The scalability and resolution of conventional optical manipulation techniques are limited by diffraction. We circumvent the diffraction limit and describe a method of optically transporting nanoparticles across a chip using a gold surface patterned with a path of closely spaced C-shaped plasmonic resonators.
The technique of using focused laser beams to trap and exert forces on small particles has enabled many pivotal discoveries in the nanoscale biological and physical sciences over the past few decades. The progress made in this field invites further study of even smaller systems and at a larger scale, with tools that could be distributed more easily and made more widely available. Unfortunately, the fundamental laws of diffraction limit the minimum size of the focal spot of a laser beam, which makes particles smaller than a half-wavelength in diameter hard to trap and generally prevents an operator from discriminating between particles which are closer together than one half-wavelength. This precludes the optical manipulation of many closely-spaced nanoparticles and limits the resolution of optical-mechanical systems. Furthermore, manipulation using focused beams requires beam-forming or steering optics, which can be very bulky and expensive. To address these limitations in the system scalability of conventional optical trapping our lab has devised an alternative technique which utilizes near-field optics to move particles across a chip. Instead of focusing laser beams in the far-field, the optical near field of plasmonic resonators produces the necessary local optical intensity enhancement to overcome the restrictions of diffraction and manipulate particles at higher resolution. Closely-spaced resonators produce strong optical traps which can be addressed to mediate the hand-off of particles from one to the next in a conveyor-belt-like fashion. Here, we describe how to design and produce a conveyor belt using a gold surface patterned with plasmonic C-shaped resonators and how to operate it with polarized laser light to achieve super-resolution nanoparticle manipulation and transport. The nano-optical conveyor belt chip can be produced using lithography techniques and easily packaged and distributed.
Capture, interrogation and manipulation of single nanoparticles are of growing importance in nanotechnology. Optical tweezers have become a particularly successful manipulation technique for experiments in molecular biology1-4, chemistry5-7 and nano-assembly7-10, where they have enabled breakthrough experiments such as the measurement of the mechanical properties of single DNA molecules4 and the sorting of cells by their optical properties11,12. Discoveries on these frontiers open up the study of even smaller systems, and they make way for the engineering of new practically beneficial products and techniques. In turn, this trend drives the need for new techniques to manipulate smaller, more rudimentary particles. In addition, there is a push to build 'lab-on-a-chip' devices to perform these functions more cheaply and in a smaller package in order to bring chemical and biological tests out of the lab and into the field for medical and other purposes13,14.
Unfortunately, conventional optical trapping (COT) cannot meet all of nanotechnology's growing demands. COT operates on the mechanism of using a high numerical aperture (N.A.) objective lens to bring laser light to a tight focus, creating a localized peak in optical intensity and high gradients in the electromagnetic field energy. These energy density gradients exert a net force on light-scattering particles which generally draws them in towards the center of the focus. Trapping smaller particles requires higher optical power or a tighter focus. However, focused beams of light obey the principle of diffraction, which limits the minimal size of the focal spot and places an upper bound on the energy density gradient. This has two immediate consequences: COT cannot trap small objects efficiently, and COT has trouble discriminating between closely spaced particles, a trapping resolution limitation known as the 'fat fingers' problem. In addition, implementing multiple particle trapping with COT requires systems of beam-steering optics or spatial light modulators, components which drastically increase the cost and complexity of an optical trapping system.
One way to circumvent the fundamental limitations of conventional focused beams of light, said to propagate in the far field, is to instead exploit the gradients of optical electromagnetic energy in the near field. The near field decays exponentially away from sources of electromagnetic fields, which means that not only is it highly localized to these sources, but it also exhibits very high gradients in its energy density. The near fields of nano-metallic resonators, such as bowtie apertures, nano pillars, and C-shaped engravings, have been shown to exhibit extraordinary concentrations of electromagnetic energy, further enhanced by the plasmonic action of gold and silver at near-infrared and optical wavelengths. These resonators have been used to trap extremely small particles at high efficiency and resolution15-22. While this technique has proven effective at trapping small particles, it has also proven to be limited in its ability to transport particles over appreciable range, which is necessary if near-field systems are to interface with far-field systems or microfluidics.
Recently, our group has proposed a solution to this problem. When resonators are placed very close together, a particle can in principle migrate from one near-field optical trap to the next without being released from the surface. The direction of transport can be determined if adjacent traps can be turned on and off separately. A linear array of three or more addressable resonators, in which each resonator is sensitive to a polarization or wavelength of light different from that of its neighbors, works as an optical conveyor belt, transporting nanoparticles over a distance of several microns on a chip.
The so-called 'Nano-Optical Conveyor Belt' (NOCB) is unique among plasmonic resonator trapping schemes, as not only can it hold particles in place, but it can also move them at high speed along patterned tracks, gather or disperse particles, mix and queue them, and even sort them by properties such as their mobility23. All of these functions are controlled by modulating the polarization or wavelength of illumination, with no need for beam-steering optics. As a near-field optical trap, the NOCB trapping resolution is higher than that of conventional focused-beam optical traps, so it can differentiate between particles in close proximity; because it uses a metal nanostructure to concentrate light into a trapping well, it is power-efficient, and does not require expensive optical components such as a high N.A. objective. Furthermore, many NOCBs may be operated in parallel, at high packing density, on the same substrate, and 1 W of power can drive over 1200 apertures23.
We have recently demonstrated the first polarization-driven NOCB, smoothly propelling a nanoparticle back and forth along a 4.5 µm track24. In this article we present the steps necessary to design and fabricate the device, optically activate it and reproduce the transport experiment. We hope that making this technique more widely available will help bridge the size gap between microfluidics, far-field optics, and nanoscale devices and experiments.
1. Design the C-shaped Engraving (CSE) Array
Figure 1. CSE Layout. Depiction of conveyor belt repeating element. Successful transport has been achieved using dy = 320 nm and dx = 360 nm. Adjacent pairs of engravings have a 60º relative rotational offset. Please click here to view a larger version of this figure.
Figure 2. Simulation Geometry. Example of numerical simulation geometry in the commercial Finite Element Method software COMSOL. Two conveyor belt periods are simulated with dy = 320 nm and dx = 360 nm and a 500 nm diameter sphere. Shaded material regions are a) HSQ, b) polystyrene, c) gold, and d) water. Please click here to view a larger version of this figure.
Figure 3. Trapping Verification. Stable trapping can be demonstrated by plotting the optical potential of activation states. A single period of just three CSEs is analyzed for simplicity. Indeed, overall trap depth is sufficient (> 10 kBT) for stable trapping at the activated engraving for each state A, B, and C. Please click here to view a larger version of this figure.
Figure 4. Handoff Verification. Handoff can be demonstrated by plotting the optical potential of old (light red) and new (bright red) activation states in sequence. A single period of just three CSEs is analyzed for simplicity. During handoff from A to B and B to C, the potential barrier in the direction of desired motion between those two positions is both small (1 kBT) and smaller than that in the opposite direction, indicating that controlled handoff is likely. Handoff from C to A is most difficult because the inter-trap barrier remains sizable at all polarizations. Please click here to view a larger version of this figure.
2. Fabricate the CSE Array
Note: The process diagram is shown in Figure 5. This process is inspired by the work in ref. 25 and 26.
Figure 5. CSE Process. Process flow diagram of the dual-layer template-stripping process. E-beam lithography with 100 keV energy is used to expose the conveyor pattern on the HSQ resist. The thin PMMA layer underneath the HSQ is intended to facilitate the final strip-off (release) of the device from the Si substrate. Please click here to view a larger version of this figure.
3. Prepare the Specimen Sample
4. Calibrate the Focus of the Optical Columns
Note: A schematic of the apparatus can be referenced in Figure 8.
5. Trap and Manipulate Specimen with Optical Energy
Note: A schematic of the apparatus can be referenced in Figure 8.
Figure 7 is a picture of the final device. At the center of the 1 cm x 1 cm gold surface is the matrix of CSE and conveyor patterns, which can be barely seen from an angled view. Figure 6 is a scanning electron microscopy image of an example CSE pattern on the final device.
The particle motion of a 390 nm polystyrene bead traveling across a nano-optical conveyor belt 5 µm in length is shown in Figure 9. The curve shows the particle’s position as a function of laser polarization angle. As mentioned in the protocol, there may be cases where transport does not succeed or near-field trapping does not initiate. The best course of action is to try a different pattern, which may be in better condition.
Figure 6. SEM image of CSE Array. Scanning electron microscope (SEM) images of the CSE patterns. (a) shows the picture of HSQ mesas after the resist development. The sample is sputtered with 5 nm gold as a conducting layer for SEM inspection. (b)-(c) show final patterns after the sample is released from the silicon substrate. Please click here to view a larger version of this figure.
Figure 7. CSE Array Chip. Picture of the final device, roughly 10 mm x 10 mm in dimension. The picture shows the front gold surface of the device. Diffraction from grating ID markings is visible as multi-color squares near the center of the chip. Please click here to view a larger version of this figure.
Figure 8. Experimental apparatus. Schematic of the experimental apparatus. Both trapping and imaging are performed in reflection mode. The different light paths are distinguished using different colors. The red, green, dashed red, blue and yellow lines represent the light paths of optical trapping (conveyor driving), fluorescent imaging, laser imaging, fluorescent excitation and bright field illumination respectively. Please click here to view a larger version of this figure.
Figure 9. Bead Trajectory Over Double-Rail Conveyor. Position vs. polarization angle for a 390 nm bead moving on a 4.5 µm long double-rail conveyor belt. Images on the left show snapshots of the sphere after each conveyor period. The curve on the right traces the calculated position of the bead centers. Please click here to view a larger version of this figure.
The NOCB combines the strong trapping forces and small trap size of plasmonic approaches with the capability to transport particles, long available only for conventional focused-beam techniques. Unique to the NOCB, the trapping and transport properties of the system are a result of surface patterning and not of shaping the illumination beam. Provided the illumination is bright enough and its polarization or wavelength can be modulated, particles can be held or moved in complicated protocols on the surface. We have demonstrated through simulation that an NOCB can also rapidly sort particles based on their mobility23. Near-field traps can serve as small reaction volumes for single-molecule chemistry, and the inherent parallelizability of the NOCB means it can be used to set up, perform and tear down as many simultaneous experiments as can be packed onto a chip and illuminated at once.
To get the NOCB to work, the near-field optical forces that hold and hand off nanoparticles must overcome the competing forces of viscous drag, conventional optical trapping (the force of the illumination beam), thermophoresis, and contact forces with other particles and the substrate itself. The near-field optical force should be as strong as possible for a given illumination power; this requires careful nanostructure design and fabrication, but in practice we need to produce a range of structures with different characteristic sizes to select the one that works the best for the given illumination wavelength. The viscous drag and thermophoresis must be suppressed as well: while they may not be able to pull particles out of near-field optical traps, they can certainly make it difficult to get a particle onto the NOCB array in the first place.
When the sample is first placed under the microscope, particles will distribute evenly throughout the volume and very rarely come near to the CSE array. (Calculations indicate that a particle must move within a few tens of nanometers of surface contact to be trapped.) When the illumination is initially turned on, the CSE array will immediately heat up and create a thermal gradient in the water that repels particles over a distance of several hundred nanometers. This barrier is overcome by trapping a particle at a distance with the focused illumination beam, and manually dragging the particle through the thermal barrier into the trapping field of the CSE. However even this method will fail if the thermal gradient is too high. In our experience, the inclusion of the copper heat sink layer was crucial for drawing the heat away from the water and weakening the thermal forces. The copper heat sink also makes it less likely that the water will boil under normal illumination intensity.
The optical gradient force on a very small particle scales as the cube of the particle diameter. This makes it much harder to trap a 100 nm bead than a 200 nm bead, since the power must be increased eight times—increasing the substrate heating by the same amount. As a practical matter we recommend trapping larger beads first (400 nm or 500 nm diameter), optimizing the trap strength and minimizing competing forces, and then attempting trapping and transport of smaller particles.
Once the sample has been prepared, experiments can be performed as long as the particles are freely floating in water. Water exits the sample by evaporation along the edges. In our lab this puts a roughly 20 min time limit on experiments. Evaporation can also result in a competing viscous drag force as water is drawn to the edges of the sample. If the sample has rough features such as bent-up metal edges or spikes that prevent it from lying flat on the glass slide, the greater exposed surface area of the water will speed evaporation. If one side is higher than the other, the evaporation will be biased towards the side with the larger sample-slide gap and the fluid will move rapidly over the nanostructures, making it harder to see, capture and hold particles.
A single NOCB can transport particles across the width of the illumination beam but no further. As the beam intensity drops off, the restoring force from the focused beam grows stronger and the NOCB handoff force grows weaker, until polarization rotation is more likely to release the particle than move it forward. For extension to longer conveyors or more parallel conveyors, the illumination area must be increased. A powerful, defocused laser diode could power a much larger area than the laser used in these experiments. Alternatively the illumination area can be increased by rapidly scanning the beam using an acousto-optic deflector.
The authors have nothing to disclose.
The authors would like to thank Professor Yuzuru Takashima at the University of Arizona for discussions on optical imaging, Mr. Karl Urbanek for assistance with high power lasers, and Max Yuen for discussions of Brownian motion. The authors send further thanks to Professor Kenneth Crozier at Harvard University for helpful discussions on optical trapping experiments. Funding was provided in part by the United States National Science Foundation (award number 1028372).
HSQ e-beam resist | Dow Corning | XR-1541-006 | |
PMMA | MicroChem | 950A2 M230002 | |
Fast curing optical adhesive | Norland Optical Adhesive | NOA 81 | |
Fluorescent carboxyl microspheres | Bangs Laboratories | FC02F, FC03F | |
Fluorescent carboxylate-modified microspheres | Molecular Probes | F-8888 | |
Quartz slide | SPI Supplies | 1020-AB | |
Inverted fluorescent microscope | Nikon | ECLIPSE TE2000-U | |
Nd:YAG laser | Lightwave Electronics | 221-HD-V04 | |
sCMOS camera | PCO | EDGE55 | |
CCD camera | Watec | WAT-120N | |
Zero-order half-wave plate | Thorlabs | WPH05M-1064 | |
Triton X-100 | Sigma-Aldrich | T8787 | |
Distilled water | Invitrogen | 10977-023 | |
Si Wafer | Silicon Quest International | 708069 | |
Optical lenses | Thorlabs |