WO2004103568A1

WO2004103568A1 - Particle separation apparatus and method

Info

Publication number: WO2004103568A1
Application number: PCT/GB2004/002202
Authority: WO
Inventors: Kevin John Donovan; Thierry Claude Lutz
Original assignee: Queen Mary & Westfield College
Priority date: 2003-05-22
Filing date: 2004-05-21
Publication date: 2004-12-02
Also published as: GB0311826D0

Abstract

The invention provides a method and apparatus for separating metallic and semiconducting nanotubes using dielectrophoresis. A sample of nanotubes, suspended in a viscous medium, is introduced to a separation chamber between two electrodes creating a non-homogenous electric field. Due to the action of a differential dielectrophoretic force, metallic and semiconducting nanotubes are separated and can be retrieved at the corresponding electrodes.

Description

Particle Separation Apparatus And Method

The invention relates to a particle separation apparatus and method, in particular for separation of nanotubes.

Carbon nanotubes (CNT) are macromolecular particles with unique mechanical and electronic properties including charge transport properties, strength and closeness to being a one-dimensional system. They consist of cylinders made from rolled up graphene sheets, The cylinders come in a variety of diameters on the scale of 1 nanometre. Depending on their chirality and diameter, CNT's can either be semiconducting, having a finite band gap, or metallic, having a band gap of zero. They may consist of a number of concentric tubes of differing diameter. Such Multiwalled Carbon Nanotubes, MWCNT's will generally consist of metallic and semiconducting nanotubes together. Of more interest to electronic applications of these materials are the Single Walled Carbon Nanotubes, SWCNT's which will be either metallic (approximately 33%) or semiconducting.. For a review of the one-dimensional electronic structure of SWCNT, semiconducting SWCNT and SWCNT-based electronic devices see P. Avouris, Carbon nanotube electronics, Chemical Physics 281 (2002), p. 429 - 445.

A problem with existing nanotube technology, is that no method or apparatus for separating metallic and semiconducting SWCNT (referred to as "nanotube" in the remainder) is known. It is an object of the present invention to solve this problem as set out below.

The invention is set out in the independent claims.

The invention makes use of dielectrophoretic forces experienced by polarisable particles, such as nanotubes, in a non-homogenous electric field. Due to the difference in polarisability of metallic and semiconducting nanotubes arising from the very much greater free charge carrier density on metallic nanotubes compared with semiconducting nanotubes they are subject to very different dielectrophoretic forces. This difference in force experienced can be used to separate the two populations. Advantageously, the magnitude of this effect is proportional to a particle's induced dipole moment and should therefore be particularly pronounced for nanotubes, which have a length of typically one micron.

Further preferred features of the invention are set out in the independent claims.

The dielectrophoretic force depends not only on the dipole moment of the nanotubes, but also on the difference in the polarisability of the nanotubes and the medium in which they are suspended. Advantageously, by adjusting the composition of the medium, its polarisability can be adjusted in a way such as to cause semiconducting and metallic nanotubes to move in different directions when the electric field is applied. This further facilitates separation significantly. Glycerol is a useful medium in this respect, as it is miscible with water, allowing the medium's polarisability to be adjusted by simply adding water. Furthermore, surfactants like gum arabic act to separate ropes of SWCNT into individual nanotubes.

As is well known, the polarisability of a particle and consequently the dielectrophoretic force depends not only on the difference in polarisability, but also on the frequency of the electric field. Another advantage of the invention therefore is that the separation can be aided by tuning the frequency of the electric field. In one specific embodiment, the electric field preferably oscillates at a first frequency during a first period and at a second frequency during a second period and the first frequency is preferably higher than the second frequency. Introducing a first, high-frequency period has the advantage that it tends to align the nanotubes, and hence the corresponding dipole moments.

Preferred embodiments of the invention are now described by way of example only and with reference to the accompanying drawings, whereby:

Figure 1 shows an electrode arrangement of an embodiment; Figure 2 shows a nanotube separation apparatus;

Figure 3 shows an alternative electrode arrangement;

Figure 4 shows another alternative electrode arrangement;

Figure 5 shows yet another alternative electrode arrangement;

Figure 6 shows a further alternative electrode arrangement; Figure 7 shows yet a further electrode configuration;

Figure 8 shows a schematic representation of a continuous flow separation apparatus;

Figure 9 shows a cross-sectional view of the apparatus shown in Figure

8; and Figure 10 shows an alternative embodiment of the apparatus shown in

Figure 8.

