Sie sind auf Seite 1von 9

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 57, NO.

2, FEBRUARY 2009

499

A Comparison of the Focusing Properties of a Fresnel Zone Plate With a Doubly-Hyperbolic Lens for Application in a Free-Space, Focused-Beam Measurement System
David R. Reid, Member, IEEE, and Glenn S. Smith, Fellow, IEEE
AbstractA Fresnel zone plate is designed to perform the function of a doubly-hyperbolic lens. The focusing characteristics of the zone plate and the lens are compared using both measured data and body-of-revolution nite-difference time-domain simulations. The results from this comparison are used to evaluate the use of zone plates as focusing elements in a free-space, focused-beam measurement system. A set of two zone plates is used to perform measurements of the material properties of a dielectric sheet, and the results are compared to the results from a standard system which uses two doubly-hyperbolic lenses. Index TermsDielectric measurements, lenses, zone plate.

Fig. 1. Schematic diagram of a focused-beam measurement system.

I. INTRODUCTION

ETAILED studies have been performed for Fresnel zone plates designed to replace hyperbolic lenses. These zone plates were designed to collimate radiation from a point source in the transmit mode, or to focus an incident plane wave in the receive mode [1][6]. There are some applications, however, for which both the source and the focal point are near the zone plate. For these applications, a zone plate can be designed to replace the functionality of a doubly-hyperbolic lens, and this is the subject that will be addressed in this paper. A practical application for such a zone plate is as a focusing element in a free-space, focused-beam measurement system. A typical focused-beam system, like the one shown in Fig. 1, consists of two horn antennas and two focusing elements. The eld radiated by the transmit horn is converted by a lens into a Gaussian beam with the waist at the point where the sample is placed. Another lens focuses the eld that is transmitted through the sample to the receive horn, where it is measured. In addition, the transmit horn is used to measure the eld reected from the sample. At the waist of a Gaussian beam, where the sample is

Manuscript received August 13, 2008; revised October 15, 2008. Current version published March 20, 2009. This work was supported in part by a Shackelford Graduate Fellowship from the Signature Technology Laboratory, Georgia Tech Research Institute and in part by the John Pippin Chair in Electromagnetics, School of Electrical and Computer Engineering, Georgia Institute of Technology. The authors are with the School of Electrical and Computer Engineering, Georgia Institute of Technology, Atlanta, GA 30332-0250 USA (e-mail: david. reid@gatech.edu, glenn.smith@ece.gatech.edu). Color versions of one or more of the gures in this paper are available online at http://ieeexplore.ieee.org. Digital Object Identier 10.1109/TAP.2008.2011392

placed, the phase is uniform and the amplitude has a Gaussian taper [7]. For a sample that is large enough that the eld of the beam is insignicant at its edges, the beam can be approximated as a plane wave. Using this assumption, the measured values of the reection and transmission coefcients can be used to calculate the complex permittivity and permeability of the sample [8][10]. For measurements at low frequencies, the size, weight, and cost of the dielectric lenses can become extreme. For example, the lenses in a system designed for 800 MHz measurements can have a diameter of six feet and weigh thousands of pounds each. In addition to being difcult to manipulate, these lenses present manufacturing challenges: It is difcult to nd a thick enough slab of material from which to machine the lenses, while molding the lenses also presents issues, including the presence of air bubbles in the completed lenses. Zone plates are lighter and thinner than traditional lenses and could mitigate these problems. Because of these advantages, zone plates may allow the operating frequency of a focused-beam system to be lowered further, to regions where lenses are prohibitively large. Zone plates are not without their drawbacks, however. The most well-documented of these is their limited bandwidth [11]. As diffractive devices, zone plates are designed to work at one particular frequency, and performance degrades away from this frequency. In addition, for a focused-beam system, the eld at the sample must have phase uniformity and a specied amplitude taper. The prole of a zone plate is discontinuous, and there is concern about how this will affect these qualities. In this paper, a zone plate designed to replace a doubly-hyperbolic lens is presented, and the focusing characteristics of both are compared using measurements and nite-difference time-domain (FDTD) simulations. The tradeoffs of using a zone plate are evaluated, and the application of zone plates to a focused-beam system is examined and experimentally veried.

