Bright upconversion over extended temperatures enabled by an organic surface layer

Nanocrystal synthesis and modification

This study employed a pyridine-2-carboxylic acid (2PA)-capped nanostructure developed by Liu and co-workers for luminescence enhancement³⁶. To maximize the 2PA-mediated effect, we first optimized the synthetic protocol by refining the reaction conditions to enable the firm bonding between the nanoparticles and capped ligand molecules. Since the direct exchange of OA with 2PA ligand is not feasible, our synthesis involves the preparation of ligand-free UCNPs followed by treatment with 2PA ligands in ethanol (Fig. 2a). In a typical procedure, hexagonal-phase NaGdF₄:Yb³⁺/Tm³⁺ nanoparticles were synthesized through a co-precipitation method as a model platform³⁷, which were subsequently treated in concentrated HCl to obtain the ligand-free UCNPs (Supplementary Fig. 2). The ligand-free nanoparticles were positively charged and favored the adsorption of negatively charged 2PA molecules in the weakly basic condition (Supplementary Fig. 3)³⁸. Accordingly, we dissolved the 2PA ligands in ethanol solution with sodium hydroxide at pH = 9 to facilitate the deprotonation of –COOH, rendering the ligands more accessible to the nanoparticles. It is worth noting that the 2PA coating process hardly affected the spherical shape and uniform size of the original nanoparticles (Fig. 2a).

a Schematic of the synthetic procedure of 2PA-capped UCNPs, along with the TEM images and size distribution statistics of ligand-free and 2PA-coated NaGdF₄:Yb³⁺/Tm³⁺ (49/1%) nanoparticles. b FT-IR transmission spectra of ligand-free UCNPs, 2PA ligand and 2PA-capped UCNPs. Insets show the chemical structure model of 2PA ligand before and after surface coordination with Yb³⁺. c TG curves of 2PA-capped UCNPs prepared under different experimental conditions, including stirring at 78 ^oC (ST/78 ^oC), ultrasonic treatment at RT (US/RT) and ultrasonic treatment at 78 ^oC (US/78 ^oC). d Upconversion emission spectra of NaGdF₄:Yb³⁺/Tm³⁺ (49/1%) colloidal nanoparticles dispersed in ethanol before and after 2PA modification. Insets show the luminescence images of ethanol-dispersed and powder samples under 980 nm excitation. Note that these nanoparticles can be well-dispersed in ethanol with stable luminescence during the measurement. e Enhancement factors at different Tm³⁺ emission wavelengths induced by 2PA modification under different experimental conditions, along with the factors derived from overall integrated intensity. Note that the 2PA-induced enhancement factors at 450 nm in powder samples were shown as rhombic symbols. The error bars represent the standard deviation of 2PA-induced enhancement factor from three sets of repeated measurements.

The successful coordination of 2PA ligands on the nanoparticle surface was confirmed by Fourier transform infrared (FT-IR) spectroscopy (Fig. 2b). The apparent peak shifts related to ring deformation (676 → 704 cm^-1, 749 → 756 cm^-1), heterocyclic skeleton vibration (1452 → 1445 cm^-1, 1524 → 1476 cm^-1), C = N and C = O stretching vibration (1593 → 1566 cm^-1, 1654 → 1625 cm^-1) indicated the involvement of the pyridine ring and –COOH group in the coordination between 2PA and surface lanthanide ions³⁹. By contrast, the ligands showed ineffective attachment to the nanoparticles in the absence of pH adjustment due to their limited ionization (Supplementary Fig. 3c). The slightly suppressed O–H stretching vibration at around 3400 cm^-1 upon ligand coating suggested that the strong coordination of rigid ligands can partially obstruct the adsorption of water molecules to the nanoparticle surface.