In overview the invention seeks to separate nanotubes dependent on their polarisability by virtue of the non-homogenous electric field used in the embodiment of this technique. In particular, the approach uses the technique of dielectrophoresis.

Dielectrophoresis is a technique for manipulating polarisable molecules or particles widely used in microbiological applications. The skilled reader will be familiar with the relevant techniques, but in summary, dielectrophoresis is based on the electrostatic force arising from the interaction of a polarisable particle and an applied non-homogenous electric field. The first effect of the electric field is to induce a dipole moment in the particle, attracting positive charges towards the negative pole of the field and vice versa and hence creating a torque that will tend to align the particle with the field lines along a longitudinal axis.

In some electrode geometries where the field gradient is aligned with the field lines, due to the non-homogeneity of the electric field, parts of the induced dipole that reside in a region of space where the electric field is stronger will experience a higher electrostatic force than parts of the dipole residing in a region of where the electric field is weaker. As a result, the particle experiences a net-force in the direction from a low field to a high field along the field gradient. Electrode geometries giving rise to such fields are illustrated in Figures 3, 4 and 6.

Other electrode geometries such as the one shown in Figure 1 give rise to a situation where the field lines and the field gradient are perpendicular to each other, for example where the field lines are concentric circles. In this situation, the dipoles will still align tangentially along the direction of the field lines, but the dielectrophoretic force they experience arises differently than for the situation just described, since the gradient is now perpendicular to the field lines. In this situation, the electrostatic force aligning the dipole tangentially is effectively pulling each end of the dipole in a direction which is tangential at the respective end of the dipole. Since the dipole is of finite length, the two respective forces will not cancel even if the dipole is perfectly aligned, but form an angle. This is due to the fact that the two forces are tangential to the field lines at two different angular positions with respect to the centre of the concentric field lines. As a result, there is a resultant force acting on the dipole and this force is pointing radially towards the centre of the concentric field lines. The magnitude of the resulting force is proportional to the strength of the field and to q-sin(-E/2r), where q is the dipole charge, I is the dipole length and r is the distance from the centre of the concentric field lines. If t is small compared to r, q-sin( /2r) ~ ql = p, the dipole moment. Thus, the force is proportional to the dipole moment.

Figure 1 shows considerations relevant to the application of dielectrophoresis in relation to separation of carbon nanotubes, illustrating schematically a dipole induced in nanotube 10 aligned along the field lines 38. An electrode arrangement comprises a first 30 and second rectangular electrode 32 shows in plan view in Figure 1, whose longitudinal axes (A,B) intersect at a point remote of the electrode configuration. In other words, the distance between the two electrodes is larger at a first, wide end 34 of the electrode configuration than at a second, narrow end 36. As a result, the field lines are "compressed" towards the end of the electrode configuration that has the smaller distance between the two electrodes, resulting in a field strength that increases from the wider end 34 to the narrower end 36. If the electrodes would be spaced infinitely closed together at the intersection of lines A and B, the field lines would be concentric circles with the intersection as centre. However, as long as the gap between the electrodes of the narrow end is small compared to the overall dimensions, the field lines will remain more or less concentric circles. This electrode configuration results in a non-homogenous electric field that gives rise to a dielectrophoretic force (C) pointing radially towards the centre of the concentric field lines. At the centre of the cell the force thus points along the direction of the axis of symmetry of the electrode configuration as described for concentric field lines above. It should be noted that in this example the direction of the dielectrophoretic force depends solely on the gradient of the field and not its direction, that is the force would be the same if the polarities of the first and second electrode were reversed.

Figure 2 shows in more detail a nanotube separating apparatus 20 including the electrode configuration of Figure 1. A separation chamber 22 contains a polarisable viscous medium 24 and a first 30 and second 32 electrode at least partially submerged in the viscous medium 24. The separation chamber 22 follows the outer contour of the first and second electrode.

In operation, the first 30 and second 32 electrode are connected to a suitable power supply 26 via leads 28. A sample of nanotubes 29 to be separated is introduced into the separation chamber in a separating region at a position between the first 14 and second 16 electrodes. Figure 2 shows a nanotube sample 29 introduced into the viscous medium at a position on the axis of symmetry of the electrode arrangement, although introducing the sample at other positions might prove beneficial. A voltage is applied to the electrodes 30 and 32 using the power supply 26, creating a non-homogenous electric field that gives rise to dielectrophoretic forces acting on the nanotubes within the sample 29.