0018-926X/$25.00 2009 IEEE

500

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 57, NO. 2, FEBRUARY 2009

(20 subzones total). When the subzones are numbered consecutively starting at the center of the zone plate, the outer radius of the th subzone is

(1) in which is the free-space wavelength. The thickness of is then calculated to satisfy the zone plate prole for each Fermats principle of least time
Fig. 2. Geometry for (a) a doubly-hyperbolic lens and (b) a zone plate.

(2) in which is the outer radius of the rst zone. In (2), takes on a value from 0 to , depending on which zone the th subzone is in. For example, when , when . A base of thickness and is added to each for structural support. The number of subzones per zone, , in (1) and (2) determines how many phase corrections are performed within the zone plate. In [1], it was shown that the value of does not signicantly affect the power delivered to the focal point once . In this study we let , which gives the smooth prole seen in Fig. 2(b). A zone plate with this prole is frequently referred to as a zoned Fresnel lens in reference to its similar appearance to the classical Fresnel lens [6]. Both the zone plate and the lens were modeled using the body-of-revolution nite-difference time-domain (BOR-FDTD) method. This method exploits the rotational symmetry of the devices to reduce the computational cost. By dependence analytically, the accounting for the azimuthal complete three-dimensional solution can be obtained using a two-dimensional simulation. The implementation of the BOR-FDTD method used for this study is similar to the implementation described in [12], with the exception of the model for the horn antenna. The eld of the isolated horn antenna was measured on the plane at which the front of the lens (zone plate) will be placed (input plane in Fig. 2). These measurements of the eld were then t to the amplitude taper and phase variation of the far-eld of a Gaussian beam. The eld from this Gaussian-beam model for the horn was input into the BOR-FDTD simulation using the total-eld scattered-eld technique [13]. On the surface corresponding to the input plane, the electric and magnetic elds were added or subtracted in a manner that created a beam that propagated towards the lens (zone plate). BOR-FDTD simulations were used as part of the design process for the zone plate. This allowed the zone plate prole

Fig. 3. Prole for a zone plate with

N = 5 and P = 4.

II. DESIGN AND MODELING The size and weight savings offered by zone plates make them especially well-suited for use at low frequencies. In the interest of cost savings, however, the comparison of a zone plate and a doubly-hyperbolic lens was performed using scale models at frequencies in the range 812 GHz. The horn used in this study was a pyramidal standard gain horn (Narda model 640). This particular horn was chosen for its linear polarization and its relatively symmetric radiated eld. Cross-sectional proles for both the lens and the zone plate are in Fig. 2. The lens is part of an existing focused-beam system at the Georgia Tech Research Institute, and the zone plate was designed with parameters similar to those of the lens to allow for a direct comparison. Both the zone with a plate and the lens were made from Rexolite radius of . The focal lengths used to calculate the proles for the hyperbolic lens were and . The design frequency for the zone plate was chosen to be the . The cross center of the frequency band of interest, section of the zone plate in Fig. 2(b) shows that one side of the zone plate is planar. This was an additional design requirement that was imposed to help simplify the manufacture of the zone plate. The prole for the other side of the zone plate was designed using ray-tracing techniques, similar to the procedure in zones that were each [4]. The zone plate was divided into subzones. These padivided into subzones, for a total of rameters are illustrated in Fig. 3, which shows the prole for a zones and subzones per zone zone plate with

REID AND SMITH: A COMPARISON OF THE FOCUSING PROPERTIES OF A FRESNEL ZONE PLATE

501

Fig. 4. Schematic diagrams of (a) the measurement setup and (b) the dipole probe.

