Plasmon-enhanced photoacoustic oscillation for micro laser motor actuation

Laser motors are essential components in micro-opto-electro-mechanical systems for a wide range of applications, and they have the advantages of non-contact actuation, easy miniaturization, and strong resistance to electromagnetic interference. Actuation based on optical radiation,^1,2 photophoretic force,^3,4 and photochemical force^5,6 are well-known light-driven methods. However, the light forces from these methods are too small (∼pN)^2,7,8 to drive motors that are several-hundred micrometers in size, especially when operating in non-liquid environments.^9,10 Photothermal (PT) techniques based on nonradiative conversion of absorbed light energy into heat have shown a promising way for laser actuation by providing greater energy transfer depth and breadth. PT actuation can provide large forces and deformations; for example, Palagi et al. demonstrated a liquid-crystal elastomer robot of 1 mm in length and 200–300 µm in diameter that could move with a crawling deformation of 110 µm under laser irradiation pulsed at 2 Hz.¹¹ Zhang et al. investigated a series of polymeric and metallic materials for a PT actuator with a maximum deformation of 12–20 µm and a dynamic response of 1–20 Hz.^12–15 Ma et al. investigated all-inorganic PT actuators using a phase-change material to deliver a very large deformation of over 800 µm and fast response of up to 100 Hz, despite a restricted working temperature.¹⁶ Other PT actuators/motors using graphene and carbon nanotubes have also been studied.^17–23 However, a problem highlighted by these studies is that the dynamic responses of workable PT actuators/motors are no more than 100 Hz, as presented in Table I; here, we refer to this issue as the frequency response limitation for laser motor actuation (for details, see the supplementary material, Note S1 and Fig. S1).

The PT method with sharp pulsed laser is capable accompanied by photoacoustic (PA) wave generation, which has been widely used in non-destructive testing²⁵ and microscopy,²⁶ and it was recently exploited as a new power source for laser motor research.^9,10 Laser motors actuated by the PA method react much faster in the high-frequency domain because they utilize acoustic energy rather than heat²⁷ and do not suffer from the frequency response limitation. However, in most cases, the accompanying PA response is extremely weak in a laser PT process, and thus, if a strong PA deformation is desirable such as for laser motor applications, much higher energy laser sources must be used.^10,28,29 The high-power laser could damage the laser motor and limit its use in most areas such as biological applications.

To overcome these limitations, we theoretically and experimentally investigated the thermal-, electric-, and acoustic-near field interactions in the PT/PA process and the significant improvement in PA responses, including the PA waves and their accompanied PA oscillations (for details about the PA oscillation, see the supplementary material, Note S2 and Fig. S2), by strong localized surface plasmon resonance (LSPR). A function layer was created by depositing Au nanoparticles (AuNPs) on the laser motor, and the AuNPs layer acts as a light absorption enhancer and could potentially address the issues of energy deficiency for strong PA wave generation. The enhancement may be due to two factors: first the AuNPs exhibit a strong LSPR effect, and this allows them to effectively absorb an incident laser.³⁰ Second, because of the close proximity of these nanoabsorbers when in clusters, their laser-induced electric- and acoustic-near field response may overlap spatially and temporally, causing further enhancement of PA effects. The energy carried by PA waves could induce a fast and large oscillation upon the motor beyond the frequency response limitation, which could reach tens of kilohertz. The broadband response of the AuNP-PA laser motor could benefit bionic research on insect flapping, vocal-cord vibration, and muscular movement. Importantly, no electrical or chemical energy is required because the driving force is derived from acoustic dissipation excited by the light source. This device could be used in a wide range of applications in fields including micro/nano physics, biochemistry, and clinical medicine.