The dipole is induced in the nanotubes because of their inherent charge carrier mobility and hence the polarisability of the nanotubes depends on the number of carriers free to respond to the electric field. It is this second factor that leads to the substantial increase in the effect on metals over that on semiconductors. The more polarisable a nanotube 10 with respect to the polarisability of a suspending medium, the stronger the dielectrophoretic force it experiences. In particular, because of their greater free electron density, metallic nanotubes are substantially more polarisable than semiconducting nanotubes and therefore a given non-uniform (inhomogeneous) field gives rise to a different dielectrophoretic force. Moreover, nanotubes are of a highly elongated shape, which, everything else being equal, results in a higher charge separation and dipole moment, than, for example, for a spherical particle. The larger dipole moment results in a higher dielectric force overall and consequently a higher difference in polarisability between metallic and semiconducting particles, as compared to a spherical particle. Accordingly, dielectrophoresis is ideally suited to separate metallic and semiconducting nanotubes, since the difference in dielectric force can cause metallic and semiconducting nanotubes to move at different velocities or even in different directions and this effect is amplified by the tube-shaped geometry of nanotubes.

Because the nanotube 10 is suspended in a viscous medium, which is itself polarisable, not only the magnitude but also the direction of the dielectrophoretic force depends on the difference in polarisability between the nanotube and the medium. In particular, the direction of the dielectrophoretic force depends on the sign of the difference in polarisability: if the nanotube is more polarisable than the medium, the direction of force is in the direction of the field gradient; if, however, the nanotube is less polarisable than the medium, the direction of force is reversed and the dipole will tend to move in a direction against the field gradient. Nanotubes placed between the two electrodes will therefore travel towards the narrow end 36 of the electrode configuration if their polarisability is larger than the polarisability of the surrounding medium. If their polarisability is lower than that of the medium, they will travel towards the wide end 34.

A possible base for the medium is glycerol, which has the advantage that it is miscible with water, allowing the polarisability of the medium to be tuned. In one embodiment, the polarisability of the medium is tuned to lie in between the polarisability of metallic and semiconducting nanotubes, causing the two types of nanotubes to travel in different directions. The polarisability of the medium may be adjusted by adding an experimentally determined amount of water to the medium.

In addition to tuning the medium's polarisability, the viscosity of the medium has to be tuned as well to allow for sufficient mobility of the nanotubes, while minimising the effect of Brownian motion. This represents a trade-off between increasing the viscosity in order to dampen out Brownian motion and the need for the nanotube to be mobile enough in order to separate in a reasonable amount of time. A possible additive to the medium is gum arabic, serving as a surfactant that induces ropes of nanotubes to dissolve into individual strands, thus facilitating separation. Other surfactants include polyvinyl pyrrolidone, PVP, and sodium dodecylbenzene sulphonate, SDS.

The electric field generated with the power source 26 can either be a dc or an ac field. However, using an ac field offers a number of advantages. First, any charged impurities present in the sample would not be subject to a significant net force, averaged over one cycle of oscillation of the field and would hence not be affected by the ac field. In contrast, in a dc field, the charged impurities would experience a strong net force and be attracted to the respective electrodes much more rapidly than the neutral (uncharged) nanotubes. Since the magnitude and direction of the dielectrophoretic force depend not only on the polarisability of the medium, but also on the frequency of oscillation of the applied field, the frequency of the field also needs to be tuned experimentally.

An optional first phase of pre-treatment at high frequencies and low voltage may be applied prior to a subsequent second phase of separation at a lower frequency and higher voltage, as this pre-treatment may help to de-tangle and orient the nanotubes in the sample. In one example the field therefore oscillates with an amplitude of 20 V at a frequency of 1 MHz for 2 hours during a pre- treatment phase. With or without pre-treatment, the separation phase is run at 800 V and 3 kHz. With this set-up, significant separation occurs within 15 minutes and separation saturates within 1 hour. Other values of voltage and frequency are of course also possible, in particular the first frequency can be in the range of O.lMHz - 1MHz and the second frequency can be in the range of 100Hz - 10kHz and the voltage can be in the range of 10V - 100V in the first period and in the range of 100V - 1000V in the second period. The dimensions of the system will play a large role in determining the voltage to be used as a larger system will require higher voltages to attain the same fields and therefore forces.

For the embodiment shown in Figure 1, assuming an angle of 60° and a potential difference of lkV between the electrodes, the magnitude of the field at 10 cm, from the virtual apex (the intersection at lines A and B) is 9550 V/m and 10 times larger at 1 cm from the virtual apex. The field gradient is proportional to the inverse square of the distance from the virtual apex and is thus 100 times larger at 1 cm than at 10 cm. The absolute value of the dielectric force depends on the polarisability of the medium and of the nanotube experiencing the force.

