Lights,
Action: Optical Tweezers
Justin E
Molloy (1) and Miles J Padgett (2)
(1)
Department of Biology, The University of York, York YO1 5YW
(2) Department
of Physics and Astronomy, University of Glasgow, Glasgow G12 8QQ
Abstract
Optical tweezers were first
realised 15 years ago by Arthur Ashkin and co-workers at the Bell Telephone
Laboratories. Since that time
there has been a steady stream of developments and applications, particularly
in the biological field. In the
last five years the flow of work using optical tweezers has increased
significantly, and it seems as if they are set to become a mainstream tool
within biological and nanotechnological fields. In this article we seek to explain the underpinning
mechanism behind optical tweezers, review the main applications of optical
tweezers to date, present some recent technological advances and speculate on
future applications both within biological and non-biological fields.
1. Introduction
It is fifteen years since Ashkin et al. [[1]]
published their seminal paper “Observation of a single-beam gradient force
optical trap for dielectric particles”.
The technique is now referred to as “optical tweezers” or “optical
trapping” and their original paper has received 400 citations – half of which
during the last five years. In
essence, optical tweezers rely upon the extremely high gradient in the electric
field produced near the beam waist of a tightly focussed laser beam, which
creates a force sufficient to trap micron-sized dielectric particles in three
dimensions. Commercial tweezers
systems are now available [[2]],
and although originally devised by physicists, it is mainly biologists who have
put optical tweezers to use.
However, technology does not stand still and tweezing techniques are
presently undergoing a further spate of development. The future of this cross-disciplinary field is bright.
2. The mechanism behind optical tweezers
Thirty years ago, Ashkin started
experimenting with optical traps [[3]]
he realised that an unfocussed laser beam would draw objects of high refractive
index towards the centre of the beam and propel them in the direction of
propagation. An arrangement of two
counter-propagating beams allowed objects to be trapped in three-dimensions. In these experiments he was able observe
the effects of radiation pressure and overcome the usually much larger
radiometric (heating) effects of light by using relatively transparent objects
in a relatively transparent medium.
He later discovered that a single, tightly focussed, laser beam could be
used to capture small dielectric particles in three-dimensions. This technique enables small particles
to be picked up and moved at will using a beam of visible light; hence was
christened optical tweezers. The
forces acting within optical tweezers can be understood either in terms of
light momentum and ray optics or the force associated with the optical electric
field. As every photon carries a momentum 
it is
straightforward to state that, if absorbed by an object, momentum will be
transferred from the light to the object.
For
particles smaller wavelength of the laser beam, the ray optical approach is
less satisfactory and it is better to consider the forces in terms of the
electric field near the trapped particle.
Forces can then be divided into those arising from scattering of the
light and those arising from an intensity gradient. Under the correct
conditions, gradient forces dominate and bring about optical trapping.
A standard
tweezers configuration: Although
optical tweezers can be commercially obtained they are also surprisingly simple
to assemble by anyone with an expertise in building optical systems. Given that many of the components are
in common with a high magnification microscope, such instruments represent a
good starting point for custom-built optical tweezers.
Obtaining
a tightly focussed beam: Most optical tweezers use diode-pumped YAG lasers;
chosen for their good beam pointing stability and low absorption of the near
infra-red laser line (1064nm) by biological materials. To produce a diffraction-limited focal
spot, a high-magnification microscope objective lens with a high numerical
aperture is used. For many applications it is useful to be able to steer the
beam around, thereby manipulating the trapped object. Clearly beam steering of this kind can be automated and used
in conjunction with an automated sample stage, enabling both absolute and
relative movement of the trapped object with respect to its surroundings.
Multiple
tweezers: Optical tweezers have also been configured using multiple
beams to simultaneously trap more than one particle. These have been implemented by the rapid scanning of a single beam
between two or more trap positions [[4],
[5]],
splitting the beam early in the optical circuit to produce multiple, separate,
light paths which are later recombined before entering the microscope [[6]]
and using computer-generated holograms to give multiple beams simultaneously [[7]]. Such multiple beam traps allow
independent positioning of different objects, or parts thereof.
Scanning: If the
laser beam is moved with sufficient speed then multiple objects can be held and
manipulated by time-sharing a single beam [8]. Multiple optical tweezers can be created in this way because
viscous drag on the trapped objects is sufficiently high to provide positional
“persistence” while the laser beam is elsewhere, servicing another object.
Multibeam
approach: A conceptually simpler but perhaps less versatile method is
to produce dual optical tweezers by dividing the laser beam into two separate
paths [10].
