring to the surface [45].

Simultaneous excitation of two Yb³⁺ ions can produce Yb³⁺ dimers with higher excitation energy, which could upconvert photons to Tb³⁺. To study the composition-dependent emission lifetimes and the effect on the energy transfer efficiency, Yan et al. employed the Tb³⁺-Yb³⁺-Nd³⁺ co-doped NaGdF₄: 80%Yb,10%Tb@NaGdF₄: 50%Nd,10%Yb nanoparticles with varied doping concentrations as the study model. The lifetime of Tb³⁺ reached 1.76 ms when the content of Yb³⁺ rose from 20% to 80% because more Yb³⁺ facilitated the formation of Yb³⁺ dimer. The lifetime at 542 nm is slightly prolonged by 0.06 ms when the proportion of Tb³⁺ increased from 5% to 10% due to promoted energy transfer from Yb³⁺ dimer to Tb³⁺. In addition, when the content of Nd³⁺ increased from 10% to 50%, near-infrared absorption intensity improved, the lifetime of Tb³⁺ at 542 nm and Yb³⁺ at 1000 nm both increased, indicating more energy was transferred to Tb³⁺ and Yb³⁺ [46].

A constant lifetime can be obtained when the doping content of the sensitizer is changed. To investigate the relationship between lifetime decay behavior and luminescence emission intensity, core–shell structure of NaYF₄@NaYF₄: x%Yb,1%Tm@ NaYF₄: y%Yb@NaYF₄ nanoparticles was developed by Zhang et al. Changing the mole content of Yb³⁺ in the first shell from 20% up to 80%, the emissive intensity changed while the luminescence lifetime at 475 nm kept constant, suggesting a constant lifetime with different emissive intensity could be obtained. When the concentration of Yb³⁺ in the first and second shell layers was changed, the varied lifetime (1256 ms to 310 ms) was obtained with a constant emission intensity (Figure 3b) [47]. For the mentioned nanoparticles, the declined Yb³⁺ concentration increased the mean value of the distance between Yb³⁺ and Tm³⁺ ions, leading to a longer lifetime due to the weakened back energy transfer process. The Yb³⁺ concentration in the energy transfer upconversion layers was decreased from 99%, 70%, 50%, 40%, 20% to 10%, resulting in the lifetime at 808 nm being increased from 1282, 1315, 1481, 1618, 1721 to 2157 μs, respectively [41]. Besides, the Tm³⁺ could be served as sensitizers and transfer energy to Yb³⁺, Ho³⁺ and Er³⁺ when excited at 808 nm and 1208 nm, respectively. The lifetime of Yb³⁺ (980 nm), Ho³⁺ (1180 nm) and Er³⁺ (1525 nm) decreased with the increase of the Tm³⁺ molar ratio [48].

The decay process of an excited state is inversely proportional to the energy transfer rate and the radiative and non-radiative transition rates. Tri-doped nanoparticles of NaYF₄@NaYF₄: Er³⁺/Yb³⁺/Mn²⁺@NaYF₄ were synthesized by Mao et al., exhibiting the new energy transfer process between Mn²⁺ and Er³⁺. With a non-radiative energy transfer channel created between the level of Mn²⁺ (⁴T₁) and Er³⁺ (²H_11/2 and ⁴S_3/2), the overall transition rate increased owing to the resonance energy transfer, leading to a reduced lifetime of Er³⁺ at 550 nm. The presence of Mn²⁺ ions promoted the relaxation of the ⁴S_3/2 energy level, and the red emission of Er³⁺ increased with the shortening of the decay time. Therefore, the increased population density of Mn²⁺ caused the decreased radiative transition rate of Er³⁺ (⁴S_3/2 and ²H_11/2) turning down to ground state, resulting in an enhancement of red emission due to energy transfer from Mn²⁺ (⁴T₁) to Er³⁺ (⁴F_9/2) (Figure 4a). Meanwhile, the increased lifetime of Er³⁺ at 650 nm also verified the role of Mn²⁺ in energy transfer trace according to the decay curves of various content of Mn²⁺(0%, 10%, 20%, 30%) (Figure 4b) [34].

