WO2012033409A1

WO2012033409A1 - Optical trap, chip, sensor system and method for manufacturing an optical trap

Info

Publication number: WO2012033409A1
Application number: PCT/NL2011/050619
Authority: WO
Inventors: Matheus Maria Van Leest; Felipe Andrés Bernal ARANGO; Jaap Caro
Original assignee: Stichting Wetsus Centre Of Excellence For Sustainable Water Technology
Priority date: 2010-09-09
Filing date: 2011-09-09
Publication date: 2012-03-15
Also published as: NL2006087C2

Abstract

The invention relates to a optical trap (2), chip and sensor system comprising the optical trap (2) and chip and method for manufacturing thereof. The optical trap (2) comprises: -a substrate; and -at least two waveguides (4, 6) deposited on the substrate by a microfabrication process, each waveguide (4, 6) having an exiting surface (12, 14) from which a beam can exit the waveguide (4, 6) when in use, wherein the exiting surfaces (12, 14) of the waveguides (4, 6) face each other and are separated by a distance (L).

Description

OPTICAL TRAP , CHIP , SENSOR SYSTEM AND METHOD FOR

MANUFACTURING AN OPTICAL TRAP The present invention relates to an optical trap.

Optical traps use light to trap a particle or biological object, such as a cell.

Optical traps have a number of applications. For example, they can be used to trap a particle for performing spectroscopy measurement, such as Raman spectroscopy or infrared spectroscopy.

Ashkin, "acceleration and trapping of particles by radiation pressure", Physical review letters, 24, 156

(1970), discloses an experimental setup for trapping of particles using light. Ashkin uses a dual beam to trap a particle. The scattering force acts on the particle in the axial direction of the beam and the gradient force in the transverse direction. Although this experimental setup is an important proof-of-principle, it does not lend itself to high volume production of a lab-on-chip technology, which is a requirement for widespread application in areas such as water-quality monitoring and medical diagnostics.

Furthermore, the use of lenses makes the setup relatively complicated .

The state of the art comprises several alternative experimental setups which share the same disadvantage as described above: these are all experimental designs which do not lend themselves for widespread application.

An object of the invention is to overcome or reduce these disadvantages and to provide an optical trap for effective and efficient industrial and/or other application.

This object is achieved with the optical trap according to the invention, the optical trap comprising - a substrate; and

- at least two waveguides deposited on the substrate by a microfabrication process, each waveguide having an exiting surface from which a beam can exit the

waveguide when in use,

wherein the exiting surfaces of the waveguides face each other and are separated by a distance.

The substrate acts as a support for the waveguides and can for example consist of a glass plate or a silicon waver.

Microfabrication processes, also known as

nanofabrication processes, comprise processes for the fabrication of small structures, typically micrometer size or smaller. These processes are commonly used for the fabrication of integrated circuits and microfluidic devices, amongst others. For example, microfabrication comprises thin film deposition, patterning, such as photolithography, etching and/or doping.

In use, two counterpropagating beams are launched into the gap between the waveguide exiting surfaces. Preferably, the waveguides are oriented substantially collinear. The gapsize, i.e. the distance between the at least two

waveguides, is preferably in the range of 0-100 μιτι, more preferably 0-20 μιη and most preferably about 5 μιη.

The fact that the at least two waveguides are deposited by a microfabrication process has several advantages.

Firstly, it enables the fabrication of an integrated design of waveguides and a lab-on-chip. The fabrication is

compatible with standard glass-based microfluidic

technology, ensuring reproducibility of the fabrication process and robustness of the devices.

A second advantage is that microfabrication of the waveguides enables optical traps of smaller size than conventional optical traps, which for instance include lenses .

By depositing the waveguides using a microfabrication process the optical trap is cheap, robust and the different components can be integrated in a precise and accurate manner. Furthermore, a large scale production is possible.

A further advantage is that the optical trap does not necessarily comprise several mechanical elements which have to be combined. In the state of the art the use of fibers and lenses is known. However, these have the disadvantage that they cannot easily be integrated into an optical trap.

Furthermore, the adhesion of the waveguide is stronger and more robust than devices which use waveguides which are deposited in a different manner.

