WO2006019776A2

WO2006019776A2 - A method for coupling terahertz pulses into a coaxial waveguide

Info

Publication number: WO2006019776A2
Application number: PCT/US2005/024783
Authority: WO
Inventors: Daniel Mittleman; Kanglin Wang
Original assignee: William Marsh Rice University
Priority date: 2004-07-14
Filing date: 2005-07-13
Publication date: 2006-02-23
Also published as: US20080309577A1; US9178282B2; WO2006019776A3

Abstract

A system for coupling teraherz (THz) radiation to a coaxial waveguide comprises an antenna that generates THz radiation having a mode that matches the mode of the waveguide. The antenna may comprise a pair of concentric electrodes, at least one of which may be affixed to or formed by one end of the waveguide. The radiation may have wavelengths between approximately 30 µm and 3 mm. The waveguide may comprise an inner core and an outer wall defining an annular region. A terahertz sensor system may comprise a terahertz antenna comprising first and second concentric electrodes, means for generating a field across the trodes and means for triggering the emission of terahertz radiation, a first waveguide having first and second ends, said first end being coupled to said antenna so as to receive at least a portion of said terahertz radiation, and a sensor for detecting said terahertz radiation.

Description

A METHOD FOR COUPLING TERAHERTZ PULSES INTO A COAXIAL

WAVEGUIDE

TECHNICAL FIELD OF THE INVENTION

The present invention provides a method for guiding terahertz pulses over practical distances with low loss and low dispersion.

BACKGROUND OF THE INVENTION

Rapid advances in laser technology have enabled various techniques for the generation and detection of electromagnetic radiation in the terahertz region (spanning from -100 GHz to -10 THz, or wavelengths between -30 μm and -3 mm). As a result, numerous uses of terahertz radiation have been explored, including trace gas detection, medical diagnosis, security screening, and defect analysis in complex materials such as space shuttle tiles. Many of these studies have relied on terahertz time-domain spectroscopy, a technique for generating sub- picosecond pulses with spectral content spanning much of the THz band.

However, progress is limited by the overwhelming reliance on free-space transport of the terahertz beam, using bulk optical components. In many real-world situations, the sample or region to be studied may not be readily accessible to a line-of-sight beam. Hence, common devices that operate at other wavelengths, such as optical fiber-based sensors or medical endoscopes rely on the guided wave delivery of light to the remote sensing location. In addition, while THz waves can be transmitted simply by free space propagation, free space propagation requires bulk optical components, which are difficult to align.

Thus, in order to expand the usefulness of THz radiation, it is desirable to provide optimized guided wave devices that operate at THz frequencies. The development of practical THz waveguides would dramatically expand the application of THz-TDS in areas such as gas sensing and nanometer thin-film measurements. Heretofore, the development of THz waveguides has been hindered by the material properties and the application requirements in this spectral range. On the one hand, the characteristics of materials at THz frequencies make it extremely difficult to build a fiber to guide THz beams over a long distance. The most transparent materials for this range are crystalline (e.g., high resistivity silicon), and thus are costly, fragile, and challenging to form into specific geometries for waveguide configurations. Other materials, such as low-loss polymers or glasses, are more malleable but exhibit prohibitively high absorption losses for propagation distances of more than a few centimeters. For this reason, THz waveguides generally must rely on propagation in air, rather than via dielectric confinement as in an optical fiber.

On the other hand, many THz applications rely on the use of broadband pulses for time- domain analysis and spectroscopic applications. To avoid pulse reshaping during propagation, low dispersion is required. But for many conventional metal waveguides (e.g., metal tubes), pulse reshaping in propagation is difficult to avoid, due to the extreme dispersion near the waveguide cutoff frequencies. Furthermore, finite conductivity of metals can lead to considerable losses in the wave propagation.

Great efforts have been devoted to finding useful THz waveguides within the last few years, and various guides with quasi-optical coupling have been demonstrated. Most of these THz waveguides have been based on conventional guiding structures, such as metal tubes, plastic ribbons, or dielectric fibers. There have also been reports on the application of the latest technology of photonic crystal fibers to THz radiation. In all of these cases, the utility for transport of THz pulses is limited by group velocity dispersion of the guided waves.

The most promising studies have reported dispersionless propagation in parallel metal plate waveguides. One type of dispersionless waveguide design has been discussed recently by Grischkowsky and co-workers. This design is a ribbon waveguide, which is dispersionless and low-loss. In the Grischkowsky design, the loss is attributable to two factors: (1) lateral spreading due to the fact that the mode is unconfined in one of the two transverse dimensions, and (2) the finite conductivity of the metal used to confine the mode, which in this case results in a reported attenuation of -80 dB/m.

Coaxial waveguides, have not heretofore been considered, due to the difficulties in coupling the radiation into the guide. This is because linearly or circularly polarized light cannot be effectively coupled into a coaxial waveguide. In particular, coaxial waveguides have not been used previously at frequencies above a few GHz, because of the difficulties in coupling the radiation into the waveguide efficiently. Hence, it remains desired to provide a waveguide that is effective at THz frequencies. SUMMARY OF THE INVENTION

The present invention provides an effective THz waveguide and a method and apparatus that allow very efficient coupling of THz energy into the waveguide. The present devices are compatible with existing terahertz generation and detection techniques. In some embodiments, the invention includes the use of a coaxial antenna and a coaxial or uniaxial waveguide. Because the present invention in some embodiments uses a mode-matched detector as a mode filter to eliminate the influences of group velocity dispersion, it can provide coupling and transmission efficiencies that have not heretofore been available.

The improved coupling is expected to be sufficiently effective to make coaxial waveguides feasible. This is quite significant because coaxial guides are one of the very few options that can propagate pulses with no group velocity dispersion.