to compensate for the shortening of the focal length discussed in [1]. Simulations were used to iteratively increase the values and in (1)(2), adjusting the zone plate prole until of the actual focal lengths of the zone plate matched the focal lengths of the doubly-hyperbolic lens. The proles in Fig. 2 are drawn to scale and demonstrate the difference in thickness between the two devices. The maximum thickness of the zone plate is 25% of the thickness of the lens, and the weight of the zone plate is only 32% of that of the lens, excluding the mounting brackets. The reduced weight relaxes the mechanical requirements for the mounting bracket, which can therefore also be made lighter, further increasing the weight advantage of the zone plate. III. EXPERIMENTAL VERIFICATION Measurements of the electric eld were made throughout the volume surrounding the focal point for both the zone plate and the doubly-hyperbolic lens shown in Fig. 2. A diagram of the experimental setup used for measuring the eld is shown in Fig. 4(a). The coordinate system in this gure will be used for the remainder of the paper. The horn was oriented so that the electric eld was polarized in the -direction. A diagram of the probe used to measure the electric eld is in Fig. 4(b). The probe was an electrically-short dipole, made from a pair of semi-rigid coaxial lines. The exposed center conductor of each line formed one arm of the dipole, and the two lines were run through a plug in the end of a WR90 waveguide. Inside the waveguide, the exposed center conductors of the lines were connected to form a loop which coupled to the magnetic eld in the waveguide. Notice that the probe is a balanced structure. The probe could be oriented to measure the - or -component of the electric eld. The -component could also be measured, although the accuracy and the physical range for the measurements were restricted in this conguration by the presence of the coaxial lines. A system of three linear positioners allowed the probe to

be moved in the -, -, and -directions with respect to the focusing element. A network analyzer connected to the horn and the probe was used to measure the magnitude and phase of the desired eld component. These measurements were used to validate the BOR-FDTD are simulations, and results at the frequency presented as a representative example of the comparisons. The magnitude of the -component of the electric eld for the zone plate is plotted in dB in Fig. 5(a) for two cuts. The red lines and the blue lines are for verare for horizontal cuts , all on the focal plane. The solid lines are tical cuts for the BOR-FDTD simulations, and the dashed lines are for the measurements. The gure shows good agreement between the two, both in the main beam and in the sidelobes. The phase for these two cuts is plotted in Fig. 5(b), and in this gure, the vertical dashed black lines indicate the extent of the main beam (the region between the rst nulls in the magnitude plot). Again, there is good agreement between the simulated and measured results, particularly in the main beam. This gure demonstrates the relative uniformity of the phase over the main beam, a feature that is necessary to satisfy the assumptions made in the focused-beam system. For comparison, the same plots are presented for the doubly-hyperbolic lens in Fig. 5(c) and (d). Again, there is good agreement between the simulations and measurements. The trends for the lens are similar to those for the zone plate. Note, however, that the sidelobes for the lens are generally lower than for the zone plate. Because of space limitations, only results for the zone plate will be presented in the remainder of this section. in dB for the zone plate Fig. 6 shows the magnitude of for two planar cuts through the focal point: a horizonal cut ( -plane) in (a), and a vertical cut ( -plane) in (b). Measured results are plotted on the left, and FDTD results are plotted on the right, with red representing the highest intensity and blue representing the lowest intensity. The eld was measured apart; this at points spaced causes the pixelation in the gures for the measured eld. Both plots show good agreement between the simulations and measurements. In Fig. 6(a), the focusing of the beam can be seen: and then expands The eld tapers to a waist near with increasing distance from the zone plate. The cuts through the waist in Fig. 6(b) clearly show the focal spot produced by the zone plate. Collectively, Figs. 5 and 6 conrm that the zone plate is focusing energy in the manner of a Gaussian beam. Measurements of the other components of the electric eld also conrm the accuracy of the FDTD simulations. The magin the -plane and of in a -plane through nitudes of are plotted in Fig. 7(a) and (b), respectively. From Figs. 6(b) and 7(a), it can be seen that the maximum value is about 20 dB less than for . The agreement befor is reasonably tween the measured and simulated values of good, although is slightly higher for the measured data. This difference can be attributed to factors in the measurement system, including imperfections in the construction and aligncomponent from ment of the dipole probe as well as a small the feed horn. Fig. 7(b) shows the presence of a -component of the electric eld in both the simulations and the measurements. Although

502

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 57, NO. 2, FEBRUARY 2009

Fig. 5. Measured and simulated (FDTD) results for the electric eld E on the focal plane. Magnitude and phase of E at f and (b), and for the lens (c) and (d). The dashed black lines in (b) and (d) indicate the extent of the main beam.