The reaction temperature is crucial for efficient capping of ligands on nanoparticles. A high temperature may favor the reaction kinetics by facilitating molecular diffusion and collision frequency, thereby contributing to a more efficient and robust bonding between the surface lanthanide ions and capping ligands. This assumption was supported by thermogravimetric (TG) analysis of 2PA-capped nanoparticles obtained at varied reaction conditions (Fig. 2c and Supplementary Fig. 4). As expected, we found an increased detachment of 2PA molecules from nanoparticles prepared at elevated reaction temperatures. The stronger intensity of FT-IR peaks related to the ligand groups further revealed a higher concentration of surface-capped 2PA molecules under a higher reaction temperature (Supplementary Fig. 3c). In addition, magnetic stirring was preferred over ultrasonic treatment as it offers better mixing homogeneity and thus enhanced ligand accessibility, enabling a higher ligand-capping efficiency.

Upconversion enhancement at RT

Upon surface modification with 2PA ligands, the upconversion luminescence of NaGdF₄:Yb³⁺/Tm³⁺ nanoparticles dispersed in ethanol was substantially enhanced under 980 nm excitation (Fig. 2d, Supplementary Fig. 5 and Supplementary Note 1). Through optimizing the 2PA concentration and capping conditions (i.e., 0.08 mmol of 2PA under stirring for 8 hours at 78 ^oC), we recorded a remarkable enhancement factor of about 2100 and 160 folds for the UV emission (¹D₂ → ³H₆) using ligand-free and OA-coated nanoparticles as controls, respectively (Fig. 2e, Supplementary Figs. 6–9 and Supplementary Table 1). Note that a high loading density of 2PA ligands on the nanoparticle surface typically rendered a significant upconversion enhancement effect (Supplementary Fig. 10 and Supplementary Table 2). Under the optimized 2PA coating conditions, the enhancement phenomenon of the nanoparticles was well preserved even in the powder form, indicating the excellent stability and effectiveness of the nanoparticle-ligand interaction (Supplementary Figs. 11–12). Similar improvement of upconversion luminescence spanning visible and NIR regions was also observed in other lanthanide activators (e.g., Er³⁺, Ho³⁺ and Nd³⁺), demonstrating the versatility of the 2PA modification in enhancing upconversion (Supplementary Figs. 13–15).

Previously, the 2PA-induced upconversion enhancement was primarily ascribed to the energy-level reconstruction of surface Yb³⁺ sensitizers, which improves energy resonance between surface and inner Yb³⁺ ions, thereby facilitating energy migration within the sensitizer sublattice and subsequent energy transfer to activators³⁶. However, our investigation suggests that the effective preservation of interfacial quenching effect of Yb³⁺-harvested excitation energy after 2PA coordination may dominate the upconversion enhancement. As shown in Fig. 3a, the upconversion emission of ligand-free nanocrystals was boosted by 43 folds if the nanocrystals were subjected to a prior solution annealing process in OA/ODE containing Y³⁺-OA complexes, which mainly passivates surface defects of the nanoparticles without altering their size and morphology⁴⁰ (Supplementary Figs. 16, 17). Further removal of high-energy surface oscillations (i.e., O–H) in a dry argon atmosphere could induce another 42-fold enhancement of the upconversion emission, rendering even higher emission intensity than 2PA-capped counterparts in ambient air.