The present invention further includes a metal waveguide with very simple geometry that can be used to guide broadband THz pulses with outstanding performance, including low loss and negligible group velocity dispersion. The guided propagation of THz pulses on a metal wire behaves similarly to the transverse electromagnetic (TEM) mode of a conventional coaxial waveguide. Since the exposed surface area of a wire is much smaller than that of any previously reported metal waveguide, the attenuation due to conductivity losses is extremely low for this configuration. The efficacy and structural simplicity of the present wire waveguide present great advantages in the manipulation of guided THz radiation. BRIEF DESCRIPTION OF THE DRAWINGS

For a detailed description of exemplary embodiments of the invention, reference will now be made to the accompanying drawings, in which:

Figure 1 is a schematic diagram of a coaxial antenna constructed in accordance with a first embodiment of the invention; Figure 2 is a schematic diagram of the coaxial antenna of Figure 1 showing the mode of generation of radiation

Figure 3 is a schematic diagram of a coaxial antenna coupled to a coaxial waveguide;

Figure 4 is a schematic diagram of an alternative embodiment of an antenna coupled to a coaxial waveguide; Figure 5 is a schematic diagram of an alternative embodiment of a coaxial waveguide;

Figure 6 is a schematic diagram of a coaxial antenna coupled to a uniaxial waveguide;

Figure 7 is a schematic diagram of one embodiment of a coaxial waveguide adapted to serve as a sensor;

Figure 8 is a schematic diagram of one embodiment of a THz probe; Figure 9 is a schematic diagram of one embodiment of a THz endoscope;

Figure 10 is a schematic diagram of one embodiment of a THz near-field scanning optical microscope;

Figure 11 is a schematic diagram of a second embodiment of a THz waveguide;

Figure 12 is a plot of arrival time as a function of incident position for a series of time- domain THz pulses propagating along a bare metal probe (squares) and a probe wrapped with a 0.5 mm PVC insulation layer (triangles), as detected in a THz apertureless near-field scanning optical microscopy system;

Figure 13 is a schematic diagram of an experimental setup for characterization of a THz wire waveguide; Figure 14 is two plots showing the spatial profile of the propagating mode on a metal wire waveguide;

Figure 15 is two plots illustrating the group velocity of propagating THz pulses on a metal wire waveguide; Figure 16 is a plot of the amplitude of a THz pulses as a function of the vertical displacement of the receiver, measured at different propagation distances;

Figure 17 is two plots illustrating the attenuation characteristic of the guided wave on a metal wire waveguide;

Figure 18 is a plot illustrating THz wave propagation on a 21 cm long wire waveguide with different bend radii R;

Figure 19 is a schematic diagram of a simple Y-splitter structure, comprising a straight waveguide and a curved waveguide in contact with each other; and

Figure 20 is a plot of the THz waveforms detected at points A, B and C in Figure 19.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

Certain terms are used throughout the following description and claims to refer to particular system components. In the following discussion and in the claims, the terms "including" and "comprising" are used in an open-ended fashion, and thus should be interpreted to mean "including, but not limited to... ." As used herein, a waveguide is intended to refer to a device designed to confine and direct the propagation of electromagnetic waves. Antenna

In order to efficiently couple radiation into a waveguide, it is desirable to match the spatial pattern (the mode) of the incident radiation to the mode of the waveguide. For a coaxial waveguide, this requires that the incident beam be radially polarized, since this is the characteristic of the waveguide mode with the lowest loss. Such a polarization profile is typically very difficult to generate for free-space radiation, which is why coaxial waveguides have not often been used at high frequencies. The present invention uses components similar to those used in time-domain spectroscopy (TDS) to generate a radial mode profile through the use of a novel, radially symmetric antenna.

A first preferred antenna configuration is illustrated in Figure 1. A first, relatively narrow lead 10 extends in a first direction and terminates in a small circular electrode 12. A second lead 20 extends in an opposite direction and terminates in an annular electrode 22 that is concentric with circular electrode 12. While electrodes 12, 22 are shown as circular, they could alternatively be elliptical, oblong, polygonal, or any other non-circular shape. Likewise, lead 10 could extend farther, or even across the diameter of annular electrode 22, so as to provide another axis of symmetry in the antenna.

Electrodes 12, 22 are preferably formed on or adjacent to a semiconducting substrate so as to create a switch. A dc bias is placed across the antenna by applying a voltage across leads 10, 20 and an ultrashort pump-laser pulse (<100 femtosecond (fs)) is focused on the annular region 30 between electrodes 12, 22. The bias-laser pulse combination allows electrons to rapidly jump the gap, so when this region is excited by the application of a fs laser pulse the resulting current in the semiconductor produces a terahertz electromagnetic wave. This antenna operates in the same manner as a typical THz emitter antenna (commonly used in THz-TDS), except that the generated mode preserves the cylindrical symmetry of the antenna pattern and is therefore radially polarized, as illustrated by the arrows in Figure 2. This radial mode is predicted to couple extremely efficiently (up to at least -56%) into an optimized coaxial waveguide.! Coaxial waveguides are discussed in detail below. In this context, optimization of the waveguide entails selecting the dimensions of the coaxial waveguide (i.e., the inner and outer radii) to match the size of the radially polarized mode. This selection is preferably based on calculations of the mode matching for a given THz beam spot size.

It is possible to use a similar antenna design for photoconductive detection of the THz pulses emerging from a waveguide. This has the unique advantage that only the fundamental transverse electromagnetic (TEM) mode of the guide are detectable, so even if the launched wave were propagating in a multi-mode regime, this would not lead to measurable group velocity dispersion but only to increased propagation losses. Such considerations are important for any spectroscopic or imaging applications in which coaxial guides would be employed. Coaxial Waveguide

As shown in Figure 3, a coaxial waveguide 40 is preferably constructed with an inner core 46 and an outer wall 48 constructed from a material that is substantially opaque to terahertz radiation such as copper or any other highly conductive material. The annular region 44 between inner core 46 and outer wall 48 is preferably either evacuated or filled with a substance that is transparent or substantially transparent to terahertz radiation, such as a gas. Further details relating to a preferred construction for the waveguide are set out in detail below.