An alternative electrode arrangement 11 shown in Figure 3 includes a first electrode 14 defining a region of high field gradient and a second electrode 16 defining a region of low field gradient. The electrode arrangement induces a dipole in a nanotube 10 oriented along electric field lines 12. In the example shown in Figure 3, the dipole experiences a net force shown by arrow 18 and moves in the direction of high field gradient towards the first electrode 14, if the medium was less polarisable than the nanotube. If the nanotube 10 in Figure 1 was surrounded by a medium more polarisable than the nanotube, it would experience a force pulling it towards the second electrode 16, rather than the first electrode 14.

In the embodiment shown in Fig. 3, assuming a radius of 30 mm for the second electrode 16 and a radius of 2 mm for the first electrode 14, an applied potential difference of 1000 V results in a field of 1.23 x 10⁴ V/m at the second electrode 16 and of 1.85 x 10⁵ V/m at the first electrode 14. The ratio of the fields at the two electrodes varies with the ratio of the two associated radii, i.e. a factor of 15 in the example. The dielectric force is proportional to the gradient of the square of the field and is approximately 3400 times bigger at the smaller electrode than at the bigger electrode, varying in proportion to the cube of the radii. The absolute value of the dielectric force depends on the polarisability of the medium and of the nanotube experiencing the force.

Figure 4 shows another alternative electrode configuration consisting of a first straight and elongated rectangular electrode 40 and a second small disc shaped electrode 42 facing it. The diameter of the disc shaped electrode is substantially smaller than the length of the strip electrode, preferably by a factor of at least ten This configuration gives rise to a non-homogenous electric field where field lines are compressed towards the small disc-shaped electrode. This results in a higher gradient close to the second electrode 42 than near the first electrode 40 and, accordingly, nanotubes with a polarisability higher than the medium will travel towards the second electrode 42 while nanotubes with a polarisability lower than the medium will travel towards the first electrode 40.

Figure 5 shows yet another electrode configuration consisting of first 50 and second 52 comb shaped electrodes having intermeshing respective sets of teeth. This configuration creates a set of wide regions 54 formed between the base and a set of adjacent teeth of one comb on three sides and the end of a tooth of the opposite comb on the forth side. This configuration also creates a set of narrow regions 56, formed between two adjacent teeth, one from each comb.

Figure 6 shows a further alternative electrode configuration consisting of a first rectangular and elongated electrode 60 and a second electrode 62 in the shape of an isosceles triangle with its apex pointing towards the first electrode. In the embodiment shown in Figure 6 the apex of the second electrode 62 is perpendicular to the first electrode 60 and the apical angle is 30°. In general, sharp points result in greater inhomogeneity and hence larger field gradients, consequently achieving higher dielectric forces. Therefore, the apex of the second electrode 62 preferably forms a generally sharp point.

Yet a further electrode configuration 70 is shown in Figure 7, providing a further development of the electrode arrangement shown in Figure 1. The electrode arrangement comprises two trapezoid electrodes, each having a long side 72 and a short parallel side 74, which are facing each other such that the short sides 74 are adjacent, preferably with the respective axes of symmetry coinciding. This electrode configuration defines two tapered or truncated V or triangle regions 76 creating a non-homogenous field as in Figure 1, connected to each other at the narrow end by a channel 78 in which the field is more or less homogeneous. A sample placed within the non-homogeneous region 76 is separated according to polarisability as discussed for the Figures 1 configuration. Nanotubes (or any other polarisable particle) with polarisability larger than the polarisability of the surrounding medium will tend to travel towards the narrow end of the V-shaped region 76, that is towards the channel 78 defining a more or less homogeneous region of the field. A nanotube entering the homogeneous region in the channel 78 between the two electrodes will experience no further or at least a significantly reduced dielectric force (which is proportional to the field gradient). Thus, nanotubes having a polarisability higher than the polarisability of the surrounding medium will tend to aggregate in the channel between the two electrodes.

This electrode configuration is a further optimisation of the Figure 1 configuration, since the symmetrical double V shape prevents end effects at the narrow end of the electrode configuration. These end effects might otherwise cause the nanotubes to move to the edge of the electrodes and thus cause disruption to the field when metallic nanotubes come into contact with the electrodes. The Figure 7 configuration solves this problem, since the tapered or V-shaped regions will act co-operatively to keep nanotubes within the channel 78 once they have entered the channel 78. Channel 78 acts as a region of more or less uniform field forming a trap for the nanotubes and thus avoiding the problem of the nanotubes contacting the electrodes as may happen for the Figure 3 configuration.

