Skip to main content

Thank you for visiting You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

Ballistic supercavitating nanoparticles driven by single Gaussian beam optical pushing and pulling forces


Directed high-speed motion of nanoscale objects in fluids can have a wide range of applications like molecular machinery, nano robotics, and material assembly. Here, we report ballistic plasmonic Au nanoparticle (NP) swimmers with unprecedented speeds (~336,000 μm s−1) realized by not only optical pushing but also pulling forces from a single Gaussian laser beam. Both the optical pulling and high speeds are made possible by a unique NP-laser interaction. The Au NP excited by the laser at the surface plasmon resonance peak can generate a nanoscale bubble, which can encapsulate the NP (i.e., supercavitation) to create a virtually frictionless environment for it to move, like the Leidenfrost effect. Certain NP-in-bubble configurations can lead to the optical pulling of NP against the photon stream. The demonstrated ultra-fast, light-driven NP movement may benefit a wide range of nano- and bio-applications and provide new insights to the field of optical pulling force.


Speed is important to the functionality of nanoswimmers/microswimmers in fluids because it determines the efficiency of swimmers1,2,3,4. Directed movement is also important as the swimmers are usually tasked to reach a target. However, directed movement and high speed in the nanoscale are rarely compatible5,6,7. In laminar fluid flow, enabling fast-moving swimmers requires large propulsion force, as it needs to counter the drag force which is proportional to the speed8. Usually, nanoswimmers/microswimmers driven by on-board propulsion (e.g., bubble repellence5,6,7) exhibit higher speeds compared to other mechanisms (e.g., mechanical motion9,10, pressure11 and thermal12,13, gradients, chemical phoresies14,15,16, and optical force17,18). The highest speed of bubble-repelled swimmers is ~105 body-length s−1 (see ref. 7), where the unit (the speed to body-length ratio) is commonly used to allow a fair comparison between swimmers with different dimensions. However, it is very difficult to control their moving directions because of the random nature of the on-board repellence force.

Optical forces on an object due to the exchange of photon momentum with the object17,18,19,20,21,22,23,24,25,26,27 can move nanoswimmers/microswimmers along the beam direction. Light propulsion has unique advantages like wireless control, high spatial and temporal precision, and instant response28. Usually, light applies a pushing force on an object in the light propagating direction, which has worked together with the optical gradient force to enable the optical tweezer effect17,19,20. Objects with high scattering efficiency like plasmonic NPs can enable stronger optical pushing forces19,20, but the achieved highest swimmer speed is merely ~103 body-length s−1 (see ref. 17), since the optical forces are very weak (10−12 to 10−14 N).

Light may also pull an object in specific optical conditions18,21,22,23,24,25,26,27. For example, two planewaves irradiating mirror-symmetrically with a large incidence angle towards an object can lead to a strongly focused forward photon scattering momentum, resulting in a net “negative” optical force22,23,24. For a spheroidal object much larger than the wavelength of the incident light and placed at a heterogenous dielectric interface, the object can be pulled by the light due to an increase in the photon linear momentum across the dielectric interface18. However, the achieved moving speed is only ~0.5 body-length s−1, and the same strategy may not be applicable to objects in homogenous media, which are the cases for most nanoswimmers/microswimmers. Theoretically, an object in homogenous medium can also be pulled by a single planewave if the object has certain unique optical configurations to enable either optical gain25,26 or near-field electromagnetic coupling between dielectric-metallic NP dimers27. However, the experimental demonstration in homogenous media is still lacking.

In this work, we observe optical pushing and pulling of Au NP swimmers exhibiting extremely high speeds (>106 body-length s−1) in water, and we show that the unusual optical pulling and high-speed movement are only made possible by the nano-bubbles generated around the laser-excited plasmonic NPs, which lead to supercavitation.


Ballistic motion of supercavitating nanoparticle

The experimental setup is shown in Fig. 1a. We disperse Au core–shell NPs made of a silica core (~100 nm) and an Au shell (~10 nm) in deionized water. A femtosecond pulsed laser (linear-polarized Gaussian beam) with a repetition rate of 80.7 MHz and a wavelength of 800 nm is focused in the NP-water suspension using a 20× objective lens (numerical aperture ~0.42). At the focal plane, the minimum waist of the Gaussian beam is ~6 μm. The laser wavelength matches the surface plasmon resonance (SPR) peak of the Au NPs in water. A high-speed camera captures the scattered light from the Au NPs to track their positions, and they are shown as glowing dots (Fig. 1b, c).

Fig. 1: Observation of extremely fast ballistic movements of Au NPs.

a Schematic of the experimental system to monitor the dynamics of Au NPs dispersed in water. The red solid lines after the 20× lens indicate the Gaussian beam envelope. Au NPs scatter the incident light (red dots), which is captured by a high-speed camera with the field of view depicted by the blue solid lines. The blue and magenta arrows depict the Au NPs moving in the +z and −z directions, respectively referred to as the “positive motion” and the “negative motion”. b, c Dark-field optical images (scale bars –50 μm) of ultra-fast ballistic Au NPs with b the positive motion and c the negative motion. The cyan triangle in b and the magenta triangle in c indicate the positions of the selected Au NPs at t = 0. The cyan dotted line in b and the magenta dotted line in c are guide for the eye. The power of the laser is 690 mW, corresponding to 12 mW μm−2 at the focal plane. d, e Representative d positive and e negative motion trajectories of the ballistic Au NPs projected onto the x − z plane. Each color corresponds to a ballistic Au NPs. The time interval between two adjacent symbols is 400 μs. The black arrows at the top of the figures indicate the focal plane location. In adkz is the wavevector of the incident laser light. f, g Images from the pump-probe optical scattering imaging experiment for f without pump beam and g with pump beam. Each green spot corresponds to the diffraction-limited scattered probe light from the Au NPs with nanobubbles.

