GB2409034A - Interaction-free object detection - Google Patents

Abstract

An interferometer includes a beam splitter BS1 which splits radiation into two paths U, D. The first path U includes a prism TIRU which reflects radiation by total internal reflection creating an evanescent wave at a sensing surface. Radiation from the paths U, D is combined by a second beam splitter BS2 which directs light towards one of two detectors Db, Dd. In the absence of an object O, total destructive interference of radiation from the paths U, D occurs at one detector Dd. The presence of an object O near the sensing surface of the first prism TIRu scatters the evanescent wave resulting in photons tunnelling out of the first prism TIRu This disturbs the total destructive interference at the detector Dd which detects photons with non-zero probability.

Description

1 2409034 Device for Detecting an Object The present invention relates to

a device for detecting an object, and in preferred embodiments may be applied to the digital imaging of an object or surface.

In recent times, digital imaging techniques have become increasingly important.

Applications of such techniques include non-contact and non-destructive surface profiling in industries where monitoring of surface topography and surface textures is of fundamental importance for critically finished surfaces, such as lenses and moulds, optical and magnetic disks, semiconductor wafers, data storage media, biomedical devices, automotive engineering, micromachining, plastics and polymers. Other applications include forensic imaging, microscopy, chemical and biosensing and medical Imaging.

The basic technique of all of these methods is the illumination of the target object with radiation of some kind, for example light, x-rays, gamma rays, microwaves, radio waves, ultrasound, electron or neutron beams. The radiation is made to interact with the object under investigation and the reflected or transmitted light is digitally recorded.

Information about the object and often its image is finally extracted from the recorded data and displayed on a screen with the help of computer software. However, such methods have an inherent limitation in that they cannot detect or image an object that is black, i.e. all absorbing, or photosensitive without damaging it.

This limitation has been addressed by utilising the properties of quantum radiation, for example a single photon beam. The basic idea is to detect the presence of objects and image them without the photon interacting with the object and being absorbed. Such methods have been termed "interaction-free" or "quantum interrogation" methods.

The ability to image an object without interacting with the object is of great significance for imaging highly absorbing and photosensitive materials, such as biological cells, emulsions for photography and holography, archival photosensitive materials such as paintings and ultracold materials such as Bose-Einstein Condensates, for example.

The only known "interaction-free" method is the Elitzur-Vaidman method (A. C. Elitzur and L. Vaidman, Foundations of Physics, Vol. 23 (7), 987-997 ( 1993)).

According to this method, a single photon is introduced into a MachZehnder interferometer and is directed by a beam splitter into one of two arms of the interferometer. The optical paths of the arms are recombined at a second beam splitter which directs the received photons to either a dark detector or a bright detector. In the absence of an object, the wave functions of the photons travelling along the two arms reach the dark detector in anti-phase and there is total destructive interference. If one of the arms of the interferometer is blocked by an object, the destructive interference is disturbed and photons are able to reach the dark detector. When single photons are introduced into the interferometer one at a time, the detection of a photon at the dark detector unambiguously indicates the presence of an object in one of the arms of the interferometer. However, the photon has not touched the object, because if it had it would have been absorbed. Hence, this detection method is considered "interaction-free".

In the Elitzur-Vaidman method the object to be imaged must be opaque and black (fully absorbing). This is because any light diffracted by an opaque object in one arm of the interferometer can enter the other arm, interfere with the light there and produce a phase-shift, invalidating the process. Furthermore, the object must be present inside the interferometer. This makes the method unsuitable for most practical applications. The intrinsic efficiency of the Elitzur-Vaidman method is at most 50% in principle, and this can only be increased by using the quantum Zeno effect, which is difficult to implement (see P. G. Kwiat, A. G. White, A. R. Mitchell, O. Nairz, G. Weihs, H. Weinfurter and A. Zeilinger, Physical Review Letters, Vol. 83 (23), 4725- 4728 (1999)).