In one specific embodiment, the properties of the medium are tuned so that, in operation, metallic and semiconducting nanotubes respectively move in different directions towards different ends of the electrode configuration. Consequently, metallic and semiconducting nanotubes can each be obtained by retrieving the suspending medium 24 at the appropriate end. The highly polarisable metallic nanotubes are attracted by the region of high field gradient adjacent to the narrow end 36 or channel 78 and the less polarisable semiconducting nanotubes are attracted to the wide end 34 and metallic and semiconducting nanotubes can therefore be retrieved in solution near the first and second electrode, respectively, using, for example, a pipette.

Separation and retrieval of metallic and semiconducting nanotubes can be automated by a continuous feed and retrieval system. Such a system could, for example, continually introduce medium containing assorted nanotubes at a smaller flow rate and at an appropriate location between the narrow and wide end. In such a continuous feed and retrieval system, a suction device would be placed at the narrow and/or wide end in order to continually retrieve medium containing separated metallic and/or semiconducting nanotubes. The flow rate of the feed and suction are adjusted such that the overall volume of medium inside the separating chamber stays constant.

In an alternative embodiment, the viscous medium and field frequency are tuned such that metallic and semiconducting nanotubes move in the same direction towards the same region of the electric field. Separation can be achieved by making use of the fact that the dielectrophoretic force experienced by the metallic nanotubes is, everything else being equal, much larger than the force experienced by the semiconducting nanotubes. As a result of Stokes law, metallic nanotubes therefore have a higher terminal velocity than semiconducting nanotubes.

This can be exploited for separation, for example by retrieving samples at different times or by applying a force, for example gravity, perpendicular to the direction in which the nanotubes are moving. In the first example, samples retrieved in a region of high field density early during the separation process will contain a higher concentration of metallic nanotubes than samples retrieved later. In the second example, the nanotubes would be spread out spatially along the direction of movement, with metallic nanotubes being in front in the direction of movement. This is because the faster metallic nanotubes travel further in the time it takes them to "fall" a given height under the action of gravity.

The apparatus and method of the preferred embodiment can be extended to separation of nanotubes by band gap in general. In this case the apparatus and method are not used to separate metallic and semiconducting nanotubes, but rather to separate the nanotubes into two populations having band gaps below and above a pre-determined threshold. Low band gap nanotubes will have a larger thermal population of free carriers in the conduction band and will therefore be more polarisable and thus acted upon more strongly by the non- uniform electric field. The threshold is determined by adjusting the composition of the viscous medium and the frequency of the electric field such that, in the preferred embodiment, nanotubes having a band gap above the threshold move to the second electrode and nanotubes having a band gap below the threshold move to the first electrode. One or both of the populations of nanotubes can then be retrieved at the respective electrode, as described above.

Using the alternative embodiment, having nanotubes moving in the same direction, a band gap separation technique working on a finer scale can be implemented. This effectively results in an apparatus and method for sorting the nanotubes by band gap. Band gap sorting can be achieved temporally, by retrieving nanotubes at a specified location of high field density in the electric field at different times, as described above. In this case, nanotubes having smaller band gaps tend to arrive earlier at the retrieval location than nanotubes having larger band gaps. This is because a larger band gap means that at any given temperature less charge carriers are available. This results in less charge separation for a given electric field and hence a lower polarisability.

Similarly, band gap sorting can be achieved spatially, by applying a perpendicular force such as gravity as described above. The location along the direction of movement can be calibrated with respect to the associated band gap and the calibration can be used to retrieve nanotubes of the desired band gap. A further embodiment of the invention, described below, is particularly useful as a continuous feed and retrieval system, as outlined above. As shown in Figures 8 and 9, two spaced, opposed plate electrodes 80 are arranged at an angle to one another along a first dimension (x direction) and parallel along a second dimension (y direction) perpendicular to the first. Essentially, this is a three dimensional representation of a version of the Figure 1 electrode configuration. Figure 9 shows a plan view along the y direction (pointing into the plane of the paper), the plane being defined by the x and z direction (perpendicular to both the x and y direction), as indicated in the figure. The electrode configuration provides a volume containing a non-homogeneous field, the volume having a cross section in the x-z plane similar to the Figure 1 electrode configuration, that is a truncated V or triangular or tapered shape having a narrow end 90 and a wide end 92, the plates 80 extending along the y direction.