The experimental setup is shown in Figs. 1(a) and 1(b), and the setup (for details, see the supplementary material, Fig. S3) was constructed on the technical platform of a laser-Doppler vibrometer microscope (LV-S01-M, Sunny Optical Co., Ltd., China), onto which the sample and the detectors were incorporated. A solid-state pulsed laser (IDOL-S, Xiton Photonics GmbH, Germany) with a wavelength of 480–600 nm, a pulse duration of 50 ns, an average laser power of 0.2 W, and a repetition frequency of 1 Hz to 20 kHz was used to irradiate the sample, and the sample is fixed by two microscope slides. A 590 nm narrow-band pass filter was used to selectively allow only the LSPR wavelength to pass. Micrographs of the laser motor in the y–z plane are shown in Fig. 1(c), and its thickness (the x-axis) was 120 µm. The laser irradiated along the x-direction, as shown in the left-half of the bottom picture of Fig. 1(c), and the laser spot was focused to a diameter of 100 µm. An electromagnetic acoustic transducer (EMAT) cooperated with the laser-Doppler vibrometer microscope was used to detect the PT deformation, PA oscillation, and PA waves. The three signals were identified based on their features in response frequencies, curve shapes, and deformation directions, as summarized in Table II, and see the supplementary material, Note S2 and Fig. S2. The EMAT sensor was made according to Ref. 31, and its magnetic-field direction was adjusted to correct the electromagnetic induction during detection. Unless otherwise stated, both the PT and PA experiments used the same laser conditions, that is, the laser was swept at different repetition frequencies with an identical average laser power, laser wavelength, pulse duration, and illumination position. An objective lens with a charge-coupled device (CCD) camera (MC-D500U, Phoenix Optical Co., Ltd., China) was used to visualize and capture the experimental results.

A plasmonic–acoustic-based PA actuation method was introduced to improve the performance of the laser motor in the high-frequency domain, as shown in Figs. 3(a) and 3(b), the PA mechanism converts the energy from optical irradiation into a localized temperature increase, which generates a mechanical acoustic wave through the thermal elastic effect.^34,35 Material is a crucial factor in the PA principle, where a higher optical energy absorption capability can excite stronger PA waves and larger PA oscillations. For this research, AuNPs with 100 nm diameter were chosen as the PA-generation material due to their high optical energy absorption capabilities at plasmon-resonant frequencies. An absorption band results when the incident light frequency is resonant with the collective oscillation of the conduction-band electrons of the AuNPs, that is, when LSPR occurs.^36–38 To facilitate explaining the mechanism that LSPR enhances PA excitation, we present near-field distributions that include electric displacement vectors filling electric fields [Fig. 3(c-i)] and thermal energy field distribution [Fig. 3(c-ii)] simulated using COMSOL software. Surface charges oscillate in the electric field, which is induced by reverse polarity surface charges on surfaces of the nanosphere and the electric field of light (linearly polarized along the z-direction). A large amount of thermal loss caused by electron oscillation is deposited at the interface between the Au nanosphere and nickel substrate, which forms a hotspot. The PA generation is quite different from the slow thermal expansion and contraction process of the PT mechanism,^39,40 and it follows the principle of laser based thermal-elastic mechanism. When the hotspot is formed on the surface of the laser motor, the energy is transferred to acoustic deformation through the thermal elastic process within a very short time (∼ns), Fig. 3(d) shows the variation of the acoustic field after the laser irradiation in 25 ns [Fig. 3(d-i)] and 50 ns [Fig. 3(d-ii)] respectively, and the high energy hotspot rapidly heats a localized volume, in which an impulsive thermal expansion induces lattice oscillations and launches acoustic pulses (i.e., PA waves) into the overlying sample.

On the other hand, the wavelength of the PA wave was calculated to be ∼249 µm based on the equation b = (a² + v²τ²)^1/2,⁴¹ where a is the radius of the laser intensity distribution (50 µm), v is the speed of the acoustic wave (∼4900 m/s, see the supplementary material, Note. S3 and Fig. S4), τ is the laser pulse duration (50 ns), and b is the characteristic wavelength of the PA wave. A relationship is drawn that the wavelength of the PA wave is much larger than the overlying sample thickness (120 µm); for the ultrathin plate structure, the radiation force of such PA wave could cause the metal plate-arm of the laser motor to oscillate (for details, see the supplementary material, Note S4 and Fig. S5). The PA generation does not lose efficiency just by changing the laser repetition frequency; the laser motor reacts much faster under the PA mechanism and can surpass the frequency response limitation in the PT response. By comparison, although the same nanosecond laser is used, the laser motor response under the PT mechanism is invalid with a high repetition frequency laser input (e.g., >100 Hz). The different response for PT and PA signals at high frequencies is mainly because global heat conduction cannot transfer as fast as localized thermal interactions.