= 10 GHz for the zone plate (a)

for an ideal plane wave, the eld for an ideal Gaussian is therefore a beam has a small (non-zero) -component [7]. particularly interesting metric for evaluating the accuracy of the BOR-FDTD simulations and the sensitivity of the measurement system. Because of the probe design, the measurements for are not as accurate as they are for and . However, the results in Fig. 7(b) agree well enough to indicate that the FDTD model is correctly accounting for . The results in Figs. 6 and 7 show that the eld at the beam waist in nearly linearly polarized, with the - and -components being about 20 dB below the eld of the -component. The results presented in Figs. 57 are a small sample of the data collected. Measurements were taken for all three compo, 10, and 12 GHz, for both the nents of the electric eld at zone plate and the lens. In general, the agreement between the simulations and the measurements for all of these cases was as good as for the case presented here. IV. COMPARISON OF THE ZONE PLATE WITH THE DOUBLY-HYPERBOLIC LENS The previous section has shown that the BOR-FDTD simulations accurately predict the focusing properties of the zone plate and doubly-hyperbolic lens. The next step in evaluating the usefulness of the zone plate for focusing applications is to compare

its focusing characteristics with those of the lens. Looking back look qualitatively simat Fig. 5, the magnitude and phase of ilar in the focal plane for the zone plate and the lens. This section describes quantitative measures that further this comparison. A. Time-Average Power Delivered to the Focal Spot When comparing the zone plate and the lens, one metric of interest is the amount of power focused by each. This can be quantied by looking at how much power is intercepted by a disc of radius on the focal plane of each device. To compute the time-average power received, the -component of the real is integrated over the part of the complex Poynting vector disc

(3) In Fig. 8, is plotted as a function of for both the lens and . For reference, the vertical dashed the zone plate at black line indicates the point at which the radius of the disc of integration equals the radius of the lens or zone plate. For the is normalized by , the injected power, which is gure, the total time-average power supplied by the horn. If all of the

REID AND SMITH: A COMPARISON OF THE FOCUSING PROPERTIES OF A FRESNEL ZONE PLATE

503

Fig. 6. Magnitude of E at f for the zone plate. (a) Horizontal cut (xz -plane), and (b) vertical cut (xy -plane) through the focus.

= 10 GHz

Fig. 7. The magnitude of (a) E in the xy -plane through the focus, and (b) E in a yz -plane through x : for the zone plate at f .

= 00 64 cm

= 10 GHz

power passed though the focal plane, then as . The green (dashed) line shows the power received for the lens. Most of the power is concentrated into a small region, . Beyond this region, the plot levels off at . Power lost to reection and over-illumination of the lens accounts for the fact that not all of the power supplied by the never reaches 1). horn is delivered to the focal spot ( The red (solid) line shows that the zone plate does not deliver as much power to the focal spot as the lens; it delivers slightly more than 60% of the injected power to the focal spot, . Subsequent increase in results in a gradual, step-like increase in received power. This behavior can be understood by looking at the -component of the Poynting vector for the lens and the zone plate, which is plotted in Fig. 9. The dashed white line indicates the location of the focal plane. These results show that both devices focus a majority of the incident power to a small spot on the focal plane. The zone plate, however, clearly directs more power away from this spot than the lens; this is indicated by the diffraction bands that occur outside the spot. The bands are green in color and are roughly 30 dB below the point of highest power (at the center of the focal spot). These for the bands account for the step-like increases in

Fig. 8. Fraction of the power supplied by the horn that passes through a disc of radius  in the focal plane for the lens and the zone plate at f .

= 10 GHz

zone plate in Fig. 8: As the radius is increased, more of this misdirected power is included in the integral, (3). In Fig. 9, the solid white lines outline the metal mounting bracket that surrounds the lens. The scattering from the bracket is visible in the gure. For the zone plate, the bracket was made to reduce the amount of scattering. The from Delrin bracket for the zone plate is not shown in Fig. 9.

504

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 57, NO. 2, FEBRUARY 2009

~ ) in dB on a cross section in the yz -plane for the lens and the Fig. 9. z ^ 1 Re(S zone plate at f = 10 GHz.