In view of the above, the present invention provides a device for detecting an object, the device comprising a beam splitter arranged to split a beam of radiation from a radiation source between a first path and a second path, and a beam combiner arranged to recombine radiation from the first path with radiation from the second path and to direct the combined radiation along a detection path, wherein the first path comprises a refractive component arranged to reflect radiation from the beam splitter by total internal reflection at a sensing surface of the refractive component, thereby creating an evanescent wave at the sensing surface, wherein the refractive component is arranged such that, in use, an object proximate the sensing surface of the refractive component scatters the evanescent wave, and wherein the second path is configured such that radiation from the first path and radiation from the second path interfere when combined by the beam combiner, whereby a characteristic interference pattern is generated when the evanescent wave is scattered.

According to the invention, the reduction in amplitude and phase-shift of the reflected beam caused by frustrated total internal reflection at the sensing surface can be used to detect an object outside the device. Since the object is outside the device and not in one of the paths, as in the conventional Elitzur-Vaidman interferometer, the object can therefore be of a range of sizes and shapes. Furthermore, the evanescent waves which exist only very close to the sensing surface, allow sub-wavelength (or super-) resolution.

The object need only have a refractive index different from that of the surrounding medium, and need not be black (totally absorbing), as is necessary in the conventional Elitzur-Vaidman interferometer. Moreover, detection can occur without any contact between the device and the object.

Preferably, the second path is configured such that, in the absence of scattering of IS the evanescent wave, radiation from the first path and radiation from the second path interfere destructively along the detection path when combined by the beam combiner. In this case, the detection path forms a dark port which has a non-zero probability of receiving radiation only when the evanescent wave is scattered. This has the advantage that the detection of any radiation at all at the dark port is an indication of the occurrence of a refractive object in the proximity of the sensing surface, which greatly simplifies detection because there is effectively no background signal.

In order to achieve total destructive interference of the radiation from the two paths, the radiation from the two paths should be in anti-phase. Conveniently, the second path may comprise a refractive component arranged to reflect radiation from the beam splitter by total internal reflection. In this way, in the absence of an object, radiation in the second path undergoes compensating phase changes. Preferably, therefore, the refractive component of the second path is substantially identical to the refractive component of the first path.

The refractive component of the first path and/or the refractive component of the second path may be any suitable component. For example, the component(s) may be a prism or an optical fibre. One or each of the paths may include lenses, mirrors or other optical components.

In a convenient configuration, the beam combiner comprises a second beam splitter. The beam splitter and/or the second beam splitter may be a partially reflective mirror. In one embodiment, the device is configured as a modified Mach-Zehnder interferometer, in which the mirrors are replaced by total internal reflectors. Other interferometer arrangements may be used.

The device may comprise a radiation source or be arranged for connection to a radiation source. The radiation may be any suitable form of radiation, for example light or any form of electromagnetic radiation, an electron beam, a neutron beam or other particle beam.

The radiation source may be configured to release single quanta of the radiation, for example single photons. In this case, the detection of the object can be "interaction free". For practical purposes, the device can function well with both single quantum light and laser light. The interaction with the evanescent wave can be controlled and made as small as required and confined to a skin depth of the object, thus reducing power-induced damages. In one embodiment, the radiation source may be a laser, such as a continuous wave interrogating laser beam. The laser beam may be attenuated. In another embodiment, the radiation source may be a pulse laser. In this case, the pulses may be attenuated to contain less than one photon per pulse on average. For example, the pulses may contain on average less than 0.7, less than 0.5, less than 0.3 or even less than 0.1 photons per pulse.

The laser beam may have a relatively large diameter in order to obtain an image of the object without rastering the beam. Where the "interactionfree" nature of the device is not critical, unattenuated laser beams can be used to improve resolution. Any power- induced radiation damage may be minimised by locating the object in the tail of the exponentially damped evanescent wave.

The device may comprise a detector arranged to detect radiation on the detection path. Alternatively, the device may comprise a connection for a detector. A suitable detector is a charge coupled device (CCD).

The device may comprise a transport mechanism arranged to transport the sensing surface across the surface of the object. In this way, the device may be configured to image objects larger than the sensing surface.

