The principles and comparison of Tm3+/Yb3+-doped UCPP and Ho3+/Yb3+-doped UCNPs-TTA CUCPP are illustrated in Fig. 1(a). The UCPP process occurs under NIR excitation, where Tm3+/Yb3+-doped UCNPs directly radiate UV light inside the photoresist. The upconverted UV light is directly used to initiate photopolymerization. The intensity of the UV light decreases with the distance from the Tm3+/Yb3+-doped UCNP light source, and above a certain light intensity threshold, the photoinitiator generates sufficient radicals to trigger the photopolymerization of monomers. The underlying challenge stems from the fact that, upon NIR excitation, UCNPs emit upconverted light isotropically inside the photoresist. Once the intensity of this upconverted light surpasses the intensity threshold for initiating photopolymerization, crosslinking reactions occur, leading to the formation of polymerized regions. However, this isotropic nature of the upconverted emission complicates the precise control of the photopolymerized resolution, resulting in gradual enlargement polymerized zone and pronounced edge roughness over time. In contrast, CUCPP is the combined effect of two upconversion processes: (1) the Ho3+/Yb3+-doped UCNPs upconvert the 980 nm excitation and emit light at 645 nm; (2) This upconverted 645 nm light is rather used to directly initiate photopolymerization but absorbed by the TTA molecules, which undergoes TTA upconversion process and emits UV light. Finally, Energy is transferred to UV photoinitiator, thus inducing it to generate free radicals to initiate the photopolymerization. Notably, the upconversion of TTA is a two-photon process, and the intensity of UV emission is approximately proportional to the square of the excitation light intensity, and only the TTA molecules in the region near the Ho3+/Yb3+-doped UCNPs can emit sufficiently strong UV light to initiate the photopolymerization. Therefore, the photopolymerization region is controlled in a very small range. Figures 1(b, c) demonstrate the actual luminescence effects of Tm3+/Yb3+ and Ho3+/Yb3+ -doped UCNPs within photoresist, as well as the changes in the photopolymerization region diameters of the two methods over time under the same 980 nm light intensity. The full video can be referred to in Supplementary Media 1. Post-development scanning electron microscopy (SEM) results, as shown in Fig. 1(e), allow for a more intuitive comparison. It can be observed that the line structure produced by UCPP exhibits uneven thickness and an apparent diffuse shape. In contrast, CUCPP demonstrates a clear advantage in overcoming dispersion. Videos of both photolithography processes are available in Supplementary Media 2.
Highly-concentration core-shell-shell β-NaYF4: Ho3+/Yb3+@NaYbF4@NaYF4 UCNPs were selected as the internal red light emission sources in the photoresist. The energy level structure and entire energy transfer process of CUCPP is shown in Fig. 2(a). Yb3+ acts as a sensitizer with a large absorption cross-section at 980 nm, efficiently absorbing NIR light at 980 nm and transferring energy to Ho3+. Ho3+ undergoes excited state absorption and emits anti-Stokes light at 645 nm. The TTA system, which absorbs light at 645 nm and emits UV light, was chosen as 2,3-butanedione (biacetyl, sensitizer)/2,5-diphenylazole (PPO, annihilator). The electrons in the ground state of sensitizer absorb photons emitted by UCNPs, and enter an excited state. The excited electrons relax back to the ground state through nonradiative relaxation, spontaneous emission, or intersystem crossing (ISC). The ground state electrons of the annihilator are sequentially excited twice by absorbing the ISC energy of the photosensitizer. Similarly, the excited electrons relax back to the ground state through nonradiative relaxation or spontaneous emission. Due to the presence of similar energy gaps, the relaxation energy can be transferred to the photoinitiator. Under the same excitation and relaxation conditions, the excited photoinitiator is converted to a triplet state through ISC. The photoinitiator in the triplet state decomposes into free radicals, which initiate the photopolymerization of monomers.
Ho3+ exhibits distinct upconversion fluorescence peaks at 540 nm, 645 nm, and 750 nm. Figure 2(b) tested the TTA system under excitation with these three monochromatic wavelengths, and a blank sample containing only photoinitiator and monomer was prepared as a contrast at 645 nm. The breaks in the lines indicate the harmonic wavelengths. Only the 645 nm light can efficiently excite the production of UV light, confirming that the energy from Ho3+/Yb3+-doped UCNPs is transferred to TTA molecules via 645 nm light. Figure 2(c) shows the emission spectrum of the TTA system at 365 nm and the photoluminescence (PL) of core-shell-shell β-NaYF4: Ho3+/Yb3+@NaYbF4@NaYF4 UCNPs. The TTA system can absorb light within the range of 600–760 nm and emit UV light. The peak at 645 nm for Ho3+ corresponds to anti-Stokes light emitted from the electronic transition 5F5-5I8, which falls within the absorption range of the TTA system. Figure 2(d) shows the absorption spectrum of the photoinitiator 184 and photoluminescence spectrum of the TTA system. Both have a favorable overlap region in the UV range, enabling efficient transmission of energy.