Significantly, the enhancement might not only occur in the particle–substrate junctions but also in the particle–particle gap. The distribution of AuNPs aggregates on the laser motor, as the SEM picture shown in the inset of Fig. 4(a) is partly like a heptamer, thus a simulated model of an AuNPs heptamer situated on the laser motor substrate is systematically analyzed including the absorption spectra, the plasmonic modes, and their contributions to the formations of hot spots by considering the far field, near field, and their relevance. The cross section of absorption compared to scattering and extinction is shown in Fig. 4(a); for the AuNPs heptamer-substrate system, a broad resonant peak of the simulated absorption spectrum centered at 590 nm originates from hybridized LSPR modes generated on the AuNPs and the nickel substrate. The electric field enhancements would be more complicated because of the competition and synergistic interactions of LSPR of AuNPs, near-field coupling effect of particle–particle and particle–substrate, and the propagating surface plasmons (PSPs) on the substrate.⁴² First, the nonradiative PSPs on the substrate make a significant contribution to near field enhancement rather than the scattering toward the far field.⁴³ More importantly, the dipoles and higher order multipoles that existed between AuNP aggregates critically influence the near field enhancement. The electric field enhancement (defined as lg |E/E₀|) is over 13 [Fig. 4(b-i)] in the particle–particle gaps at 590 nm exciting wavelength, compared to about 8 at the particle–substrate junction. These results can be concluded as the dipole coupling mode between AuNPs heptamer, which belongs to bounded surface charge density wave originating from the collective oscillation of free electrons induced by the incident light,⁴⁴ that is, there are more surface charges accumulate at interparticle junctions compared to particle–substrate junctions. These enhanced-field distributions corresponding to hybridized LSPR modes show strong energy hot spots at the particle–particle gaps and particle–substrate junctions [Fig. 4(b-ii)], enjoying over 1 × 10⁵ of maximum energy enhancement compared to bare nickel substrate excitation.

The energy enhancement of the AuNPs function layer is confirmed by the PA wave signals detected by the EMAT, and the laser-Doppler vibrometer microscope was also used for displacement amplitude calibration. The 100 nm AuNPs were prepared by a seed-mediated growth method, for details see Sec. VI. Then, the AuNPs were anchored on the y–z surface of the laser motor through a continuous layer deposition process, for details see Sec. VI. In order to better analyze the PA wave signals, a one-shot laser pulse (pulse energy = 100 µJ) was employed. It can be seen from Fig. 4(c) that both the amplitude and relaxation time of the wave excited on the AuNP-PA laser motor were much larger—∼60 nm and 1 µs for the AuNP-PA laser motor [Fig. 4(c-i)] compared to 20 nm and 0.2 µs for the undecorated motor [Fig. 4(c-ii)], respectively. This meant that the acoustic field of the AuNP-PA laser motor contained more energy and a greater proportion of the incident light was converted into acoustic energy. The pulse width τ_s of the PA wave was ∼50.8 ns based on equation τ_s = b/v,³⁹ which was consistent with the laser pulse width and definitively proved that the PA wave was induced by the laser.

The frequency response and amplitude of PT deformation and PA oscillation of the laser motor were detected and compared in the experiment, as shown in Figure 5(a); in order to eliminate the influence of the system vibration noise, all the detection processes were conducted on an optical platform, and the signals were filtered. The PT deformation decreased exponentially with increasing laser repetition frequency, from 17.4 to 0.7 µm at frequencies of 0–100 Hz, respectively. There was no detectable PT deformation at higher frequencies. By comparison, the laser motor under AuNP-PA excitation had strong oscillations of greater than 5 µm across a broad range of frequencies from 1 Hz to 20 kHz. A mechanical resonance with a maximum output of ∼18.8 µm occurred near 4 kHz when the laser repetition frequency matched the mechanical resonant frequency of the laser motor. The real-time response of the PT deformation and PA oscillation was also compared and analyzed. It can be seen from the x–y views of the laser motor under the CCD camera that, when the laser irradiated the laser motor at a repetition frequency of 1 Hz, it swelled [Fig. 5(b-i), multimedia view] because the PT deformation of the laser motor beam along the z axis caused the viewing window to become out-of-focus during optical monitoring. While the repetition rate of the laser increased to a higher frequency of 10 kHz, there was no swelling could be observed any longer. In contrast, a rapid and directional forced PA oscillation was visualized at 10 kHz along the y axis [Fig. 5(b-ii), multimedia view]. The dynamic response of the laser motor under the PA and PT mechanisms was quite different even under identical irradiation conditions. In general cases, the PT deformation and PA oscillation would overlap with each other, but their strengths may vary significantly, and they would rarely be noticed at the same time because the PA oscillation is always too weak, as shown by the curve with blue triangles in Fig. 5(a). The undecorated laser motor cannot generate a detectable PA oscillation in most frequency bands, such as 0–1000 Hz and >10 kHz. With proper NP decoration, the AuNP-PA laser motor could produce an output with an amplitude that is comparable to that of the PT deformation but with a much broader frequency response. The reproducibility was also evaluated in the experiments, a batch of three AuNP-PA laser motors was prepared following the same fabrication route, the SEM images of the assembled AuNPs on the laser motor are shown in Fig. 5(c), and the result confirms that similar distribution of AuNPs with only a few relative position variances is achieved on the laser motor. The reproducibility is characterized by the amplitude deviation of PA oscillations, as the error bars shown in Fig. 5(a); by dividing the standard deviation by the mean value, the variation coefficients are calculated to be less than 8%, confirming a high reproducibility of the AuNP-PA laser motor. The fast reaction and large amplitude in the high frequency domain with good stability and high reproducibility demonstrate the effectiveness of this AuNP-PA laser motor in solving the frequency response limitation and being promising for finding applications in micro/nano physics, biochemistry, clinical medicine, and so on.