In a preferred embodiment according to the invention, at least one waveguide is a solid core waveguide.

Solid core waveguides can be produced relatively easy by a microfabrication process. These waveguides are less fragile and less sensitive to pollution.

The waveguides can comprise any suitable material, i.e. any material which can be deposited by a microfabrication process and is suitable for operating as (part of) a

waveguide. For example, a semiconductor compound can be used as waveguide material.

In a preferred embodiment of the present invention, at least one waveguide comprises a silicon compound.

Many silicon compounds are suitable for deposition on a substrate by a microfabrication process. Furthermore, these compounds often have suitable waveguiding capabilities.

In a further preferred embodiment of the present invention, at least one waveguide comprises S1₃N₄ and/or

SiON. S1₃N₄, also known as silicon nitride, has a good index contrast with S1O₂ (n S13N4 = 1.9, nS102 = 1.45), which can be used as a cladding material for the waveguides. Other cladding materials are also possible.

Furthermore, silicon nitride has the advantage that it is suitable for operation both at the near-visible and visible wavelength used in Raman spectroscopy.

Another advantage of silicon nitride is that it is compatible with glass-based micro fluidic technology.

S1₃N₄ can be deposited on the substrate using the aforementioned microfabrication process. In another example, the waveguide comprises a core of S1O₂ and a S1₃N₄ cladding.

This type of waveguide provides a beam profile which

conforms with the requirements of optical traps.

Alternatively, the waveguide comprises SiON, also known as silicon oxynitride, or a combination of silicon nitride and silicon oxynitride.

In a preferred embodiment according to the invention a fluid channel is provided which passes between the exiting surfaces of the waveguides.

The fluid channel is arranged to pass through the gap provided by the distance between the waveguides. In use, the fluid channel acts as a transportation channel of particles which are comprised in the fluid. The fluid can be a gas or a liquid, for example. Preferably, a liquid is provided in the fluid channel in use.

By providing a fluid channel, the optical trap can be fed with a liquid or gas containing particles, such as biological cells. By trapping these particles, the optical trap enables studying of the particles, such as by means of

Raman spectroscopy or infrared spectroscopy.

The fluid channel enables the use of the optical trap as a part of a sensor system with which an on-line and/or at-line measurement can be performed by taking samples from a fluid. Furthermore, the optical trap can perform a filter function due to the fluid channel, wherein particles from a fluid in the channel are retained by the optical trap.

Preferably, the fluid channel is formed by a micro fabrication process. This enables producing both the fluid channel and the waveguides using microfabrication and thereby simplifies production of the optical trap.

In a preferred embodiment according to the invention, the fluid channel comprises a trench of which the walls are in substantially the same plane as the exiting surfaces of the waveguides.

This enables a less complex production by micro

fabrication than when using, for instance, a hollow

waveguide as a fluid channel as in conventional systems. For example, a trench is created by an etching process.

In a further preferred embodiment, at least a part of the walls of the trench form a cladding for the waveguides.

This further simplifies the construction of the optical trap according to the invention, since the cladding of the waveguides and the walls of the trench can be formed in a single micro fabrication step.

For example, the fluid channel comprises a trench formed in a Si02 layer which also serves a cladding for the waveguides. Other materials may be used as well.

In a preferred embodiment of the invention, the optical trap comprises an optical source.

The optical source can provide the beam for the optical trapping of particles by the trap. Furthermore, the optical source can enable probing a trapped particle, for instance through spectroscopy. Preferably the optical source is a laser. The laser can be used as a excitation source for performing Raman

spectroscopy .

This enables trapping and excitation of particles by the same waveguides. No additional waveguides have to be provided for excitation. This leads to a simple device which is advantageous for mass production.

In a further preferred embodiment, the optical trap comprises means for controlling the relative phase of the beams exiting the at least two waveguides in use.

By enabling controlling the relative phase of the beams of the trap, the trapped object can be moved in the

longitudinal direction of the beams. This will be further explained below, in respect to simulations. The direction in which the trapped object can be moved is the X-direction as indicated in figure 1.

The waveguides are coupled with an optical source, for example a laser, and the relative phase between the beams exiting the waveguides are controlled using said means.