Introducing transition metal ions with a long lifetime into conventional UCNPs is particularly attractive. The lifetime of Mn²⁺ ions could be modulated by crystal-site engineering. Liu et al. tuned the luminescence properties of Mn²⁺ in core–multishell nanoparticles by doping Ca²⁺ or Mg²⁺, changing the output color from green to yellow and prolonging its lifetime [49]. The spin-forbidden transition of Mn²⁺ occurs between ⁴T₁→⁶A₁, allowing a longer fluorescence lifetime than lanthanide emitters. Because of the larger energy mismatch between Yb³⁺ and Mn²⁺, the Yb³⁺-Mn²⁺ dimer is difficult to form. However, Zhang et al. reported the successful preparation of Yb³⁺-Mn²⁺ dimers, obtaining a substantial long lifetime of Eu³⁺ (91 ms), while the normal UC lifetime of Eu³⁺ is only about 7 ms. In addition, there is a dynamic population balance between the energy state |²F_7/2,⁴T₁(4G)⟩(Yb³⁺–Mn²⁺ dimers) and ⁵D₀ (Eu³⁺), causing the sustained energy transferring from Yb³⁺–Mn²⁺ dimers to Eu³⁺ [50].

The long-lived Mn²⁺ integrated with the short-lived lanthanide particle platform could establish a new energy transfer pathway, and then affect the whole decay process. In a NaGdF₄: 30%Mn@NaGdF₄: 49%Yb,1%Tm@NaYF₄ nanoparticle, Gd sublattice-mediated energy migration facilitates Mn²⁺ upconversion luminescence, leading to a decrease in the lifetime of Gd³⁺ at 311 nm (⁶P_7/2→⁸S_7/2) from 6.5 to 4 ms (Figure 4c) [51]. As a result, the lifetime of lanthanide ions may be affected when the external ions introduced and interfered with the energy transfer channels.

Doping various amounts of Gd³⁺ into the NaYF₄ host nanocrystals could regulate the upconversion photoluminescence lifetimes. Xie et al. prolonged the lifetime of Er³⁺ (⁴S_3/2→⁴I_15/2, ⁴F_9/2→⁴I_15/2) by utilizing the Gd³⁺, substituting for Y³⁺ and Yb³⁺ in the crystal lattice of NaYF₄ host, which attributed to the energy transfer rate decrease caused by the average sensitizer-activator separation increasing [52].

An efficient energy transfer pathway could be established in fluorescence dye-loaded rare-earth nanocrystals, which enables luminescence lifetime tuning [53]. In contrast to conventional molecular donor–acceptor pairs, the energy transfer efficiency is related to the distances between lanthanide-doped nanoparticles, and thus significantly depends on the nanoparticle diameter. Hirsch et al. synthesized NaYF₄: 20%Yb,2%Er with the precisely controlled size spanning from 10 to 43 nm, and coated with sulforhodamine B and rose bengal by ligand exchange. The nanoparticles with a mean diameter ranging from 20 to 25 nm possessed an optimum efficiency of 50–60%. The lifetime of Er³⁺ at 600 nm decreased primarily due to the competition of non-radiative surface deactivation at the smaller surface-to-volume ratios (Figure 5a) [54]. Li et al. loaded the IR-820 on NaYF₄: Tm to construct a FRET system. Luminescence decay from ³H₆→³H₄ transition was used as a detection signal. When the energy accepter (IR-820) was attached to the donor (Tm³⁺: ³H₄) under 785 nm excitation, the lifetime of Tm³⁺ at 800 nm decreased because of luminescence resonance energy transfer (Figure 5b) [55]. Su et al. loaded the IR-806 on the NaGdF₄: 49%Yb,1%Tm@NaYF₄: 20%Yb@NaGdF₄: 50%Nd,10%Yb@NaGdF₄ nanoparticles to improve the ultraviolet luminescence intensity. With the back energy transferred from the nanoparticles to dye molecules, the decreased lifetimes of Gd³⁺ and Tm³⁺ ions were observed at 253, 276, 290, 310, 360 and 475 nm [56]. An organic fluorescent dye as an antenna could be used to broaden and increase absorption for UCNPs, allowing the energy to flow to dye molecules. Meanwhile, the hybrid system between dye molecules and UCNPs creates a new energy diffusion pathway, increasing the radiative transition process. Li et al. added the Cy3-SO₃ into a NaYF₄: 20%Yb,2%Er@CaF₂ solution to construct an energy dissipation channel, which could transfer energy to the dye by the radiative transition. As a result, the luminescence lifetime of Er³⁺ at 488 nm decreased with the concentration of Cy3-SO₃ increase (0.67, 2, 4, and 5.33 μM). The reduced lifetime value verified the non-radiative energy transfer process between Er³⁺ and Cy3-SO₃ [57].