Thus the device allows optical imaging of highly absorbing objects and photosensitive objects with the theoretically minimum possible damage and with super s resolution. It is ideal for safe medical diagnostics, archival imaging, non-contact, non- reflective surface profiling and testing in industry, and imaging in biomedical research and at ultra low temperatures.

An embodiment of the invention will now be described by way of example only and with reference to the accompanying drawings, in which: Figure I is a schematic representation of a Mach-Zehnder interferometer

according to the prior art;

Figure 2 is a schematic representation of an arrangement for the interaction-free detection of an object according to the Elitzur-Vaidman method of the prior art; Figure 3 is a schematic representation of a modified Mach-Zehnder interferometer according to an embodiment of the invention; Figure 4 is a schematic representation of the interferometer of Figure 3 in the presence of an object; Figure Sa is a schematic representation of the interferometer of Figure 3 incorporated in a chip; and Figure Sb is a schematic representation of the chip of Figure Sa including an input laser and output cameras.

According to the known Elitzur-Vaidman method for interaction-free detection (A. C. Elitzur and L. Vaidman, Foundations of Physics, Vol. 23 (7), 987- 997 (1993)), a Mach-Zehnder interferometer as shown in Figure 1 comprises a first beam splitter BS1 which receives input photons from a single- photon source, indicated by a star. The beam splitter BS1 either reflects the photons along a first path (or arm) U or transmits the photons along a second path (or arm) D. The probability of a photon being reflected or transmitted is in each case 50%. Along each path, the photons are reflected by respective mirrors Mu MD, towards a second beam splitter BS2. The second beam splitter BS2 either reflects photons from the first path U towards a dark detector (or port) Dd or transmits these photons towards a bright detector (or port) Db. Similarly, the second beam splitter BS2 either transmits photons from the second path D towards the dark detector Dd or reflects these photons towards the bright detector Db. In each case, the probability of a photon being reflected or transmitted is 50%, as for the first beam splitter BS1. In the absence of any obstructions in the arms U. D, photons reach the bright detector Db in phase and the dark detector Dd in anti-phase, so that one port Db is bright and the other port Dd is dark due to constructive and destructive interference, respectively.

If an opaque and black object O is present in one of the arms U of the interferometer, as shown in Figure 2, the object blocks photons in that arm by absorption and destroys the interference at the second beam splitter BS2. Consequently, the dark port Do will no longer remain dark but will record 25% of the photons, as only 50% of the photons traverse the unblocked arm D. Since only single photons are sent in, one at a time, a photon arriving at Dd cannot have traversed the path containing the object and interacted with it, for then it would have been absorbed by the object. Consequently, the photon must have traversed the path D and yet, it unambiguously indicates the presence of the object O in the other arm U. The method fails in 50% of cases in which the photon traverses the path U and is absorbed or scattered by the object O. A photon arriving at the bright port Db does not give any information about the presence of the object O. because photons arrive at the bright port Db even when the object is absent from the path U (see Figure 1). Thus, even in principle, with lossless mirrors and 100% efficient detectors, the efficiency of the method is only 25%, i.e. the fraction of input photons that give interaction -free detection.

In mathematical terms, after passage through the first 50-50 beamsplitter BS 1 the initial normalized radiation state loci> becomes: lair> = (1/312) [ei|ru> + eiX |1rD> ] due to a phase shift of on reflection and X on transmission, with <flu lyrU> = <lard lard> = 1 and <gru lord> = 0. In the absence of the object, this state evolves into: ( l /2) [e 2 i (e 2 i X |r,l> + e i % |grb>) + e i (em) (e i x Or> + e i |grb>)] = (1/2) [e % (l -1) lab> |Dd> + e x ( 1 + 1) | art> |Dh>] = <orb |grb> < Db | Do> This shows that the port Db is bright due to constructive interference and the port Do is dark due to destructive interference.