It is noted that a further advantage of the optical trap according to the invention is that it enables a compact embodiment. This is for instance enabled by the formation of the channel and waveguides in a single micro fabrication step and using the waveguide for both trapping and

excitation. For example, chips incorporating the optical trap of the invention therefore have sufficient space to fit additional components.

The invention further relates to a chip and a sensor system comprising the optical trap as described above.

The same effects and advantages apply in respect to the chip and the sensor system as those described in respect to the optical trap. The invention further relates to a method for

manufacturing an optical trap, comprising the steps of:

- providing a substrate; and

- depositing at least two waveguides on the substrate

such that the waveguides each have an exiting surface and the exiting surfaces face each other, wherein depositing comprises a microfabrication process.

The same effects and advantages as described for the optical trap above apply to the method according to the invention.

Preferably the depositing of the waveguides on the substrate comprises thin film deposition. More preferably vapour deposition is used, for example chemical vapour deposition or physical vapour deposition.

Preferably the material deposited on the substrate for use as a waveguide is silicon nitride and/or silicon

oxynitride .

Preferably silicon oxide is used as a cladding material for at least one waveguide.

Preferably the method according to the invention is used to manufacture an optical trap, chip or sensor system as described above.

Further advantages, features and details of the

invention are elucidated using the accompanying drawings, wherein

- Figure 1 shows the geometry of an optical trap

according to the invention as used for the simulations described below;

- Figure 2 shows the simulated intensity distribution of the electrical field in the xy-plane for the 1 μη² cross-section square waveguide (a) , and transverse (b) and axial (c) intensity profile for the waveguide. In figure 2c, A indicates the facet position; - Figure 3 shows the averaged intensity distribution for the dual-beam trap obtained from the simulation based on 1 μη² cross-section square waveguides, for the yz- plane without bead (a) , xy-plane without bead (b) , yz- plane with a 1 μιη diameter polystyrene bead centered at point (0, 1, 0) (c) and the xy-plane with a 1 μιη diameter polystyrene bead centered at point (0, 1, 0) (d) ;

- Figure 4 shows the calculated force component as a

function of bead position on the x- , y- , and z-axis, respectively (a-c) , for several bead diameters. The insets in (b, c) are blow ups of the force curves for positive bead positions for the 0.2 and 0.5 μιη diameter beads. The curves in (a-c) are piecewise cubic spline interpolations between the data points; and

- Figure 5 shows the trapping potentials for the x- direction (a) and y-direction (b) , as deduced from the force curves in Fig. 4a and 4b, respectively; and

- Figures 6A-C show the optofluidic design of a dual- waveguide trap according to the invention, wherein Fig.

6A shows an overview of the trap, Fig. 6B shows a close-up of a detail of 6A and Fig. 6C shows even more detail . Simulations

To establish the specifications of the dual-waveguide optical trap according to the invention we perform finite- difference time-domain (FDTD) simulations of the profile of the free-space beam exiting the facet of a single square waveguide, in dependence of the waveguide transverse

dimensions. As a further step we perform simulations of the trapping forces for the dual waveguide situations, and from that obtain the trapping potential. All simulations are carried out for the often used Raman excitation wavelength λ=785 nm, so that a realization of the device will not only trap the (biological) object under consideration, but at the same time will generate a Raman spectrum measurable with a standard spectrometer. We envision that our design will be very suitable for trapping and Raman identification of bacteria in water. It is noted that different wavelengths may be used as well. Preferably a wavelength in the range of 700-1400 nm is used, more preferably in the range of 700- 1000 nm, more preferably 700-800 nm and most preferably 700- 850 nm.

Approach

Optical trap 2 comprises two collinear truncated Si₃N₄ waveguides 4, 6 of thickness t and width w, from which two counter-propagating beams 8, 10 are launched into the gap L between the waveguide facets 12, 14 (Fig. 1) . Si₃N₄ is chosen as waveguide material since it is suitable for operation both in the near-visible and visible wavelength ranges used in Raman spectroscopy and in view of its compatibility with glass-based microfluidic technology or CMOS-compatibility. At the same time Si 3N4 has a g^ood index contrast with SiO2

(n Si3N4=1.9 versus nSi02=1.45), which is used as cladding^ material 16, 18 for the waveguides, as depicted in Fig. 1.