The embodiment comprises means for generating a flow, preferably laminar, in the y direction, thus defining an entry end where the flow enters the volume between the two plate electrodes and an exit end where the flow exits. A nanotube sample source 86 is provided in close proximity to both the entry and wide end and sample retrieval means are situated at the narrow end. The sample retrieval means comprise one or more separators 82 and one or more retrieval ports 84. If several separators/retrieval ports are used, these may be spaced out along the y direction.

In operation, the flow means create a flow in the y direction inside the volume between the plate electrodes, and nanotubes are introduced into the flow from sample source 86. An ac or dc voltage source is connected to the plate electrodes 80, thus creating a non-homogeneous, time-varying field in the volume between the plate electrodes 80, which field induces a dielectrophoretic force acting on the nanotubes. For a nanotube where polarisability is larger than that of the flow medium the dielectric force points in the x direction, from the wide end 92 to the narrow end 90.

As the nanotubes travel towards the narrow end, their flow will be interrupted by a separator 82 and the further dielectric force acting on nanotubes caught by separator 82 will cause the respective nanotube to travel along the separator towards the retrieval ports 84 where the nanotube may be retrieved from the flow medium. As described above, depending on the magnitude of dielectrophoretic force (which in turn depends on the relative polarisability of the medium and the nanotubes), dependent on their bandgap nanotubes will achieve different terminal velocity in the x direction and will therefore take different amounts of time to reach the narrow end 90 from the point of insertion at sample source 86. As the nanotubes travel at constant velocity with the flow in the y direction, the different times of arrival translate into different y locations at which a nanotube will reach the narrow end 90, depending on their bandgap. Thus, by positioning the retrieval means 82, 84 and in particular the separator 82 at a given y location, nanotubes above a corresponding bandgap value may be retrieved. As such, this may be used for separating metallic and semiconducting nanotubes in a continuous flow arrangement whereby the y- location of separator 82 defines the cut-off value of bandgap.

If more than one retrieval port 84 and/or separator 82 are used, nanotubes having bandgaps falling within different bandgap regions may be separated simultaneously, whereby the y location of the separators 82 adjacent to each retrieval exit 84 determines the limits of the bandgap region of nanotubes retrieved by that retrieval port 84. In this case, the leading separator (in the flow direction) defines the lower bound of the bandgap region and the trailing separator defines the upper bound of the bandgap region. A further electrode configuration for use with the embodiment of Figures 8 and

9 is shown in Figure 10, analogous to the double V electrode configuration of Figure 7. In this electrode arrangement, the plate electrodes are profiled such as to define an inter-plate volume extending along the y direction, which has a cross section in the x-z plane defining two tapered or truncated V or triangle shapes connected by a linear channel 112, similar to an hourglass shape and to the Figure 7 electrode configuration. With this electrode configuration, one or two sample sources 86 may be used, each placed at the respective wide ends of the electrode configuration, that is at the outer most ends of the electrode configuration. The separators 82 are provided within the channel 112 connecting the two truncated V- volumes defined by the plate electrodes at their narrow ends. Retrieval port 84 may be provided through an aperture 120 in the linear portion 110 of the plate electrode defining the channel 112.

A person skilled in the art will further be aware of the specific details of the construction of the embodiments discussed above with reference to Figures 8 to

10 and any appropriate elements or components may be used. It will be understood by a person skilled in the art that any of the continuous flow embodiments described with reference to Figures 8 to 10 may also be provided with retrieval means at the wide end of the plate electrodes 80 or 100 in order to retrieve nanotubes having a polarisability lower than the flow medium. It is also understood that any of the electrode arrangements described above are equally suitable for use in the dielectric separation of any polarised particle, not only nanotubes. The invention may also be supplied as a general polarisability filter and is not restricted to a bandgap filter.

It will be appreciated that individual features of the embodiments can be varied, interchanged or juxtaposed as necessary. The nanotubes separating apparatus can be scaled up in a number of ways. One way to increase the amount of nanotubes that can be separated in a given time, is to simply scale up the overall dimensions of the electrode arrangement. However, this approach is hmited by the magmtude of the potential difference applied between the electrodes, needed to achieve a sufficiently high field gradient in the separating chamber. This potential difference increases as the dimensions of the electrode arrangement are increased. An alternative approach to increasing the separation capacity of the disclosed apparatus, would be to provide a number of separation chambers in parallel. This way the dimensions of the individual chambers can be kept relatively low achieving higher field gradients for a given applied potential difference.