a Normalized integral intensity of overall emissions (300–900 nm) in 2PA-capped (blue), ligand-free (gray) and solution-annealed NaGdF₄:Yb³⁺/Tm³⁺ (49/1%) nanoparticles (green) measured in ambient air, along with the solution-annealed UCNPs measured in dry argon (yellow). Note that the solution-annealed sample is denoted as Y-OA. b Normalized integral intensity of untreated and solution-annealed UCNPs after 2PA modification measured in ambient air (gray and green) and dry argon (orange and yellow), respectively. Note that the light-blue areas highlight the corresponding enhancement factors induced by 2PA ligands. c Multiphonon upconversion emissions (450, 475 and 800 nm) in NaGdF₄:Yb³⁺/Tm³⁺ (49/1%) before and after 2PA modification as a function of the power density of 980 nm laser in the 1.5–96 W/cm² range. Note that the slope of the fitting line reflects the number of photons (n) required to populate the relevant emitting state according to the relation between emission intensity and pump power: I∝Pⁿ. d TG curves of ligand-free and 2PA-capped UCNPs showing the weight loss from RT to 170 ^oC due to surface dehydration. e Optimized structures and the calculated adsorption energies of 2PA-alone (E_ads−1), H₂O-alone (E_ads-2) and H₂O-2PA (E_ads-3) on the GdYbF-terminated surfaces of NaGd_0.5Yb_0.5F₄ crystal. Note that the interactions between the ligands and nanoparticles were investigated on the stable (0001) facet. f Comparison of emission intensity loss for ligand-free and 2PA-capped UCNPs in ethanol solutions containing different amounts of water. The error bars represent the standard deviation of emission intensity from three sets of repeated measurements.

After the surface quenching effect was minimized beforehand, the 2PA-mediated upconversion enhancement is largely compromised. As summarized in Fig. 3b, ligand coordination on solution-annealed nanoparticles induced a minor enhancement of upconversion emission by 13 folds, in contrast to 92 folds for untreated nanoparticles. Likewise, we only recorded a 10-fold emission enhancement of nanoparticles due to 2PA coordination in dry argon flow (Supplementary Fig. 18 and Supplementary Note 2). Notably, the solution-annealed nanoparticles in a dry environment can merely be marginally enhanced by 2.3 folds upon 2PA coating (Supplementary Fig. 19). Similar diminishing of 2PA-induced enhancement effect was also detected in the absence of Yb³⁺ ion, stemming from suppressed long-distance energy migration and thus alleviated surface quenching (Supplementary Fig. 20). These observations validate that 2PA achieves its major role in enhancing upconversion by preventing interfacial quenching of the excitation energy, that is, passivating surface defects and isolating high-energy oscillations of surface water molecules. In the absence of surface quenching, 2PA lost most of its power to boost the upconversion emission.

The effect of the 2PA-coordinated surface layer on preserving energy was further supported by time decay studies, which revealed a positive correlation of the upconversion enhancement factor with the increase in Yb³⁺ lifetimes (Supplementary Fig. 21). The reduced quenching of excitation energy facilitates the utilization of the absorbed photons, thereby lowering the power density threshold of intensity saturation⁴¹ (Fig. 3c). This saturation effect resulted in the attenuation of 2PA-induced enhancement factors under high excitation power densities (>20 W/cm², Supplementary Fig. 22). Accordingly, the 2PA ligand offers a significant enhancement effect at a relatively low power density before saturation occurs. Moreover, the preservation of excitation energy in Yb³⁺ ions also favours multiphoton processes, resulting in a greater enhancement factor of a higher-order upconversion luminescence (Fig. 2e). It is worth noting that 2PA coordination induced negligible changes in the absorption band of Yb³⁺, indicating that the local crystal structure was minimally altered (Supplementary Fig. 23)⁴². Therefore, the effective protection of excitation energy at the inorganic-organic interface played the major role in enhancing the upconversion process, which was backed by improved quantum efficiency (QY, Supplementary Table 3).

The comparison in Fig. 3b implies that 2PA coordination is highly effective in rejecting water-induced quenching. Mechanistic investigations by FT-IR and TG analysis revealed a suppression in O–H stretching vibration and a slight reduction in water adsorption upon 2PA modification (Fig. 3d). Therefore, partial rejection of water adsorption could be identified as a contributing factor to the mitigated surface quenching. However, the minor reduction of surface water content (~ 16%) is unlikely to take full responsibility for the nearly two orders of magnitude enhancement of upconversion emission. We thus speculate that the 2PA-coordinated surface layer may shield the surface lanthanide ions from the vibrations of water molecules.