B. Size and Axial Position of the Focal Spot In addition to comparing the amount of power delivered to the focal spot, it is also important to look at the size and axial position of this spot. For a plane parallel to the -plane at the distance from the output plane of the lens or zone plate, the radius at which the power has decreased 10 dB from the on-axis . This radius is smallest at value was determined and called the waist of the beam, and its value at the waist indicates how is plotted for the lens tightly the beam is focused. In Fig. 10, (a) and zone plate (b) for three frequencies: (dashed black line), 10 GHz (solid red line), and 12 GHz (dash-dot blue line). For the lens, the beam waist is located at approximately the same distance from the output plane for all three frequencies. This distance is marked by the vertical dashed black line. As the frequency is lowered, the radius of the focal spot increases (dash-dot blue line) to 4.6 cm for from 3.4 cm for (dashed black line). The results for the zone plate show a different trend. At the , the plot of for the zone plate design frequency, looks similar to that for the lens (solid red lines). However, as the frequency changes, the waist of the beam moves signicantly. (dashed black line), the waist gets closer to the At zone plate, while it moves farther from the zone plate at (dash-dot blue line). At , the region where does not change much with changing is smaller in the axial . However, the general form direction than it is at remains similar to that of the lens: As is increased, the of spot becomes narrow, levels out at a well-dened waist, and then , the waist becomes less wellexpands again. At dened, indicating that the focusing mechanism is beginning to is break down. This is clear from Fig. 11, in which . Contrasting plotted in dB for the zone plate for this with the same quantity for in Fig. 9 (right side), we see that at 12 GHz there is no clearly dened beam waist, and that the eld looks more like that of a collimated beam. C. Phase Uniformity Over the Focal Spot Another metric that is important in many focusing applications is the uniformity of the phase over the focal spot. This is quantied by computing the root mean square (RMS) deviation
Fig. 10. Radius of the 010 dB spot for (a) the lens and (b) the zone plate for f = 8, 10, and 12 GHz. Here, z is the distance from the output plane of the lens or zone plate.

~ ) in dB on a cross section in the yz -plane for the zone plate Fig. 11. z ^ 1 Re(S at f = 12 GHz.

in phase from a uniform value over the disc of radius at each axial position . The values used for in this calculation are those plotted in Fig. 10. For an ideal plane wave, the phase deviation would be zero for all , and for a Gaussian beam, the phase deviation would be only exactly zero at the beam waist [7]. Fig. 12 shows the phase deviation as a function of axial dis, tance from both the lens (a) and the zone plate (b) at 10, and 12 GHz. The phase deviation is very close to zero at one point for each of the frequencies for both devices. At this point, the deviation for the lens ranges from 0.2 to 0.6 over the frequency range. The deviation for the zone plate ranges from 0.2 to 0.9 . The locations of these points correspond to in Fig. 10 and, therefore, follow the locations of minimum the same trends: The waist of the beam for the lens stays at a xed distance as the frequency is varied, while it moves for the zone plate. Simulation results, not shown here, indicate that the choice of has little effect on the size and location of the beam waist. The spot size and the phase deviation over this

REID AND SMITH: A COMPARISON OF THE FOCUSING PROPERTIES OF A FRESNEL ZONE PLATE

505

Fig. 13. Schematic diagram of a focused-beam measurement system with zone plate focusing elements. The spacing between the zone plates and the sample is adjusted for three frequencies.

Fig. 12. Deviation in phase over the 10 dB spot for (a) the lens and (b) the zone plate at f = 8, 10, and 12 GHz.