The intensity of anti-Stokes luminescence follows a power-law relationship with the excitation light intensity. Figures 3(b), (e), and (h) show double-logarithmic curves of excitation light intensity versus emission light intensity for three cases: 980 nm excitation of Ho3+/Yb3+-doped UCNPs producing 645 nm light, 645 nm excitation of the TTA system producing 365 nm light, and 980 nm excitation of the UCNPs-TTA hybrid system producing 365 nm light. Under 980 nm NIR light irradiation, the electron transition of Ho3+ from 5F5 to 5I8 can be described by a power index n = 1.406 of the excitation photon energy (980 nm), which is necessary for the electron to reach the 5F5 excited state. That is to say, the intensity of the emission at 645 nm is approximately proportional to the 1.406th power of the excitation light intensity. Likewise, under 645 nm light irradiation, the emission intensity of the TTA system at 365 nm is approximately proportional to the 1.797th power of the excitation light intensity. When the Ho3+/Yb3+-doped UCNPs-TTA hybrid system is directly irradiated with 980 nm NIR light, the luminescence intensity at 365 nm is proportional to the 3.102nd power of the NIR excitation light intensity. Since 3.102 > 3, and the overall process's power index is approximately equal to the sum of the previous two process's power index, the entire process cannot be a three-photon process, but a four-photon process.
Numerical simulations of the photopolymerization area's evolution under laser irradiation over time further demonstrate the advantages of achieving low resolution via CUCPP. The detailed numerical simulations procedure can be found in the Supplementary Numerical Simulation Method. Figure 3(c) and 3(f) respectively show the double logarithmic relationships of 645 nm upconversion light from Ho3+ and 365nm light from the TTA system with the intensity of the 980 nm and 645 nm excitation light. Both processes are consistent with the results obtained from actual measurements. By combining the two processes, it is possible to derive the change in UV light emission from the Ho3+/Yb3+-doped UCNPs-TTA hybrid system with the intensity of the 980 nm excitation light, as shown in Fig. 3(i). The double logarithmic curve has a value of 3.2, which is close to the actual measurement results. Using a 980 nm Gaussian beam to irradiate a 60 µm × 60 µm area, and after obtaining the distribution of 645 nm light intensity, the point spread function is used to connect the luminescence processes of UCNPs and TTA. This enables determination of the polymerization area size at specific irradiation times and calculation of the relationship between the radius of the photopolymerization region and irradiation time, intensity of the 980 nm laser. As shown in Fig. 3(j), the UCPP region diameter continuously increases over time. In contrast, during the first second, the CUCPP region diameter increases with time and then stabilizes. This characteristic indicates that the UV light emitted by the Ho3+/Yb3+-doped UCNPs-TTA hybrid system is more concentrated at the irradiation center, effectively overcoming the diffusion. The numerical simulation results are in good agreement with the experimental measurements. By comparing Fig. 3(k), it can also be observed that the CUCPP diameter increases more slowly with the increase of excitation light intensity.
High-concentration doped core-shell-shell β-NaYF4: Ho3+/Yb3+@NaYbF4@NaYF4 UCNPs were synthesized via two-step hydrothermal method, and successfully incorporated into the photoresist after surface modification. Other materials used include: dipentaerythritol pentaacrylate (SR399), 1-hydroxycyclohexyl phenyl ketone (photoinitiator 184), 2,3-butanedione (biacetyl), 2,5-diphenylazole (PPO), and 6-Mercapto-1-hexanol. The molecular structures of all materials are shown in Fig. 4(a, b). All materials were used as received without further purification.
The synthesis method for the β-NaYF4:20%Yb,10%Ho cores can be found in the Supplementary Methods. Growing two shell layers outside the nanoparticle core significantly enhances the luminescence efficiency of Ho3+. Here is the method for growing the shell layers. Adding a water solution containing NaCF3COO (4.0 mmol) and Yb(CF3COO)3 (4.0 mmol) to a mixture of 8 mL oleic acid (OA, 90%) and 10 mL 1-octadecene (ODE, > 90%), both of which were placed in a 50 mL flask. Nanoparticles were previously dispersed in cyclohexane. The mixture was then heated at 100°C for 20 minutes and further heated at 130°C for 30 minutes to form Yb-oleate complex. After cooling to room temperature, the NaYF4:20%Yb,10%Ho cores were added to a flask, and the mixture was heated at 100 ℃ for 20 minutes to evaporate off cyclohexane. Finally, the mixture was heated to 320°C and maintained under vigorous stirring in Ar flow for 45 minutes. The core-shell UCNPs were precipitated by adding ethanol and collected by centrifugation at 10,000 rpm for 10 minutes. By repeating the aforementioned steps using freshly collected β-NaYF4: Ho3+/Yb3+@NaYbF4 UCNPs as cores and substituting Yb(CF3COO)3 with Y(CF3COO)3, an additional shell layer was grown to obtain core-shell-shell β-NaYF4: Ho3+/Yb3+@NaYbF4@NaYF4 UCNPs.