The triplet excitons could be trapped by inter- or intra-molecular interactions and prolong organic phosphorescence. For example, Yb³⁺ luminescence could be generated by organic Yb³⁺ complexes and hybrid organic-conjugated Yb³⁺-doped nanoparticles. Ye et al. prepared a composite thin film, in which the Yb³⁺ ions are incorporated with tetrakis-(pentafluorophenyl)imidodiphosphinate to form the Yb(F-TPIP)₃ chelate, while zinc salt of 2-(tetrafluoro-2-hydroxyphenyl)tetrafluorobenzothiazole (Zn(F-BTZ)₂) served as the organic chromophore. The Zn(F-BTZ)₂ possessed the emission ranging from 450 nm to 900 nm under 405 nm excitation, and gave rise of the Yb³⁺ emission centered at 975 nm from the transition of Yb³⁺: ²F_5/2→²F_7/2. Note that the intrinsic lifetime of Yb³⁺ at 1 μm is about ~1 ms. The lifetime of sensitized organic Yb³⁺ compounds could prolong up to ~0.3 s. The prolonged emission lifetime was demonstrated by dynamic equilibrium due to the energy transfer process from long-lived organic triplet excitons [58].

The surface ligand coordination could reconstruct the crystal-field splitting and orbital hybridization, and narrow the gap of the 4d orbitals between inner and surface lanthanide sensitizers. For example, after the bidentate picolinic acid (2PA) molecules coordinated to NaGdF₄: Yb nanoparticles, a longer lifetime (289 μs) at 980 nm was observed. The results confirmed that 2PA molecules could activate the dark surface layers and facilitate energy migration in the Yb³⁺ sublattices. The Yb³⁺-2PA coupling facilitated energy migration by 4f-orbital energy resonance within the ytterbium sublattice, which can reduce surface defects to hinder energy diffusion. Density functional theory (DFT) verified that 2PA coating could narrow the gap between the superficial and inner Yb³⁺ by lowering the empty 4f levels. Rigid ligands also stabilized the excited state of superficial Yb³⁺, prevented the superficial lanthanide ions from the solvent and fluoride vacancy-induced quenching, and thus significantly suppressing multiphonon non-radiative decay (Figure 5c) [59].

The synthesized process would affect the crystallinity, phase state and volume of nanoparticles and thus influence the decay behaviors. Vatsa et al. studied the decay process of GdVO₄: Dy³⁺ nanoparticles after heat treatment. When heated from 500 °C to 900 °C, the lifetime of Dy³⁺ (⁴F_9/2 level) extended from 114 μs to 260 μs due to the reduction of the non-radiative process by surface inhomogeneities. This increase in lifetime can also be ascribed to the decrease in the surface defects with the particle size increases in the heat treatment process (Figure 6a) [60]. For YVO₄: Ln³⁺ (Ln³⁺ = Dy³⁺, Eu³⁺) nanoparticles, the increase in covalent bond interaction caused by heat treatment led to a red shift in V–O charge transfer (CT). Similarly, the lifetimes of Dy³⁺ at ⁴F_9/2 and Eu³⁺ at ⁵D₀ increase with temperature from 500 °C to 900 °C due to the reduction of the non-radiative process on the surface of the particles [61].