By observing the glowing dots (Supplementary Video 1), we find some fast NPs moving along the beam propagation direction (“positive motion”) and some in the opposite direction (“negative motion”). NPs moving in both directions can have very high speeds, in stark contrast to the rest NPs experiencing Brownian motion. We single out several representative cases of ballistic NPs moving in either direction (see examples in Supplementary Video 2 for the positive motion and Supplementary Video 3 for the negative motion), track their positions as a function of time, and calculate their instantaneous speeds. We find that the maximum speed for the positive motion is 336,000 μm s−1 and that for the negative motion is 245,000 μm s−1 (see Fig. 1b, c). We also find that the average speed of the observed representative positive motions (204,000 μm s−1) is higher than that of the negative motion (109,000 μm s−1).

When the Gaussian beam illuminates the NP, there are mainly three types of possible forces, namely the optical force (radiative pressure)17,19,20, the optical gradient force17,19,20, and the photothermal gradient force12,13,29. Gravity and buoyancy are not considered because they are perpendicular to the optical axis. To identify the dominant force driving the ballistic NPs, we analyze several representative ballistic trajectories around the focal plane as shown in Fig. 1d, e. It is found that the NPs can cross the focal plane (black arrow in Fig. 1d, e) regardless of their moving directions. In addition, we can see that to the left (or right) side of the focal plane, ballistic NPs can move in both directions. These indicate that the ballistic NPs are not driven by any gradient (optical or photothermal) forces, since such forces are symmetric about the focal plane, which should converge the NPs from both sides of the focal plane towards it. We have also calculated the optical force field on the NP under the Gaussian beam used in the experiment (see Supplementary Note 1). We find that there is only positive force on the NP and the amplitude is almost symmetric to the focal plane. If gradient optical force is significant, it should have broken such a symmetry since it changes sign across the focal plane. We can thus infer that the gradient optical force is negligible compared to the dissipative optical forces. At lower NP concentrations, we also find ballistic NPs with both negative (Supplementary Video 4) and positive motions (Supplementary Video 5), suggesting that the ballistic movements should not be the result of the environment (e.g., scattered light from surrounding NPs or thermalization of water due to NP heating). Additionally, when we deposited the ballistic NPs on a quartz substrate and then characterized their geometrical configurations using a scanning electron microscope, it is found that the deposited NPs are isolated single NPs (see Supplementary Note 2). This also confirms that the ballistic movements are not related to the aggregation of NPs. These analyses leave the dissipative optical force (radiative pressure) as the only possible driver of the ballistic NPs.

To fully explain the observed ballistic NP movements, we need to understand how the NPs can move in speeds much higher than that of a typical optical force can sustain according to Stokes’ law and why there are both optical pushing and pulling phenomena on some specific NPs. As discussed later, the optical force from our laser cannot sustain the observed high speed unless the effective viscosity experienced by the NPs is 100× lower than that of liquid water. The optical pulling is also impossible with our Gaussian beam irradiating an Au NP in a homogenous medium, unless there is a nearby dielectric structure optically coupled to the NP like the ones proposed in refs. 21,27.

According to several studies30,31,32,33, when an Au NP is irradiated by a laser at the SPR peak, it can instantaneously generate a nano-bubble around it. Our heat transfer analysis also estimates that the plasmonic heating can rise the temperature of an NP inside a nanobubble to ~850 K (see Supplementary Note 3), which is much higher than the critical temperature of water (647 K). This calculated temperature from our simple model is also in reasonable agreement with that inferred from experiments (~1000 K)34. The hot NP thus can instantly evaporate water molecules, like the Leidenfrost effect, and can be fully encapsulated by vapor30,31,32,33. As long as the NP is illuminated by the laser, the NP can keep the high temperature. Using a pump-probe optical scattering imaging setup, we confirmed the existence of nanobubbles around the plasmonic Au NPs in our experiments when excited by the pump laser (Fig. 1f, g). This pump-probe method detects the intensity change of the scattered probe light (533 nm) at a certain solid angle due to the formation of nanobubble on Au NP under the illumination of the pump beam at the SPR peak (800 nm) (see Supplementary Note 4 for details). We note that the lifetimes of the plasmonic nanobubbles are reported to be ~200 ns (see refs. 30,31,32,33). In our NP movement experiments, the time interval between laser pulses is ~12 ns, allowing the nanobubbles to shrink due to cooling by the surrounding water30,31,32,33,35,36, but they may not completely collapse since the pulse period (~12 ns) is shorter than the bubble lifetime (~200 ns). Dissolved gas in water can also contribute to the nanobubble volume, making it more stable than a pure vapor bubble37.