If an opaque and dark object is present in one arm, say U. the state prior to the second beam splitter BS2 is transformed into: lo> = ( 1/42) [e 2 i Or> Lao> + e i (x + I) |grD> ] Since Leo> and Lou> are orthogonal, interference is destroyed. The state then evolves to: (1/42)ei|u>l4O> +(1/12) ei(%+)|D> on reflection on MD, which leads to |> = (1/42) e Lou> Lao> + (1/2) e (x) [e lath> |Dh> + e % |d> |Dd>] after passage through the second 50-50 beam splitter BS2. Thus,

<+ 1 +, = (1/2) <flu 110> <TO two> + (1/4) [<pub | cab> < Db | Db> + <d | d> < Dd | Dd>] This shows that the photon is absorbed 50% of the time by the opaque object. The port Db remains bright with reduced probability and gives no information about the presence of the object because it fires whether or not the object is present. However, the port Dd is no longer dark, and every photon that is registered in this port unambiguously indicates the presence of the object even though it did not travel along the path U and interact directly with the object; if it did so, it would have been absorbed. The firing of the detector at the dark port Dd constitutes so-called "interaction-free detection", and its probability Ping= 0.25.

The efficiency of such detection is defined by 11 = Pifm / (Pifm + Pabs) where Pats is the probability that the photon is absorbed or scattered by the object.

The photons registered by Db do not count. If Rat and R2 are the reflectance and To, T2 the transmittance of the beam splitters BS I and BS2 respectively, then assuming a lossless beam splitter, Ping = To T2 and Pabs = Rat.

Therefore, q=T T2/(T T2+R) Since Ri + Ti = 1 (i = 1, 2) and one can choose T2 = R', = To / (1 + To) = l - R/(2 - Rat) This shows that q 0.5 as Rat 0. This is the maximum possible efficiency that can be obtained by adjusting the reflectance Ret. However, for a balanced interferometer where R' = T2 = 0. 25 and the intensities in the two arms are equal, the probability of an interaction-free measurement is at a maximum, Pifm = 0.25 and the efficiency is only 0.33.

It must be noted that it is possible to make the dark port Dd bright and the bright port Db dark by simply placing a completely transparent phaseshifter in one arm of the interferometer that introduces an additional phase-shift of it. Thus, the brightness of Dd does not by itself imply interaction-free detection. The object to be detected in an interaction-free way must be totally non-transparent or opaque and black for the Elitzur- Vaidman method to work. In practice, even an opaque object might scatter or diffract light that may interfere with light from the other arm of the interferometer to cause a phase-shift. Thus, as soon as there is some probability that a photon can be transmitted or diffracted by the object, it is no longer sensible to describe the measurement according to the Elitzur-Vaidman method as truly interaction-free (see A. G. White, J. R. Mitchell, O. Nairz and P. G. Kwiat, Physical Review A, Vol. 58, 605-613 (1998)).

According to an embodiment of the invention, as shown in Figure 3, the object does not need to be in one of the arms of the interferometer, nor need it be opaque and black. The interferometer according to this embodiment is modified relative to the interferometer of Figure 1 by replacing the two mirrors MU, MD in the two arms of the interferometer by two identical total internal reflectors TIRU and TIRD. In the absence of any object in the vicinity of the interferometer, the port Db will be bright and the port Dd will be dark, as in the Elitzur-Vaidman method. If an object O with a refractive index different from that of its surroundings is present outside the interferometer but close enough to one total internal reflector, say TIRE, as shown in Figure 4, the evanescent wave associated with the total internal reflection will interact with the object and a fraction of the photons will tunnel out of the interferometer. This will cause a phase- shift and reduction in amplitude of the reflected photon inside the interferometer, and the dark port Dd will no longer be dark. The object should be within the 'penetration depth' of the evanescent wave, which is of the order of the wavelength of the photons. Since single photons are sent in, one at a time, a photon arriving at Dd could not have interacted with the object and tunnelled out, for then it would have been lost from the interferometer. Yet the detection of a photon at the dark port Dd unambiguously indicates the presence of the object. The important point here is that although the photon wave function splits into an evanescent part and a reflected part at the total internal reflector, a single photon either interacts with the object and tunnels out, or it is entirely reflected. These are mutually exclusive alternatives. In the latter case, whose probability can be arranged to be very high, light practically does not interact with the object. This is impossible with multi- photon states or classical light because with such light reflection and evanescent wave interaction are concurrent and not mutually exclusive. Detection with such light states is not therefore "interaction-free".