As is well known, the decay time constant is inversely proportional to the radiative and non-radiative transition rates in the cross-relaxation process. The luminescence lifetime decreases with the increase of ambient temperature in most cases. The decay time with specific emissions produced by radiative transition rarely varies with temperature, while the non-radiative decay rate changes significantly with temperature [62]. For example, the lifetime of Yb³⁺ at 1000 nm reduced from 470 ± 11 μs to 390 ± 12 μs in NaYF₄: Nd³⁺, Yb³⁺ nanoparticles as the temperature rose from 25 °C to 45 °C. While the thermal coefficient α_τ calculated by the TGI system was almost unchanged (−0.0092~−0.010 °C⁻¹) (Figure 6b,e) [63]. Moreover, cross-relaxation between Tm³⁺ (¹G₄) usually occurs when raising the emitter concentration or temperature. Yu et al. compared the sensitivity of β-PbF₂: Tm³⁺/Yb³⁺ with different Tm³⁺ doping concentrations. They found that the relative sensitivity maximum values of ¹G₄ state lifetime in 0.0005Tm, 0.01Tm and 0.05Tm nanoparticles are 0.16%K⁻¹, 0.26%K⁻¹ and 0.46%K⁻¹ at 488K, respectively, indicating the potential ability as an indicator of upconversion luminescence lifetime-based thermometer [64].

In addition, the host matrix has a significant effect on thermal sensitivity. Díaz et al. found that oxide materials are more sensitive than fluoride ones by comparing the decay curves of NaYF₄: Er,Yb and NaY₂F₅O: Er,Yb nanoparticles at room temperature and 60 °C. [65]. The temperature dependency of Yb³⁺ emission lifetime in NaYF₄@NaYF₄: Yb³⁺,Nd³⁺@CaF₂ nanoparticles was determined by the energy transfer and back energy transfer rate, the energy migration process (among Yb³⁺), as well as radiative and non-radiative transition. Both the concentration of Nd³⁺ and Yb³⁺ affected the temperature sensitivity by changing the distance of Yb³⁺-Yb³⁺ and Yb³⁺-Nd³⁺, which in turn affected the back energy transfer processes from Yb³⁺ to Nd³⁺ and energy migration between Yb³⁺ ions. NaYF₄@NaYF₄: 20%Yb³⁺,60%Nd³⁺@CaF₂ as the nanoprobe possessed optimum thermal sensitivity through varying doping concentrations of Yb³⁺ and Nd³⁺, in which the lifetime of Yb³⁺ at 980 nm descended from 898 μs to 450 μs when the temperature increased from 10 °C to 64 °C. (Figure 6c,f) [66].

Interestingly, Li et al. demonstrate the lifetime compensation with temperature in NaErF₄@NaGdF₄ core–shell nanoparticles. The temperature-independent lifetime is attributed to the balance between lattice expansion (prolonging the lifetime) and thermal quenching (shortening the lifetime). A considerable energy migration process occurs in the high-doping concentration of Er³⁺, and the efficiency is proportional inversely to the average donor-acceptor distance with sixth order of magnitude. As a consequence, elevated temperature induces the lattice to expand, leading to a longer transfer distance, and ultimately prolonging the lifetime of Er³⁺. However, the prolonged lifetime caused by lattice expansion compensated for the difference value of the shorter lifetime aroused by thermal quenching, resulting in the temperature-independent lifetime (Figure 6d) [67].