It is understood, that the invention extends to embodiments in which only the frequency of the electric field or only the composition of the medium is adjusted to achieve the desired separation. Any appropriate electrode arrangement can be adopted, for example the relative position of the two electrodes is not restricted to the ones discussed; the first electrode may he on or off an axis of symmetry of the second electrode. Furthermore the separation chamber 22 is of course not restricted to the particular shape described, although preferably it allows the two electrodes 14 and 16 to be at least partially submerged and in fluidic communication once the separation chamber 22 is filled with the viscous medium 24. Similarly, the electrodes are not restricted to a particular shape. Electrodes having sharp or substantially sharp points, for example, are advantageous as providing high non-uniformity and hence field gradient. The electrodes may preferably have appropriate dimensions to extend along at least the entire depth of the medium in the separation chamber. Alternatively, one or both electrodes may be adapted to extend along only part of the depth of the medium in the separation chamber, thus being completely submerged inside the medium when the separation chamber is filled. The voltage, frequency and frequency variation can be adjusted as appropriate. Finally, the medium need not be restricted to the three ingredients mentioned above. For example, the medium might contain any or a combination of glycerol, water, hydrocarbon oil, silicone oil, polyorganosiloxane or dichloroethane. The medium may further contain a variety of surfactants such as for example gum arabic, polyvinyl pyrrolidone, PVP, and sodium dodecylbenzene sulphonate, SDS.

Claims

1. A nanotube separating apparatus for separating metallic and semiconducting nanotubes, having a separating region and first and second electrodes arranged to produce a non-homogenous electric field in the separating region.

2. An apparatus as claimed in claim 1, further including a viscous medium inside the separating region.

3. An apparatus as claimed in claim 2, wherein the viscous medium contains any from the group of glycerol, water, hydrocarbon oil, silicone oil, polyorganosiloxane and dichloroethane.

4. An apparatus as claimed in claim 2 or 3, wherein the viscous medium further contains a surfactant.

5. An apparatus as claimed in claim 4, wherein the surfactant is gum arabic polyvinyl pyrrolidone, PVP, or sodium dodecylbenzene sulphonate, SDS.

6. An apparatus as claimed in any of claims 2 to 5, wherein the viscous medium is adapted to cause metallic and semiconducting nanotubes to move in different directions when the electric field is applied.

7. An apparatus as claimed in any of the preceding claims, wherein the field is fixed in time.

8. An apparatus as claimed in any of claims 1 to 6, wherein the field is time varying.

9. An apparatus as claimed in claim 8, wherein the field oscillates at a fixed frequency.

10. An apparatus as claimed in claim 9, wherein the frequency is adapted to cause metallic and semiconducting nanotubes to move in different directions when the electric field is applied.

11. An apparatus as claimed in claim 8, wherein the field oscillates at a first frequency during a first period and at a second frequency during a second period.

12. An apparatus as claimed in claim 11, wherein the first frequency is higher than the second frequency.

13. An apparatus as claimed in claim 12, wherein the first frequency is in the range of O.lMHz - 1MHz and the second frequency is in the range of lOOHz -lOkHz

14. An apparatus as claimed in claims 11 or 12, wherein the second frequency is adapted to cause metallic and semiconducting nanotubes to move in different directions when the electric field is applied.

15. An apparatus as claimed in claims 11 to 14, wherein the electric field is created by a potential difference applied between the first and second electrodes, the potential difference for the first period being different than during the second period.

16. An apparatus as claimed in claim 15 in which the potential difference is smaller during the first period than during the second period.

17. An apparatus as claimed in claim 15, wherein the potential difference is in the range of 10V - 100V in the first period and in the range of 100V -

1000V in the second period.

18. An apparatus as claimed in any of the preceding claims, wherein the first electrode is elongated along a first dimension and the second electrode is spaced apart from the first electrode along a second dimension, perpendicular to the first dimension, and is smaller than the first electrode along the first dimension.

19. An apparatus as claimed in claim 18, wherein the first electrode is of rectangular shape.

20. An apparatus as claimed in claim 18, wherein the first electrode is in the form of an annular section which is concave with respect to the second electrode.

21. An apparatus as claimed in any of claims 18 to 20, wherein the second electrode is disc-shaped.

22. An apparatus as claimed in claim 21, in which the second electrode is situated on a radius of the first electrode.

23. An apparatus as claimed in claim 22 in which the second electrode is concentric with the first electrode.

24. An apparatus as claimed in claim 21, in which the second electrode is situated off a radius of the first electrode.

25. An apparatus as claimed in claim 21, wherein the radius of the second electrode is smaller than the distance between the first and second electrodes by a factor of more than 10.