Unlike in the Elitzur-Vaidman method, only the photons in the arm U near which the object must be present are involved in the actual detection, and not those in the arm D. The absorption or scattering of these photons can be made arbitrarily small in principle by placing the object near the tail end of the exponentially damped evanescent wave. The method is therefore nearly 100% efficient in principle with ideal mirrors and detectors. It is highly sensitive to the presence of objects within the evanescent wave because it makes use of high visibility interference phase-shifts. It is also highly sensitive to surface topography and the refractive index profile because of the exponentially varying nature of the evanescent wave. Furthermore, a very important property of evanescent waves is that they are able to distinguish very fine spatial variations in an object because of high momentum components of the evanescent wave parallel to the surface. They are said to give sub-wavelength or super- resolution.

Because frustrated total internal reflection is used in one arm of the interferometer, the leakage (or tunnelling) of light through the exponentially damped evanescent wave outside the sensing surface can be controlled and made arbitrarily small by placing the object at a suitable distance from the sensing surface. Thus, power induced damage to the object can be controlled and minimised. The probe radiation can be controlled not to enter into the bulk of the materials to be detected or imaged but to stay confined near the surface of the object within a skin depth of the order of the wavelength. Thus, power-induced radiation damage can be controlled and eliminated for all practical purposes, making the method safe and attractive for medical diagnostics.

The properties of evanescent waves will be explained in more detail in the following. Considering two materials with refractive indices nj and no with nit, no, it is well-known that total internal reflection (TIR) occurs at the boundary of these materials when light is incident on the boundary from the material with the higher refractive index nj at an angle Oi greater than the critical angle: Oc= sin -I no / nil.

The wave does not, however, vanish in the second medium with refractive index no but is exponentially damped: VEv(X'y) = llr(x,O) e with the 'penetration depth' 5, (ni, no, \) = Hi / 2 <(nit sin2 Gi - nit) = ki / 2 ni l( sin2 Oi - sin2 Oc) taking the x-axis along the boundary and the penetration in the transverse direction y. The wave in the second medium, with lower refractive index is called an 'evanescent wave'. The electric and magnetic vectors E and B in an evanescent wave are in time quadrature, and so the Poynting vector (c/4) (E A B) vanishes.

If a material with a refractive index n, comes within the 'penetration depth' it, of the evanescent wave, it scatters the wave, i.e. the electric and magnetic vectors are no longer in time quadrature, a part of the energy leaks (tunnels) out across the boundary and propagates parallel to the boundary, frustrating total internal reflection. Thus, for fixed Bi and ni, any roughness of the surface of the material with lower refractive index (variation in y) or inhomogeneity in its refractive index (variation of n, along x) will be reflected as intensity variations in the beam cross-sections recorded by the ports of the modified Mach-Zehnder interferometer according to the invention. Por this purpose, attenuated continuous wave laser beams should be used for imaging.

There are two significant features of evanescent waves in the context of the present invention. Firstly, the component of the momentum perpendicular to the boundary surface is imaginary, which is why the wave is exponentially damped and non radiating. This implies through momentum conservation across the boundary that the momentum components of the evanescent wave parallel to the surface are large. High momentum components imply small spatial dimensions and high resolution. Traditional microscopes, on the other hand, operate in far-field conditions where diffraction effects limit the resolution to a maximum of half a wavelength.

Secondly, the expression above for the penetration depth it, shows that for a fixed wavelength X, the penetration depth increases indefinitely as Di approaches Dc. Thus, it is possible to adjust the penetration depth by varying the angle of incidence and make it sufficiently large when required. This is particularly simple when prisms are used as total internal reflectors. Optical fibres may be specially modified to take advantage of this feature.