Optical forces on nanoparticle in nanobubble

To understand how the nanobubbles may influence the optical force direction, we calculate the optical force in the beam propagation direction (Fz) on an NP with a nanobubble. We consider a bubble with a radius rnb < 120 nm attached to the surface (θ, φ) of an Au core–shell NP (θ and φ are polar and azimuthal angles, Fig. 2a). When the bubble keeps growing, it will eventually encapsulate the NP, and we assume such a case for rnb > 120 nm (i.e., inset in Fig. 2b). We have to note that the exact bubble radius for the attach-to-encapsulate transition will not influence the physics inferred from our analysis. Our assumption of the nanobubble nucleation and the encapsulation of NP is based on experimental studies, where Fu et al.30 showed that upon laser excitation, a nanobubble can nucleate at a certain location on the Au NP surface and Lachaine et al.32 showed that the nanobubble can fully encapsulate the NP if the fluence of the femtosecond laser pulse is higher than a threshold. For the NP used in this study, the threshold is around ~7 mJ cm−2 (see ref. 32), and our laser fluences are of 9–15 mJ cm−2. The encapsulation of NP by the nanobubble is also intuitive given the very high temperature of the NP (>850 K), which is above the water critical temperature and can instantaneously evaporate liquid in contact with the hot NP surface. Since the beam waist (~6 μm) at the focal plane is much larger than the NP size (~120 nm), and the ballistic movements start at locations away from the focal plane, a linearly polarized planewave as an incident light is a good approximation38 for estimating Fz. In addition, since our laser pulses have an ultra-short duration (~94 fs), the time-averaged optical force by the pulsed laser can be approximated39 by that from a continuous laser with the same central frequency and power density (see Supplementary Note 5). We calculate electromagnetic field distributions at various rnb, θ, and φ using the finite element method (FEM) and then calculate Fz using the Maxwell stress tensor20,27. More details are provided in the “Method” section.

Fig. 2: Optical force on an Au NP with a nanobubble.

a Schematic of an Au NP with a nanobubble. The Au NP (yellow sphere) consists of a 100 nm-diameter silica core and a 10 nm-thick Au shell. The nanobubble with a radius of rnb is attached to the surface (θ, φ) of the Au NP, where θ and φ are, respectively, the polar and the azimuthal angles in the polar coordinate with the origin (o) at the center of the Au NP. Ex, Hy, and kz depict the electric field, the magnetic field, and the wavevector of the incident planewave, respectively. b The calculated optical force in the z-direction (Fz) as a function of rnb. Here, θ = 0o, φ = 0o and the amplitude of Ex is 2.6 × 106 V m−1, corresponding to the laser in the experiment. The insets illustrate the schematic configurations of Au NP with a nanobubble with different rnb. c The calculated Fz as a function of θ and φ. Here, rnb = 130 nm and the amplitude of Ex is 2.6 × 106 V m−1. On top of the contour, schematic configurations of Au NP with a nanobubble as a function of θ are illustrated. df The directions and normalized magnitudes of Poynting vectors of scattered fields Ssca for the structures of d rnb = 130 nm and θ = 180o, e without a nanobubble, and f rnb = 130 nm and θ = 0o. Directions of Fz are also shown.

We find certain geometrical windows that induce either positive or negative Fz (Fig. 2b, c). For rnb > 90 nm (Fig. 2b) and θ < 75o (Fig. 2c), negative Fz can be achieved, where the nanobubble locates at the back side of the NP with respect to the incident light direction (kz). Otherwise, Fz is positive. Fz is found insensitive to the azimuthal angle φ (Fig. 2c). While the positive Fz is intuitive, we need to understand the negative Fz on the NP under illumination of the single planewave. The optical force on the NP in the nanobubble can be decomposed into two parts40: one from the interaction between scattered fields themselves \(({\mathbf{F}}_{\mathrm{z}}^{{\mathrm{ss}}})\), and the other from the interaction between the scattered field and the incident field \(({\mathbf{F}}_{\mathrm{z}}^{{\mathrm{si}}})\), and \({\mathbf{F}}_{\mathrm{z}} = {\mathbf{F}}_{\mathrm{z}}^{{\mathrm{ss}}} + {\mathbf{F}}_{\mathrm{z}}^{{\mathrm{si}}}\). The sign and magnitude of each decomposed part can unveil which light field contributes dominantly to determining Fz. For a representative case of negative motion (rnb = 130 nm and θ = 0o), it is found that \({\mathbf{F}}_{\mathrm{z}}^{{\mathrm{ss}}}\) is responsible for the optical pulling force, where \({\mathbf{F}}_{\mathrm{z}}^{{\mathrm{ss}}}\) is −1.254 × 10−12 N, but \({\mathbf{F}}_{\mathrm{z}}^{{\mathrm{si}}}\) is only 4.99 × 10−13 N. Since \({\mathbf{F}}_{\mathrm{z}}^{{\mathrm{ss}}}\) is due to the net momentum the NP can gain by compensating the net scattered photon momentum leaving the NP, if the scattering is isotropic, \({\mathbf{F}}_{\mathrm{z}}^{{\mathrm{ss}}}\) should be zero. Thus, the negative \({\mathbf{F}}_{\mathrm{z}}^{{\mathrm{ss}}}\) implies that there is uneven radiative scattering from the NP. We have investigated the energy flow pattern of scattered photons with the Poynting vector and found that at the backside of the NP, there is a “saddle” singular point, which induces unusual energy flows around the NP (see the magenta triangle in Fig. 2d). The saddle point turns the energy flow scattered from the NP to the back side of the NP and leads the energy to flow around the “vortex” singular points located at the sides of the NP (see the cyan circles in Fig. 2d). It is this circulation of the scattered energy that allows the NP to possess negative momentum. Our result corresponds well to an analysis introduced in ref. 41, where a particle irradiated by a Bessel beam can receive an optical pulling force when the pairs of saddle and vortex singular points, which are located at the backside of the NP, redirect and focus the energy flow into the particle. Without a nanobubble, the saddle point and the circulation of energy flow do not exist (see Fig. 2e), and the scattering is symmetric about the light polarization axis, leading to \({\mathbf{F}}_{\mathrm{z}}^{{\mathrm{ss}}}\) = 0. This indicates that the nanobubble is essential for forming the saddle singularity. We believe that the formation of a saddle point is due to the curvature of the nanobubble surface, as it works as an optical mirror to reflect the scattered light from the NP. We have also investigated the energy flow of scattered light for the representative case of positive motion (rnb = 130 nm and θ = 180o). As the NP is now placed close to the right nanobubble surface, the saddle point is sandwiched between the NP and the nanobubble (see Fig. 2f). While there are still two vortex points which can circulate the energy flow to be redirected into the NP, the strength of circulation is much suppressed in comparison to the negative motion case. As a result, the magnitude of optical pulling force that the NP receives is much weaker than the pushing force, leading to a positive \({\mathbf{F}}_{\mathrm{z}}^{{\mathrm{ss}}}\) of 1.068 × 10−12 N. We have also investigated the optical stress tensor profile on the surface of the Au NP for the representative cases (see Supplementary Note 6). Both cases show negative stress at the light-incoming side and positive stress at the rear side, and the sign of stress reverses as across the equator of the NP. For the negative motion case, it is observed that the negative stress around the side poles not only has a higher magnitude but also covers a larger area than those of the positive stress, thus leading to a net negative \({\mathbf{F}}_{\mathrm{z}}^{{\mathrm{ss}}}\). We believe that the stronger negative stress is due to the strong energy circulation around the vortex singular point at the side poles of NP (see Fig. 2d). Meanwhile, for the positive motion case, the area and the magnitude of negative stress around the side poles are suppressed compared to that of the negative motion case, which yields a net positive \({\mathbf{F}}_{\mathrm{z}}^{{\mathrm{ss}}}\). It is due to the weakened energy circulation as the saddle point is squeezed between the surfaces of the NP and the bubble (see Fig. 2f). We have also calculated the far-field scattering patterns (Supplementary Note 7). It noted that the far-field scattering patterns do not intuitively correspond to the directions of optical forces on the NPs, which means that the optical force on the NP inside the nanobubble is the result of near-field scattering.

