Touchless tweezing

Illustration by Terry Miura

Imagine being able to pick up and move a single cell without physically touching it. Sound like science fiction? By leveraging special properties of laser beams, optical tweezers can do just that. Optical tweezers are noninvasive tools that use a laser beam, or beams, to generate piconewton forces powerful enough to manipulate microscopic matter. This ability is of increasing interest in an array of disparate subject areas that include studies of biological cells and molecular motors, micromachines, microfluidics, colloid physics, and properties of light beams. Recent research involving a special class of light beam—one that does not diffract—is opening up intriguing new avenues of research.

Optical tweezers were first demonstrated by researchers at Bell Labs (Murray Hill, NJ) in 1986.¹ The basic principles can be explained in terms of Newton's laws. Because light carries a momentum, changing the direction of the light means that there must be a force associated with that change. So if we take a laser and shine it through a small particle, the light will be refracted as it moves through the particle. The force involved in this change of direction acts on the particle such that the particle moves toward the most intense part of the laser beam.

A laser beam has a Gaussian profile, so the most intense part of the beam lies at the center of the beam axis. The force, therefore, confines the particle to the beam axis, and as the focus of the beam is the most intense part of the beam along the beam propagation direction, it draws the particle toward the focus. The particle becomes trapped in three dimensions. To create large enough forces to achieve this 3-D trapping effect, we do not need very much power (typically a minimum of a few milliwatts). We do, however, need high-intensity gradients, and so we focus the beam down to a spot only a few microns in diameter.

Trapping particles in this way enables a wide variety of studies. It is possible to measure the elastic properties of DNA by grabbing hold of beads attached to the ends of the molecules and stretching them. Similarly, the force-producing properties of molecular motors, such as kinesin, may be studied with optical tweezers, as can the unfolding of proteins. Particles are easier to probe if they are trapped. Optical tweezers can be combined with Raman spectroscopy, two-photon spectroscopy, and confocal microscopy, among other tools.

By combining optical tweezers with other laser beams, researchers can perform microsurgery on particles. For instance, they can grab chromosomes and then cut them into small pieces for further analysis using an IR (1064 nm) trapping beam and a green (532 nm) cutting laser, known as optical scissors. This is possible because most biological matter does not absorb strongly in the IR spectral region but does at green wavelengths.

Optical tweezers constitute a powerful method with great versatility, but they have some limitations. One of these exists because we use a tightly focused beam to act as the optical trap. The more tightly we focus a laser, the more rapidly it diverges. This means that the force exerted by the beam drops off very quickly as we move away from the trapping region. It drops off so quickly, in fact, that by the time we are a few tens of microns away from the beam focus, the forces are insufficient to trap particles anymore. A single-beam trap is, to all intents and purposes, a single-particle trap and is certainly only really useful near the focal region.

Another limitation is that if you place an object in the path of a laser beam, the beam behind the object will bear little resemblance to the original due to effects such as diffraction, refraction, and reflection/absorption. This again restricts the distance over which a Gaussian beam can act as optical tweezers.

The divergence that a tightly focused beam suffers is caused by diffraction. The smaller the aperture encountered by a beam, the more widely the beam spreads. In the case of a tightly focused beam, the focused spot essentially acts like an aperture for the beam.

It is a fact of physics that all light beams suffer from diffraction; it was not thought that much could be done about it. Then in 1987, Durnin, Miceli, and Eberly showed a class of light beam exists that actually remains diffraction-free.² The amplitude of such beams is proportional to a Bessel function, and so we call them Bessel beams. A Bessel beam looks like a bright spot surrounded by a set of concentric rings. Unfortunately, for the beam to remain diffraction-free over an infinite distance, it would have to have an infinite number of rings. That would require the beam to carry an infinite amount of power, which is clearly impossible. In the lab, therefore, we can only realize an approximation, a quasi-Bessel beam.

Figure 1. In optical tweezers, the focused beam moves the sphere to the most intense part of the beam. The Gaussian beam diverges away from the focus.

A very efficient method for making a Bessel beam is to illuminate a conical lens—an axicon—with a Gaussian beam (see figure 1). Other methods include the use of holograms or spatial light modulators. The modulators can act like dynamic, computer-configurable holograms, thus offering the prospect of generating more advanced light patterns such as arrays of Bessel beams.

Figure 2. An input Gaussian beam focused by the axicon forms a Bessel beam over the propagation distance Z_max. Plane waves (dotted lines) form a cone, and the interference of the beams on the axis of the cone forms the central, non-diffracting spot of the Bessel beam (image).

