The optical diffraction barrier has always been an ultimate constrain for researchers in exploring the nanoscale world. The emergence of stimulated emission depletion (STED) technology has proven to be a deterministic functional technique in providing a subversive way to erupt this limitation^[1–8] and has a wide-range application in the fields of laser direct writing^[9–12], optical data storage^[13–15], and biology^[16–19]. The spatial resolution of STED microscopy is determined by the three-dimensional (3D) coincidence degree of the donut depletion beam to the solid excitation beam to a great extent.

Conventional dual-beam alignment is usually accomplished by scanning gold nanoparticles with two laser beams individually and then matching their positions by adaptive position adjustment. This adaptive adjustment requires multiple scans of two laser beams, which is certainly time-consuming, and there is always a trade-off between the scan speed and the resolution of the image. Besides, there are several disadvantages such as mismatches between the scattering and the fluorescence modes and drift errors caused by non-real-time observation. Adaptive optics with the inclusion of a spatial light modulator is implemented for automatic alignment of STED microscopies by directly imaging fluorescent samples^[20]. The fluorescence lifetime distribution alignment method is further developed and can reduce the drift error by single-frame imaging^[21]. Although the alignment methods using adaptive optics^[20] and fluorescence lifetime distribution^[21] are employed to overcome the above issues of mismatch modes and drift errors, they are still in the raster scanning mode, which is time-consuming, especially for potential scans of (more than two) multiple beams. Besides, the latter one is only applied to a pulsed excitation beam. The dual-wedge phase plate method can achieve the spatial coincidence of the two beams directly without raster scanning alignment using a single optical path^[22]. However, each designed phase plate only corresponds to a specific group of excitation and depletion beam wavelengths.

To solve the aforementioned problems, this study proposes a novel, fast dual-beam adjustment method for STED microscopies, which directly measures the aggregation-induced emission (AIE) fluorescence spectrum from the labeled spots in the tetraphenylethene (TPE)-doped resin. The position of the first excitation beam is recorded by fabricating the labeled spot. Thus, it does not require scanning the first excitation beam. Furthermore, randomly distributed nanoparticles in conventional methods are easy to aggregate, while our method overcomes it by manually fabricating the deterministic labeled spots. We choose AIE dye resin, as it can form deterministic AIE fluorescence spots. Besides, multiple combinations of excitation and depletion beam wavelengths are possible. Furthermore, the fast-dynamic observation of the fluorescence spectrum of labeled spots can eliminate drift errors^[23]. This adjustment method can promise ultra-high adjustment resolutions of 22 nm in the $x$ relative displacement and 27 nm in the $z$ relative displacement for spatial matching of two laser beams. The demonstrated method is simple and universal, and will apply to scenarios if three or more beams or fields of view of different objectives need to be matched in space.

The STED system of dual-beams’ adjustment is illustrated in Fig. 1(a). The Gaussian-shaped excitation beam is a 515 nm femtosecond (fs) laser at a repetition rate of 40 MHz with 500 fs pulse width. The other inhibition beam is a 405 nm CW laser with a doughnut shape. The process of alignment of two beams of a laser is to initially write a fluorescent-labeled spot inside the material based on the polymerization reaction of the photo-initiator in the resin and the 515 nm fs excitation laser. This labeled spot contains the exact 3D position information of the excitation beam [Fig. 1(b)] and generates the AIE effect on the curing sample. The polymerization process is the reaction of the first laser beam and photo-initiator, which is a prerequisite for the fluorescence emission of the aggregated AIE excited by the first and second laser beams of the STED system for the dual-beams’ adjustment in the later steps. The type of the first laser is not limited to a fs pulsed laser, but also can be other pulsed or CW lasers, as long as the wavelength of the first laser falls in the absorption band of the selected photo-initiator. Then, the second inhibition beam requires adjustment, resulting in reduced adjustment time. Finally, the 3D coincident adjustment is achieved expeditiously by detecting the variation of fluorescence intensity through the spectrometer instead of raster scanning spatial distributions of the two laser beams. The camera shown on the left side of Fig. 1(a) is used for coarse pre-alignment first. The fluorescence intensity will be at its peak when there is no spatial mismatch between the excitation and the inhibition laser beams, both along the lateral direction of the $x/y$ relative displacement and $z$ relative displacement axes [Fig. 1(c)]. It is worth noting that for dual-beam spatial adjustment, inhibition beams of STED microscopes are not used to inhibit but to excite the spontaneous emission of the AIE-based fluorescent sample with TPE. Our AIE-based fluorescent material is for beam alignment, and the imaging sample for STED microscopies will be changed to other different types of materials.