Still referring to Figure 3, the radially polarized beam 32 generated by an antenna, such as that described above, can be coupled to a coaxial waveguide in any suitable manner. In one embodiment, beam 32 may be collimated and focused on one end 42 of coaxial waveguide 40. Beam 32 is preferably focused by a lens 50 such that a significant portion of its energy is incident on annular transmitting region 44. It is preferred that beam 32 not be focused too tightly, as that will result in too much of the energy being incident on core 46. Hence, it is preferred that the degree of focus be optimized. This optimization can be performed experimentally or mathematically.

In alternative embodiments, one or both of the electrodes of the antenna may be constructed on, attached directly to, or integral with the end of the waveguide. In these embodiments, circular electrode 12 may be attached to or formed by the end of the core of the waveguide and/or annular electrode 22 may be attached to or formed by the end of the outer wall of the waveguide. These embodiments avoid the need for collimation and/or focusing while providing coupling at a very high efficiency.

In an exemplary embodiment, shown in Figure 4, an annular electrode 20 is spanned by a web of semiconducting material 25 and a lead 10. is provided to the end of core 46, which serves as the inner electrode. The fs laser pulse is applied to semiconductor 25 while a voltage is applied across the two electrodes. In other variations (not shown), the inner electrode rather than the outer electrode may be separate from the waveguide and may support a radially extending semiconducting web. Regardless of whether one, neither, or both of the electrodes is separate from the waveguide, a sheet or web of semiconducting material extends normal to the longitudinal axis of the waveguide and spans its cross-sectional area. Also, regardless of the configuration of the electrodes and waveguide, it is not necessary that the dimensions of electrode 12 coincide with those of the core of the waveguide or that the dimensions of annular electrode 22 coincide with those of the outer wall of the waveguide.

The emission properties of the antenna (the mode pattern and the fraction of the radiation coupled into the high-dielectric substrate) are preferably optimized with respect to the geometrical factors (inner and outer radii, line thicknesses, etc). This optimization can be performed numerically, using commercially available finite element analysis software or any other suitable algorithms.

Fabrication techniques suitable for making the present antennas or antenna/waveguide combinations are similar to the techniques used to make conventional THz dipole antennas. Once fabricated, each antenna is coupled to an femtosecond optical excitation beam, preferably using fiber delivery (as has been done with some dipole antennas). An additional step may involve coupling or attaching these antennas directly to the input and output faces of the waveguide as described above, in order to achieve the highest possible levels of energy coupling into the guided mode. Some embodiments of the waveguide include a core and outer wall, as described above. In addition, because some embodiments include an evacuated or gas-filled annulus, it is preferred in those embodiments to provide means for maintaining the core at the center of the waveguide. It is preferred that the inner conductor be positioned as precisely as possible along the axis of the outer cylinder so as to maximize symmetry of the waveguide and efficiency of transmission. This can be accomplished in any of several ways.

One approach, shown in phantom in Figure 5, is to use a plurality of cylindrical plugs 60, spaced apart along the length of the waveguide cylinder, with holes 62 drilled in their centers through which the inner conductor is inserted. If these holes are positioned accurately and plugs 60 are sized correctly, plugs 60 will hold inner conductor 46 in precise coaxial alignment with outer wall 48. Plugs 60 are preferably composed of a material that has a low refractive index and is as nearly invisible to THz radiation as possible, such as rigid polystyrene foam, other polymeric foams, or, less preferably, non-foamed materials. Other means for supporting core 46 may be used, including but not limited to braces, fins, and legs. In preferred embodiments, the waveguide is substantially straight. In alternative embodiments, the axis of the wave guide may include one or more curves, however it is expected that a curvilinear waveguide will result in some loss of efficiency, as some of the transmitted energy is lost to modes that are not compatible with the waveguide. More specifically, curves in the waveguide tend to convert from the TEM mode to other modes. These other modes are not detected by the radial receiver antenna, with the result that the detected energy is lower. Thus, bends introduce loss. It may be possible for the bend to also introduce dispersion, which is typically undesirable. Dispersion would involve coupling from the TEM mode into other modes, and then back again. Uniaxial Waveguide In still another embodiment, a waveguide for THz radiation may comprise only a conductive core 46. Because a significant portion of the losses in a coaxial waveguide are the result of the finite conductivity of the outer wall, while only a portion of the losses are due to dispersion (diffractive spreading), a waveguide comprising only a core 46 may transmit THz radiation sufficiently for some purposes, particularly over short distances (less than a few meters). It is further expected that such a uniaxial waveguide may include some degree of curvilinearity. Uniaxial waveguides in accordance with the invention may be constructed of any suitably conductive material, including but not limited to bare metal wire. In some embodiments, a bare metal wire with a thin dielectric coating may be advantageous in confining the mode closer to the surface of the wire. The material and diameter of the core may be optimized according to the desired operating parameters such as wavelength and transmission distance but is preferably between 0.1 and 30 mm in diameter, more preferably between 0.1 and 20 mm, and still more preferably 0.5 and 10 mm in diameter. The uniaxial waveguides may also be used in plural, in a THz sensor such as that shown in Figure 8 and described below. A uniaxial waveguide in accordance with the present invention, sometimes referred to herein as a "wire waveguide," may be coupled with a THz antenna in the manner(s) described herein. For example, the center one of a pair of concentric electrodes can be affixed to the end of the waveguide or a radially polarized beam can be focused on the end of the waveguide. The wire waveguide could also be coupled to the THz antenna through the use of a substrate lens, similar to the type of lens typically used in terahertz time-domain spectroscopy. Alternatively, a linearly polarized THz beam may be focused on a scattering object placed near the waveguide. If the beam is sufficiently intense, enough scattered energy will be coupled into the waveguide modes and transported along the waveguide to serve as a detectable signal. This latter technique is illustrated in Figure 12. Finally, one could couple free-space radiation onto the waveguide through the use of a grating coupling scheme, in which a periodic modulation of some property of the wire (e.g., the diameter) is used to diffract incident radiation along the wire axis, thus exciting the guided mode. Applications