Low dynamic viscosity experienced by ballistic nanoparticles

Figure 3a shows the calculated Fz on a supercavitating NP along the central axis of the Gaussian beam for two representative cases: one for the positive motion (rnb = 130 nm and θ = 180°) and one for the negative motion (rnb = 130 nm and θ = 0°). The NP is found to experience Fz on the order of ~10−12 N around the focal plane (Fig. 3a), with the negative Fz uniformly smaller than the positive Fz in amplitude. It is noted that the profiles of the optical forces, which correspond to the light intensity profile, confirm that the optical gradient forces are negligible, which would have changed the sign of the force across the focal plane. Meanwhile, the Stokes’ friction force in a flow with a low Reynolds number (Re), which is Re = ~10−3 <<1 in our case, allows us to estimate the order of magnitude of the driving force (Fdriving) on the ballistic NP given their measured speeds as: Fdriving = 6πηrv ≈ 6π × 8.9 × 10−4 (kg m−1 s−1, η, viscosity of liquid water) × 6 × 10−8 (m, r, radius of Au NP) × 1 × 10−1 (m s−1, v, speed of ballistic NP) = ~10−10 N. This is two orders of magnitude larger than our calculated |Fz| since we used liquid water viscosity. As the plasmonic Au NP is intensely excited by the laser to instantaneously evaporate the surrounding water and generate a nanobubble to encapsulate itself, i.e., supercavitation, the dynamic viscosity experienced by the NP should be different. Using the calculated optical force (~10−12 N) and Stokes’ law, we can extract an effective dynamic viscosity (ηeff) experienced by the ballistic NPs as:

$$\eta _{{\mathrm{eff}}} = \frac{{(z_2 - z_1)}}{{6{\uppi}rv}}\left[ \int_{z1}^{z2} \frac{1}{{{\mathbf{F}}_{\mathrm{z}}(z)}}{\mathrm{d}}z \right]^{ - 1},$$

where v is the experimentally determined average speed between positions \(z_1\) and \(z_2\) where the ballistic movement occurs. The averaged ηeff (for the representative cases in Fig. 1) is ~1.27 × 10−5 kg m−1 s−1 for the negative motion and ~1.03 × 10−5 kg m−1 s−1 for the positive motion. These values are much smaller than the dynamic viscosity of liquid water but very close to that of vapor steam (~1 × 10−5 kg m−1 s−1), suggesting that the ballistic NPs are moving in a gaseous environment. This can be interpreted as the following: as the laser-excited Au NP encapsulated by a bubble moves forward, it keeps evaporating water, maintaining a vapor cushion in front of it and extends the bubble boundary forward (Fig. 3b). This is very similar to the observed near-zero drag force on a hot metal sphere enclosed by a gas cavity due to the Leidenfrost effect42 and in great analogy to the supercavitating torpedoes, which realized 5× faster speed than conventional ones43. As the NP moves, the trailing end of the bubble, which becomes further away from the hot NP, cools and vapor condenses back to liquid (Fig. 3b). This interpretation is well backed by the results of our molecular dynamics (MD) simulations. In Fig. 3c, we clearly see that a hot NP encapsulated by a vapor bubble can continuously evaporate liquid molecules in front of it as it moves forward with a speed of ~13 m s−1, so that the bubble boundary is extended and the NP is always enclosed in a gaseous environment (see Supplementary Notes 8 and 9 for simulation details). While this simulation uses a Lennard-Jones argon model system, the physics should be the same and analogy can be drawn with our NP-in-water experiment. We note that the 13 m s−1 speed simulated is much higher than our experimental observation (0.1–0.3 m s−1), but the NP is still found able to instantaneously extend the bubble front while moving. We also note that the bubble can experience optical forces as well, but they are usually positive with amplitude calculated to be around 6.0 × 10−13 N. However, with such forces, the nanobubble can only move with a speed more than two orders of magnitude slower (e.g., Stokes’s law yields a speed of ~410 μm s−1) than the NPs. Therefore, it should be the instantaneous evaporation that keeps the NP inside a gaseous environment, instead of the synchronization between the dynamical motions of the NP and the bubble. We mention that there is also electrostriction44 forces that can deform the nanobubble, but in our scenario it should not be important as new interfaces of nanobubble is formed by the hot moving NP.