A Bessel beam may be thought of as the interference of a set of plane waves traveling along a cone, with the central maximum of the Bessel beam formed by the interference of these waves on the optical axis. This analogy allows us to understand one of the other remarkable properties of the Bessel beam—that of reconstruction. Unlike a Gaussian beam, which is distorted after encountering a particle, a Bessel beam is able to reconstruct itself around an object (see figure 2). Because the Bessel beam consists of interfering waves traveling on a cone, it is able to re-form itself after encountering an obstruction, so long as some of the waves are able to move past the obstruction; the new center of the beam consists of the interfering waves that move past the outside of the obstruction. This property allows us to overcome the limitation in Gaussian-beam optical tweezers of only being able to trap particles that are very close together.

In our recent work, we showed that a Bessel beam can be used to tweeze particles that were both aligned and 3 mm apart and that resided in completely distinct sample cells.^3,4 This was possible for two reasons: first because the Bessel beam is diffraction-free—its central core looks like a rod of light whose intensity remains constant—and second, because the beam is able to re-form itself after encountering an obstacle.

Figure 3. In the experimental setup for Bessel beam optical tweezers, the 1 W, 1064-nm beam from an Nd:YVO₄ laser is passed through an axicon and telescoped back down before passing through the sample stage. We capture the data profiles and their cross sections (inset) using a 100X microscope objective and a CCD camera.

We used a single neodymium-doped vanadate (Nd:YVO₄) laser at 1064 nm to create a Bessel beam with 19 rings and a central spot (see figure 3). Total power was 700 mW, giving approximately 35 mW as the central spot. We trapped a hollow sphere approximately 5 µm in diameter between the central spot and the first ring of the Bessel beam. The sphere distorted the beam, which re-formed after a short distance and was able to stack three 5-µm-diameter solid silica spheres. Passing those, the beam re-formed once more (see figure 4).

Figure 4. The system can manipulate particles that have large spatial separations, such as these cells 3 mm apart and 100 mm deep. We observe the beam a short distance above (a). Here the beam has been distorted by the particle. Some small distance above the first sample cell the beam has reformed (b,c) and is no longer distorted. The beam enters the second cell and is able to stack three 5-µm-diameter solid silica spheres (d). We can observe the beam profile above the stack of particles (e, f).  (J. ARLT, D. P. RHODES)

picking up the pieces

Another feature of the reconstructing Bessel beam tweezers is that they are able to capture a variety of particles not easily trapped by a single Gaussian beam. For instance, we have trapped a solid silica sphere in one cell, a hollow sphere in a second cell, and a birefringent particle in a third. Hollow spheres have a lower refractive index than the surrounding water in the sample cell and so are repelled from regions of high light intensity. We can trap them with the Bessel beam using the dark regions between the bright rings. We are also able to create sets of 2-D arrays with the Bessel beam, with each array lying in a different plane. The particles in the sample cell fill the rings of the Bessel beam but do not completely block it. Therefore, the beam is able to re-form itself and can then trap another array of spheres.

Applications for optical tweezers and optical manipulation are varied and include biological applications (such as cell sorting, cell dissection, cell manipulation, and studies of molecular motors) and colloid physics (such as particle interaction measurements).⁵ In microfluidics, the tool can optically control gates and pumps. In nanotechnology, it can drive nano- and micromachines and create 3-D structures. Optical tweezers can also measure spin and orbital angular momentum measurements in studies of the properties of light beams.

The Bessel beam tweezers may allow some of these techniques to be enhanced. For instance, we can drive single microcogs using optical tweezers, but if we wanted to drive two microcogs connected by an axle, we would have more trouble. Using reconstructing Bessel beams, we can use the beam to drive one of the cogs and force it to re-form to drive the second cog, thus increasing the efficiency of the micromachine.

We also foresee applications using arrays of similarly prepared samples, perhaps lab-on-a-chip devices. A Gaussian beam can drive a single micropump. With a Bessel beam we should be able to drive a whole stack of aligned micropumps simultaneously. Using a control beam able to interact with each pump in the stack will reduce complexity and allow more efficient automation of such devices. In colloid physics, Bessel beams may be useful for studying arrays of particles, including simultaneously looking at multiple sample cells that each have slightly different initial conditions.

Optical tweezing is entering a new phase, with maturation of existing techniques and with the creation of sophisticated light patterns becoming easier thanks to new technologies such as spatial light modulators.⁶ Whole new capabilities are being brought to the subject, allowing flexible, dynamic control of beam shapes and positions. In addition, we envisage that optical tweezing will make a significant contribution in many interdisciplinary subjects over the coming years. oe

References

1. A. Ashkin, J.M. Dziedzic, et al., Opt. Lett. 11, 288 (1986).

2. J. Durnin, J. Miceli, et al., Phys. Rev. Lett. 58, 1499 (1987).

3. J. Arlt, V. Garcés-Chávez, et al., Optics Communications 197, 239-245 (2001).

4. V. Garcés-Chávez, D. McGloin, et al., Nature 419, 145 (2002).

5. For a respresentative list of research groups using optical tweezers, go to www.st-andrews.ac.uk/~atomtrap/Resources.htm .

6. Dholakia, G. Spalding, et al., Physics World 15, 31 (2002).