Holograms:
Computer generated holograms are used widely to convert the fundamental
Gaussian mode emitted from most commercial lasers into a beam with a different
intensity and phase structure [[8],
[9]]. In its simplest form, a
computer-generated hologram is produced from the calculated interference
pattern that results when the desired beam intersects the illuminating laser
beam. Although such holograms can be fabricated to produce any beam or
combinations thereof, recent advances in display technology and computing power
means that holographic patterns can be calculated in real time and transferred
to spatial light modulators to give a dynamically adjustable holographic
element. Excitingly such a technique seems to combine the simplicity of rapid
scanning using modulators with the simultaneous trapping of the multi-beam
approach
3. Biological applications of optical tweezers
Biologists were quick to take
advantage of optical tweezers as a tool for purposes such as measuring the
compliance of bacterial tails [[10]],
the forces exerted by single motor proteins kinesin, myosin and
RNApolymerase stretching of single DNA molecules [[11]]. In order to convert optical
tweezers into an instrument for measuring small forces a sensor capable of
measuring the position of the trapped objects with nanometre precision is
required. Because proteins and DNA are so small, modern studies use plastic
microspheres as handles to manipulate the material under test indirectly. The
position of such uniform microspheres, held in optical tweezers, can be
determined using a four-quadrant photosensor to measure the “centre-of-gravity”
of the object, either using conventional imaging (bright field, dark field, or
phase contrast) or by interferometry.
One of the
best-studied, and most interesting problems in biophysics is study of molecular motors, proteins that take chemical energy and convert this to
mechanical work. They perform a multitude of functions that are essential to
life; from RNA transcription and protein synthesis to cell division, vesicle
trafficking, cell locomotion, endocytosis and of course, the best known
example, muscle contraction.
Rotary
motors: Berry and Berg [[12]]
recently made a detailed study of the, proton powered, bacterial rotary motor
mechanism by measuring the force produced by both its forward and reverse
rotation. They used a plastic bead
held in an optical tweezers as a handle to push against the bacterium and
rotate it around its fixed flagellum.
They found that, the torque required to push it backwards is only a
little more than that which stops it.
This indicates either that the motor can slip, or that the mechanism is
reversible and the transition from motor to pump (dynamo) is seamless. Such studies give great insights into
the molecular mechanism of this biological electric motor.
Kinesin: A landmark
in the study of single motor proteins came in 1993 when Svoboda, Block and
others [[13]]
measured the individual steps taken by the molecular motor, kinesin, as it
walked along a fixed microtubule track. The crucial observation made, and
indeed a testament to the incredible sensitivity of their method, was that they
could identify discrete, 8nm, steps taken by the kinesin molecule. Rather than
moving smoothly like an ensemble would do, the single molecule moved in a
stochastic, jerky fashion. The excitement of these measurements should not be
forgotten! Molecular scale motions
powered by just 1 ATP molecule (equivalent to less than 1/10th the
energy of a single photon) were being observed in real-time, with no signal
averaging required
Myosin: In 1994,
Finer, Spudich and Simmons [17] devised a way to record
the activity of single myosins. Their method required a dual optical tweezers
system. Using the twin tweezers, to hold a single actin filament suspended
between two plastic microspheres the filament could then be placed in the
vicinity of a third, fixed, microsphere that had been sparsely coated with
myosin. The resulting movements were
measured from the position of one of the trapped beads. A year later, Molloy et al. [24] performed a similar
study but they realised that because the intermittent movements produced were
so small (only about 5nm) the starting position for each interaction observed
was being randomised by the thermal vibrations of the beads in the
tweezers. By monitoring the variance of the position signal they
were able to measure off many hundreds of myosin binding interactions and then
later deconvolve the myosin induced motion from the background thermal vibrations.
RNApolymerase: With the
ability to manipulate DNA and a proven track record of working with motor
proteins like kinesin, Block and colleagues [[14]]
set claim to another first when they measured the force produced by RNA
polymerase as it transcribed a DNA gene.
Using optical tweezers, they attached a plastic bead to one end of a DNA
strand and allowed the other end to bind to an immobilised molecule of RNA
polymerase. The bead then measured the pulling force produced by the
RNApolymerase. The force was about
5 times larger (25pN) than that produced by either kinesin or myosin.
Protein
unfolding: Tskhovrebova, Sleep and other [[15]]
used optical tweezers to pull on a giant, structural, protein called titin.
They attached one end of titin to a plastic bead (the optical tweezers handle)
and the other to a microscope coverslip.
When they pulled on the protein with a force of about 100pN individual
protein domains unfolded catastrophically to relax the tension in a step-wise
fashion. Unfolding events could be
quantified in terms of their lifetimes and by the extension produced as the
tertiay folded structure unwound to give a linear chain of amino acids. The excitement of these measurements is
that in the future it may be possible to follow the folding and unfolding of
different amino acid sequences and so deduce the energy profile of the folding
process. Such studies will give
great insight into the protein folding mechanism and it is crucial that we
understand this fully, if we are to make sense of the new genomic data.
4. Alternative tweezers
configurations
Clearly the biological
applications of optical tweezers have been wide ranging, however, optical
tweezers themselves have been undergoing interesting developments in a number
of areas.