The Si0₂ layer has a trench of which the walls are in the same vertical plane as the facets, thus defining a fluidic channel 20. This is illustrated by grid 22. In practice, the floor and ceiling of channel 20 will be the substrate and a sealing layer (not shown) bonded to the Si0₂ cladding 16, 18, each positioned at such a distance that the propagation of the beams and the formation of the trap are not hindered. The width of the fluidic channel 20 and thus the gap between the facets is chosen to be 5 μιη. This size is compatible with various types of biological objects, for instance bacteria. Further, a 5 μιη gap enables a

straightforward choice of the transverse dimensions of the waveguide on the basis of the preferred type of intensity distribution of the beam emitted from the waveguide, as discussed below. It is noted that proper trap operation is not limited to this gap size.

Waveguide transverse dimensions

Many biological objects studied in optical trapping have a close to circular shape, which suggests a trapping potential of high symmetry. Therefore, we adopt a square cross section for the waveguide, of which the side is further denoted by w.

Since analytical solutions for the mode structure of this type of waveguide and the connected beam profile launched into free space do not exist, we simulate the behavior with the three-dimensional finite-difference time- domain (FDTD) method using the freely available software package Meep.

Simulation

We choose w = Ιμιη. A S1₃N₄ waveguide of this size imposes modest requirements on the fabrication technology, while it can accommodate up to 8 modes per polarization direction for the operation wavelength A=785 nm.

In this simulation the dipole source exciting

continuous waves in the waveguide is polarized parallel to the y-direction, thus populating modes with field components E_y and H_z. The optical power delivered by the dipole source to the waveguide is 0.5 W. For the medium in the fluidic channel we take water, using n_H2o=l-33. The size of the calculation cell in the x- , y- and z- direction is 17 μιτι, 11 μιη and 11 μιτι, respectively.

Perfectly matched layers (PMLs) of 2 μm thickness are included to prevent artificial reflections from the

calculation cell boundaries. Simulation results are

presented for a vacuum wavelength of λ=785 nm, a standard wavelength in Raman spectroscopy.

In Fig. 2a-c we show three types of plots characterizing the simulated beam exiting a single waveguide 6 (for which we take the left one in Fig. 1, thus defining position and orientation with respect to the coordinate system) . Figure 2 shows simulated intensity distributions of the electrical field in the xy-plane for the 1 μη² cross-section square waveguide (a) , and transverse (b) and axial (c) intensity profile for this waveguide. In (a) the normalization is such that the full grey scale is used. In this simulation the source-facet distance is 4 μιη. In Fig. 2a the grey scale represents the intensity I of the electric field E(I∞E²) in the xy-plane, obtained by averaging the intensity

distribution over the last four periods of the simulation

(=4x785 nm/c; c is the speed of light in vacuum.), where the solution has converged. This averaging removes the half- wavelength-period modulation of the intensity distribution in the water region, as would be present in a snapshot of the propagating beam. Figs. 2b and 2c show the averaged transverse intensity profile on the y-axis (axis through the trap centre and parallel to the fluidic channel) and the averaged intensity profile on the x-axis (axis of the waveguide) , respectively.

The most salient feature in Fig. 2a is the spreading of the beam, indicative of diffraction of modes at the

wavelength-size facet of the waveguide. For the dual-beam optical trap this spreading leads to a non-zero scattering force, which drives a particle to the centre of the trap in case of an axial displacement away from the centre. Further, beam divergence helps to generate a wide region where the gradient force opposes radial displacements, thus enabling trapping of larger objects. The intensity pattern inside the waveguide shows a short period modulation on top of a wide maximum, as seen more clearly in Fig. 2c. The former one is an interference effect resulting from superposition of the incoming wave from the source and the wave partially

reflected at the waveguide facet. The wide maximum is part of the pattern resulting from the interference of multiple modes. This multimode interference (MMI) occurs as a result of excitation of several modes in the waveguide, which arises from the character of the source we use. It has a homogeneous current distribution across its plane and therefore is not tailored to selectively excite the zeroth order mode. For the transverse profile and the axial

profile, a bell-shaped curve similar to a Gaussian and a decaying curve are observed, respectively, as expected for a spreading beam. Before it decays, in the axial direction the beam shows a maximum, which is the continuation of the MMI pattern inside the waveguide. The detailed shape of the transverse and axial curves depends on the source-facet distance, which influences the MMI pattern. However, the main characteristics leading to a well defined optical trap, viz. the Gaussian-like shape and the decay, are seen for each source position.