Fig. 3: Positive and negative optical forces and the nanoscale Leidenfrost effect.

a The calculated Fz on an Au NP with a nanobubble along the central axis (z-direction) of the Gaussian beam. A beam waist and intensity at the focal plane (z = 0) are 6 μm and 12 mW μm−2, respectively. The red line represents the positive force for rnb = 130 nm, θ = 180o, and the blue line represents the negative force rnb = 130 nm, θ = 0o. The insets illustrate (left top) the positive motion and (right bottom) the negative motion. b Schematic illustration of a supercavitating ballistic Au NP. Under the laser illumination, the plasmonic Au NP is intensely heated to generate a nanobubble which encapsulates it. When the Au NP moves forward, it keeps evaporating water and maintaining a vapor cushion in front of it, which effectively makes the NP moving in a virtually frictionless environment, like the Leidenfrost effect. In the meantime, vapor at the trailing end of the bubble cools and condenses back to liquid as the hot NP moves forward. c Nanoscale Leidenfrost effect of a hot moving NP simulated by molecular dynamic (MD) simulations. The hot NP is thermostated at 1000 K (Supplementary Note 9 for justification), and the blue contour visualizes the isosurface of the critical density of liquid argon (0.536 g cm−3), which represents the nanobubble surface. The NP moves in the positive x-direction with a speed of 13 m s−1.


Since nanobubble dynamics is stochastic, not all Au NPs are enclosed by nanobubbles to exhibit the ballistic movement. We can see in Supplementary Video 1 that the ballistic motion generally occurs within 300 μm at either side of the focal plane with a significant portion of the NPs in this region exhibiting ballistic motion. This is because the laser intensity is sufficiently high in this region. NPs outside of this region are mainly seen drifting due to thermal convection perpendicular to the laser beam direction. However, even within this region, not every NP is guaranteed to move ballistically since bubble generation is inherently statistical and the probability may depend on factors like defects on the NP surface and geometry of the NP (e.g., defects and sharp corners are known to change bubble nucleation threshold45). If a nanobubble is generated and the NP is encapsulated, moving forward or backward still would depend on the relative position of the NP inside the bubble (as discussed in Fig. 2). In addition, Fig. 2b, c show that the geometrical window for the negative Fz is narrower than that for the positive Fz, which may explain our observation that more ballistic NPs move positively than negatively (Supplementary Video 1). The magnitude of the positive Fz being generally larger than the negative Fz (Fig. 3a) also explains the observed higher average speed of the positive motion than that of the negative motion (Fig. 1). Although Fz depends on the instantaneous geometrical configuration of the NP-bubble structure and thus the forces should be dynamic, we observe that the NPs move almost strictly along the beam axis (Fig. 1d, e). Moving in a virtually frictionless supercavitation, the ballistic NPs exhibit speed-to-body length ratios greater than 106. This is  three to seven orders of magnitude higher than other directed nanoswimmers/microswimmers and one to five orders of magnitude higher than random microswimmers (Fig. 4).

Fig. 4: Comparison of the ballistic Au NP swimmers with other swimmers.

Top-four highest speeds (measured in body-length s−1) for the negative and the positive motion compared to those of the nanoswimmers/microswimmers in the literature. The filled symbols represent nanoswimmers/microswimmers whose movement can be directed. The hollow symbols represent those moving in random directions. The legend indicates propulsion mechanisms. The gray dash lines are contour lines of the magnitude of the force needed to achieve the corresponding speed of a swimmer in liquid water.

In theory, any nanoscale objects that can have supercavitation after laser excitation should exhibit similar ballistic movements. To confirm that this effect is generalizable, we have performed experiments with Au nanorods (NRs), which also have a SPR peak of 800 nm. It is observed that Au NRs also exhibit similar ballistic movements with an observed speed of ~13,000 μm s−1 (Supplementary Note 10 and Supplementary Video 6). Since the SPR characteristic of Au NR is anisotropic due to its shape, only the NRs that align with the incident electric field direction can experience strong SPR coupling and lead to supercavitation. Thus, the ballistic movements of Au NRs may easily be interrupted, because the different forces (e.g., optical force and drag force) can make NRs to misalign with the electric field. However, in Supplementary Note 10, we can clearly see a glowing dot moving for ~13 μm within 1 ms along the beam propagating direction (see also Supplementary Video 6). This ballistic Au NR shows a peak speed of 260,000 body-length s−1, which is still ~102 to 105 times faster than the reported nanoswimmers/microswimmers driven by optical forces17,18. With the calculated optical force (1 × 10−13 N–7 × 10−13 N) on the Au NR (see Supplementary Note 10), Stokes’ law implies a dynamic viscosity of 4.4 × 10−5 kg m−1s−1, which is much lower than that of liquid water and on the same order of magnitude as that of vapor. We have not observed any negative motion for Au nanorods, most likely due to the stringent alignment requirements that limit the probability of enabling supercavitation and negative optical forces. Nevertheless, this result demonstrates that the ballistic movement can be potentially generalized to other kinds of plasmonic NPs with different composition, geometry and dimensions. We believe that with proper NP design, both the ballistic motion and optical force direction can be manipulated and optimized.