To gain mechanistic insight into the role of the organic surface layer in mediating nanoparticle-water interaction, we calculated the adsorption energy (E_ads) of different surface configurations using the density functional theory (DFT) through the following equation⁴³:

where E_{slab+molecule}, E_slab and E_molecule indicate the energy of the nanohybrid system, bare surface and the isolated ligand molecule, respectively. The theoretical results reveal a strong interaction between 2PA and surface metal ions due to the formation of a stable five-membered chelate ring (Fig. 3e and Supplementary Fig. 24). Consequently, the attachment of water molecules on the nanoparticle is mainly bridged by 2PA ligands through weak intermolecular interactions³⁹, resulting in increased water-nanoparticle separation. In line with calculation results, the 2PA-capped nanoparticles displayed appreciably higher resistance to water-induced quenching than the ligand-free counterparts (Fig. 3f and Supplementary Fig. 25). It is worth noting that, despite being less effective than OA ligand in isolating surface water molecules (Supplementary Fig. 4), 2PA performs much better than OA in defect passivation (vide infra), thereby offering a superior overall upconversion enhancement effect (Supplementary Fig. 9). As an added benefit, the 2PA ligands significantly improved the dispersibility, luminescence intensity and stability of the nanoparticles in diverse polar solvents (Supplementary Figs. 26, 27), greatly expanding their utility in advanced photonic applications.

Bright upconversion in a thermal field

The main mechanism of thermal quenching is the promotion of multiphonon relaxation and phonon-assisted energy transfer to quenching centers at elevated temperatures. Accordingly, the 2PA-coordinated surface layer with an energy-preserving effect is potentially useful for combating thermal quenching. To test our hypothesis, we first measured the thermal stability of the upconverson luminescence in dry argon flow (Supplementary Figs. 28, 29), which could exclude dehydration-induced anti-thermal quenching effect widely observed in upconversion nanoparticles^34,35. As anticipated, the upconversion emission of 2PA-capped nanoparticles exhibited greatly improved thermal resistance compared to the uncoated counterparts (Fig. 4a). These observations are ascribed to the reduced defect formation in the 2PA-coordinated surface layer, which effectively suppresses non-radiative dissipation of excitation energy in a thermal field (Fig. 4b)⁴⁴. As a result, we observed a significant amplification of the 2PA-induced upconversion enhancement effect as the temperature increased (Fig. 4c). On a separate note, in-situ powder X-ray diffraction (XRD) measurements revealed an unaltered thermal expansion coefficient of the nanoparticles after 2PA modification (Supplementary Fig. 30), which ruled out the effect of anomalous lattice expansion on the thermal-resistant upconversion.

a Thermal stability of blue emission in ligand-free and 2PA-capped NaGdF₄:Yb³⁺/Tm³⁺ (49/1%) powder sample after the complete removal of water molecules in dry argon. Inset shows the schematic of the effect of the organic surface layer on preserving excitation energy. Note that the integral intensity of blue emission (440–500 nm) at 443 K preserved 32% and 91% of the initial intensity at 303 K for ligand-free and 2PA-capped samples, respectively. b Luminescence decay curves of Yb³⁺: ²F_5/2 → ²F_7/2 transition in ligand-free (top panel) and 2PA-capped UCNPs (bottom panel) as a function of temperature measured in dry argon. c The 2PA-induced enhancement factor in dry argon for blue emission as a function of temperature. Note that the factor denotes the integral intensity (440–500 nm) ratio of nanoparticles with and without 2PA coating at a given temperature, which increased from 37 to 104 folds as the temperature increased from 303 to 443 K. d Upconversion emission spectra of 2PA-capped UCNPs as a function of temperature in ambient air, along with the optical images at 303 and 443 K under 980 nm excitation. e The enhancement factors at different Tm³⁺ emission wavelengths versus the temperature in 2PA-capped UCNPs. Note that the inconsistent enhancement factors recorded for 450 and 360 nm emissions, which both originate from the ¹D₂ level, can be ascribed to measurement errors caused by the low signal-to-noise ratio at 360 nm. f Normalized integral intensity of blue emission as a function of temperature in ligand-free NaGdF₄:Yb³⁺/Tm³⁺ (49/1%), NaGdF₄:Yb³⁺/Tm³⁺ (49/1%)@NaYF₄:Yb³⁺ (49%), NaGdF₄:Yb³⁺/Tm³⁺ (49/1%)@NaYF₄ and 2PA-capped NaGdF₄:Yb³⁺/Tm³⁺ (49/1%) powder nanoparticles. Note that the amount of powder nanoparticles was all set the same, and these spectra were acquired under identical conditions for quantitative comparison. The blue emission of 2PA-capped UCNPs was 4,400, 570 and 40 folds higher than that of ligand-free UCNPs, UCNPs@active shell and UCNPs@inert shell at 443 K, respectively.

