Optically-driven organic nano-step actuator for reconfigurable photonic circuits

SJ_Zhang
Nov. 16, 2025

Abstract

The assembling and reconfiguration of the integrated devices are of great importance to extend the capability of photonic chips based on top-down fabrication approaches. Here, we demonstrate a fully-programmable organic micro-actuator for precise manipulation of on-chip microstructures. Controlled by a low-power laser, the micro-actuator achieves a 30 nm motion step size, and shows the capability to traverse various chip substrates, overcome obstacles, and push microspheres to target locations. The micro-actuator is applied to fine-tune the microcavity and shift the resonance by three linewidths without compromising its quality factor. This optically-driven micro-actuator offers a unique approach for post-fabrication assembly and reconfiguration of photonic circuits, paving the way for adaptive, multifunctional photonic systems.

Introduction

In recent decades, integrated circuits have achieved remarkable success, revolutionizing various aspects of our daily lives^1,2,3,4. In particular, with the development of complementary metal-oxide-semiconductor (CMOS) technology, photonic integrated circuits (PICs) have advanced rapidly and have been applied in various fields, including telecommunications⁵, sensing^6,7, computing^8,9, photonic accelerated machine learning^10,11,12, and quantum technologies^13,14,15. Standard top-down fabrication processes allow for the design and fabrication of various photonic microstructures through etching or depositing materials layer-by-layer on planar substrates. This approach enables the monolithic fabrication of compact and efficient photonic devices, including waveguides, modulators, resonators, and detectors, and offers unprecedented functionalities in manipulating photons at the micro- and nano-scales. However, conventional top-down fabrication approaches limit the available optical materials and reconfigurability of the fabricated photonic structures, hindering the flexibility and adaptability of PICs⁸. To address these limitations, bottom-up assembly of micro-/nano-structures on a chip provides a promising alternative approach^16,17,18. By providing basic building blocks of micro-/nano-structures, functional devices can potentially be assembled through the arrangement of these building blocks, and the fabricated devices can be dynamically reconfigured in situ via precise re-arrangement of the building blocks within the circuits.

Various techniques have been explored for the manipulation and assembly of micro-/nano-structures. Near-field probes, such as atomic force microscopy tips or scanning tunneling microscopy tips, offer one possible realization of high-precision positioning and manipulation^19,20,21. However, the required nanometer positioning and mechanical stabilization are highly complex, and direct contact with materials may cause damage. Optical tweezers, which utilize focused laser beams to trap and manipulate particles, have been widely employed for assembling microstructures^22,23. While optical tweezers provide far-field non-contact manipulation, they typically require an aqueous environment, presenting significant challenges in maintaining PICs from contamination when immersed in liquid. Recently, micro- and nano-actuators, which are devices capable of generating mechanical motion or forces at small scales, have emerged as promising tools for manipulating and assembling micro- and nano-structures on a substrate^{24,25,26,27,28}. Various non-contacting stimuli, such as acoustic fields²⁴, magnetic fields^29,30,31, or optical fields^{32,33,34,35,36} have been applied to drive these actuators.

Although most demonstrations are limited to aqueous environments, significant advances have been made in optically driven micro-actuators operating in ambient conditions. Early pioneering works demonstrated light-induced crawling of organic crystals on solid surfaces³⁷ and light-fueled microscopic walkers³⁸, establishing the fundamental possibility of optical actuation in non-aqueous environments. More recently, Tang et al.³⁹ demonstrated precise opto-thermo-mechanical actuation of gold microplates with nanometer-level step resolution and programmable directional control, and the photothermal shock approach is further developed in subsequent studies^40,41,42. Similarly, Linghu et al.⁴³ achieved nanowire manipulation with sub-nanometer precision through plasmon-driven actuation on solid surfaces, while Chen et al.⁴⁴ advanced this field by developing photoacoustic actuators with two-dimensional control capabilities. Our previous work^45,46 established the fundamental mechanism of light-driven organic crystal actuators, revealing how laser-induced transient elastic lattice deformation enables controlled motion. However, achieving high precision and a fully programmable arrangement of these actuators remains a challenge.

Here, a fully programmable optical-driven organic micro-actuator operating in ambient conditions is experimentally demonstrated on a photonic chip. The micro-actuator can be activated with a laser beam when its power exceeds a threshold, exhibiting repetitive forward and backward motion as well as left and right turning by controlling the trajectory of the scanning laser beam. We achieve a motion precision of 30.2 ± 1.1 μm per 1000 control steps (30 nm per step). We apply the programmable micro-actuator to a photonic chip to precisely tune photonic structures, demonstrating the ability to move across barrier waveguides and shift the resonance of a microring by up to three linewidths without compromising the optical loss. Our demonstration provides an approach for controlling, assembling, and reconfiguring integrated devices on a chip in a non-aqueous environment, and holds great potential for bottom-up construction of functional integrated circuits.