Second, the optimization for NP should be considered; in general, the PA signal is proportional to laser energy, the absorption cross section, and the coefficient of thermal expansion;²⁷ metal nanoparticles, such as AuNPs and AgNPs, are commonly used for PA excitation due to their strong light absorption ability and a high coefficient of thermal expansion. In addition, the size of NP is also a crucial factor because it can influence light absorption by the LSPR effect. The absorption spectrum with different diameter AuNPs shows that the absorption peak shifts to longer wavelength with bigger size NPs, e.g., the absorption peak is located at a wavelength of about 530 nm for 20 nm AuNPs and a wavelength of about 590 nm for 100 nm AuNPs, see the supplementary material, Fig. S6. The maximum absorptivity is roughly the same for AuNPs smaller than 100 nm; thus, the only considered optimum condition for these size NPs is whether the LSPR wavelength matches the laser wavelength. However, in applications, the smaller the NPs the shorter the laser pulse duration required based on the criteria τ ≤ τ_T; it will be challenging for technological as well as safety constraints. The NPs larger than 100 nm are not quite suited to the PA method in laser motor application because the absorptivity at LSPR wavelength is relatively much lower compared to NPs smaller than 100 nm, which results in low efficiency of photoacoustic conversion; nevertheless, it may be an option for high laser energy pumped PA motor because smaller NPs are more likely to be melted by high energy laser.⁴⁸ 

The geometry of the AuNP-PA laser motor is also an important factor that needs to be considered; in this study, the laser motor is specially designed with a short-thin arm and a long-wide arm, of which the short-thin arm acts as a PA wave launcher while the long-wide arm functions as an elastically suspended proof-mass to produce an inertial force. The displacement of the PA oscillation is roughly proportional to the motor length based on the geometric relationship, see the supplementary material, Note S4 and Fig. S5, so the long-wide arm can provide larger PA oscillation. On the other hand, benefiting from the ultra-high frequency of the PA oscillation, a huge acceleration can be obtained by the AuNP-PA laser motor, and the acceleration produced by the AuNP-PA laser motor can be up to 3155–94 652 m/s² depending on the laser driving frequency and oscillation modes (for details, see the supplementary material, Note S5, Note S6, Figs. S7 and S8). Based on such advanced features of the AuNPs laser motor, we look forward to it being designed as a kind of nano biomimetic robot as shown in Fig. 6. The robot can obtain a big push force by the high frequency flapping of the mechanical wings and move with a fast speed; it can be used to remove blood clots and so on. We believe that there will be good application prospects for the AuNP-PA laser motor in the future.

Laser induced PT and accompanying PA signals, including PT deformation, PA oscillation, and PA wave, are detected separately. I, detection method for PT deformation. Align the detecting light of the laser Doppler vibrometer along the z direction, when the PT deformation moves along the z axis based on the experimental system, and it can be detected and identified separately because the PA oscillation (move along the y direction) and PA wave (move along the x direction) do not have any motion relative to the laser Doppler vibrometer; thus, they cannot be detected. In addition, the PT deformation can also be identified by aligning the magnetic field direction of the EMAT along the y direction; in this arrangement, only the PT deformation can cut the magnetic induction line and produce signal currents. II, detection method for the PA oscillation. Align the direction of the detecting light of the laser Doppler vibrometer along y and the magnetic field direction of the EMAT along x, and the PA oscillation was identified. III, detection method for the PA wave. Align the direction of the detecting light of the laser Doppler vibrometer along x and the magnetic field direction of the EMAT along z, and the PA waves were identified. All the PT and PA signals were further evaluated using the Fast Fourier transform method according to their frequency response features.