The present invention has many possible uses. First, the use of guided terahertz pulses eliminates the need for free-space optical components, which vastly simplifies the alignment of a terahertz spectrometer. This makes the use of a terahertz system far more realistic for many applications, particularly those for which sensitive alignment is problematic. A good point of comparison in this case is the Fourier Transform Infrared (FTIR) spectrometer, a device found in virtually all undergraduate analytical laboratories, at every university in the US. If a terahertz system were as easy to use as an FTIR, one could imagine that it could also be used as a teaching tool in similar fashion.

In the present coaxial system, lateral spreading is eliminated because the mode is confined, but the finite conductivity of the metal remains as limitation on the efficiency of the waveguide. This constraint could be minimized by using materials with very high conductivity such as copper or silver, or some other lower-conductivity material coated with a thin layer of copper or silver. In the latter case it is preferred that the thin layer be thicker than the skin depth of the high-conductivity material, which is roughly 1 micron. In addition, because the present coaxial waveguide can have a small cross-section and because of the simplicity of its design, the present coaxial waveguide is more compatible than other recently demonstrated THz waveguides with many envisioned applications, such as endoscopy.

Many terahertz imaging and sensing applications will require transmitting terahertz radiation to and receiving it from a sample that is difficult to reach. Many of these applications require an endoscope-type configuration, to guide the terahertz waves to the sample, and then guide the reflected radiation back to a detector. The present waveguides are perfectly suited for this purpose. For example, a waveguide may be used to transmit THz radiation through openings that are too small for effective transmission of an unguided wave. The wave transmitted in this manner can be received at a remote receiver or by a receiver mounted on the endoscope.

In other applications, the waveguide itself can serve as the sampling container. In one such embodiment, illustrated in Figure 7, the outer wall 48 of the waveguide may be perforated with one or more openings 70 so as to allow the passage of a fluid (liquid or gas) into or through the waveguide, as indicated by the arrows. If desired, two or more plugs 60 can be used to define an inner chamber 72 and limit the axial flow of fluid through the waveguide. In alternative embodiments (not shown), plugs 60 are provided with holes and allow fluid flow along the length of the waveguide, in addition to, or instead of the transverse flow illustrated in Figure 6. Probe or Sensor System

Figure 8 shows a THz probe or sensor system comprising a first waveguide 110 and a second waveguide 112. Waveguides 110, 112 may be coaxial but are preferably uniaxial waveguides. First waveguide 110 may be used to transmit a THz signal, such as may be provided by a coupled coaxial antenna, to a desired object of interest 120. Second waveguide 112 receives the echo/reflected portion of the signal and transmits it to a receiver (not shown). The received signal contains information about the object of interest 120. A system such as this can be used to facilitate inspection of enclosed spaces or other regions, or of objects that cannot be viewed using line-of-sight techniques or optical inspection systems. The system shown schematically in Figure 8 may, in some embodiments, be combined with a reflector 114, as shown in Figure 8. Reflector 130 reflects the incident signal from waveguide 110 off an adjacent surface or object 122, from which the signal returns and is reflected into waveguide 112, as indicated by the arrows in Figure 9. This combination allows inspection of a confined area of interest and is therefore particularly useful for endoscopy. Among other applications, THz endoscopy is proving particularly useful for cancer detection. Recent studies have shown that THz is very useful for skin cancer detection, and it is being investigated for detection of cancers of the colon, esophagus, etc. THz Apertureless Near-field Scanning Optical Microscopy

According to another embodiment of the invention, a THz waveguide is used as the emitting tip of a near-field scanning optical microscope (NSOM). In this technique, light is scattered off a subwavelength-sized metal tip which is held close to a surface. The scattered light is collected in the far-field, giving subwavelength resolution in the immediate neighborhood of the tip apex. One embodiment of a THz NSOM system is illustrated in Figure 10; in other embodiments the waveguide could be coupled to a coaxial THz antenna. Spectral Analysis

A still further embodiment of the invention is achieved when a uniaxial waveguide includes a plurality of variations in its diameter, as shown in Figure 11. These form ridges 140, which function as a diffraction grating, causing the THz signal to spread into its spectral components and thus allowing filtering of the signal.

The desired diffraction can be achieved using a variety of techniques. For example, instead of varying the diameter of the metal wire, the exterior surface of the wire with can be coated with another material. The coating may vary in thickness, or may be periodic along the length of the antenna. Similarly, the spectral spreading is not restricted to uniaxial waveguides and may likewise be implemented in the coaxial geometry. Experimental

The following experiments show how a metal waveguide with very simple geometry, namely a bare wire, can be used to guide broadband THz pulses with outstanding performance, including low loss and negligible group velocity dispersion. Propagation effects in near-field optical antennas