spot (shown in Figs. 10(b) and 12(b) for ) are similar for zone plates with a nite number of phase corrections per zone, . even as low as In summary, at the design frequency, the zone plate behaves like the doubly-hyperbolic lens: The focal spot is of similar size and location, and the uniformity of the phase at the beam waist is comparable. Away from the design frequency, the location of the beam waist for the zone plate is signicantly more frequency-dependent than it is for the lens. In addition, the focusing mechanism of the zone plate begins to break down as the frequency moves farther away from the design frequency. The importance of these bandwidth limitations, of course, depends on the application. The particular application of zone plates to a focused-beam system will be considered in Section V. V. APPLICATION TO A FOCUSED-BEAM SYSTEM The results from the comparison in the previous section can be applied to study the feasibility of replacing lenses with zone plates in a focused-beam measurement system. The limitation on the power delivered to the focal spot of the zone plate, discussed in Section IV-A is, by itself, not signicant. Calibration procedures for the focused-beam system allow the measurements to be independent of the amount of power delivered to the sample. However, the fact that a signicant amount of power can be sent in directions away from the focal spot can become important. One of the assumptions of the focused-beam technique is that the edges of the sample are outside the region that is strongly illuminated. If zone plates are used and an edge of the sample is located at a point that receives signicant power, such as in a diffraction band, some of this power may be scattered back to one or both of the horns. A combination of techniques can be used to solve this problem. Absorber can be placed around the edges of the sample to reduce unwanted scattering. The signals received by the horns can also be time gated to remove unwanted reections from either an edge of the sample or

from elsewhere in the system. The time-gating procedure is not a unique requirement of the zone plate system; it is also necessary for the lens system. The frequency-dependant shift in the beam waist is a concern that must be accounted for in a focused-beam system that uses zone plates. If a sample is placed at the beam waist for , it will not be at the waist for other frequencies. Consequently, the phase will not be as uniform as possible, and the size of the focused spot will be larger than intended. One approach to overcome this problem is to compensate for the shift in the beam waist by moving the zone plates with respect to the sample at each frequency. The drawback of this approach is that some of the simplicity of the focused-beam system is lost and the system becomes less convenient to use. A good compromise between optimizing the position of the sample at each frequency and ease-of-use is to choose a few different spacings and then measure a band of frequencies for each of these spacings. This idea is illustrated for three frequencies in Fig. 13: The zone plates are closest together for measurements at 8 GHz and farthest apart for measurements at 12 GHz. The spacings are determined by the locations of minimum phase deviation; for the zone plate considered in Fig. 12 these spacings for , for , are: for . and As a nal test, a focused-beam system using two zone plates was constructed and used to measure a 0.32 cm thick sheet of Rexolite. Measurements of the transmitted and reected eld were used to calculate the permittivity of this sheet. The system was calibrated using the response/isolation method, and details of this technique are given in an Appendix. The calibrated, frequency-domain measurements were time gated in post-processing by convolution with a Kaiser window to remove unwanted reections, including multiple reections between the zone plate or lens and the sample. The permittivity was then computed using a method similar to the inversion algorithm for the transmission/reection method for waveguide or coaxial line measurements in [14]. The measured permittivity was compared to results from a focused-beam system that uses two doubly-hyperbolic lenses and to published values. The values of the relative permittivity for Rexolite

506

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 57, NO. 2, FEBRUARY 2009

Fig. 14. The relative permittivity for a sheet of Rexolite determined by focusedbeam measurements. Measurements were made using two lenses and using two zone plates with the spacing for the zone plates set for f .

= 10 GHz

Fig. 15. The relative permittivity for a sheet of Rexolite determined by focusedbeam measurements. Measurements were made using two zone plates with the , 10, and 12 GHz. spacing set for f

=8

found in the literature are in the range

, [15][18]. When is low, as it is for Rexolite, it cannot be accurately measured is considered. with a free-space system; therefore, only In Fig. 14, the green (dash-dot-dot) line is the measured result using two doubly-hyperbolic lenses, and the red (solid) line is the measured result using two zone plates with the spacing set . The gray box shows the range of values found for in the literature. The RMS deviation from the average value from the liter, was computed for both the lens and ature, the zone plate over the frequency range 812 GHz. The deviation for the lens was 0.18%, while for the zone plate it was 0.53%. From Fig. 14, it is clear that the most accurate measurements made with the zone plate system are around the design . frequency, If the technique proposed earlier is implemented, and the zone , plate system is recongured with the correct spacing for 10, and 12 GHz as in Fig. 13, then the agreement between the measurements and the published value improves in the region around these respective frequencies. This suggests that the three measurements made with zone plates could be stitched together to get a more accurate measurement across the entire frequency band. This point is illustrated in Fig. 15. The black (dashed), red (solid), and blue (dash-dot) lines are for the permittivity mea, 10, sured using two zone plates with the spacing set for and 12 GHz, respectively. Comparing this measurement with spacing alone (solid red the measurement for the line) in Fig. 14, a clear improvement can be seen. The RMS deviation from the published value improves from 0.53% for the measurement in Fig. 14 to 0.31% for the measurement in Fig. 15. VI. CONCLUSION A zone plate was designed to replace a doubly-hyperbolic lens. The focusing characteristics of the zone plate and the lens were compared using measurements and body-of-revolution nite-difference time-domain simulations. At its design frequency, the focusing of the zone plate was shown to be qualitatively similar to the lens, despite the fact that the zone plate delivers less power to its focal spot. The location of the beam waist for the zone plate, however, was shown to move as a function of frequency.