The above analysis also applies to photons which are quantum mechanical objects (see P. Ghose and M. K. Samal, Physical Review B. Vol. 64, 036620 (2001)). In this case there is tunnelling across a barrier, which is a well- understood quantum mechanical phenomenon. Evanescent wave calculations for finite size beams and references to related results are reported by R. A. Cornelussen, A. H. van Amerongen, B. T. Wolschrijn, R. J. C. Sprecuw and H. B. van den Henvell, arXiv:physics/0205014 v2 (2002) Moreover, a method of amplifying evanescent waves using negative refractive index materials has been reported by J. B. Pendry, Physical Review Letters, Vol. 85, 3966 (2000) (see also X. S. Rao and C. K. Ong, arXiv:cond-mat/0304133 v2 (2003) ), in which evanescent waves grow exponentially. This method may be used to extend the penetration depth 6, considerably.

In mathematical terms, according to the embodiment of the present invention, as in the Elitzur-Vaidman case, after passage through the first beam splitter BS 1 the state of the photon is |> = (1/42) [e i MU> + e x IY[D> ] On the total internal reflectors in the two arms this evolves to: (1/42) e b [e i Blue > + e i x |14DE > ] + (1/112) e i a [e i Lou > + e i x |VD> ] where Lou E > and IT'D E > are evanescent wave states and I a 12 + lb 12 = 1.

Since the evanescent waves disappear when the photon moves on, the state evolves to: ( l //2) e i [e i Lou> + e i x |\rD> ] - |≥(1/2) ei[ei (em> +eiX |b>)+ eiX (eiXlJd> + e|>)] after passage through the second 50-50 beam splitter BS2 = (1 /2) em [e2iX ( 1 - 1) Aid> - ei x (1 + 1) Fib> ] Thus, <+ | l> <lab | lob> < Db | Db> = 1 Thus the port Db will be bright and the port Dd dark because of interference. This is the initial balanced state of the interferometer.

However, when an object O with a refractive index is near one of the total internal reflectors, say in arm U. the evanescent wave interacts with the object and a fraction of the photons leak or tunnel out of the interferometer. The amount of leakage (or tunnelling) is determined by the distance d of the object from the totally internally reflecting surface and is exponentially damped: b=exp(-d/(ni,nt,\)) In this case, total internal reflection is partially frustrated in the arm U. and the state let> becomes (e i (T + a) b /42) let's> Lao> + (1/12) e i [e i a Lou> + e i x |D> ] after the total internal reflectors ( l Do> being the object state) and l+ > = (eider A) b/12)|yruE>|O> + (1/2) em [em a (e2iX laid> + eix lath>) + eiX (e|> + em |b>) ] after passage through the second 50-50 beam splitter BS2, = (eider A) |b/2) |uE> No> + (1/2) em [e2iX ( 1 - a) |d> - ei x (a+ 1) |b> ] Thus, <if I l> = ( I b I /2) (< Mu 1< No | Vu | No > + ( 1 /4) [I 1-a I <Vd | d> < Dd | Dd> + | 1 + a I <fib | pub> < Db | Db>] Thus, the initial balance is lost and the dark port Dd registers photons that did not tunnel out of the interferometer and interact with the object, and yet unambiguously indicates the presence of the object outside the interferometer.

It is possible to make b O by placing the object to be imaged near the tail of the evanescent wave (d >> 6,), increasing the efficiency of the method, in principle, as close to 100% as desired.

The visibility at the port Do is unity because V (Dd) = (Imax-Imin) / (Imax + Imin) = 1 simply because Imjn = 0 in the ideal case. This formula for the visibility of interference fringes is not applicable to the Elitzur-Vaidman method because of the complete destruction of interference. Hence the necessity to define efficiency of the method in terms of the fraction T1 of photons that give IFM. In terms of such a definition also the efficiency of the device according to the invention can be made close to 100% by making Pabg go to zero by having the object near the tail end of the evanescent wave.

If coherent radiation is used as the interrogating beam, phase sensitivity is only bounded by the so-called Shot-noise limit Vp x 1 /4N where N is the mean photon number of the interrogating beam. This limitation on the accuracy of the phase-shift measurement can be improved by increasing the number of paths. The phase sensitivity rescales as V<px 1/M<N where M isthe number of paths and the non-unit quantum efficiency at the detectors only decreases the effective M by a factor of (see G. M. D'Ariaro and M. G. A. Paris, Physical Review A, Vol. 55, 2267- 2271 (1997) ).