As a further step, the averaged intensity distribution is obtained from simulations of the complete dual-beam trap based on 1 μη³ cross-section square waveguides, without bead and with a 1 μιη diameter polystyrene bead 24 centered at the point (0, 1, 0) . We use again 4 μιη as source-facet distance in the two waveguides. The sources each deliver 0.5 W of optical power to a waveguide (i.e. total input power Jo is 1 W) and they oscillate in phase. The results are shown in Fig. 3, for the yz-plane without bead (a), for the xy-plane without bead (b) , for the yz-plane with a 1 μιη diameter polystyrene bead centered at (0,1,0) (c) and for the xy-plane with bead (d) . In the images the normalization of the distributions is such that the full grey scale is used.

Contrary to Fig. 2a, the averaging now does not remove the half-wavelength periodicity in the water region, since in this region interference of the beams gives a stable standing wave pattern. The period of the resulting intensity modulation in Fig. 3b amounts to 297 nm, in agreement with the half-wavelength value 785 nm/ (2x1.33) =295 nm in water. The figure clearly shows that the combined effect of the two beams indeed yields an overall intensity distribution that should give a gradient force for radial displacements and a scattering force for axial displacements. In Figs. 3b, 3d the effect of spreading of the individual beams is also clearly observed in the overall intensity distribution. When comparing Figs. 3a, 3b with Figs. 3c, 3d, the effect of the bead on the field distribution is apparent. As a result of the higher index of the bead (-n_poi_yst. =l .6) , it concentrates the electrical field inside its volume, thus perturbing the distribution in its vicinity. In Fig. 3d. this effect is also reflected in the somewhat shorter period of the pattern inside the bead. From simulations performed for various phase differences between the exciting sources, we find that the global intensity distribution does not depend on the phase between the beams. The exact positioning of the periodicity in the gap, however, does depend on the relative phase. As a final check we also perform several simulations for the polarization parallel to the z-direction. The resulting intensity distributions are very close to those of Figs. 2 and 3, as expected on the basis of symmetry.

Trapping characteristics

The geometry being set, we determine the trapping capabilities of the device with calculations of the optical force and the trapping potential on the basis of FDTD simulations of the three-dimensional distribution of the electrical and magnetic fields. This is done for positions along the three axes of the coordinate system (see Fig. 1) and using polystyrene spherical probe particles 24, i.e. the field distributions are calculated in the presence of the probe. The diameters of the polystyrene beads are 0.2, 0.5, 0.8, 1.0 and 1.4 μιτι, while their refractive index is taken as n _{i t} =1.6. The optical force is evaluated by integrating the Maxwell stress tensor T over a cube with surface S around the bead:

(F)= _s (T) ^■ dS (1)

In Eq. (1) the angular brackets denote time averaging.

Omitting the time averaging, the Cartesian component F± (i=x, y, z) of the force can be written as F_± = §_s TijUjdS (2) th _→

Here ri_j is the j component of the normal n to the surface S, while the tensor components Ti_j(i, j=x, y, z) are given by Ti_j = DiE^* _j - - δ₁₃Ε⁾·Ε^* + Β_±Η% - -5±_j H · B^* . (3) In Eq. (3) , E, D, H and B are the electrical field, the dielectric displacement field, the magnetic field and the magnetic induction, respectively, 5±j is the Kronecker delta and the asterix indicates complex conjugation. We use a cube of side a = 2R+res) , symmetrically placed around the bead for evaluation of the surface integral. Here R is the bead radius and res is the spatial resolution of the FDTD grid. For the fields in the stress tensor at a certain time, we use the values exported for two successive grid times of the FDTD simulation, synchronized with the synchronization tool of Meep for the highest accuracy. The time-averaged force is determined by taking the average of the time-dependent forces calculated for the grid times in the last four wave periods, where it has reached the steady state.