For the preparation of the AIE TPE-doped resin, the five-functional monomer dipentaerythritol pentaacrylate (DTPA), TPE dye, 2-hydroxy-2-merthylpropiophenone (HMPP), and photo-initiator 1173 were all provided by Sigma. The monomer DTPA, TPE dye, HMPP, and photo-initiator 1173 were dissolved in acetone and later mixed uniformly by ultrasonic vibration. To remove the acetone, the sample material was kept in the oven at 50°C for 12 h. Afterward, the sample material was assembled by dropping about 0.5 mL resist onto the slide and compressing another slide to clamp the photoresist between them, which were separated by a 40 µm thick tape spacer. Thereafter, the material was cured by a high-power (200 W) hand-held UV irradiation for 20 s.

Different from the conventional dyes, the AIE-based fluorescent radiation’s intensity increases with the increase of aggregation state^[24–26]. The excitation spectrum of the sample is from 250 to 410 nm (red solid line). Comparing the curing samples with (yellow solid line) and without TPE dye (blue solid line), it is evident that the samples with TPE dye can radiate intense fluorescence from 400 to 600 nm [Fig. 2(a)]. Figure 2(b) shows different wavelength ranges of inhibition beams corresponding to the AIE dye selection. Except for our specifically designed sample with TPE, other potential AIE dyes can also be used as the sample material for dual-beam alignment. For example, distyreneanthracene (DAS) (TTVP dyes) can be selected when the inhibition beam range is 350 to 500 nm (390 to 620 nm)^[27,28]. This means that the AIE-induced dye resin sample is applicable for the dual-beam adjustment process in many cases as long as excitation and inhibition beams of STED microscopes fall in the spectral range of 250–620 nm (i.e., single-photon excitation of AIE-based fluorescence) or, in the spectral range of 500–1240 nm, covering the entire visible and part of the near-infrared range (i.e., two-photon excitation of AIE-based fluorescence).

For the first step, as the power of the 515 nm fs laser beam increases, the feature size of the labeled spot will also increase. Lecia SP8 confocal microscopy was performed at the National Center for Protein Science Shanghai to get the measurement result in Fig. 3(a). As the power of the excitation beam reaches 0.2 mW, the labeled spot size is 230 nm. When the power is increased to 0.3 mW, the size is 271 nm. The power of the excitation beam is further increased to 0.7 mW, and the size could reach 396 nm [Fig. 3(a)]. Thus, for the alignment process, the excitation beam power is chosen to be 0.3 mW since it can form an appropriate feature size of the labeled spot, which should not be too small. If the lateral $(x,y)$ size is too small, the alignment accuracy will be enhanced, but the observation of the fluorescence intensity will be challenging due to the rather weak signal. Meanwhile, the accuracy of the dynamic position adjustment of the two beams depends on the lateral size of the labeled spot, which should not be overlarge. The scanning electron microscope (SEM) image of the labeled spot into the resin sample with 0.3 mW of the beam power shows that the lateral $(x,y)$ dimension is 271 nm and the axial ($z$) dimension is 992 nm [Fig. 3(b)].

For the second step, it is obligatory to carry out lateral $(x,y)$ coarse pre-alignment to move the inhibition beam to the position of the 515 nm excitation beam, as shown in Fig. 4(a). Once the two beams are adjusted to enter the image view of the camera in parallel, their relative initial position deviates slightly [Fig. 4(a-1)]. It should be mentioned that the 515 nm fs excitation laser and the 405 nm CW laser are blocked by the dichroic mirror in Fig. 1(a). The small green spot (excited by the 515 nm fs laser) and large red spot (excited by the 405 nm CW laser) in Fig. 4(a-1) are the emitted fluorescence light of the AIE dye resin sample with TPE imaged by the camera on the top side of Fig. 1(a). The 515 nm fs laser has a relatively small intensity in order to generate a small-sized recorded spot, which enhances the adjustment resolution. Thus, the green emission spot is relatively small, which is the fluorescence emission of the recorded spot excited by the 515 nm laser. Then, the position of the 405 nm CW laser beam is adjusted to approach the labeled spot gradually [Fig. 4(a-2)] and to move forward further away from the labeled spot [Fig. 4(a-3)]. The alteration in fluorescence intensity at various positions validates that the closer the inhibition beam is to the center of the labeled spot, the stronger the detected fluorescence intensity is [Fig. 4(b)]. The process of coarse pre-alignment of two beams (solid beams) in the lateral $(x,y)$ direction is achieved by spectrometer dynamic observation, reducing the drift error caused by delayed imaging of raster scanning in other conventional methods. Both directions of the $x/y$ relative displacement and the $z$ relative displacement are required to be adjusted to ensure the spatial coincidence of the two beams in three dimensions. The adjustment in the lateral $(x,y)$ direction is the first step before the fluorescence observation of the labeled spot for the coincidence adjustment process in the next direction of the $z$ relative displacement.