The propagation of THz radiation along bare metal wires was first observed in the demonstration of apertureless near-field scanning optical microscopy (NSOM) using THz-TDS. The experimental setup is depicted in the inset of Fig. 12. The broadband single-cycle pulses of free-space THz radiation are generated using a photoconductive transmitter and are focused onto a beryllium-copper probe acting as an apertureless near-field optical antenna. The probe has a tip of about 25 μm radius and a shaft of 0.5 mm diameter coated with a thin layer of tin to prevent oxidation. The sample is a featureless gold-coated silicon wafer, placed in close vicinity of the tip. The mean distance between the tip and the gold surface, d ~ 350 nm, is precisely controlled by a piezoelectric transducer. In such a configuration, the tip strongly interacts with its image in the metal surface, and converts the localized evanescent field around the tip to propagating radiation through a scattering process. The electric field of the scattered THz pulses is detected in the time-domain by a photoconductive receiver which is located near the tip. In the measurement, the probe tip is vibrating normal to the surface at 750 Hz to modulate the scattered radiation, and the detected signal is demodulated using a lock-in amplifier. We can observe the propagation of THz pulses by moving the incident focal spot up along the shaft of the probe, as shown by the dashed arrow. In this case, some of the incident THz radiation is coupled into a propagating mode on the shaft. This propagating mode moves down the shaft and excites the tip, producing a scattered wave which is detected by the receiver. Different incident positions lead to different propagation times and therefore different time delays in the detected time-domain waveforms. A piece of metal perpendicular to the needle is placed close to the shaft at the incident spot to provide a sharp start point of the propagation. Scattering of the THz radiation at the edge of this metal helps to couple the incident wave into a propagating mode on the shaft. The THz transmitter, the focusing lenses, and this metal scatterer are all mounted on a movable stage so that the incident position along the shaft can be precisely controlled.

The propagation effect is evident from the relative delay of the waveforms obtained by moving the transmitter stage along the shaft of the needle in steps of 1.5 mm. As the point of incidence moves- away from the tip, the pulse takes longer to propagate along the shaft, and its amplitude decreases. The propagation is largely nondispersive, since the shape of the time- domain waveform does not depend strongly on propagation distance. The group velocity of the propagation mode can be extracted from the time-domain waveforms. The relative time delay of these waveforms shows a linear dependence on the propagation distance, as depicted by the squares in Fig. 12. A least-squares fit to these data yields the group velocity v_g = (3.00 ± 0.01) x 10⁸ m/s, the free-space velocity of light. A similar measurement is performed in which the needle is wrapped with a 0.5 mm PVC layer. The existence of the insulator layer distorts the detected waveforms, and also reduces the group velocity of the propagation to 0.8c, as depicted by the triangles in Fig. 12. This result indicates that the propagation is confined and guided along the surface of the probe.

Besides the measurements with a bare needle and an insulated needle, we have also performed the propagation measurements with a circular aluminum barrier situated on the probe, and observed the disturbance and the reflection of the propagation mode. These results revealed the possibility of a new method for THz wave guiding and manipulating. However, the waveforms we detect in these experiments are not the electric field of the propagating THz pulses, but the scattered radiation from the probe tip. To eliminate the spectral filtering effects introduced by the probe tip, a new experimental configuration for direct measurement of the THz propagation on bare metal wires was required. This new configuration permitted us to fully characterize the propagating mode along the wire waveguide. Direct characterization of the THz wire waveguide

For a better observation and characterization of the guided THz propagation on metal wires, we changed the experimental setup from the NSOM configuration to a new configuration in which the electric field of the guided mode is directly detected at the end of the waveguide. With the fiber-coupled transmitter and receiver, we can change the incident position (the start point of the propagation) and the detection position of the THz pulses, to observe the spatial profile of the guided mode. A long stainless steel wire with a smooth surface, rather than the tiny tapered probe in the NSOM experiments, was used as the waveguide for the new measurements. Experimental setup

A schematic illustration of the new experimental setup is shown in Figure 13. As in the NSOM experiments, the broadband single-cycle pulses of free-space THz radiation are generated and coherently detected using ultrafast photoconductive sampling. The horizontally polarized THz pulses are focused onto the stainless steel waveguide. Another stainless steel wire is placed at the focal spot, oriented perpendicular to the waveguide (the y direction in Fig. 13). This second wire serves as an input coupler. As such, it could be replaced with a full or partial foil or mesh screen, or any other device capable of coupling the mode of the radiation to the mode of the waveguide. Scattering of the input THz radiation at the intersection structure helps to excite the radially polarized mode which can propagate along the waveguide. Both the waveguide and the coupler are 0.9 mm in diameter, and the separation between them is 0.5 mm. The receiver is placed at the end of the waveguide and is oriented to detect only the vertically polarized component of the electric field in order to eliminate the possibility of detecting directly scattered radiation which would interfere with the detection of the guided mode. The incident THz beam is modulated by a chopper in front of the transmitter and a lock-in amplifier is used for detecting the induced photocurrent in the receiver. The THz transmitter, the focusing lenses, and the coupler are all mounted on a movable stage so that the incident position along the waveguide can be controlled. The THz receiver is mounted on a three-axis stage for detection at various positions with respect to the end of the waveguide. Spatial profile

As the first step in characterizing the propagating mode on the wire waveguide, we measured the spatial profile of the electric field around the waveguide by vertically scanning the THz receiver at the end of the waveguide. Fig. 14(a) shows typical time-domain electric field waveforms, for two different receiver positions located 3 mm above and 3 mm below the wire waveguide. These waves are vertically (y) polarized, perpendicular to the horizontally (x) polarized input beam. The polarity reversal as the detector scans across the wire clearly shows the radial nature of the guided mode. The peak-to-peak amplitude of the waveform as a function of the vertical displacement of the receiver is depicted by the squares in Fig. 14(b). The amplitude decreases with the transverse displacement approximately as \lr. Since the polarization response of the photoconductive receiver antenna is not perfectly symmetric, the measured electric field is not precisely zero at the center point in the experiment. This can also explain the slight asymmetry in the amplitude profile of the detected waveforms. The observed behavior can be understood in terms of either the TEM mode of a coaxial waveguide or in terms of a Sommerfeld wave. The TEM mode in a coaxial waveguide is radially polarized, and the electric field varies as the inverse of the radial position, as