From the calculated average force the trapping

potential at position r₀ can simply be obtained by

integration according to

U(r₀) = - f (F(f ) - dr (4)

The integration path is along either of the three axes of the system.

Before carrying out a complete set of force

calculations, we perform several experiments to establish the grid size needed to reach sufficient accuracy for the calculated force. We take the calculated force for an empty cube (i.e. polystyrene is replaced by water, so that the force should be zero) as a measure of the numerical error in the force calculated for the particular bead size and the particular grid size. In these experiments the bead is positioned at a site of lower symmetry, to avoid

cancellation effects of integrals over symmetrically

positioned cube faces. In the trade-off between grid size and simulation time we arrive at res=l μιη/20 for force calculations on the x- axis and res=l μιη/30 for force calculations on the y- and z- axis, giving an estimated average error in the force of about 20

Force characteristics of the dual beam trap

In view of the symmetry of the problem, we only

calculate the optical forces for positive positions of the bead on the X, y- and z-axis and obtain the forces for negative positions from sign reversal.

In Fig. 4 we present the resulting curves, for

positions of the bead on the three axes and for various bead diameters. The insets (4b, c) are blow ups of the force curves for positive bead positions for the 0.2 and 0.5 μιη diameter beads. The curves in (4a-c) are piecewise cubic spline interpolations between the data points.

On the x-axis the force behaves completely different than on the y- and z-axis. On the x-axis the force shows a composite behaviour: strong oscillations superimposed on a straight line of a weak slope through the origin. On the other two axes it shows the usual restoring character, i.e. the force is systematically directed such as to counteract a displacement from the equilibrium position.

For the y- and z-direction, for each bead size, the force is linear around the trap centre, indicating behaviour according to Hooke's law. With increasing distance to the centre, the force goes through an extremum of which the position increases with increasing bead size. Further, the extremum occurs for a bead position such that the whole bead is on one side of the trap centre. This occurrence of a bead-size dependent extremum clearly indicates operation outside the Rayleigh regime, where the bead is much smaller than the wavelength, so that it virtually does not affect the distribution of the electrical field. This finding is consistent with the influence of the particle on the field as seen in Figs. 3c and 3d: the particle disturbs the field distribution. Clear enough, in this regime of large bead sizes, the disturbance will depend on the bead size, as we see reflected in Figs. 4b and 4c.

The period of the force oscillations in the x-direction is 297 nm, in very good agreement with the period of the intensity modulation in Fig. 3b. Therefore, the oscillating force relates to the interference of the coherent output beams of the waveguides. We argue that in the gap between the waveguides a one-dimensional optical lattice is formed, which is the dual-waveguide equivalent of an optical lattice formed with interfering laser beams and used for the study of periodic arrangements of trapped atoms. We first limit the discussion to bead diameters of 0.2, 0.5, 0.8 and 1.0 μιη, for which the straight line on which the oscillations are superimposed has the usual negative slope of the

scattering force. As expected, this slope increases with increasing bead size. We now focus on the behaviour close to x=0, where an intensity maximum occurs, to understand the oscillatory behaviour. At x=0 the oscillatory curve for 0.2 and 0.8 μιη has a negative slope, in agreement with the expectation that the gradient force pushes the particles to the intensity maximum. For 0.5 and 1.0 μιτι, on the contrary, the slope is positive and the particles are pushed to an intensity minimum. This behaviour, including the spatial oscillatory character, agrees completely with the

theoretical result that the axial optical force depends periodically on the particle size, in such a way that as a function of its size the equilibrium position of the

particle alternates between the intensity maxima and minima. As for the oscillatory behaviour, the 1.4 μιη bead falls in the same class as the 0.5 and 1.0 μιη beads. The line on which the oscillations are superimposed, however, has a positive slope. This behaviour so far is not properly analysed. Possibly, to grasp the complete picture, in addition to the scattering force a gradient force related to the intensity maximum close to the facet of the individual beams (see Fig. 3c) has to be involved. This intensity maximum plays the role of the focus of a laser beam and thus tends to attract the particle. In practice, whether this sign reversal of the background slope occurs, will depend on the possibility to dominantly excite and propagate only the zeroth order mode (in simulations this can be done by using a tailored source) , to avoid MMI and thus the maximum just outside the facet.