Rotation
within optical tweezers (optical spanners): Physicists have used
optical tweezers as a tool to study the transfer of angular momentum from light
to particles. For particles
trapped on the beam axis, both the spin and orbital angular momentum of a light
beam have been shown to cause rotation of birefringent [[16]]
and absorbing [[17], [18]] particles, respectively. For practical
applications, rotation within optical tweezers has a history dating back to
1991 when Sato used the gradient forces associated with a rotating high-order
Hermite-Gaussian mode to induce the rotation of red blood cells [[19]]
(figure 9c).
Perhaps
the most obvious way to generate an asymmetric beam is to introduce an aperture
into the optical path (figure 9e).
Mounting the aperture within a rotation stage that is itself mounted on
an x-y translation stage means that its rotation axis is easily aligned to the
beam axis. This simple method for
rotational control does not require high-order modes, interferometric precision
or computer-controlled optical modulators and can be easily applied to, or
removed from, existing tweezers apparatus.
Interferometric
optical tweezers: One new approach to optical tweezers was reported in
1997 when Choui et al. [[20]]
used two optical beams to create an interference pattern, the intensity profile
within which gave rise to a gradient force confining both spherical and
“rod-like” particles. Changing the
path difference between the beams caused a translation of the interference pattern
and the particles trapped within. Most recently this interference technique has
been applied to the simultaneous trapping of both low and high index particles
[[21]],
where the high-index and low-index particles are confined to the regions of
high and low intensity respectively.
Optical
tweezers with high-order Gaussian modes and other beams: High-order,
Hermitte-Gaussian modes can be used to force the specific
alignment of an asymmetric object within the optical trap. The annular nature of these beams also
enables objects of low refractive index to be trapped. A further recent
development in optical tweezers is the use of Bessel light
beams. Over a limited range, the
central region of the beam propagates without changing shape, creating an
intensity distribution which has no gradient in the propagation direction. This year, Arlt et al. [[22]]
used Bessel beams to laterally trap objects in two-dimensions and most
excitingly to obtain stacking and guiding of the trapped objects along the
bright central core of the beam, see figure 12. In this case the Bessel beam was created from a standard
Gaussian beam using a specially fabricated glass axicon, although computer
generated holograms could also be used for the same purpose.
Optical
tweezers using diode lasers: If optical tweezers are to become
a routine tool, the transition to diode lasers is clearly attractive. The first use of a laser diode within
optical tweezers was reported in 1991 [[23]]. In that work a 1300nm InGaAsP diode was
shown to trap both silica spheres and yeast cells with an efficiency comparable
to a conventional laser. One
advantage offered by diodes is that most biological samples are highly
transparent in the infra-red and the absence of light absorption and the
corresponding heating reduces the risk of inducing cell death (“opticution”).
Optical
tweezers for metal particles: Metallic particles scatter,
reflect and absorb much more strongly than the transparent particles normally
held within optical tweezers.
Consequently, the balance between the gradient and scattering force is different,
making the trapping of metal partilces much more difficult. However several method have been
developed. Rayleigh particles (10-50nm diameter) have been trapped in three
dimensions as in this size regime scattering from metal and dielectric
particles is similar [[24]]. Larger Mie particles have been trapped
in two dimensions, the forces arising from a creeping wave induced by a surface
plasmon [[25]]. Metal particles have also been confined
by annular rings [[26]]
or rotating beams [[27]],
which rely on light scattering to produce a repulsive force and trap the
particle in two dimensions.
Three-dimensional confinement has also been achieved to an annular
region around the outside of a Laguerre-Gaussian mode [[28]];
when the trapping beam is directed upwards the scattering force providing both
vertical and radial restoring forces.
5. Possible future uses of optical tweezers
Regarding the technical
development of optical tweezers it seems as if the ongoing development of
spatial light modulators will have a significant impact. Not only will these allow the building
of simple multi beam tweezers without problems associated with limited dwell
times but they will also replace the static holograms currently used for
alternative beam generation. In
the latter case, a real time addressable light modulator will allow the user to
switch between different trapping beams allowing independent control of the
axial and lateral trapping force.
Physics/engineering: Optical
forces associated with scattering, the gradient force, or angular momentum
transfer have been postulated as a means to driving micro-machines and so far
the scattering force in particular has looked promising. Another area in which optical tweezers
have not been exploited is in the assembly of the micro-machines themselves, or
microstructures such as photonic crystals.
Chemistry: Optical
tweezers have only recently been applied to colloid chemistry and clearly this
is an area of great future potential. Controlled chemical reactions, single
molecule assays, femto-litre stopped- and quench-flow studies should all be
possible with high time resolution.
Biological:
applications of optical tweezers are moving to a new level. Biologists thirst for more detailed
mechanical and biochemical information on how single molecules work. Recent advances by the Yanagida group
in which optical tweezers have been combined with single-molecule fluorescence
imaging is a very exciting development [[29]]. However, they also need advances in the
tweezers design itself. They now
want to grapple with single biological molecules and sense the vibrations of
individual domains and side chains with high time and high spatial
resolution. The desire for higher
resolution data throws down the gauntlet to physicists to devise new breeds of
optical tweezers that are sharper and more dextrous than those in use today.
References