In practice, the entire interferometer can be incorporated in a chip with the help of integrated optoelectronic techniques, as shown in Figures 5a and 5b. Chips of this l 0 general kind are known, for example, for biosensing (see B. Schneider, E. Dickinson et al, Biosensors Bioelectron, Vol. 15, 597 (2000)).

An optical interferometric circuit may be fabricated on a small chip using techniques of integrated optics involving thin film deposition on a silicon or glass substrate. Freely propagating light waves can be replaced by guided waves in planar single-mode thin (of approx 40 nm thickness) waveguiding films deposited on a silicon wafer substrate with first a buffer layer, for example SiO2, then a core (SiO2/GeO2) and finally a cladding layer (SiOJB/P). The guided mode of light possesses an evanescent field distribution that decays exponentially into the buffer below and cladding above. The' widths of these three layers can be controlled and made to design depending on the end use. Integrated optical total internal reflectors and 50-50 beam-splitters can also be inserted into the circuit. Further, in order to couple laser light into the waveguides, couplers' are integrated into the supporting substrate. The input coupler couples laser light from a diode laser to the waveguide core (approx 4-6 microns x 4-6 microns) with typically 80% efficiency. The light travelling within the core is finally scattered by the output coupler into the space above the waveguide surface. This light is then picked up by CCD cameras placed a few millimetres from the waveguide surfaces. Measurement of propagation losses in such waveguides with light injection by grating couplers shows only 0.5 dB/cm loss in excess ofthe loss by light leakage to the substrate (A. Koster, D. Pascal and S. Laval, IEP, University Paris Sud). Other coupling devices may also be used such as an integrated microlens or a pism (evanescent wave coupling).

As shown in Figure 5b, pulses from a laser pass through neutral density filters so as to contain no more than 0.1 to 0.3 photons per pulse. The light is then coupled to the interferometer circuit with the help of an input coupler. The edge of the silicon wafer containing one total internal reflector TIRE is coated with a highly reflecting thin film of thickness less than the penetration depth 5, of the evanescent photons, leaving the face of the total internal reflector TIRU exposed to act as the sensor. The precise thickness of this coating should be varied to ensure that the object to be detected and imaged is at the right distance from the TIRU face to reduce photon loss through tunnelling to the optimum level and also reduce the power induced damage to a photosensitive material sample whenever necessary. This coating also helps shield the chip from stray background radiation entering into the optical circuit.

The photons from the output couplers are detected by CCD cameras and sent to a computer for conversion to images with the help of available software packages.

Software is already available to convert interference data into quantitative 3-D images, for example the Windows NT-based MetroPro_ comprehensive metrology analysis software available from Zygo Corporation of Middlefield, CT, USA.

The assembly may be provided with precision vertical (z) and horizontal (x-y) scanning facilities to cover the entire sample. Existing methods of surface profiling that make use of reflected light can be adapted to incorporate the device of the invention to provide them with superresolution. With the device of the invention, highly absorbing and nonreflecting materials can be imaged. Existing commercial products include the ZYGO NewView 5000 3-D Surface Profilers available from Zygo Corporation of Middlefield, CT, USA and the Chapman MP3100 and MP2000 Plus Surface Profilers available from Chapman Instruments of Rochester, NY, USA.

The device according to the invention has a large number of possible applications.