Results

Figure 1a, b presents a schematic illustration of the optical-driven organic micro-actuator that enables precise manipulation and control of on-chip microstructures in a non-aqueous environment. The micro-actuator is made from organic molecular microcrystals based on indigo-based donor-acceptor (D-A) molecules (Fig. 1c, and see Supplementary Note 1 for more details). A far-field laser beam can actuate the micro-actuator to move or rotate on a flat substrate. As illustrated in Fig. 1b, these micro-actuators offer a versatile toolset for microscale manipulation, opening up different possibilities for bottom-up assembly and reconfiguration of PICs. The organic microcrystals are prepared through self-assembly in a dichloromethane/ethanol mixture, with lengths ranging from 10 to 200 μm, widths from 1 to 10 μm, and thicknesses between 1 and 2 μm. A scanning electron microscopy image (Fig. 1d) reveals a typical micro-actuator with a bar-shaped morphology, measuring 9.7 μm in length and 1.2 μm in width. Single-crystal X-ray analysis reveals that the organic micro-actuator is part of the monoclinic crystal system, characterized by a space group of P2_1/c. The unit cell dimensions are 15.83 Å, 15.46 Å, and 8.92 Å along the a, b, and c axes, respectively (see Supplementary Table 1 for more details). The absorption spectrum shows that the microcrystals have a wide absorption range from 300 to 700 nm with negligible fluorescence emission (quantum yield 0.5%, see Supplementary Fig. 3 for more details). When exposed to a focused laser beam within this wavelength range, the strong absorption cross-section of the indigo-based microcrystals leads to lattice deformation⁴⁶.

Fig. 1: Experimental setup and demonstration of organic micro-actuators.

a The experimental setup for manipulating and imaging the micro-actuators on a chip. The scanning galvo system is the key component for generating programmable, arbitrary laser trajectories to optically drive the actuator. BS: beam splitter, CCD: charge-coupled device. b Schematic illustration of the actuator’s movements due to the scanning laser beam, showing the assembling of the letter “U" after an actuator has moved in a straight line, turned around, and crossed a waveguide. The white dashed curve represents the movement trajectory of the actuator. c Molecular structure of the micro-actuator. d SEM image of a typical micro-actuator on a glass substrate, with the length and width being 9.7 μm and 1.2 μm, respectively. Scale bar: 2 μm. e Measurement of the micro-actuator’s displacement as a function of laser power, revealing a threshold behavior.

To realize the programmable actuation of the microcrystals, we employ a scanning galvo system shown in Fig. 1a to directionally scan the laser spot across the sample along its long side (see Supplementary Note 2, Section A for details). The sequential transient expansion and restoration of neighboring lattices synchronize, resulting in a directional motion of the micro-actuator. This optically-induced actuation can generate a driving force on the micro-newton (μN) scale⁴⁶, sufficient to propel the micro-actuator and its cargo across the substrate. In our experiments, the micro-actuators are driven by a continuous wave 405 nm laser, with the waist of the Gaussian laser spot size of around 0.9 μm on the micro-actuator and a fixed scanning speed of 62 μm/ms (see Supplementary Note 2, Section B for details). To characterize the nanometer-scale motion of the micro-actuator, we first investigate its displacement on a glass substrate with superhydrophobic treatment by scanning the laser spot and measuring the accumulated spatial displacement over 1000 steps. Figure 1e shows typical results of this displacement characterization with various laser powers, revealing that the micro-actuator exhibits significant displacement only when the power exceeds a threshold value (see Supplementary Note 2, Section E for details).

As previously studied in ref. ⁴⁶, the actuator’s directional motion stems from spatially and temporally coordinated lattice deformations along the long axis (c-axis) of the micro-crystal induced by the scanning laser. While the underlying microscopic mechanism, whether the photothermal effects or other photoinduced processes, remains under investigation, the localized laser-material interaction clearly induces localized lattice deformation that drives the controlled motion. As the laser spot illuminates successive regions along the actuator, it creates a propagating wave of expansion and contraction, reminiscent of a caterpillar’s locomotion, where coordinated muscle contractions produce net displacement. This motion preferentially occurs along the micro-actuator’s long axis due to the anisotropic crystal structure with different elastic moduli along different crystallographic directions. We conjecture that precise and programmable motion of the micro-actuator can be achieved by carefully controlling the scanning trajectory of the laser spot. For instance, by selectively scanning one edge of the micro-actuator, we can create asymmetric deformation across its width, generating torque that enables controlled rotation in addition to straight motion.