A consequence of the independence of the global

intensity distribution in the gap on the relative phase of the beams is that the global force distribution shows this property as well. The phase of the oscillations of the force in the x-direction, however, does depend on the relative phase. This phase dependence enables moving the particle in the x-direction by changing the relative phase, assuming that the coherence of the laser source driving the

waveguides allows for building up of a stable interference pattern .

By nature of the trap, the trapping force is strongest in the y- and z-direction, where it takes maximum values of 13.5 and 16 pN/W for the 1.4 μιη bead. On the basis of symmetry, one would expect these values to be the same. That they differ in this way is indicative of the accuracy of the force calculation for res=l μιη/30. Using these numbers, the average trap stiffness in the transverse is direction is 0.03 pN/nm/W, a value very suitable for trapping of μιη-sized particles . Trapping potentials

From the force curves in Fig. 4 we obtain the trapping potential by integration. Using Eq. (4), for the y-direction the trapping potential U_y(y₀) at position y₀ (y₀≥0; as before we obtain the negative branch using symmetry) is given by

U_yiyo)≡ U(0,y_o,0) = - £°<_y(y)> dy (5)

To evaluate the integral for U (y₀), we set the lower integration limit to 1.75, which we estimate "by eye", by extrapolating for each bead size the saturating force curve, to be the y-value for which F(y) would be zero. The same procedure applies to evaluation of the integral for U_z(z₀) . For the x-direction the force field is limited to the gap, while in the simulations the size of the probe particle limits this range even further. In calculating the potential in the x-direction we integrate over the range for which the bead-size dependent forces are displayed in Fig. 4. We normalize U_x(x) by applying a scaling factor which makes U_x(0) equal to £/ (0) .

In Fig. 5 we show the resulting trapping potentials U_x(x) and U (y) , normalized to kT (T=300 K) . Since U (y) is similar to U_z{z) , we omit it here. For the 1.4 μιη bead U_x{x) is not shown, in view of the deviating slope of the background force for this size, as discussed above.

For U_x{x) , plotted in Fig. 5a for a power of 100 mW, we observe a parabolic shape, of which the curvature increases with increasing bead size and on which oscillations are superimposed corresponding to the force oscillations in Fig. 4a. Although the trap stiffness is weaker than that of a conventional laser trap, stable optical trapping without bouncing at the channel walls can be induced for this laser power. This is deduced from the criterion U/kT >10 for stable trapping. In Fig. 5a, the horizontal dashed lines in each potential well correspond to U/kT =10. For bead sizes 1.0, 0.8 and 0.5 μιη the position of the line indicates that the criterion is fulfilled. The average displacement of the trapped particle from the potential minimum is relatively large, as indicated by the crossings of the dashed lines with the potential. For the 0.2 μιη bead stable optical trapping does not occur for this power. Instead, for this size the channel walls take care of the confinement. At higher laser power, where the depth of the minima of the oscillatory potential can exceed 10kT, trapping may occur in a local minimum, similar to the situation of particle trapping in an optical lattice. Note that the applied optical power IQ is 100 and 15 mW in (a) and (b) ,

respectively, while stable optical trapping can still be induced, as indicated by the position of the horizontal dashed lines. These are lOkT above the bottom of the

corresponding potential.

For U(y), plotted in Fig. 5b for a power of 15 mW, the parabolic well is deep enough to give stable trapping for

0.8, 1.0 and 1.4 μιη beads, as again indicated by the dashed horizontal lines.

Fabrication process

The waveguides are patterned in a S1₃N₄ layer deposited on a glass wafer by chemical vapor deposition (CVD) , yet omitting etching of the gap. The waveguides connect to the output branches of a Y-junction, which is patterned in the same layer and of which the input will be driven by a laser. For complete embedding of the waveguides, an S1O₂ overlayer is deposited by CVD. After planarization of this layer, the fluidic channel and the gap between the waveguides are formed in a combined patterning step. The upper seal of the channel is formed by bonding a second wafer on top of the assembly, into which access holes to the fluidic channel are etched .