I, preparation for the nickel laser motor. The laser motor was fabricated using LIGA technology, and the material is pure nickel (99.5%); for the procedures of LIGA technology, see the supplementary material, Fig. S9. II, preparation for the AuNPs. The 100 nm Au nanoparticles (AuNPs) were synthesized using a seed-mediated growth method.⁴⁹ In the first step, add 6 ml of 1 wt. % sodium citrate (CA) solution into 200 ml of 0.01 wt. % boiling HAuCl₄ solution and then 16 nm Au seeds were synthesized. In the second step, add 3 ml of 16 nm Au seeds, 0.6 ml of 1 wt. % ascorbic acid (AA), and 0.2 ml of 1 wt. % CA into 20 ml deionized water and then drop 1.1 ml of 0.01 wt. % HAuCl₄ solution into it and then the 50 nm AuNPs were synthesized. The third step is to repeat the above procedures with 2 ml of 50 nm Au seeds, 0.4 ml of AA, 0.1 ml of CA, and 0.654 ml of HAuCl₄ solution to synthesize the 100 nm of AuNPs. III, preparation for the AuNPs laser motor. The AuNPs function layer was prepared and anchored onto the laser motor according to the literature,⁵⁰ and 1.5 ml of cyclohexane was dropped into 6 ml of AuNPs colloids to form an immiscible water/cyclohexane interface to prepare monolayered AuNPs. Then, ethanol was added drop by drop until a specular reflection occurred and a shiny gold color appeared. The nickel laser motor was then immersed in the AuNPs monolayer film at a small angle and slowly pulled out; after the cyclohexane was naturally evaporated, the AuNPs monolayer film was successfully transferred to the surface of the laser motor, and the whole process was demonstrated in the supplementary material, Fig. S10.

Laser motors based on the PT actuation method have the advantage of providing large forces and displacements. However, their dynamic responses are severely restricted by the frequency response limitation. Studies on the frequency response limitation in a PT actuation process show that under high-repetition-frequency laser irradiation, the temperature difference of the laser motor becomes smaller and smaller, and the thermal deformation tends to be zero, which causes a failure of PT actuation. By comparison, the AuNP-PA actuation method performed better in the high-frequency domain because the PA oscillations utilized acoustic energy rather than heat. The results showed that the AuNP-PA laser motor could generate PA oscillations of over 5 µm by pulse laser with repetition frequencies of 1 Hz to 20 kHz, exhibiting a maximum value of 18.8 µm at the mechanical resonant frequency. The variation coefficients of different AuNP-PA laser motors following the same fabrication route were calculated to be less than 8%, confirming the high reproducibility of this kind of PA laser motor. The AuNP-PA laser motor effectively offset the shortcomings of PT actuation in the high-frequency domain, yielding comparable output amplitudes. The fast reaction and large amplitude in the high frequency domain with good stability and high reproducibility demonstrate the effectiveness of this method in solving the frequency response limitation. Such an AuNP-PA laser motor requires no electrical or chemical energy and could benefit bionic research in areas including insect flapping, vocal-cord vibration, and muscular movement. These devices could be used in a wide range of applications in fields including micro/nano physics, biochemistry, and clinical medicine.

See the supplementary material for the complete characteristics and performances of the studied AuNP-PA laser motor and the relative experimental instruments.

Financial support from the National Natural Science Foundation of China (Grant Nos. 52105595, 51972292, and 52375550) and the Natural Science Foundation of Zhejiang Province, China (Grant No. LQ22A040002) are gratefully acknowledged.

The authors have no conflicts to disclose.

Fanghao Li: Conceptualization (equal); Data curation (equal); Formal analysis (equal); Funding acquisition (equal); Investigation (equal); Methodology (equal); Visualization (equal); Writing – original draft (equal). Mengru Zhang: Data curation (equal); Software (equal); Validation (equal). Cuixiang Pei: Data curation (equal); Resources (equal); Validation (equal). Xinyao Yu: Data curation (equal); Software (equal). Li Jiang: Funding acquisition (equal); Resources (equal). Yadong Zhou: Resources (equal); Software (equal); Supervision (equal). Fanli Zhang: Methodology (equal); Validation (equal). Yunfeng Song: Project administration (equal). Jian Chen: Conceptualization (equal); Funding acquisition (equal); Project administration (equal); Writing – review & editing (equal).

The data that support the findings of this study are available from the corresponding authors upon reasonable request.