Sample preparation

A quartz cuvette (Hellma, Sigma-Aldrich, 10 mm × 10 mm) with four windows is pre-cleaned with acetone, isopropyl alcohol and deionized water in an ultrasonic bath. A silica-Au core–shell NP-water suspension (Auroshell, Nanospectra Biosciences, Inc) with a NP number density of 2 × 1015 m−3 or an Au nanorod NP-water suspension (NanoXact, nanoComposix) with a NP number density of 2.6 × 1017 m−3 is mixed with deionized water to achieve different number densities. The NP number densities are: Fig. 1b–e and Supplementary Videos 13 (for the core–shell NP): 4 × 1014 m−3, Supplementary Video 4 (for the core-shell NP): 5 × 1013 m−3, Supplementary Video 5 (for the core–shell NP): 1 × 1013 m−3, and Supplementary Video 6 (for the Au NR): 1.3 × 1017 m−3. The cleaned cuvettes filled with these NP-water suspensions are used for the experiments.

Ultrafast laser experiment and characterization

Mode-locked femtosecond laser pulses (center wavelength of 800 nm and a full-width half maximum of 10 nm) are emitted from a Ti:sapphire crystal in an optical cavity (Spectra Physics, Tsunami). The laser is collimated and a linearly polarized TEM00 mode beam. The time interval between pulses is 12.4 ns (80.7 MHz), and the pulse duration is ~94 fs. The power of the laser is fixed to 690 mW. The femtosecond laser passes through a 20× objective lens and focuses on the NP-water suspension in the cuvette. A beam profiler (Thorlab BP104-UV) measures the minimum waist of the Gaussian beam (~6 μm). We note ballistic movements were still observed with a 10× objective lens (Supplementary Video 2) or a laser power as high as ~1 W. The scattered lights from the Au NPs passes through another objective lens and focuses on the image sensors of the high-speed camera (HX-7, NAC). The recorded images in the camera are then analyzed using a customized image processing software in MATLAB.

Calculation of optical force

The FEM software (RF module, COMSOL Multiphysics) is used to calculate the optical forces on different Au NPs with nanobubbles (structures illustrated in Fig. 2a). The Au NP consists of a silica core (radius of 50 nm) and an Au shell (thickness of 10 nm). It is modeled that an Au NP with a nanobubble is immersed in an infinite volume of water. In the simulations, for the infinite volume of water, a spherical volume of water (radius of 400 nm) is enclosed by a perfectly matched laser with a thickness of 130 nm. The Au NP with a nanobubble is located at the center of the water sphere. The refractive indices of water, air, and silica at the wavelength of 800 nm are 1.33, 1.00, and 1.50, respectively. The real and complex parts of the refractive index of the Au shell at the wavelength of 800 nm are 0.154 and 4.91, respectively. A linearly polarized planewave is modeled as the incident light for Fig. 2b–f. A Gaussian beam given by the paraxial approximation38 is used for Fig. 3a and extracting ηeff, where it has a minimum waist of 6 μm and the maximum light intensity of 12 mW μm−2, corresponding to the experimental conditions. After solving the electromagnetic field distribution of the simulation domain, optical force (F) is calculated as:

$${\mathbf{F}} = {\oiint} T {\cdot} {\mathbf{n}}\,{\mathrm{da}},$$

where T is the time-averaged Maxwell stress tensor20,27, and n is the normal vector of the surface of the Au NP. We use the time-averaged Maxwell stress tensor to calculate F in the z-direction on the surface of the Au NP.

Data availability

All datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.

Code availability

The code used for data analysis as well as for display of the data is available from the corresponding author upon reasonable request.


  1. 1.

    Strebhardt, K. & Ullrich, A. Poul Ehrlich’s magic bullet concept: 100 years of progress. Nat. Rev. Cancer 8, 473–480 (2008).

    CAS  PubMed  Article  PubMed Central  Google Scholar 

  2. 2.

    Sanchez, M. M. & Schmidt, O. G. Medical microbots need better imaging and control. Nature 545, 406–408 (2017).

    ADS  Article  CAS  Google Scholar 

  3. 3.

    Luo, M., Feng, T., Wang, T. & Guan, J. Micro-/nanobots at work in active drug delivery. Adv. Funct. Mater. 28, 1706100 (2018).

    Article  CAS  Google Scholar 

  4. 4.

    Gao, W., Sattayasmitsathit, S. & Wang, J. Catalytically propelled micro-/nanomotors: how fast can they move? Chem. Rec. 12, 224–231 (2012).

    CAS  PubMed  Article  PubMed Central  Google Scholar 

  5. 5.

    Manjare, M., Yang, B. & Zhao, Y.-P. Bubble driven quasioscillatory translational motion of catalytic micromotors. Phy. Rev. Lett. 109, 128305 (2012).