We next assessed the temperature-dependent upconversion emissions under ambient air conditions, which is the typical operational environment for many applications. Due to the involvement of moisture, thermally enhanced emission was detected for both ligand-free and 2PA-coated nanoparticles through thermal dehydration³⁴ (Fig. 4d and Supplementary Note 3). Intriguingly, we observed a more prominent thermal enhancement effect of upconversion emissions for 2PA-capped nanoparticles compared to the ligand-free counterparts (25 versus 7.2 folds for overall intensity and 240 versus 2.0 folds for 450 nm emission, Fig. 4e and Supplementary Fig. 31). Similar reinforced enhancement of multicolor luminescence was also recorded at the elevated temperature for Er³⁺, Ho³⁺ and Nd³⁺ activators after 2PA coordination (Supplementary Fig. 32). Although the well-established core–shell structures can render slightly brighter upconversion than the 2PA-coated nanoparticles at RT due to more effective shielding against surface water molecules, they fail to maintain the initial emission intensity as the temperature increased (Fig. 4f and Supplementary Figs. 33–35). Notably, we detected much stronger luminescence in 2PA-capped nanoparticles than core–shell samples after dehydration at elevated temperatures or in dry argon (Supplementary Figs. 36, 37). This phenomenon likely arises from the combined effects of reduced defect formation, minimized light scattering and improved energy-level alignment upon 2PA modification³⁶. Therefore, surface coordination with 2PA ligand provides a general approach for enhancing upconversion across a wide temperature range, with the advantages of unaltered particle size and convenient synthesis.

We ascribe the significant thermal enhancement effect in 2PA-capped nanoparticles to effective surface dehydration mediated by the organic interfacial layer that weakens the water-nanoparticle interactions. Notably, we recorded a progressive increment of the emission intensity in the ligand-free nanoparticles by continuously heating at 443 K, which was better pronounced under a lower humidity condition (Fig. 5a and Supplementary Fig. 38). In sharp contrast, 2PA-coated nanoparticles quickly attained intensity maximum and then remained stable under continuous laser irradiation for 5 hours at the high temperature, regardless of the humidity. These observations aligned well with the thermal evolutions of decay times (Fig. 5b). Specifically, the lifetime of naked nanoparticles kept prolonging upon continuous heating at 443 K, while the lifetime maximum was rapidly established in the 2PA-coated counterparts (Supplementary Fig. 39). These results demonstrated that the ligand-coordinated surface layer greatly promotes the thermal desorption of surface-adsorbed water molecules.