^E' = —a ^m rln— b where a and b are the radii of the outer and inner conductors, respectively, and V is a position- independent voltage. Although providing a qualitative picture, this description cannot be extended to cover the case of interest here because this expression vanishes in the relevant limit, flf -> ∞ . A more accurate picture can be obtained from Sommerfeld' s description of an electromagnetic wave propagating along the surface of a cylindrical conductor, a so-called Sommerfeld wire wave. In this case, it has been shown that the important propagating solution is an axially symmetric TM wave. Outside the metal, the variation of the radial electric field component (the dominant component) is described by a Hankel function, H_\ (γr), where γ is defined in terms of the propagation constant k of the field outside the wire according to γ² = ω²/c² - k² . For a perfectly conducting wire, γ= 0 and the field propagates with a velocity determined solely by the external medium (in our case, air). For large but finite conductivity, γ is small and the approximate form for the Hankel function can be used, appropriate for small argument:

Thus, a Sommerfeld wire wave also exhibits XIr decay, within a distance r₀ «|l/^| of the wire surface. The Sommerfeld description can be used to estimate the distance that the wave extends from the metal surface, for a metal of finite conductivity. To do so, one must determine γ by solving the transcendental equation which results from the boundary conditions at the wire surface. Following the method described by Goubau, we compute the amplitude of the wave as a function of radial distance, for the case of a 0.9 mm-diameter stainless steel (type 304) cylinder, with a conductivity of 1.39x10⁶ mho/m, about 2.4% of the conductivity of copper. To account for the finite aperture of our detector, we convolve this Hankel function with a Gaussian of 6 mm full-width at half-maximum. The resulting profile is shown as a solid curve in figure 13(b). We can also calculate the radius inside of which 50% of the power is guided. At a frequency of 0.3 THz, half of the power is transmitted through an area extending roughly 1.2 millimeters from the axis of the wire. The reasonably good agreement between the experimental results and the calculations suggests that the surface wave picture is an appropriate model for our experimental situation. However, as discussed below, the frequency-dependent attenuation is still lacking a quantitative description. Propagation characteristics

The propagation characteristics of the guided mode are studied by moving the incident position of the THz beam along the waveguide. In this way, we can obtain the time-domain waveforms as a function of propagation distance. There is no evident change in the temporal shape of the waveforms for propagation up to 24 cm, the limit of our optical delay line. This shows that the propagation is largely dispersionless. As in the NSOM experiment, we determined the broadband group velocity of the propagation mode by analyzing the dependence of the relative time delay of the waveforms on the propagation distance. A least-squares linear fit to these data yields the group velocity v_g = (2.995 + 0.001) x 10 m/s, as shown in Fig. 15(a). Fig. 15 (a) shows the arrival time as a function of incident position for a series of time-domain THz pulses detected using the setup illustrated in Fig. 13 (squares). The spectrum-weighted average group velocity of the guided mode is obtained from the least-squares linear fit to these data. The THz waveforms detected after 4 cm and 24 cm of propagation are shown in the insets. To study the group velocity dispersion, we extract the group velocity for different frequency components by analyzing the spectra of these waveforms, using

^v' = — ^c i^ ⁽³⁾

where w^is defined as

C n_eff {(0) = Aφ(ω)-~ cod (4)

Aψ(ώ) is the phase change for propagation distance d at angular frequency ω. Fig. 15(b) shows the extracted data, confirming that there is no measurable group velocity dispersion throughout the accessible spectral range. This is to be expected, given that the Sommerfeld surface wave model predicts a group velocity deviating from c by less than one part in 10⁴, for our experimental situation.

In order to study the evolution of the guided mode in propagation, we compared the spatial profile of the guided mode detected at different propagation distances, each obtained in the same manner as that in Fig. 14(b). These profiles are depicted by the curves in Fig. 16. It is immediately clear that the electric field is more closely confined to the surface of the wire for the shortest propagation distances. Subsequently, the guided mode spreads laterally, especially during the first several centimeters of propagation, and approaches a spatial profile described roughly by \lr.

For each propagation distance, we extract the waveform with the maximum peak-to-peak amplitude. Except for the few shortest propagation distances, these were obtained at a fixed receiver offset of roughly 3 mm (see Fig. 16). These amplitudes were plotted as a function of propagation distance in Fig. 17(a). The amplitude attenuation coefficient a of the wire waveguide can be extracted from these data, simply by fitting the dependence of the pulse amplitude E on the propagation distance x to: fc_*{x) = h_ϋe /-^N

The value we obtain, a ^~ 0.03 cm^"1, is the lowest of any waveguide for broadband THz pulses reported to date. This method can give us the spectrum-weighted amplitude attenuation coefficient, but a more detailed characterization is required to obtain the frequency dependence of the loss. We extract the attenuation coefficient of each frequency component from the amplitude spectra of the THz waveforms detected at various propagation distances. The spectrum of the attenuation coefficient is shown in Fig. 17(b). We note that the attenuation decreases with increasing frequency.

The low attenuation obtained here emphasizes one unique aspect of the wire waveguide. Compared to other waveguide geometries, a metal wire has much smaller surface area interacting with the electromagnetic field, so the propagation loss due to finite conductivity of the metal is negligible. This is consistent with Sommerfeld's wire wave model, which predicts a very small propagation loss due to the finite conductivity of the metal wire. However, the spectrum of the attenuation obtained in our experiment can not be described simply by the Sommerfeld model, as shown in Fig. 17(b). The predicted losses increase with increasing frequency, similar to other THz waveguides where the attenuation is dominated by ohmic effects. In contrast, the observed losses decrease with increasing frequency. This indicates that much of the measured losses arise from other sources, such as diffractive spreading of the propagating mode in the lateral dimensions, as seen in Figure 16. The significance of this loss mechanism for Sommerfeld waves has been discussed previously. By moving the receiver away from the end of the waveguide, we observe a sharper drop in the amplitude of the detected pulses, as depicted by the hollow squares in Fig. 17(a), indicating an increasing divergence when the mode propagates off the end of the waveguide.