    ADS  Article  CAS  Google Scholar 

  6. 6.

    Baylis, J. R. et al. Self-propelled particles that transport cargo through flowing blood and halt hemorrhage. Sci. Adv 1, e1500379 (2015).

    ADS  PubMed  PubMed Central  Article  CAS  Google Scholar 

  7. 7.

    Kagan, D. et al. Acoustic droplet vaporization and propulsion of perfluorocarbon-loaded microbullets for targeted tissue penetration and deformation. Angew. Chem. Int. Ed. 51, 7519–7522 (2012).

    CAS  Article  Google Scholar 

  8. 8.

    Purcell, E. M. Life at low Reynolds number. Am. J. Phys. 45, 3–11 (1977).

    ADS  Article  Google Scholar 

  9. 9.

    Sing, C. E., Schmid, L., Schneider, M. F., Franke, T. & Kats, A. A. Controlled surface-induced flows from the motion of self-assembled colloidal walkers. Proc. Natl. Acad. Sci. 107, 535–540 (2010).

    ADS  CAS  PubMed  Article  PubMed Central  Google Scholar 

  10. 10.

    Ghosh, A. & Fischer, P. Controlled propulsion of artificial magnetic nanostructured propellers. Nano Lett. 9, 2243–2245 (2009).

    ADS  CAS  PubMed  Article  PubMed Central  Google Scholar 

  11. 11.

    Wu, Z. et al. Turning erythrocytes into functional micromotors. ACS Nano 8, 12401–12048 (2014).

    Google Scholar 

  12. 12.

    Xuan, M. et al. Near infrared light-powered janus mesoporous silica nanoparticle motors. J. Am. Chem. Soc. 138, 6492–6497 (2016).

    CAS  PubMed  Article  PubMed Central  Google Scholar 

  13. 13.

    Xuan, M. et al. Self-propelled nanomotors for thermomechanically percolating cell membranes. Angew. Chem. Int. Ed. 57, 12463–12467 (2018).

    CAS  Article  Google Scholar 

  14. 14.

    Chen, C. et al. Light-steered isotropic semiconductor micromotors. Adv. Mater. 29, 1603374 (2017).

    Article  CAS  Google Scholar 

  15. 15.

    Wang, Y. et al. Bipolar electrochemical mechanism for the propulsion of catalytic nanomotors in hydrogen peroxide solutions. Langmuir 22, 10451–10456 (2006).

    CAS  PubMed  Article  PubMed Central  Google Scholar 

  16. 16.

    Ibele, M., Mallouk, T. E. & Sen, A. Schooling behavior of light-powered autonomous micromotors in water. Angew. Chem. Int. Ed. 48, 3308–3312 (2009).

    CAS  Article  Google Scholar 

  17. 17.

    Konlger, A. & Kohler, W. Optical funnerling and trapping of gold colloids in convergent laser beam. ASC Nano 6, 4400–4409 (2012).

    Article  CAS  Google Scholar 

  18. 18.

    Kajorndejnukul, V., Ding, W., Sukhov, S., Qiu, C.-Q. & Dogariu, A. Linear momentum increase and negative optical forces at dielectric interface. Nat. Photon. 7, 787–790 (2013).

    ADS  CAS  Article  Google Scholar 

  19. 19.

    Ashkin, A., Dziedzic, J. M., Bjorkholm, J. E. & Chu, S. Observation of a single-beam gradient force optical trap for dielectric particles. Opt. Lett. 11, 288–290 (1986).

    ADS  CAS  PubMed  Article  PubMed Central  Google Scholar 

  20. 20.

    Lehmuskero, A., Johansson, P., Dunlop, H. R., Tong, L. & Kall, M. Laser trapping of colloidal metal nanoparticles. ACS Nano 9, 3453–3469 (2015).

    CAS  PubMed  Article  PubMed Central  Google Scholar 

  21. 21.

    Dogariu, A., Sukhov, S. & Saenz, J. J. Optically induced ‘negative forces’. Nat. Photon. 7, 24–27 (2013).

    ADS  CAS  Article  Google Scholar 

  22. 22.

    Chen, J., Ng, J., Lin, Z. & Chan, C. T. Optical pulling force. Nat. Photon. 5, 531–534 (2011).

    ADS  CAS  Article  Google Scholar 

  23. 23.

    Bzrobohaty, O. et al. Experimental demonstration of optical transport, sorting and self-arrangement using a ‘tractor beam’. Nat. Photon. 7, 123–127 (2013).

    ADS  Article  CAS  Google Scholar 

  24. 24.

    Damkova, J. et al. Enhancement of the ‘tractor-beam’ pulling force on an optically bound structure. Light Sci. Appl. 7, 17135 (2018).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  25. 25.

    Gao, D., Shi, R., Huang, Y. & Gao, L. Fano-enhanced pulling and pushing optical force on active plasmonic particles. Phys. Rev. A 96, 043826 (2017).

    ADS  Article  Google Scholar 

  26. 26.

    Mizrahi, A. & Fainman, Y. Negative radiation pressure on gain medium structures. Opt. Lett. 35, 3405–3407 (2010).

    ADS  PubMed  Article  PubMed Central  Google Scholar 

  27. 27.

    Guo, G., Feng, T. & Xu, Y. Tunable optical pulling force mediated by resonant electromagnetic coupling. Opt. Lett. 43, 4961–4964 (2018).

    ADS  CAS  PubMed  Article  PubMed Central  Google Scholar 

  28. 28.