Fig. 16. Schematic diagram of the four congurations necessary for a response and isolation calibration for the focused-beam system.

Using the results of this comparison, the application of zone plates to a free-space, focused-beam measurement system was examined. Two focused-beam systems were constructed: one using traditional doubly-hyperbolic lenses and the other using zone plates. Measurements of a dielectric sample over the range of frequencies showed that the zone plate system can produce results that are comparable to the lens system, but that care must exercised in interpreting the results, particularly away from the design frequency. A technique that accounted for the frequency-dependent shift of the focal point was presented and shown to offer the potential to improve results. The calculations and measurements from this research give us condence that zone plates can be used in a focused-beam measurement system designed to work at low frequencies, as discussed in the introduction.

REID AND SMITH: A COMPARISON OF THE FOCUSING PROPERTIES OF A FRESNEL ZONE PLATE

507

APPENDIX CALIBRATION PROCEDURE The calibration is applied to the measured, focused-beam data using the same procedures that are used for GTRIs standard focused-beam system: A response and isolation calibration is made for each of the S-parameters. This requires eight distinct measurements in four different congurations. Fig. 16 is a schematic diagram of these four congurations. Using the conguration in Fig. 16(a), the transmission response and are measured. A metal plate of known coefcients thickness is placed in the sample holder, as shown in Fig. 16(b), and and the and the reection response coefcients transmission isolation coefcients and are measured. For all four of these measurements, the metal plate is left in place. Finally, the two congurations in Fig. 16(c) and (d) are used to measure the reection isolation coefcients and , respectively. In these congurations, a piece of broad-band foam absorber placed in front of one of the horns is treated as a matched load. These coefcients can then be used to calibrate the S-parameters measured with the sample in place

(4) , 12, 21, or 22. Additionally, the thickness of in which is accounted for by multiplying and the metal plate by to adjust the reference plane for the measurements. to account for All four -parameters are multiplied by the space taken up by the sample to be measured, which has a thickness of . ACKNOWLEDGMENT The authors would like to thank J. Schultz, S. Blalock, and P. Friederich of the Signature Technology Laboratory at GTRI for sharing their measurement system and for helpful discussions throughout every phase of this project. REFERENCES
[1] D. R. Reid and G. S. Smith, A full electromagnetic analysis of grooved-dielectric Fresnel zone plate antennas for microwave and millimeter-wave applications, IEEE Trans. Antennas Propag., vol. 55, no. 8, pp. 21382146, Aug. 2007. [2] J. M van Houten and M. H. A. J. Herben, Analysis of a phase-correcting Fresnel-zone plate antenna with dielectric/transparent zones, J. Electromagn. Waves Applicat., vol. 8, no. 7, pp. 847858, 1994. [3] Y. J. Guo and S. K. Barton, Fresnel Zone Antennas. Dordrecht, The Netherlands: Kluwer Academic, 2002. [4] H. D. Hristov, Fresnel Zones in Wireless Links, Zone Plate Lenses, and Antennas. Boston, MA: Artech House, 2000. [5] A. Petosa, N. Gagnon, and A. Ittipiboon, Optimization of dielectric Fresnel lens thickness for maximizing gain, in Proc. 12th Int. Symp. Antenna Tech. and Appl. EM and URSI Conf., Jul. 2006, pp. 519522. [6] O. V. Minin and I. V. Minin, Diffractional Optics of Millimetre Waves. Bristol, U.K.: Institute of Physics, 2004. [7] G. S. Smith, An Introduction to Classical Electromagnetic Radiation. Cambridge, U.K.: Cambridge Univ. Press, 1997, ch. 3.