We note that our measurements do not reflect the losses associated with the coupling of the linearly polarized free-space THz beam to the guided mode. In the experiment described here, only about one percent of the power is coupled to the radially polarized waveguide mode from the free-space incident beam. The more effective mode-matching needed to improve the input coupling can be obtained using the mode-matched antennas described above. Manipulation of the guided pulses

We next explore the manipulation of the guided mode. The ability to direct radiation along curves is one of the most important features for a practical waveguide. We compare the amplitude of THz pulses after propagating on a waveguide bent with different radii. The results are shown by the hollow triangles in Fig. 17. The propagation distance is 21 cm, and the radius of curvature R is varied from 90 cm down to 20 cm in steps of 10 cm. The amplitude of the electric field E' as a function of the propagation distance x along the bent waveguide is described by

EXx) = E_oe-^«'x ₍₆₎ where a' is the amplitude attenuation coefficient for a bent waveguide. By comparing equation (6) to equation (5) we find

H-) a'= a + — ≡- ^x (7)

So the amplitude attenuation coefficient for each bend radius can be extracted by comparing the amplitude of the detected THz pulse to that of a straight waveguide with the same propagation distance x. The extracted data are depicted by solid squares in Fig. 18. As we can see, even a slight bend on the waveguide can lead to a considerable increase in the loss, from 0.03 cm^"1 for a straight waveguide to nearly 0.05 cm^"1 for a bend radius of 90 cm.

The bend loss can be explained by the continuous conversion of the guided propagation into radiation modes as the wave travels around a curve. This is easy to understand by considering the wavefront of the transverse field, which must rotate around the center of the curvature during propagation. Consequently, at some distance from the center of curvature the phase velocity would exceed c, the propagation speed of the guided mode. So the portion of the field outside this point must be radiated, causing the power loss in the guided mode. This loss mechanism resembles that of a bent dielectric optical waveguide, in which the attenuation coefficient a can be described by a semi-empirical form:

where R is the radius of curvature and C₁ and C₂ are constants independent of R. A fit using equation (8) shows a good agreement with the experimental data, as seen in Fig. 18, suggesting that radiation is the dominant mechanism for the propagation loss of a bent metal wire waveguide.

From the spatial profile of the propagation (Fig. 14(b) and Fig. 16) we can see that the guided mode has a large extent compared to the cross section of the waveguide. Hence we predict that the guided mode would be easily coupled between two curved waveguides in contact with each other (or between a curved waveguide and a straight one). These features enable the Y- splitter for the wire waveguide, as illustrated in Fig. 19. The validity of this scheme has been verified by electric field measurements with such a structure. The waveforms in Fig. 20 are detected at points A, B, and C in Figure 19, namely: at the end of the straight waveguide (A), at the end of the curved waveguide (B), and at a position between them (C). The separation between A and B is 2 cm, and the waveforms are detected with the THz receiver 3 mm below the plane of the splitter structure. The plot clearly shows that part of the propagating power on the straight, waveguide is coupled to the branch waveguide by the Y-splitter. Conclusions

We have demonstrated a new type of THz waveguide with low loss, negligible group velocity dispersion and structual simplicity. This waveguide enables many new THz sensing applications. It is now possible to direct the THz pulse inside of containers or around corners, where line-of-sight optics are not practical. Besides the waveguide described above, we have also tried many other metal wires as THz waveguides. The materials for these guides include steel, aluminum, copper, zinc and nichrome. The wire diameter of these guides ranges from 0.5 mm to 6.4 mm. There is no strong difference in the performance of these waveguides, showing that THz pulses can be launched along any thin metal rod structures. In situations where the guided mode could be perturbed by other structures close to the waveguide, an outer metallic shield could be provided, forming a coaxial waveguide, as long as the additional ohmic losses could be tolerated.

With a Y-splitter structure used to separate the output wave from the input wave, and a small mirror attached at the end of the waveguide as a 90-degree output director, we have successfully demonstrated a THz endoscope, by detecting THz pulses reflected from the bottom and the side wall inside a container. Further improvement could be made by combining an endoscope with an imaging system. This may be accomplished by scanning the endoscope along the surface of the detected region, or alternatively, scanning or rotating the sample to obtain an internal THz image. One challenge for this application is the low power transmitted by the endoscope which strongly limits the data acquisition rate as well as the dynamic range. With optimization of the mode of the input beam and the coupling geometry using the invention described above, the power launched into the endoscope probe can be greatly increased.

It is also interesting to note that this waveguide naturally generates a radially polarized mode. So with a focusing lens mounted at the distal end of the endoscope, a higher resolution can be obtained than in the normal THz imaging system, due to the sub-diffraction-limited focusing of radially polarized beams. Furthermore, since the radially polarized mode is an ideal input field for a coaxial near-field probe or an apertureless near-field optical antenna, nanometer- resolved endoscopic THz imaging may be possible. This would pave the way for a wide range of new applications for terahertz sensing and imaging.

While preferred embodiments of this invention have been shown and described, modifications thereof can be made by one skilled in the art without departing from the scope of the invention. The embodiments described herein are exemplary only and are not limiting. Accordingly, the scope of protection is not limited to the embodiments described herein, but is only limited by the claims which follow, the scope of which shall include all equivalents of the subject matter of the claims.

References

The following references provide background information and are each incorporated herein by reference, except to the extent that they define terms differently than those terms are defined herein: 1. Miyaji, G; Miyanaga, N; Tsubakimoto, K; et al.