Multiple traps and channels can be implemented using multiple Y-junctions, by repeated branching out of the input waveguide. Preferential trapping of objects of a specific shape (e.g. cigar-like) can be promoted by adjusting the beam profile with a tailored waveguide cross section.

Figures 6A-C show a possible design of an optical trap according to the invention. The device 60 comprises a fluid channel 62, 64, 66, which is formed as a trench in a layer on the substrate. The channel comprises an input part 62 of a first width Wi and an output part of preferably equal width. In between the input part and the output part a part 66 with a reduced cross section is provided. Due to their relative large width, parts 62 and 64 are easily connectable to further fluid lines and/or channels.

The device 60 further comprises a waveguide 68, which comprises a Y-junction 70 where it splits into two

waveguides 72, 74, which are arranged such that their ends opposite the junction 70 are facing opposite to each other (in the region of part 66 with reduced cross section) .

It is noted that the layer in which the fluid channel 62, 64, 66 is provided, serves as cladding material for waveguides 72, 74. In this embodiment, this is a S1O₂ layer.

In the device shown, the phase of the light entering at the input side of waveguide 68 is equal in both waveguide 72 and waveguide 74. It is however also possible to provide one or both of the waveguides 72, 74 with a phase shifter, for example a heating element, such as a resistor (not shown) .

Zooming in on part 66 (figure 6B, 6C) , it can be seen that the fluid channel is even thinner in a region 76 of length L, where L is for example 5-20 ym. Preferably L is larger than the width of the region 76, This ensures a well-defined homogenous cross section over a certain length of the channel. Furthermore, the flow profile of the fluid in the channel is homogenous. For example, L is at least twice the width of region 76. The cross section of region 76 corresponds with the dimensions of the particles to be measured. Due to the fact that the fluid channel has a small cross section over only a relatively small length with respect to its total length, the flow resistance due to the narrow part 76 is relatively small.

On the basis of the results for polystyrene beads it can be expected that the dual-waveguide device according to the invention can trap biological objects such as bacteria. Fabrication of the device is compatible with standard glass- based microfluidic technology. The maximum theoretical gradient force in the radial direction is 16 pN/W,

corresponding to a trap stiffness of 0.03 pN/nm/W. The depths of the trapping potentials deduced from the force curves indicate that the dual-waveguide trap leads to stable optical trapping for practical optical powers. In case of coherent beams launched by the waveguides, we find an interesting oscillatory force for the axial direction, which can be used to radially move a particle.

The present invention is by no means limited to the above described preferred embodiments thereof. The rights sought are defined by the following claims within the scope of which many modifications can be envisaged.

Claims

1. Optical trap, comprising

- a substrate; and

waveguide when in use,

2. Optical trap according to claim 1, wherein at least one waveguide is a solid core waveguide.

3. Optical trap according to claim 1 or 2, wherein at least one waveguide comprises a silicon compound.

4. Optical trap according to claim 3, wherein at least one waveguide comprises S1₃N₄ and/or SiON.

5. Optical trap according to at least one of the claims 1-4, wherein a fluid channel is provided which passes between the exiting surfaces of the waveguides.

6. Optical trap according to claim 5, wherein the fluid channel comprises a trench of which the walls are in

substantially the same plane as the exiting surfaces of the waveguides .

7. Optical trap according to claim 6, wherein at least a part of the walls of the trench form a cladding for the waveguides .

8. Optical trap according to any of the claims 1-7,

comprising an optical source.

9. Optical trap according to claim 8, wherein the optical source is a laser.

10. Optical trap according to claims 8 or 9, comprising means for controlling the relative phase of the beams exiting the at least two waveguides in use.

11. Chip comprising the optical trap according to any of the claims 1-10.

12. Sensor system comprising the optical trap according to any of the claims 1-10.

13. Method for manufacturing an optical trap, comprising the steps of:

- providing a substrate; and

- depositing at least two waveguides on the substrate such that the waveguides each have an exiting surface and the exiting surfaces face each other, wherein depositing comprises a microfabrication process.

14. Method according to claim 13, depositing comprising thin film deposition.

15. Method according to claim 14, depositing comprising vapour deposition.