For example the measurement and control of surface textures and noncontact 3-D topography (surface profiling) are critical in many industrial applications to ensure that components perform as intended. In view of the possible super-resolution and minimum power damage capabilities, the device can be applied to the imaging of data storage and magnetic materials, optical parts and lenses, optical disks, computer hard disks, silicon wafers, ophthalmic lenses and moulds, cornea imaging, any critically finished surface, biomedical devices, automotive engineering, semiconductors and microelectronics, micromachining, plastics and polymers, materials research and surface characterization, medical imaging, holography, compact disc technology (to prolong compact disc lifetimes, imaging highly absorptive and photosensitive samples, imaging cold and ultra cold matter states such as cold trapped atoms (see R. A. Cornelussen, A. H. van Amerongen, B. T. Wolschrijn, R. J. C. Spreeuw and H. B. van den Henvell, arXiv:physics/0205014 v2 (2002)), Bose- Einstein Condensates, etc. In particular, the interference generated within the interferometer can be utilized to generate holograms.

The device is capable of providing surface roughness and refractive index profiles of an object, which is not possible with the Elitzur-Vaidman method, for example.

Moreover, no mechanical contact between the material to be detected or imaged and the device is necessary, eliminating all possibilities of contamination and damage. The device can be used to study very fast phenomena or moving biological specimens. The to device has a high signalto-noise ratio because it does not require a small aperture to produce a small spot.

In summary, a device for detecting an object includes a beam splitter which splits a beam of light from a source between a first path and a second path. The first path includes a first prism arranged to reflect light from the beam splitter by total internal lS reflection at a sensing surface of the first prism. The total internal reflection creates an evanescent wave at the sensing surface. The second path includes a corresponding second prism. A second beam splitter recombines light from the first path with light from the second path and directs the combined light towards one of two detectors. The first and second paths are configured such that light from the first path and light from the second path interfere constructively in the direction of one detector and destructively in the direction of the other detector when no object is present so that no light is received by the detector. When an object nears the sensing surface of the first prism in the first path, the object scatters the evanescent wave and allows a small fraction of photons to tunnel out of the first prism. This disturbs the total destructive interference at the dark detector, which therefore detects photons with a non-zero probability. The detection of a photon at this detector indicates the presence of an object in the proximity of the sensing surface of the first prism. When single photons are made to pass through the device one at a time, the object can be detected without the photon touching it and being scattered or absorbed.

Claims (13)

1. Clanns I. A device for detecting an object, the device comprising a beam

   splitter arranged to split a beam of radiation from a radiation source between a first path and a second path, and a beam combiner arranged to recombine radiation from the first path with radiation from the second path and to direct the combined radiation along a detection path, wherein the first path comprises a refractive component arranged to reflect radiation from the beam splitter by total internal reflection at a sensing surface of the refractive component, thereby creating an evanescent wave at the sensing surface, wherein the refractive component is arranged such that, in use, an object proximate the sensing surface of the refractive component scatters the evanescent wave, and wherein the second path is configured such that radiation from the first path and radiation from the second path interfere when combined by the beam combiner, whereby a characteristic interference pattern is generated when the evanescent wave is scattered.
2. 2. A device as claimed in claim 1, wherein the second path is configured such that, in the absence of scattering of the evanescent wave, radiation from the first path and radiation from the second path interfere destructively along the detection path when combined by the beam combiner.
3. 3. A device as claimed in claim I or 2, wherein the second path comprises a refractive component arranged to reflect radiation from the beam splitter by total internal reflection.
4. 4. A device as claimed in claim 3, wherein the refractive component of the second path is substantially identical to the refractive component of the first path.
5. 5. A device as claimed in any preceding claim, wherein the refractive component of the first path and/or the refractive component of the second path is a prism.
6. 6. A device as claimed in any preceding claim, wherein the beam combiner comprises a second beam splitter.
7. 7. A device as claimed in any preceding claim, wherein the beam splitter andlor the second beam splitter is a partially reflective mirror.
8. 8. A device as claimed in any preceding claim comprising a radiation source.
9. 9. A device as claimed in claim 8, wherein the radiation source is configured to release single quanta of the radiation.
10. 10. A device as claimed in any preceding claim, wherein the radiation is electromagnetic radiation.
11. 11. A device as claimed in claim 10, wherein the radiation is light.
12. 12. A device as claimed in any preceding claim comprising a detector arranged to detect radiation on the detection path.
13. 13. A device as claimed in any preceding claim comprising a transport mechanism arranged to transport the sensing surface across the surface of the object.