Intense Longitudinal Electric Fields Generated From Transverse Electromagnetic Waves

APPL PHYS LETT, 84 (19): 3855-3857 MAY 10 2004

2. Dorn, R; Quabis, S; Leuchs, G

Sharper Focus For A Radially Polarized Light Beam PHYS REV LETT, 91 (23): art. no.-233901 DEC 5 2003

3. Armstrong, DJ; Phillips, MC; Smith, AV

Generation Of Radially Polarized Beams With An Image-Rotating Resonator APPL OPTICS, 42 (18): 3550-3554 JUN 20 2003

4. Moshe, I; Jackel, S; Meir, A Production Of Radially Or Azimuthally Polarized Beams In Solid-State Lasers And The

Elimination Of Thermally Induced Birefringence Effects OPT LETT, 28 (10): 807-809 MAY 15 2003

5. Grosjean, T; Courjon, D; Spajer, M

An All-Fiber Device For Generating Radially And Other Polarized Light Beams OPT COMMUN, 203 (1-2): 1-5 MAR 1 2002

6. Bomzon, Z; Kleiner, V; Hasman, E

Formation Of Radially And Azimuthally Polarized Light Using Space-Variant Subwavelength Metal Stripe Gratings APPL PHYS LETT, 79 (11): 1587-1589 SEP 10 2001 7. Oron, R; BHt, S; Davidson, N; et al.

The Formation Of Laser Beams With Pure Azimuthal Or Radial Polarization APPL PHYS LETT, 77 (21): 3322-3324 NOV 20 2000

Claims

What is claimed is:

1. A system for coupling teraherz (THz) radiation to a coaxial waveguide, comprising: an antenna that generates THz radiation having a mode that matches the mode of the waveguide, said antenna being positioned such that radiation emitted by said antenna is incident upon the waveguide.

2. The system according to claim 1 wherein the antenna comprises a pair of concentric electrodes.

3. The system according to claim 2 wherein at least one of the electrodes is affixed to one end of the waveguide.

4. The system according to claim 2 wherein at least one of said electrodes is formed by an end of the waveguide.

5. The system according to claim 1 wherein the waveguide includes an annular conducting region.

6. The system according to claim 1 wherein the waveguide includes an annular conducting region and the antenna emits radially polarized radiation.

7. A system for transmitting teraherz (THz) radiation, comprising: an antenna comprising first and second electrodes, said first and second electrodes being substantially concentric and capable of emitting THz radiation in response to an application of energy; a waveguide coupled to said antenna.

8. The system according to claim 7 wherein said antenna is affixed to one end of said waveguide.

9. The system according to claim 7, further including a second antenna affixed to a second end of said waveguide.

10. The system according to claim 7 wherein said waveguide comprises an inner core and an outer wall, said inner core and said outer wall being concentric and defining therebetween an annular region.

11. The system according to claim 10 wherein said annular region is filled with a gas.

12. The system according to claim 10 wherein said annular region is substantially evacuated.

13. The system according to claim 10, further including at least one support member in said annular region supporting said core relative to said outer wall.

14. The system according to claim 13 wherein said support member includes at least one fluid flow passage therethrough.

15. The system according to claim 10 wherein said outer wall includes at least one fluid flow passage therethrough.

16. The system according to claim 10 wherein said core is between 10^"4 and 3x10^'2 meters in diameter.

17. The system according to claim 10 wherein said core is between 10^"4 and 2x10^"2 meters in diameter.

18. The system according to claim 10 wherein said core is between 5x10^"4 and 10^"2 meters in diameter.

19. The imaging system of claim 7 wherein said waveguide is at least 10 cm. in length.

20. The imaging system of claim 7 wherein said waveguide is at least 1 m. in length.

21. The imaging system of claim 7 wherein said waveguide is at least 10 m. in length.

22. A terahertz system, comprising: a terahertz antenna comprising first and second electrodes, said first and second electrodes being substantially concentric; means for generating a field across said electrodes and means for triggering the emission of terahertz radiation; a first waveguide having first and second ends, said first end being coupled to said antenna so as to receive at least a portion of said terahertz radiation; and a sensor for detecting said terahertz radiation.

23. The terahertz system of claim 22 wherein said sensor comprises a second terahertz antenna.

24. The terahertz system of claim 22, further comprising a second waveguide positioned to receive radiation from said second end of said first waveguide.

25. The terahertz system of claim 24 wherein said second waveguide is a uniaxial waveguide.

26. The terahertz system of claim 24, further including a reflector.

27. The terahertz system of claim 22 wherein said first waveguide is a uniaxial waveguide.

28. The terahertz system of claim 27 wherein said first waveguide includes at least one variation in diameter.

29. The terahertz system of claim 22 wherein said waveguide is at least 10 cm. in length.

30. The terahertz system of claim 22 wherein said waveguide is at least 1 m in length.

31. The terahertz system of claim 22 wherein said terahertz radiation has a frequency between about 0.1 and 50 THz.

32. A method for transmitting a signal, comprising: generating a radially polarized signal having wavelengths between approximately 30 μm and 3 mm; providing a waveguide having a mode that matches the mode of said beam; and coupling said beam to said waveguide such that a portion of the energy of said beam is transmitted along said waveguide.

33. The method according to claim 32 wherein said radially polarized signal has wavelengths between approximately 50 μm and 800 μm.

34. The method according to claim 32 wherein said radially polarized signal has wavelengths between approximately 100 μm and 500 μm.

35. The method according to claim 32 wherein said portion comprises at least about 50% of the energy of said radially polarized signal.

36. The method according to claim 32 wherein said transmission is essentially dispersionless.

37. The method according to claim 32 wherein said transmission is results in a loss of energy from said coupled energy portion of less than 15% per meter.

38. The method according to claim 32 wherein said transmission is results in a loss of energy from said coupled energy portion of less than 10% per meter.

39. A terahertz near-field scanning optical microscope, comprising: a terahertz transmitter; a stage for receiving a sample; a waveguide coupled to said transmitter, said waveguide including a tip in close proximity to said stage; and a receiver positioned to receive terahertz radiation emitted from said tip and scattered by said sample.

40. The terahertz near-field scanning optical microscope of claim 39 wherein said waveguide is a wire waveguide.

41. The terahertz near-field scanning optical microscope of claim 39, further including an input coupler.

42. The terahertz near-field scanning optical microscope of claim 39 wherein said transmitter is a coaxial antenna.