a Normalized integral intensity of overall emissions (300–900 nm) at 443 K against the 980 nm laser irradiation time in ligand-free and 2PA-capped NaGdF₄:Yb³⁺/Tm³⁺ (49/1%) nanoparticles under different relative humidity conditions. Note that the heating process from 303 to 443 K took around 40 min. b Luminescence decay curves of Yb³⁺: ²F_5/2 → ²F_7/2 transition in ligand-free (top panel) and 2PA-capped UCNPs (bottom panel) as a function of temperature and heating time at 443 K. c In-situ FT-IR spectra of ligand-free (top panel) and 2PA-capped (bottom panel) UCNPs during heating from 303 to 443 K. Inset shows the schematic of the effect of the 2PA-coordinated surface layer on mediating nanoparticle-water interaction. d Normalized emission intensity at 800 nm during natural cooling from 443 K in ligand-free and 2PA-capped UCNPs. e, f Reversibility test of the thermal enhancement behavior under different humidity conditions in ligand-free and 2PA-capped UCNPs. Note that the time interval was 40 min between each heating-cooling cycle. g Size-dependent 2PA-induced enhancement factors of 450 nm emission (¹D₂ → ³F₄) as a function of temperature, along with the TEM images of 2PA-capped UCNPs with average particle sizes of 10, 15 and 20 nm, respectively. h Comparison of upconversion emission spectra in the blue region for ligand-free UCNPs at RT, 2PA-capped UCNPs at RT and 2PA-capped UCNPs at 443 K under 980 nm excitation, along with the corresponding luminescence images.

In-situ FT-IR spectroscopy was employed to shed more light on the ligand-mediated dehydration process. As compiled in Fig. 5c, the O–H absorption peak in the 2PA-capped nanoparticles quickly diminished as the temperature increased from 303 to 443 K. By contrast, the absorption peak was largely preserved at elevated temperatures in the ligand-free counterpart. The results confirmed that the 2PA-coordinated surface layer expedited the thermal desorption process. Moreover, the FT-IR intensity of vibration peaks related to 2PA ligands hardly weakened as the temperature increased, suggesting the negligible thermal desorption of surface 2PA ligands. This exceptional thermal stability of the 2PA-coordinated layer is likely due to the formation of a stable five-membered chelate ring between the ligand and nanoparticle surface.

2PA ligands also favor the re-adsorption of water molecules onto the nanoparticle surface upon cooling, leading to a highly reversible modulation of upconversion by thermal stimuli. During the natural cooling process, ligand-free nanocrystals featured incomplete recovery to the initial intensity of upconversion emission and O–H stretching vibration within 3 h (Fig. 5d). This observation stemmed from the inefficient surface re-adsorption of surrounding water molecules, resulting in poor repeatability over successive thermal cycles (Fig. 5e). Upon surface modification with 2PA ligands, the emission intensity rapidly returned to the original level within 40 min after the temperature fell back to RT, in consistency with the FT-IR results (Supplementary Fig. 40). These results suggest that the strong coordination of high-rigidity 2PA ligands to surface lanthanide ions can facilitate the water re-adsorption process (Supplementary Fig. 41). Accordingly, we realized excellent repeatability of thermal enhancement performance in 2PA-capped nanocrystals under different humidity conditions (Fig. 5f).

The effect of the 2PA-coordinated surface layer on enhancing upconversion is strongly dependent on dopant concentration and nanoparticle size. Typically, an increase in Yb³⁺ concentration or a decrease in particle size resulted in a notable amplification of the enhancement effect mediated by 2PA ligands, which was more pronounced at elevated temperatures (Fig. 5g and Supplementary Figs. 42, 43). The observations can be ascribed to prevalent surface quenching in small-sized and heavily doped nanoparticles, stemming from a large surface-to-volume ratio and fast energy hopping through the crystal lattice²³. Consequently, the utilization of the 2PA-coordinated surface layer could enable a more significant recovery of radiative emissions. Remarkably, we observed an approximately 4000-fold enhancement of the 450 nm emission for 10 nm nanocrystals in the powder form at RT upon 2PA coordination (Supplementary Fig. 44). Further increase of temperature to 443 K induced additional thermal reinforcement of the blue emission by about 1,200 folds in the powder form, which was over six orders of magnitudes higher than that of the ligand-free counterparts at RT (Fig. 5h and Supplementary Fig. 45).