As shown in Fig. 4(c, d), the UCNPs exhibit a hexagonal crystal structure with a diameter of 12 nm. The powder X-ray Diffraction (XRD) pattern of β-NaYF4: Ho3+/Yb3+@NaYbF4@NaYF4 confirms the formation of hexagonal phase and is in good agreement with the JCPDS data (card no. 28–1192, a = 5.960 Å, c = 3.510 Å, Space group P63/m). Compared to core-shell structures, the spatial separation distribution of sensitizers Yb3+ and activators Ho3+ in the adjacent layers of the bilayer core-shell-shell UCNPs can significantly suppress their cross-relaxation, enabling efficient energy transfer at the interface and enhancing the emission efficiency of Ho3+ by 10 to 100 times. Furthermore, the core-shell-shell effectively suppresses concentration quenching. By experiment measurement, the upconversion luminescence efficiency of UCNPs is positively correlated with the concentration of Ho3+ in the core layer within the range of 0–20%, and obvious luminescence weakening occurs only when the concentration of Ho3+ in the core layer is increased above 30%. The sensitizers and activators used in the CUCPP were Yb3+ (20%) and Ho3+ (10%) in the core layer of UCNPs.
The surface modification of UCNPs was performed by mixing 6-Mercapto-1-hexanol into the cyclohexane solution containing UCNPs. The synthesized UCNPs have alkyl groups on their surface and are insoluble in acrylate-based photoresist monomers. 6-Mercapto-1-hexanol can break the alkyl groups at the carboxyl sites and replace them with thiol groups, resulting in high solubility of UCNPs in SR399. A mixture of surface-modified UCNPs solution and 184 was mixed with SR399 at room temperature for 1 hour, centrifuged to remove the precipitate, and then dried in a vacuum oven for 5 hours to obtain the polymer sample. As shown in Fig. 4(e), UCNPs exhibit excellent dispersion in the polymer, which facilitates uniform outward diffusion of upconverted luminescence upon NIR excitation.
CUCPP can approach the limit resolution resolution achievable with single-beam exposure. The device used for CUCPP is a self-assembled continuous laser direct writing system operating at 980 nm. The laser used is a continuous-wave 980 nm solid-state laser produced by Yuanming Laser in China, priced at $300, emitting a circular Gaussian beam. As shown in Supplementary Fig. 6, The laser beam is expanded and focused on the substrate. The substrate is controlled by custom-designed software to manipulate the stage movement and the opening, closing of the optical shutter. The software can slice 3D structures layer by layer and plan the path for each layer to achieve additive manufacturing of 3D structures. The limit resolution achievable by a single 980 nm excitation beam is approximately \(\:\frac{\lambda\:}{2NA\bullet\:\sqrt{3.1}}\approx\:200\:\text{n}\text{m}\) . Figures 5(a) show SEM images of line patterned samples with an excitation intensity of 105 W/cm2 and a laser writing speed of 80 µm/s. The structural resolution is 290 nm, which approaches the limit resolution. During the CUCPP process, the emission of red light from the UCNPs remains a diffusion-based process, leading inevitably to some blurring of the structure, which is why reaching the ultimate resolution is challenging.
CUCPP can be used to print two-dimensional patterns and 3D structures. Figure 5(b, c) shows SEM images of the two-dimensional panda pattern polymer samples that emit upconversion red light upon irradiation with 980 nm light. Figure 5(d-f) presents oblique top-view SEM images of a 3D regular dodecahedron structure and its top-view luminescent image under 980 nm light illumination at different heights. The 3D regular dodecahedron structure is produced through layer-by-layer line scanning, with each edge length of 15 µm and lateral resolution of 3 µm. Figure 5(g, h) depicts top-view SEM images of a 3D woodpile structure, where the overall design dimensions are 40 µm × 40 µm, the width between lines is 3 µm, the layer height is 1 µm, and each line has a resolution of 480 µm. Laser fabrication is achieved through repeated motion to build up the layers.