26. An apparatus as claimed in any of claims 18 to 20, where the second electrode is in the shape of a triangle, for example an isosceles triangle having an apex facing the first electrode.

27. An apparatus as claimed in any of claims 1 to 17, in which the first and second electrode are of an elongated shape and their respective longitudinal axes intersect.

28. An apparatus as claimed in any of claims 1 to 17, in which the first and second electrode define between them a tapered space having a narrow end and an opposed wide end.

29. A dielectrophoretic separation apparatus for separating or sorting particles comprising a separating region and first and second electrodes arranged to produce a non-homogenous electric field in the separating region, whereby the first and second electrode define between them a tapered space having a narrow end and an opposed wide end.

30. An apparatus as claimed in claim 28 or 29, the first and second electrode further defining a second tapered space having a narrow end and an opposed wide end, the first and second spaces being connected at the respective narrow ends by a linear channel.

31. An apparatus as claimed in claim 30, the first and second electrode comprising trapezoid-shaped conductors.

32. An apparatus as claimed in claim 28 or 29, wherein the first and second electrode each comprise a plate, the plate defining the truncated triangular shape in a cross sectional plane and extending in a direction perpendicular to the cross sectional plane.

33. An apparatus as claimed in claim 30, wherein the first and second electrodes each comprise a profiled plate defining the shape as claimed in claim 30 in a cross sectional plane and extending in a direction perpendicular to the cross sectional plane.

34. An apparatus as claimed in claim 32 or 33, the apparatus further comprising flow means arranged to produce flow of a flow medium in a direction perpendicular to the cross sectional plane, a sample source for introducing the nanotubes into the flow medium and sample retrieval means for retrieving separated nanotubes.

35. An apparatus as claimed in claim 34 when dependent on claim 2, wherein the viscous medium is the same as the flow medium.

36. An apparatus as claimed in claim 34 when dependent on claim 32, wherein the sample retrieval means are situated at the narrow end and comprise one or more separators arranged perpendicular to the direction of field flow.

37. An apparatus as claimed in claim 34 when dependent on claim 33, wherein the sample retrieval means comprise one or more separators arranged perpendicular to the direction of fluid flow and situated within the linear channel.

38. An apparatus as claimed in claim 38, wherein at least one of the first and second electrodes comprises an aperture allowing retrieval of nanotubes from the linear channel.

39. An apparatus as claimed in any of claims 34 to 38, wherein the sample source and retrieval means are arranged for, respectively, continuous injection and retrieval of nanotubes, providing a continuous separation apparatus.

40. A method of separating metallic and semiconducting nanotubes, including placing the nanotubes in a separating region; and applying a non- homogenous field in the separating region.

41. A method as claimed in claim 40, further including suspending the nanotubes in a viscous medium.

42. A method as claimed in any of claim 41, further including adapting the viscous medium to cause metallic and semiconducting nanotubes to move in different directions when the electric field is applied.

43. A method as claimed in claim 40, wherein the field is fixed in time.

44. A method as claimed in claim 40, wherein the field is time varying.

45. A method as claimed in claim 44, wherein the field oscillates at a fixed frequency.

46. A method as claimed in claim 45, further including adapting the frequency to cause metallic and semiconducting nanotubes to move in different directions when the electric field is applied.

47. A method as claimed in claim 44, wherein the electric field oscillates at a first frequency during a first period and at a second frequency during a second period.

48. A method as claimed in claim 47, wherein the first frequency is higher than the second frequency.

49. A method as claimed in claim 48, wherein the first frequency is in the range of O.lMHz - 1MHz and the second frequency is in the range of 100Hz - 10kHz

50. A method as claimed in claims 47 or 48, further including adapting the second frequency to cause metallic and semiconducting nanotubes to move in different directions when the electric field is applied.

51. A method as claimed in any of claims 47 to 50, wherein the electric field is created by a potential difference applied between the first and second electrodes, the potential for the first period being different than for the second period.

52. An apparatus as claimed in claim 51 in which the potential difference is smaller during the first period than during the second period.

53. A method as claimed in claim 51, wherein the potential difference is in the range of 10V - 100V in the first period and in the range of 100V - 1000V in the second period.

54. A nanotube separating apparatus for separating nanotubes according to band gap, having a separating region and first and second electrodes arranged to produce a non-homogenous electric field in the separating region.

55. A method for separating nanotubes according to their band gap, including the steps of placing the nanotubes in a separating region; and applying a non-homogenous field in the separating region using a first and second electrode.

56. A method or apparatus as set out in the description of the preferred embodiment and the accompanying figures.