Full-color upconversion switching by thermal stimuli

The ligand-mediated upconversion enhancement with dimensional dependence provides tremendous opportunities for full-color tuning by thermal stimuli. As a proof-of-concept, we devised a 2PA-capped core–shell–shell nanostructure for thermochromic upconversion switching. In detail, the inner core and outmost layer with distinct accessibility to surface ligands are designed to show thermally quenched and enhanced upconversion, respectively (Fig. 6a). An inert interlayer is incorporated to precisely control the impact of the 2PA-coordinated surface layer and the luminescent shell on emission behavior of the core. As a result, thermal fine-tuning of upconversion emissions across the full spectrum could be realized by using core/shell combinations with red-green-blue (RGB) emission features (Supplementary Figs. 46–49 and Supplementary Table 4). Apart from the thermo-responsive emissions, these UCNPs also preserved the inherent dependence on excitation power and wavelength. The highly variable upconversion emission holds great potential for applications in advanced information security. In an exemplary illustration, we patterned a graphic on a copper substrate using four types of 2PA-capped UCNPs, which could reveal a series of optical codes with distinguishable graphic and color features by controlling the temperature and excitation power/wavelength (Fig. 6b).

a Schematic illustration of 2PA-capped multilayer nanoparticles for color-switchable upconversion emissions by thermal stimuli, along with the luminescence photographs of representative nanoparticles under 980 nm excitation at 303, 383, 403 and 443 K. b Design principle of multicolor “butterfly” pattern composed of nanoparticles with different compositions, and photographs of the pattern under different excitation/temperature parameters. c Schematic of the pixelated hybrid textual pattern consisting of different nanoparticles, and logic decryption process based on three groups of photonic output codes under diverse excitation/temperature parameters. The output colors of red, green and blue in response to 1530 nm excitation at 303 K, 980 nm excitation at 303 K and 980 nm excitation at 443 K are designated as the binary codes of “1”, which are processed to display the cryptographic information by an AND logical operation. Nanoparticles used for generating these patterns are NP#1: NaErF₄:Tm³⁺ (0.5%)@NaYF₄@NaGdF₄:Yb³⁺/Er³⁺ (20/2%)@2PA; NP#2: NaErF₄:Tm³⁺ (0.5%)@NaYF₄@NaYbF₄:Tm³⁺ (1%)@2PA; NP#3: NaGdF₄:Yb³⁺/Er³⁺ (20/2%)@NaYF₄@NaYbF₄:Tm³⁺ (1%)@2PA; NP#4: NaGdF₄:Yb³⁺/Tm³⁺ (49/1%)@NaYF₄; NP#5: NaErF₄:Tm³⁺ (0.5%)@NaYF₄; NP#6: NaGdF₄:Yb³⁺/Er³⁺ (20/2%)@NaYF₄.

In a further set of experiments, we established a multi-level pattern for logic encryption by integrating the stimuli-responsive and excitation wavelength-dependent features of 2PA-capped UCNPs. In our design, photonic color output codes are programmable by controlling the input parameters of temperature and excitation wavelength, and the cryptographic data can be decoded via prescribed logical operations⁴⁵. As illustrated in Fig. 6c, we devised an array of hybrid textual information consisting of four sets of alphabetical codes using different UCNPs. The encrypted information of “UCNP” can only be deciphered by a specific combination of excitation wavelength and thermal conditions followed by an AND logical operation. These results substantiated the great superiorities of thermochromic upconversion for high-security data storage, with the advantages of direct visualization and easy authentication. As an added benefit, the thermal control over the upconversion emissions can also be harnessed for high-performance ratiometric thermometry, enabling superior thermal sensitivity and resolution (S_r-max = 12.0% K^-1 and δT_min = 0.06 K, Supplementary Fig. 50 and Supplementary Table 5).