    Xu, L., Mou, F., Gong, H., Luo, M. & Guan, J. Light-driven micro/nanomotors: from fundamentals to applications. Chem. Soc. Rev. 46, 6905–6926 (2017).

    CAS  PubMed  Article  PubMed Central  Google Scholar 

  29. 29.

    Seol, Y., Carpenter, A. E. & Perkins, T. T. Gold nanoparticles: enhanced optical trapping and sensitivity coupled with significant heating. Opt. Lett. 31, 2429–2431 (2006).

    ADS  CAS  PubMed  Article  PubMed Central  Google Scholar 

  30. 30.

    Fu, X., Chen, B., Tang, J. & Zewail, A. H. Photoinduced nanobubble-driven superfast diffusion of nanoparticles imaged by 4D electron microscopy. Sci. Adv. 3, e1701160 (2017).

    ADS  PubMed  PubMed Central  Article  CAS  Google Scholar 

  31. 31.

    Lukianova-Hleb, E., Volkov, A. N. & Lapotko, D. O. Laser pulse duration is critical for generation of plasmonic nanobubbles. Langmuir 30, 7425–7434 (2014).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  32. 32.

    Lachaine, R. et al. Rational design of plasmonic nanoparticles for enhanced cavitation and cell perforation. Nano Lett. 16, 3187–3194 (2016).

    ADS  CAS  PubMed  Article  PubMed Central  Google Scholar 

  33. 33.

    Metwally, K., Mensah, S. & Baffou, G. Fluence threshold for photothermal bubble generation using plasmonic nanoparticles. J. Phys. Chem. C 119, 28586–28596 (2015).

    CAS  Article  Google Scholar 

  34. 34.

    Hodak, J. H., Henglein, A., Giersig., M. & Hartland, G. V. Laser-induced inter-diffusion in AuAg core–shell nanoparticles. J. Phys. Chem. B 104, 11708–11718 (2000).

    CAS  Article  Google Scholar 

  35. 35.

    Sasikumar, K. & Keblinski, P. Molecular dynamics investigation of nanoscale cavitation dynamics. J. Chem. Phys. 141, 234508 (2014).

    ADS  PubMed  Article  CAS  PubMed Central  Google Scholar 

  36. 36.

    Merabia, S., Keblinski, P., Joly, L., Lewis, L. J. & Barrat, J.-L. Critical heat flux around strongly heated nanoparticles. Phys. Rev. E 79, 021404 (2009).

    ADS  Article  CAS  Google Scholar 

  37. 37.

    Maheshwari, S., van der Hoef, M., Prosperetti, A. & Lohse, D. Dynamics of formation of a vapor nanobubble around a heated nanoparticle. J. Phys. Chem. C 122, 20571–20580 (2018).

    CAS  Article  Google Scholar 

  38. 38.

    Trojek, J., Chvatal, L. & Zemanek, P. Optical alignment and confinement of an ellipsoidal nanorod in optical tweezers: a theoretical study. J. Opt. Soc. Am. A 29, 1224–1236 (2012).

    ADS  Article  Google Scholar 

  39. 39.

    Preez-Wilkinson, N. et al. Forces due to pulsed beams in optical tweezers: linear effects. Opt. Express 23, 7190–7208 (2015).

    ADS  PubMed  Article  CAS  PubMed Central  Google Scholar 

  40. 40.

    Salandrino, A., Fardad, S. & Christodoulides, D. N. Generalized Mie theory of optical forces. J. Opt. Soc. Am. B 29, 855–866 (2012).

    ADS  CAS  Article  Google Scholar 

  41. 41.

    Gao, D. et al. Unveiling the correlation between non-diffracting tractor beam and its singularity in Poynting vector. Laser Photon. Rev. 9, 75–82 (2015).

    ADS  Article  Google Scholar 

  42. 42.

    Vakarelski, I. U. et al. Self-determined shapes and velocities of giant near-zero drage gas cavities. Sci. Adv. 3, e1701558 (2017).

    ADS  PubMed  PubMed Central  Article  CAS  Google Scholar 

  43. 43.

    Ng, K. Overview of the ONR supercavitating high-speed bodies program. AIAA Guid. Navig. Control Conf. Exhibit. (2006).

  44. 44.

    Ellingsen, S. A. Theory of microdroplet and microbubble deformation by Gaussian laser beam. J. Opt. Soc. Am. B 30, 1694–1710 (2013).

    ADS  CAS  Article  Google Scholar 

  45. 45.

    Li, C. et al. Nanostructured copper interfaces for enhanced boiling. Small 4, 1084–1088 (2008).

    ADS  CAS  PubMed  Article  PubMed Central  Google Scholar 

Download references


This work is supported by National Science Foundation (1706039 and 1937923) and the Center for the Advancement of Science in Space (GA-2018-268).

Author information




E.L. and T.L. designed the experiments. E.L. set up and performed the experiments, and performed the optical simulations. D.H. and T.L. performed the MD simulations. All authors discussed the results and the mechanism of the ballistic Au NPs. E.L. and T.L. wrote the manuscript.

Corresponding author

Correspondence to Tengfei Luo.

Ethics declarations

Competing interests

The authors declare no competing interests.

Additional information

Peer review information Nature Communications thanks the anonymous reviewers for their contribution to the peer review of this work. Peer review reports are available.

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Lee, E., Huang, D. & Luo, T. Ballistic supercavitating nanoparticles driven by single Gaussian beam optical pushing and pulling forces. Nat Commun 11, 2404 (2020).

Download citation

Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.


Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing