Optical tweezers: 20 years on

David McGloin


In 1986, Arthur Ashkin and colleagues published a seminal paper in Optics Letters, ‘Observation of a single-beam gradient force optical trap for dielectric particles’ which outlined a technique for trapping micrometre-sized dielectric particles using a focused laser beam, a technology which is now termed optical tweezers. This paper will provide a background in optical manipulation technologies and an overview of the applications of optical tweezers. It contains some recent work on the optical manipulation of aerosols and concludes with a critical discussion of where the future might lead this maturing technology.


1. Introduction

The idea that light can trap and manipulate particles is what sold me on a career in research. It is one of those counterintuitive ideas that just seems wrong at some level, but when it is explained it makes perfect sense. The idea that light can exert forces on particles so as to push them (rather than trap them) is not so strange if we consider light as photons which possess momentum. If light can be reflected from a surface or scatter in some way then we must allow for the fact that its momentum has been changed and there must be, from Newton's second law, a force (force is proportional to the rate of change of momentum) associated with this change. This is how radiation pressure can be described. The concept of radiation pressure was considered by James Clerk Maxwell (1873) as he probed the consequences of his description of electromagnetic radiation.

In a medium in which the waves are propagated there is a pressure in the direction normal to the wave, and numerically equal to the energy contained in unit volume.(Maxwell 1873)

When we consider radiation pressure today, we tend to make use of lasers with their associated high intensity, and so it seems remarkable that P. N. Lebedev demonstrated the existence of radiation pressure using no more than a focused arc lamp (Lebedev 1901). Moreover, he did this in 1901, pioneering an area that would not see real resurgence until the early 1970s. This work would lead to two Nobel prizes (to date) allowing the laser cooling of atoms (e.g. Chu 1998) and the creation of Bose–Einstein condensates in cold atomic gases (e.g. Ketterle 2002).

The second type of force that light can exert can again be described by Newton's laws (although this explanation is only strictly valid in the case where the wavelength of the light is much smaller than the size of the particle involved) by considering what happens to the light as it traverses a dielectric particle. First, one notices that the light is refracted through the object, and as the light's direction is changed so must its momentum: thus a force must be acting on the particle (figure 1). To understand the direction of the force, we must consider the fact that the experiments that will be discussed in this paper make use of lasers. The typical profile of a laser beam is Gaussian, with the most intense part of the beam lying in the centre.

Figure 1

(a) The basic optical tweezers principle: take a laser source and focus it through a high numerical aperture microscope objective. (b) The beam paths through a dielectric sphere. The thicker line indicates a higher incident beam intensity. The imbalance in intensity between the inside and the outside of the beam means that the applied force on the bead must act towards the higher intensity part of the beam. This illustrates how a particle is confined in the transverse plane of the beam. z-trapping is not shown but works in a similar manner to the transverse trapping.

Thus, if the refractive index of the particle is greater than that of the surrounding medium, then the particle is attracted to the centre of the beam; if it is less than that of the surrounding medium, then the particle is repelled from the beam. Since the force is dependent on the intensity gradient of the beam, this type of force is called the gradient force (also called the dipole force). The assumption in this paper is that the relative refractive index (the ratio of the particle to medium refractive index) is greater than 1 and that we are working in the attractive force regime. From figure 1 we can see that particles should be relatively easy to confine in the transverse direction of the beam, but what about in the direction of beam propagation? Although it may be slightly counterintuitive when considered in light of radiation pressure we can also observe trapping in this axial direction, whereby the particle is confined very close to the beam focus, provided the gradient force is larger than the radiation pressure force. This z-trapping condition is achieved practically using high numerical aperture optics (the majority of experiments make use of oil immersion microscope objectives with NAs>1). The technique developed by Ashkin et al. (1986) in which a particle is confined in this manner, by a single laser beam, is known as optical tweezers and celebrates its twentieth birthday in 2006. The background, state of the art and future outlook in the general area of optical manipulation are the subject of this review.

2. A brief history of optical forces

After the very early pioneering work, interest in optical forces would largely die away until the development of the laser during the 1960s. Arthur Ashkin, working at Bell Labs, pioneered study in this area and produced a stream of remarkable work that laid the foundations for the field. Indeed, one could argue that the majority of work in many areas of optical manipulation is really only incremental in terms of the work carried out by Ashkin and colleagues. He initially focused on the radiation pressure of light. He demonstrated the ability of light to guide particles (Ashkin 1970), to levitate particles (radiation pressure–gravity traps; Ashkin & Dziedzic 1971), to confine particles in dual-beam radiation pressure traps (Ashkin & Dziedzic 1985), the levitation of airborne droplets (Ashkin & Dziedzic 1975), confinement in vacuum (Ashkin & Dziedzic 1976) and precision trapping via feedback (Ashkin & Dziedzic 1977a). Many of these techniques fell away from what the mainstream optical trapping community were actively working on but are now seeing a resurgence in interest. My own group, for example, works on optical levitation and guiding and is implementing dual-beam trapping methods, all primarily to trap airborne particles (which will be discussed below). Other notable work making use of radiation included the observation of whispering gallery modes in levitated droplets by Ashkin & Dzeidzic (1977b); such cavity resonance (the droplet acts as a microscopic optical cavity) can be used to experimentally verify Mie theory and size droplets very accurately.

Up until the demonstration of the single-beam trap, much of the work on optical forces had been pushing the drive towards laser cooling of atoms (e.g. Chu et al. 1985) and Ashkin's work fitted in as physics of the highest rank—he achieved just about everything one could imagine doing with radiation pressure over the course of a decade, but without a definitive focus. Also the availability of laser sources at the time may have limited work in this area by the wider community. In contrast, the optical tweezers technique would open up new areas of study in a short period of time.

The paper ‘Observation of a single-beam gradient force optical trap for dielectric particles’ (Ashkin et al. 1986) is a classic. Not only does it discuss a wholly new technique, but also it outlines exactly how the field would pan out over the next two decades.

They also open a new size regime to optical trapping encompassing macromolecules, colloids, small aerosols, and possibly biological particles. The results are of relevance to proposals for the trapping and cooling of atoms by resonance radiation pressure.(Ashkin et al. 1986)

And this is exactly what people would continue to work on. The paper also holds a few surprises. Not only are large Mie particles trapped (10 μm diameter) but also small Rayleigh particles, indeed evidence is presented demonstrating the trapping of 25 nm diameter silica beads, which still presents a real experimental challenge today and is of relevance for the developing ‘nanotechnology’ field. The paper also outlines the drag and drop technique for measuring the forces involved on holding particles, a quick and dirty method that is used in many laboratories today.

3. Techniques

The specific function of optical tweezers is to allow the non-invasive manipulation of single particles (or in more advanced set-ups many single particles simultaneously) and to carry out some kind of study on that particle. Over the past 20 years, the science enabled by optical tweezers has concentrated on the measurement of position and forces to incredible precision, primarily on force-producing molecules in biology, but also in colloid interactions studies and hydrodynamics. They have also allowed the controlled study of the properties of light beams and enabled single particle spectroscopy in a controlled manner. New variations on the original single-beam trap continue to develop, opening up new studies and allowing us to underpin our work with a better understanding of the science behind the interaction of light with matter. So how do these basic techniques work?

One of the most powerful things about optical tweezers systems is their simplicity. A functional optical tweezers that is used to carry out publication quality research can be constructed from a laser, a couple of telescopes, a few mirrors, a microscope objective and some imaging optics with a camera. This has allowed the proliferation of the technique and its introduction into undergraduate teaching labs worldwide. Indeed, a summer student in my group recently developed a ‘portable’ compact optical tweezers system which consisted of a low-power laser diode, a dichroic mirror, an aspheric lens (instead of a microscope objective) and a battery powered wireless camera and high brightness LED (for imaging). The whole system is mounted on a post and is portable in the sense that it can be lifted in one hand and moved from place to place (and still work). A schematic for a simple optical tweezers set-up is shown in figure 2.

Figure 2

A basic optical tweezers system. The beam is expanded to the desired size by the first telescope. The second telescope aids in beam alignment and beam steering. This expanded beam passes through a microscope objective into the sample. Such a system is very simple to design and build, and this simplicity is one of the optical tweezers' great selling points.

At present, one of my primary research interests is the use of holographically generated light fields for the manipulation of particles and atoms. This is one of the newer methods for the trapping and manipulation of multiple particles. These fall into three broad areas: simple beam combination techniques, scanning techniques and holographic techniques.

The beam combination techniques are best summarized by the work of Fallman & Axner (1997), in which a laser is split into two separate beams. With careful optical design, these can be combined and independently controlled in the focal plane of the microscope objective. Thus, a dual-beam optical tweezers is formed, and this type of system forms the workhorse for much of the force measurement research that currently takes place. Scanning techniques generally make use of acousto-optic deflectors (AODs) which can scan a beam from point to point at kilohertz rates. This is a very powerful and flexible technique (one of the most impressive demonstrations is that of micro-Tetris by Christoph Schmidt's group at Vrije University (http://www.nat.vu.nl/compl/index-en.html)), which works by time-sharing the light between trapping sites. So long as the beam returns to the trapped particle before it diffuses away, then the particle will remain trapped. This type of effect can be partially extended into three dimensions by some clever optics, and has been demonstrated by Alfons van Blaaderen's group (Vossen et al. 2004).

To achieve true three-dimensional control of multiple trap sites, one must move to holographic techniques. A hologram is able to control the phase of a light beam, which tells how the beam will propagate. So if one wishes to have a laser beam turn into a picture of the Royal Society crest a hologram which encodes the phase of the crest pattern must be generated. Then by reflecting our normal laser beam off the hologram, we can transform that beam into the desired pattern. The holograms can be generated by the use of well-established algorithms. For complex holograms (designs other than simple arrays of spots), we make use of iterative algorithms designed to solve the inverse problem. The concept is illustrated in figure 3.

Figure 3

Producing a hologram from an intensity pattern. To recreate an intensity image using a hologram, we first need to compute the desired phase of that object. This is done by taking the desired image ((a) the Royal Society logo) and feeding it through a computer algorithm, in this case an iterative adaptive algorithm. The computer can compute the hologram (b) and what it predicts the image will look like when ‘replayed’ by the illuminating laser beam (c).

Such work was pioneered by Fournier et al. (1995), who made use of static glass holograms to produce beams with multiple trapping sites. Follow-up work to this was done by David Grier's group (then in Chicago, now at NYU) who studied the use of glass-etched holograms for examining the dynamics of colloidal particles in these ‘extended’ light fields. This work would form the basis for later optical sorting techniques. The next step beyond the fabricated techniques was to use dynamically alterable devices, spatial light modulators. The pioneers in this area were Tizani's group at the University of Stuttgart, who outlined how computer controlled holograms (Reicherter et al. 1999; Liesener et al. 2000) could be used to generate trapping patterns that could be iterated in real time and allowed full three-dimensional control over the particles. This work was quickly built upon by the Grier group in the paper ‘Dynamic holographic optical tweezers’ (Curtis et al. 2002), and is now regarded by many as the foundation on which the growing body of work in this area is based. The 2002 paper outlined how iterative algorithms could be used to implement complicated trapping arrays and extended the work of Tiziani from a few to hundreds of trap sites. Although it is not the first paper in the field it did seem to energize the community about this new technique. The advantages of using dynamic holographic optical tweezers are that they offer full control over the spatial localization of a trapped particle. This means that each particle can be moved independently in three dimensions. Further, the use of holograms offers the potential to correct for optical aberrations in the system (Wulff et al. 2006) as well as offering a user-friendly experimental experience (if one hologram is wrong, then it is a simple matter to change it to a better one). The disadvantages of this technique tend to lie in the speed of the devices, their efficiency and image fidelity. The issue of speed is one that can easily be seen by trying to recreate the beam scanning technique of an AOD by a spatial light modulator (SLM). The parameters for each manufactures' SLM are slightly different, but we carried out a time-sharing experiment with a phase modulating Boulder nonlinear systems device (Melville et al. 2003), which should have been able to run at above video frame rates (compared with kilohertz rates for an AOD) and found that in practice, for trapping experiments, it was limited to around 10 Hz. This was enough to trap six particles via time sharing but shows that the SLM is not optimized for rapid dynamic tasks. In the experiment to demonstrate this, we also showed some of the power of the SLM, in that the particles trapped were all trapped on different z-planes, spaced approximately 1 μm apart, a trick that cannot be done with an AOD. Of course, one can just make a hologram to trap the six particles simultaneously rather than by time sharing. The dynamics of the SLM have not really been an issue in experiments to date and for those experiments that do require speed, such as atom trapping (McGloin et al. 2003), different types of spatial light modulators with much lower efficiency can be used, whereby kilohertz rates can be achieved (Boyer et al. 2004, 2005).

To date, much of the work on the SLMs has concentrated on device characterization and novel colloidal studies, with some work in the biosciences, as well as on more general beam shaping techniques. To date, two of the main players are the Grier group at NYU and Miles Padgett's group in Glasgow. The New York group has focused on colloidal manipulation, beam shaping and algorithm improvements. The power of SLMs in beam shaping techniques was shown in the study of optical vortices (optical singularities; Curtis & Grier 2003a,b) in which optical vortex beams could be created very simply in real time. They could be modulated with ease to alter the shape, while still retaining the orbital angular momentum associated with such beams. The experiments described by Curtis & Grier (2003a) are simple to try using dynamic techniques, but would require significant fabrication effort to be done ‘offline’. Such vortices have also been used to study hydrodynamic coupling between colloidal particles (Ladavac & Grier 2005) and develop microfluidics pumps (Ladavac & Grier 2004). Other colloidal work has made used of dynamically evolving light patterns to mimic peristaltic pumps, whereby temporally evolving light patterns can be used to transport particles in the absence of flow. Such techniques can also be used to achieve and study thermal ratchets (Lee & Grier 2005a, 2006; Lee et al. 2005) including the observation of flux reversal, whereby particles can be made to move in the opposite direction to the direction of the optical pattern. My own group has been working on similar mechanisms using static patterns. We have examined how particles move in Bessel beams, which show directed transport of particles owing to the intensity imbalance between the outside and the inside of the beam (Milne et al. 2005; Paterson et al. 2005). This technique can be used to sort particles, including red and white blood cells in the absence of flow. The sorting is relatively slow, but may find niche application areas.

Other notable work by the New York group in this area has been the demonstration of dual wavelength holographic optical tweezers (Lee & Grier 2005b), a technique that is likely to be of interest for future optical tweezers spectroscopic tools, the push of optical tweezers as nanotools for the manipulation of both carbon nanotubes (Plewa et al. 2004) and semiconductor nanowires (Agarwal et al. 2005), and the assembly of quasicrystals (Roichman & Grier 2005).

The other main group working on holographic traps is based at the University of Glasgow, which is the home of the signature experiment in this area. Akin to the Tetris experiment using AODs, the ‘Smallest strip the willow in the world’ (Willow) demonstrates the power of the SLM technique (and is beautifully put to music to boot). The Glasgow work has focused on complicated beam shaping, and the creation and controlled rotation of three-dimensional crystals (Bingelyte et al. 2003; Jordan et al. 2004; Leach et al. 2004a; Sinclair et al. 2004a), three-dimensional beam propagation algorithms (Sinclair et al. 2004b; Whyte & Courtial 2005) as well as the structure of light beams (Leach et al. 2004b). They have also looked at limiting values in holographic traps and aberration correction (Sinclair et al. 2004c) and are now exploring applications in microfluidics, including using holographic tweezers and video microscopy to map out fluid flow in microchannels and around rotating microobjects (Di Leonardo et al. 2006).

Another recent first for holographic tweezers includes the Raman imaging of cells (Creely et al. 2005), in which a cell is manipulated by an array of spots and scanning through the probing Raman beam. For people interested in single particle spectroscopy, optical manipulation seems to offer much: the particle of interest is localized and static and therefore is easy to probe. There has been increasing work in this area, with Raman spectroscopy being a popular choice among researchers. Thurn & Kiefer (1984) carried out work on Raman spectroscopy of levitated droplets, while Biswas et al. (1989) looked at stimulated Raman scattering (SRS) and this work built on Ashkin & Dziedzic's (1977b) whispering gallery mode work. Direct Raman spectroscopy on a trapped particle was shown by Ajito & Torimitsu (2001) on droplets in solution and on polystyrene beads.

The use of holographic tweezers in cellular microscopy (Emiliani et al. 2005) has also been recently shown, demonstrating how the technique may be used in biology to measure forces, generate forces (Emiliani et al. 2004) or to locally probe different parts of a cell simultaneously by trapping an array of particles and ‘moulding’ them around the cell.

The use of holograms in optical trapping is not limited to the use of spatial light modulators. One of the most significant papers in recent years in the field was the demonstration of particle sorting using an optical lattice (MacDonald et al. 2003). Here, a simple-etched hologram was used to create a five-beam interference pattern which was projected into a sample. As particles flow through the three-dimensional optical structure, they are separated. The separation is due to the different interaction with the light that particles with different polarizabilities have. It is also dependent on the connectivity of the lattices sites, with some light leakage between sites offering the best sorting. The sorting mechanism is due to the interplay between the Brownian motion of the particles, the optical forces and the flow-induced forces. By tailoring the relative phases and intensities of the interfering beams a sophisticated sorting sieve can be created. Such work has also been demonstrated in a slightly simpler set-up using a single line of the SLM-generated traps (Ladavac et al. 2004a). As optical sorting is passive, particles are sorted by their inherent properties and do not necessarily require labelling. Therefore, the goal would be to separate cell types, such as red and white blood cells, or cancerous and non-cancerous cells merely due to the fact that they interact with the light in slightly different ways. Work towards these goals is underway, but robust and reliable methods of routinely sorting cells, as opposed to non-biological colloidal particles, are still some way off.

Another developing technique for the manipulation of large numbers of particles is evanescent field manipulation (figure 4). This work, pioneered by Kawata (Kawata & Sugiura 1992), has seen a resurgence of late (Garcés-Chávez et al. 2005; Quidant et al. 2005) and recent work has demonstrated optically bound arrays in evanescent fields (Mellor & Bain 2006) and also large area manipulation using surface plasmon field enhancement (Garcés-Chávez et al. 2006). While this type of manipulation may have applications in colloidal crystallization studies, it is not yet evident if it offers any significant advantages over existing techniques. Further the issue of the control of individual particles within the evanescent field has yet to be seriously addressed, but it does show promise as the areas over which particles can be manipulated are significantly larger than the microscope-based techniques.

Figure 4

(a) Experimental configuration for large area evanescent manipulation. The inset shows the beam geometry at the interface of the prism and the sample. (b) Field intensity of a five-beam interference pattern used to trap particles in (c). Reprinted with permission from V. Garcès-Chávez.

The final technique to discuss is one of the most widely used in optical manipulation, the ability to detect very small position changes and the ability to sense and apply forces in the piconewton range. These techniques rely on the fact that a simple beam optical tweezers is a harmonic trap and any particle trapped in it can be considered a damped harmonic oscillator. This information allows us to calibrate the trap to work out the spring constant and using this information allows us to sense the position of a trap particle or work out the force being applied to the particle. The primary uses of this technique are in the studies of biological organisms, and in particular, the work on molecular motors (Neuman & Block 2004). In the past year or so, Steven Block's group at Stanford have developed their already world-leading studies in this area to push the position sensing capabilities of optical traps to angstrom precision (quite remarkable when you consider the diffraction limited resolution of the trap itself). They were able to observe the base pair stepping of RNA polymerase, showing that the steps averaged around 3.7 Å (Abbondanzieri et al. 2005; Greenleaf et al. 2005). Another landmark experiment in this field was also carried out last year (Rohrbach 2005) in which an optical tweezers was used to measure a controlled 25 fN force on a 533 nm latex sphere. The suggestion is that this is the smallest switchable force ever measured and illustrates the power of optical tweezers in this arena. To put this in context, the thermal forces on such a particle are likely to be in the piconewton regime and the Sun exerts around 20 fN on a 75 μm diameter dust particle floating in the atmosphere. These two results (position and force sensitivities) push the techniques into new realms that will allow us to explore ever more sensitive parameters and is likely to enable optical tweezers to get more of a handle on the nanoworld.

(a) Optical manipulation derivatives

Recent work by Ming Wu's group at Berkeley (Chiou et al. 2005) has demonstrated a convergence of two types of manipulation to offer a tantalising vision of the future of optical manipulation. The concept is essentially an extension of the trapping technique known as dielectrophoresis (DEP) in which electric fields are used instead of optical fields to trap (or repel) particles. Conventional DEP makes use of patterned electrodes to allow localization of the electric field to enable trapping. As such it is a fixed architecture technique requiring complicated patterning to enable more arbitrary functionality. The beauty of the new light-induced dielectrophoresis (LIDEP) is that a large area electrode can be ‘patterned’ by a light field, by either scanning a pattern on the electrode or a mask (such as a hologram) to project a static pattern, and the optical power level required to create an electric trap can be extremely low (microwatts) compared with optical tweezers. The power of the technique lies in combining the large area effect of DEP with the precision and reconfigurability of optics. This new technique has made many in the optical trapping community to sit up and take note—traditionally, the optical and dielectrophoretic communities work in isolation. Work in the Optical Trapping Group in St Andrews has recently produced a proof of concept LIDEP device with the ultimate goal of producing a large area, high throughput sorting device (S. N. Neale 2005, personal communication).

4. Recent work

At present, one of my research focuses is the manipulation and interrogation of droplets. Some of the application of this work is inspired by our collaborators, the group of Jonathan Reid, in the Chemistry Department at Bristol University. The focus is on the manipulation of droplets with a view to studying their size, composition and dynamics (gas uptake studies, particle coagulation, etc.), in atmospheric chemistry. Our work concentrates on areas, such as guiding of droplets over large distances using radiation pressure, designing new techniques to trap and probe droplets enabling the manipulation of many droplets simultaneously to try and create optically driven digital microfluidic chemical microreactors.

We have recently been working on the optical guiding of aerosol droplets (Summers et al. 2006). Typical work on the optical spectroscopy of single droplets relies on the probing of freely falling droplets. Therefore, single droplet manipulation techniques have great utility in allowing more accurate and longer time-scale studies. Optical tweezers is one method of achieving single droplet localization. However, for some studies, it would be useful to drive a droplet (or other airborne particle) through a number of spectroscopic beams (say, doing a fluorescence measurement followed by a Raman measurement) in a controlled fashion; we have been investigating ways to do this optically (figure 5). The simplest method is to take a Gaussian beam and levitate a particle against gravity. One can then alter the power in the beam and adjust the equilibrium position to adjust the height of the particle. Using this method, we can guide droplets (of water, ethanol and dodecane) over several hundred micrometres. However, a higher guiding distance is desired to give appropriate spatial separation between our multiple probe beams and so we make use of the non-diffracting properties of the Bessel beam. The Bessel beam has a long, thin core which does not spread in the same way as a focused Gaussian beam of similar dimension. As such we can use it to controllably guide droplets over much longer distances, of around 4 mm.

Figure 5

Optical levitation of dodecane droplets. We are able to stably trap arrays of droplets (in this case six) using a Bessel beam. We believe the arrays are optically bound, that is the position of the droplet above the one below is determined not only by the levitating beam but also by the interaction of the light with the lower droplet.

Trapping airborne particles is inherently more difficult than trapping particles in a suspending medium such as water. This is due to the reduced damping offered by the medium, and so one cannot simply go and pick up the particle one desires as in conventional fluid sample optical tweezers, rather we must wait until a particle falls into the trap and then make a decision as to whether it is the one we want. This is one reason why airborne trapping has received far less attention than trapping in the underdamped regime to date, despite much of Ashkin's early work being devoted to levitating droplets. The focus for early airborne work up until a few years ago was optical levitation, often combined with Raman spectroscopy (Thurn & Kiefer 1984; Biswas et al. 1989). Omori et al. (1997) appear to have been the first group to directly optically trap (as opposed to levitate) an airborne particle, in this case glass beads. Magome et al. (2003) achieved the same result with liquid droplets. Reid's group along with Andrew Ward at the Rutherford Appleton Laboratory (RAL) followed up this experiment with work trapping two droplets in a dual-beam trap (Hopkins et al. 2004) and presented the first demonstration of optically controlled coagulation of aerosols. Using the same system around the same time (at RAL), Martin King examined how single seawater and oleic acid droplets reacted with ozone (King et al. 2004). Both of these recent papers indicate the potential power of optical trapping techniques to elucidate mechanisms in atmospheric science.

Reid's work demonstrated the utility of using cavity-enhanced Raman scattering in sizing droplets (to within ±2 nm, limited by measurement resolution) and also outlined how tweezers may be used to study droplet dynamics by examining what happens during coagulation (figure 6a). We have recently demonstrated the use of holographic optical tweezers to trap multiple aerosol droplets (Burnham & McGloin 2006; figure 6b,c) with the intention to explore similar ideas with increasing numbers of particles. We have also shown that the refresh rate of the SLM does not seem to be a limiting factor to observe the real-time manipulation of droplets and this has allowed us to demonstrate the controlled coagulation of water droplets. We have also tentatively demonstrated the ability of optical tweezers to rotate droplets (via orbital angular momentum) and we believe that this may allow us to create micromixers for airborne chemical reactors and microfluidic devices.

Figure 6

(a) Droplet coagulation and associated Raman spectra. The sharp spectral peaks are due to the cavity enhancement provided by the droplets. The uppermost spectra are for the droplet on the right, the middle spectra are the combined signal for both drops and the lower signal for the coagulated drop. The combined volume calculated from the spectra is 3.905×10−16 m3, while the coagulated volume is calculated as 3.902×10−16 m3. (b) Array of six droplets held by holographic optical tweezers. (c) Array of four droplets in a Y configuration held by holographic tweezers. (Reproduced by permission from the PCCP Owner Societies.)

5. Future outlook

So where to now? Optical manipulation research has not been shy about jumping on bandwagons and the push towards nanotech and biotech and by implication microfluidics are probably where the application-driven focus of the work will take place. Integration of optical manipulation systems into lab-on-a-chip type microflow devices is an active area of research in my group if advanced devices of this type are economically viable, scalable and practical at whichever coal face one is interested in working (GP's surgery, medical labs, industrial labs, academic labs, etc.) and that this area is certainly worth exploiting to the full. Integration of optical sorting (which has yet to really prove itself in the biological arena beyond limited proof of concept experiments) and single particle spectroscopy techniques are also worth pursuing provided they can provide adequate throughputs and offer real advantages over existing techniques. Recent, unpublished, work at St Andrews suggests AOD sorting methods may be where the future of the technique lies. It is likely that they will find niche areas rather than replacing conventional proven techniques, such as FACS.

The advanced use of techniques, such as holographic tweezers is still in its infancy, while some interesting experiments have been carried out the real proof of the pudding is still to be shown. I am confident that this method will make a scientific impact, but it may only be in speeding up or making more convenient existing experimental systems. The application areas that look worth exploiting again lie in microfluidics, in novel force-sensing techniques in biology and colloidal interactions.

Near field methods appear to be an exciting option to try and move optical manipulation properly into the nano-regime. If people are serious about using optical tweezers to move truly nanoscale objects, then this is the way forward to overcome the diffraction limit problem. Work has already begun and is likely to prove a very fruitful avenue of research over the coming decade or so.

Both near field and holographic techniques offer the possibility of manipulating large numbers of particles simultaneously. This may allow information about large-scale colloidal dynamics but I have doubts that this will prove any better than existing techniques. In the case of large-scale manipulation by surface evanescent fields, this is mainly due to the fact that the optical forces induce troublesome thermal forces, but there remains plenty of mileage left in this field. The use of holographic fields is probably limited by the fact that they are used directly in optical tweezers (using high NA objectives) and as such have a limited field of view which represents a limited number of trapped particles. However, for smaller scale studies, these may be a good option.

Some of the pioneering experiments that have recently been reported making angstrom scale position measurement and femtonewton force measurements are just jaw dropping. The implications are for ever more precise measurements on biological motors and other single molecular systems.

To conclude, this has been a brief dash over the field of optical manipulation, which celebrates its twentieth birthday in 2006. The future for the area looks bright as this active research area matures into a set of well-defined tools, and there is no reason to think that another 20 years of active research are not about to begin.


I would like to acknowledge the Royal Society for their support as well as the RSC, NERC and EPSRC for funding some of the work mentioned above. The members of the Optical Trapping Group at St Andrews are also thanked for many stimulating discussions over the past several years.


  • One contribution of 23 to a Triennial Issue ‘Mathematics and physics’.


View Abstract