To verify the programmable micro-actuator, we designed a laser spot scanning pattern for each step to realize an arbitrary laser spot scanning trajectory. Figure 2a illustrates the trajectory in a single laser scanning step, with the trajectory consists of 10–40 rectangles, with edges are aligned to cover an area much wider than the micro-actuator, ensuring a uniform laser actuation and avoiding misalignment during motion, and scanning direction for all micro-actuators are aligned with their long axis corresponding to the crystallographic c-axis (see Supplementary Note 2, Section C). Each complete scanning cycle lasts 380 ms, with an effective actuation window of just 1.95 ms during which the laser-material interaction drives the motion. The accumulated displacements are measured from the CCD images by comparing the location of the micro-actuator using an image recognition algorithm (see Supplementary Note 2, Section D). Interestingly, the displacement exhibits a threshold behavior with respect to the laser power. Below the threshold of 31.2 μW, the displacement is negligible compared to the measurement precision of around 0.5 μm. In contrast, for a laser power of 38.9 μW, an actuation precision of about 15.4 nm per step is achieved. Above the threshold, the typical step displacement of the actuator ranges from 10 nm to 500 nm, varying with the samples and laser power. Notably, the micro-actuators can operate and show similar behaviors on various substrates or thin films commonly used in photonic chips, including silicon, silicon dioxide, silicon nitride, gallium nitride, aluminum nitride, sapphire, and lithium niobate (see Supplementary Note 3, Section C). While the specific threshold power may vary between individual micro-actuators and the supporting substrates, the threshold behavior is universal and independent of laser polarization. For powers below the threshold, the motion is not recognizable even when accumulating 10⁴ steps. Further investigations reveal that the micro-actuator’s performance depends critically on laser spot diameter, irradiation duration, and scanning step size, while remaining insensitive to laser polarization (see Supplementary Note 2, Section F for more details). A more tightly focused beam and slower scanning result in a lower threshold, with an appropriate speed range of 50–100 μm/ms for scanning. The existence of a power threshold ensures that the micro-actuator’s position remains stable under ambient conditions and is not altered by stray light. These findings lead to a simple rule-of-thumb: the micro-actuator can be driven by a focused laser beam, with the step size increasing only when the power is above the threshold.

Fig. 2: High-precision straight motion and turning of micro-actuators.

a The laser scanning trajectories (the purple traces) for manipulating the actuator’s straight movement (the red arrow). b The motion properties of the actuator are very stable and repeatable, as demonstrated by continuously driving the actuator by reversing the laser scan direction per 1000 steps. After a total of 0.32 million motion steps, a high motion precision of 30.2 ± 1.1 μm per 1000 steps is achieved. c Snapshots of the micro-actuator illustrating one round of forward and reverse movement, as described in (b). d The laser scanning trajectories for controlling the turning of micro-actuators. e Snapshots capturing the continuous turning motion of a micro-actuator. The red arrows represent the movement direction of the actuator. Scale bar: 20 μm.

Figure 2 demonstrates the exceptional controllability and reproducibility of the micro-actuator’s motion, as well as its ability to change direction by modifying the laser scanning trajectory. As indicated in Fig. 2a, the direction of the actuator’s motion is opposite to the direction of scanning laser spots, enabling it to move forward or backward by alternating the laser scanning direction. To demonstrate the long-term stability and reliability of the micro-actuator, we actuated the sample forwardly and backwardly for several days, accumulating 2000 steps in each round. Figure 2b shows a representative section of the micro-actuator’s displacement, which was analyzed by tracking its location after every 100 steps of laser scans (see Supplementary Note 2, Section G for more details). Figure 2c provides typical CCD images showing snapshots of the micro-actuator within one round of movement, as indicated in Fig. 2b. These results demonstrate that the movement of the micro-actuator is repeatable and highly stable, with a displacement of 30.2 ± 1.1 μm for 1000 steps, corresponding to a single step size of approximately 30 nm.