[8] D. K. Ghodgaonkar, V. V. Varadan, and V. K. Varadan, A free-space method for measurement of dielectric constants and loss tangents at microwave frequencies, IEEE Trans. Instrum. Meas., vol. 37, pp. 789793, Jun. 1989. cek, Microwave Measurements of Complex Permit[9] J. Musil and F.Z tivity by Free Space Methods and Their Applications. Amsterdam, The Netherlands: Elsevier, 1986. [10] L. E. R. Petersson and G. S. Smith, An estimate of the error caused by the plane-wave approximation in free-space dielectric measurement systems, IEEE Trans. Antennas Propag., vol. 50, no. 6, pp. 878887, Jun. 2002. [11] J. E. Garrett and J. C. Wiltse, Fresnel zone plate antennas at millimeter wavelengths, Int. J. Infrared Millimeter Waves, vol. 12, no. 3, pp. 195220, Mar. 1991. [12] T. G. Jurgens, J. G. Blaschak, and G. W. Saewert, Bodies of revolution, in Computational Electrodynamics: The Finite-Difference TimeDomain Method, A. Taove and S. C. Hagness, Eds., 2nd ed. Boston, MA: Artech House, 2000, ch. 12, pp. 529568. [13] A. Taove and S. C. Hagness, Computational Electrodynamics: The Finite-Difference Time-Domain Method, 2nd ed. Boston, MA: Artech House, 2000. [14] J. Baker-Jarvis, Transmission/reection and short-circuit line permittivity measurements, National Institution of Standards and Technol., Boulder, CO, 1990, Tech. Rep. NIST Tech. Note 1341. [15] Datasheet for Rexolite C-Lec Plastics Co [Online]. Available: http:// www.sdplastics.com/rexolite.html [16] A. von Hipple, Dielectric Materials and Applications. Boston, MA: Artech House, 1995, ch. 5. [17] W. R. Humbert, A new technique for measuring the electromagnetic properties of rotationally symmetric materials, Ph.D. dissertation, Georgia Institute of Technology, Atlanta, 1997. [18] Certicate of Analysis National Institution of Standards and Technology, Gaithersburg, MD, 2006 [Online]. Available: https://srmors. nist.gov/certicates/view_cert2pdf.cfm?certicate=2870 David R. Reid (S03M08) received the B.S.E.E., M.S.E.C.E, and Ph.D. degrees from the Georgia Institute of Technology, Atlanta, in 2002, 2005, and 2008, respectively. He is currently working at the Signature Technology Laboratory, Georgia Tech Research Institute. His research interests include the characterization of electromagnetic materials, Fresnel zone plate antennas, and numerical modeling with the nite-difference time-domain method.

Glenn S. Smith (S65M72SM80F86) received the B.S.E.E. degree from Tufts University, Medford, MA, in 1967 and the S.M. and Ph.D. degrees in applied physics from Harvard University, Cambridge, MA, in 1968 and 1972, respectively. From 1972 to 1975, he served as a Postdoctoral Research Fellow at Harvard University and also as a part-time Research Associate and Instructor at Northeastern University, Boston, MA. In 1975, he joined the faculty of the School of Electrical and Computer Engineering at the Georgia Institute of Technology, Atlanta, GA, where he is currently Regents Professor and John Pippin Chair in Electromagnetics. He is the author of the book An Introduction to Classical Electromagnetic Radiation (Cambridge, 1997) and coauthor of the book Antennas in Matter: Fundamentals, Theory and Applications (MIT Press 1981). He also authored the chapter Loop Antennas in the Antenna Engineering Handbook, (McGraw-Hill, 2007). His technical interests include: basic electromagnetic theory and measurements, antennas and wave propagation in materials, and the radiation and reception of pulses by antennas. Prof. Smith is a member of Tau Beta Pi, Eta Kappa Nu, and Sigma Xi, and a member of URSI Commissions A and B.

Das könnte Ihnen auch gefallen