The fine alignment of the two beams with ultra-high accuracy for conventional methods is usually based on raster scanning in the 3D space. However, it will certainly be most challenging without raster scanning. We provide an effective method to solve the above problems, as shown in Fig. 5, by slightly moving the position of the 405 nm CW laser beam to detect the fluorescence intensity change of the labeled spots. The fluorescence intensity is 583 (with the arbitrary unit) at the peak when the initial lateral $(x,y)$ distance between the two beams is $\mathrm{\Delta }x=135\mathrm{nm}$. The fluorescence intensity increases to 675 as $\mathrm{\Delta }x$ decreases to 5 nm. While the inhibition 405 nm CW laser beam moves forward, the fluorescence intensity decreases with $\mathrm{\Delta }x=-134\mathrm{nm}$ and the intensity of 586. The scale values shown in Figs. 5(b) and 5(d) are the value of the staff gauge of the movement stage for controlling the position of the beams. The division of the difference between the maximum and the minimum $\mathrm{\Delta }x$ ($135\mathrm{nm}+134\mathrm{nm}=269\mathrm{nm}$) and the corresponding difference of the scale values ($5.00\mathrm{mm}-4.77\mathrm{mm}=0.23\mathrm{mm}$) is 11 nm/0.01 mm for a minimum cell of the scale values (0.01 mm). Thus, a resolution of 22 nm for spatial matching of two beams along the lateral direction of $x/y$ relative displacement is achieved, which is limited by the sensitivity of the spectrometer. Similarly, we also calculate the matching resolution along the axial ($z$) direction. The maximum intensity at the peak is 766 when $\mathrm{\Delta }z=17\mathrm{nm}$. The division of the difference between the maximum and the minimum $\mathrm{\Delta }z$ ($496\mathrm{nm}+492\mathrm{nm}=988\mathrm{nm}$) and the corresponding difference of the scale values ($9.44\mathrm{mm}-0.67\mathrm{mm}=8.77\mathrm{mm}$) is 1 nm/0.01 mm. Thus, a resolution of 27 nm for spatial matching of two beams along the direction of $z$ relative displacement is achieved.

Normally, after the direction of $z$ relative displacement coincidence adjustment, it will induce a slight mismatch of the lateral $(x,y)$ coincidence of the two beams. Thus, the lateral $(x,y)$ alignment should follow the direction ($z$) alignment iteratively until the fluorescence intensity approaches a maximum value. Once the coincidence process is finished, we then add the vortex phase plate in the optical setup to form a doughnut shape for the inhibition beam. When the excitation beam is entirely filled inside the hollow part of the inhibition beam, the 3D coincidence of the two beams is completed.

This study demonstrates a dual-beam alignment method for STED characterized by using the AIE dye resin sample with TPE. Contrasted with conventional methods, it just requires forming a labeled spot with the 3D position information and adjusting the inhibition beam to match the excitation beam, resulting in the fast accomplishment of the adjustment. The fluorescence light intensity from the resin sample excited by the two beams contains the coincidence status of their lateral and axial positions. Resolutions of 22 nm along the lateral direction of $x/y$ relative displacement and 27 nm along the direction of $z$ relative displacement for spatial matching of two beams are achieved. Furthermore, following this strategy, many other kinds of AIE dyes covering a wide range of wavelengths can be used to fabricate the resin sample, which will be suitable for dual-beam adjustment of most STED microscopies. Except for STED microscopies, this method will be further applied in other types of microscopies or scenarios, which require the spatial matching of three or more laser beams, and the fields of view of different objectives with ultra-high precision.