To demonstrate the capability of the organic micro-actuator in changing the direction, the laser scanning trajectory is modified to cover only a portion of the micro-actuator’s surface, as shown in Fig. 2d. By selectively scanning one edge of the micro-actuator, the device can be made to turn either clockwise or counterclockwise, as the two edges experience different moving step sizes. As illustrated by the snapshots in Fig. 2e, the micro-actuator turns its motion direction while moving forward, resembling a right-turning vehicle. Similarly, the micro-actuator can turn left by covering the other edge of the micro-actuator’s surface, or by reversing the direction of laser scanning. Additionally, the movement of the micro-actuator along its long axis can be suppressed during rotation by simultaneously scanning the two edges from opposite directions. It is worth noting that not all micro-actuators followed the turning rule, with about half of them exhibiting this behavior, possibly due to variations in crystalline structure orientation or surface contact conditions. Indeed, our successful demonstration of directional turning by selectively illuminating one edge of the actuator further validates that the actuation mechanism stems from localized laser-material interactions causing asymmetric deformation patterns, rather than uniform thermal expansion.

Figure 3 highlights the advanced micro-manipulation capabilities of the optically driven micro-actuator and its potential for creating complex patterns and structures on a chip. We achieve programmable manipulation of the micro-actuator by controlling the trajectory of laser scans. To move or rotate a selected micro-actuator, we control the galvo mirror to rotate and displace the scanning trajectory, aligning the square laser trajectory with the device’s long axis and translating the trajectory to overlap with the device. During the manipulation process, the laser power and scan direction can be easily switched, enabling precise control over the micro-actuator’s motion. The versatility of this approach is demonstrated in Fig. 3a, where we construct the letter “U" by moving and rotating three micro-actuators in a specific sequence. In particular, the laser scanning trajectories shown in Figs. 3a(i–iii) illustrate the alignment of the trajectories to specific target micro-actuators, with negligible influence on other nearby micro-actuators. Following similar sequences, we can create various letters, numbers, or patterns using multiple micro-actuators. This is further exemplified in Fig. 3b, where we showcase the assembly of the letters “USTC" by randomly selecting and manipulating micro-actuators from a cluster of devices (see Supplementary Movies 1–4 for the details of the assembling process). In addition to self-manipulation, the micro-actuator can also be used to push and deliver other microstructures on the chip. Figure 3c, d presents an example of this capability, where a micro-actuator is programmed to push a silica microsphere (22 μm in diameter) along different directions and eventually deliver the microsphere to a location over a distance of 0.36 mm away. By controlling the micro-actuator’s motion, we can guide the microsphere along different directions with high precision (see Supplementary Movies 5–6). One of the most intriguing aspects of the micro-actuator is its ability to carry objects with a weight comparable to its own.

Fig. 3: Programmable structure assembling by micro-actuators.

a The self-assembling process of several randomly selected actuators to form the letter “U” by programming the motion of micro-actuators. The red arrow indicates the movement direction of the actuators. b The images of the assembled letters “USTC''. c and d Demonstration of arbitrary manipulation of a silica microsphere’s position using an optical actuator. The blue dashed circles indicate the movement destination of the actuator. Scale bar: 20 μm.

These results validate the potential of micro-actuators in arranging complex patterns and shapes with high flexibility and scalability, opening up different possibilities for creating functional structures on a chip. Figure 4a illustrates the application of the programmable micro-actuator for assembling and reconfiguring photonic devices (see Supplementary Note 3, Section A for details of the measurement setup). To utilize the micro-actuator for these applications, it is essential to transfer micro-actuators to chips and test their performance on various chip materials. One approach to prepare the micro-actuator is to directly apply a micro-droplet of solution containing the optical actuators onto idle areas of the chip. The micro-actuators can then be selectively activated and driven to the working zone over a relatively long distance, typically several hundred microns. However, this method risks introducing dust to the surface, which may degrade the qualities of devices within the chip. An alternative approach is to initially prepare the micro-actuator on a clean glass substrate and then selectively pick up and translate the activated micro-actuators to a target location on the chip with a precision of 10 μm. In our experimental setup, we opted for the latter method, using metal probes to transfer the micro-actuators from a glass substrate to the photonic chip in the vicinity of the microring resonator or waveguide²¹ (see Supplementary Note 3, Section B).

Fig. 4: Reconfiguration of photonic integrated circuits (PICs) by a micro-actuator.

a Schematic of in situ tuning of PICs by controlling micro-actuators on a photonic chip. The purple and red dashed box plots depict the situations where the actuator is on and off the microring, respectively. b and c The evolution of transmission spectra of a microring resonator as the micro-actuator crosses it. The resonance frequency of the microring is tuned over a range of about 5.2 GHz. The curve labeled with number (i) corresponds to the transmission spectrum of step 110 in (b), (ii–vii) are the transmission spectra from steps 1492 to 1497, respectively. d Snapshots of micro-actuator crossing the microring resonator, corresponding to the spectra shown in (c).

Figure 4b–d demonstrates the application of the optically driven organic micro-actuator to a PIC based on lithium niobate (LN) microstructures on a sapphire substrate⁴⁷. This emerging platform is particularly attractive for photonics applications due to its exceptional optical properties⁴⁸. Although the programmable motion of micro-actuators on a flat surface is demonstrated, their ability to overcome barriers is crucial for practical applications. The LN optical waveguides and microresonators on the chip have a wedge-type structure with 220 nm thick LN wedges. As depicted in Fig. 4d, the micro-actuator is placed near the microring resonator. The micro-actuator can indeed act as a dielectric material to reconfigure the effective optical refractive index of photonic devices, as the air cladding around the waveguides is replaced by organic material with a higher refractive index. To quantify the impact of the micro-actuator on the microring resonator, we monitor its optical transmission in real-time when the micro-actuator is driven to cross it (Supplementary Note 3, Section A). The evolution of the spectrum of a selected resonance is shown in Fig. 4b, showing a sudden resonance shift at steps around 1500.

The detailed spectra for these steps are plotted in Fig. 4c, demonstrating the ability of the micro-actuator to tune the resonance of the microring. When the micro-actuator is across the microring, the resonance shifts by around 5.2 GHz, corresponding to a shift of three linewidths or 2.5% of the resonator’s free spectral range. The maximal overlap of the micro-actuator and the microring optical path is around 30 μm, indicating a modification of the guided mode’s effective refractive index n_eff, by about 0.0013. Numerical simulations of the guided mode suggest a modification of n_eff is around 0.007 when the micro-actuator is well-fitted on the top surface of the microring, implying a non-perfect contact between the micro-actuator and the microring in current experiments. According to the experimental results, a modulation of the optical phase in the waveguide by 0.1π could be achieved with a single 30 μm-long micro-actuator, providing a powerful way to reconfigure optical circuits. Importantly, the changes in mode linewidth are less than 16% when the micro-actuator crosses the microring, and the linewidth recovers to its initial value once the micro-actuator has left the microring. Therefore, the micro-actuator exhibits low optical loss at telecom wavelengths and does not damage the surface of the microring during the crossing process.

Discussion

We have demonstrated an optically driven organic nano-step micro-actuator that offers precise, non-destructive, and low-power manipulation of microstructures on a chip. The capability of the micro-actuator to precisely tune the resonance of a microring without compromising its optical quality factor promises a universal approach to nonvolatile and reversible reconfiguration of PICs. By leveraging the micro-actuator’s ability to manipulate and position various functional materials, such as gain media^49,50, nonlinear optical materials^51,52, and phase-change materials^53,54, highly efficient, multifunctional, and intelligent photonic systems can be created for various applications, including neuromorphic computing. Moreover, the principles and techniques developed in this work could be extended to other fields, such as microelectromechanical systems, hybrid electronic-photonic devices, and nanorobotics, where precise manipulation and actuation at the micro- and nanoscale are crucial.

However, there is still room for improvement and further development of this technology. Firstly, the long-term stability and controllability of the microcrystals could be enhanced by optimizing the synthesis process and exploring alternative organic materials. Secondly, the precision of the nanometer-scale step size is currently limited by the accuracy of the laser scanning system. The single-scan stability could be improved by optimizing the optical setup, enabling even higher precision control of the micro-actuator’s motion. Additionally, the speed of micro-actuators can be significantly improved by reducing the redundancy in the laser scanning trajectories. If the single-step scanning length could be reduced to hundreds of microns, actuation speeds up to 100 steps per second would be possible. Thirdly, developing an advanced feedback control system would enable the micro-actuator to adapt to changes in its environment. By implementing real-time recognition of the micro-actuator’s location and orientation, along with dynamic adjustment of the scanning trajectory, more efficient and precise manipulation of on-chip components could be achieved. Lastly, future work should investigate the transient thermomechanical processes during laser actuation, potentially employing analytical frameworks such as those developed for photothermal shock processes⁴⁰ to quantify temperature gradients, stress-strain evolution, and resulting forces that drive the actuator’s motion. Future investigations using pulsed laser excitation, as demonstrated in ref. ⁴¹, could provide deeper insights into the relationship between thermal dynamics and mechanical response in organic micro-actuators. Moreover, alternative optical manipulation strategies such as the “trapping" method^40,41,42, which demonstrates that expanded beam profiles with appropriate thermal gradients can achieve stable trapping and continuous manipulation of micro-objects on solid surfaces, offer a potential complementary manipulation approach for our organic actuators. With the improvement and exploration of new materials and control strategies, these micro-actuators could become even more powerful tools for the assembly and reconfiguration of devices on a chip, paving the way for the development of advanced, adaptive, and intelligent photonic systems.

News