A sub-millimeter-dimension electro-optic probe that provides enhanced scanning accessibility with significantly less intrusiveness than metal-based or even other dielectric probes during electromagnetic characterization of microwave devices is presented. The quantitative and qualitative relative invasiveness of the probe on the operation of an example antenna device-under-test is explored with respect to previously demonstrated fiber and wafer electro-optic sensors. We also demonstrate that the miniaturized probe, with a diameter of 125 µm, can be used to reconstruct the three orthogonal vector components of near-electric fields without the need for different probe crystals or multiple calibration procedures. Finally, the advantages of the reduced size and invasiveness of the new micro-scale probe are demonstrated through the enhanced resolution of detailed images extracted from planar antennas, as well as the capability of reaching into circuit locations heretofore inaccessible.
© 2008 Optical Society of America
In pursuing minimally invasive detection of the electric near-field from a radiation source, electro-optic (EO) probing techniques have been the most successful and attractive options [1–3]. The significantly lower intrusiveness of EO probing originates from the dielectric nature of the EO probe structure, as it avoids the use of highly field-disruptive metal components. Although such dielectric EO-probe antennas or sensors enable significantly lower invasiveness in near-field sensing than conventional metal-based devices, the cross-sectional size and volume of EO probes, including their crystals and support structure, are the main limitation to realizing an ideal minimum-invasiveness configuration. Generally, to increase the scanning mobility and spatial resolution, EO sensors are mounted on the fiber facet, which is terminated with a bulk ferrule and/or graded-index lens [4,5]. The typical size of such a termination scheme is the ~1-2-mm diameter of a ferrule or lens onto which a smaller sensor crystal is mounted. Such a termination scheme has a limit to its size due to the geometry required to re-direct the modulated beam back through the optical fiber.
In this paper, we report what is likely the least-invasive fiber-based field probe assembled and employed to date. It eliminates all optical mounting components, resulting in a micro-EO-sensor tip, or micro-scale probe, being mounted directly at the end of a standard telecommunication fiber (model SMF-28). The significant size-reduction of the probe provides not only a diminished invasiveness, but it also eliminates the need for changing probes for other vector-field measurements, allowing a uniform probe sensitivity to be maintained. The full x-y-z vector field re-construction from a planar radiation antenna, attained using the same probe in different positions, is presented. In addition, the relative invasiveness of various scales of EO probes has been investigated via microwave network analysis and comparisons of near-field EO amplitude maps. Finally, a significantly improved spatial accessibility that is beneficial for probing fields around smaller devices and structures, especially in high frequency applications, is presented.
2. Structure of a micro-cavity EO probe
A number of fiber-based EO probes have been reported by several groups [4,5,6,7]. All of these are based on a retro-reflection process employing dielectric mirrors [4,5,6] on the probe itself or on highly reflective surfaces inherent to the device under test (DUT) . These fiber-based EO probes can be categorized into two main types: those that are directly fiber-mounted, and those that are attached to a ferrule and lens assembly, depending on how the optical beam diverges within the EO crystals [4,5,6]. The size of the EO crystal is the crucial factor in determining the mounting type, and a clear criterion was presented in Ref. 5. Crystals up to 1 mm thick can be directly mounted on a thermally-expanded-core fiber-end, because the increased numerical aperture of the expanded core allows one to achieve virtually a collimated beam within the crystal volume.
While thicker crystals increase the EO sensitivity, the coupling efficiency of the reflected beam back into the fiber core is typically degraded due to the divergence of the beam. Furthermore, an increased volume and interaction length of the crystal would nearly always be detrimental for invasiveness, primarily in the case of the former, and sensing bandwidth, in the case of the latter. Thus, the use of thin crystals that possess a small surface area is essential to minimizing invasiveness related to the probe volume. While probe sensitivity can be sacrificed when maximizing bandwidth and reducing invasiveness, the small interaction length within a thin crystal can be stretched if an etalon is created and multiple Fabry-Perot round trips inside of the cavity are used for EO-phase-retardation enhancement [8,9].
The structure of a fiber-based micro EO probe is shown in Fig. 1(a). In this case, we were able to procure a 52-µm-thick x-cut LiTaO3 wafer and coat it with five-layer high-reflection coatings on both of its larger surfaces. The high-reflection layers consisted of alternate quarter-wavelength Bragg stacks with high and low indices of ZnSe (2.41) and MgF2 (1.34) at the optical telecom wavelength (~1562 nm) of our laser. The wafer was diced in a ~0.2 mm×0.2 mm scale to fit onto the ~125-µm diameter of a standard fiber, and it was attached with ultra-violet curing optical cement. Because the crystal is very thin, the first few round-trip components of the light inside the micro-resonator can still be reasonably aligned back to the fiber without the need for an expanded-core. The use of a standard fiber and minute volume for a sensor crystal, without the need for ferrules, graded-index lenses, and glass alignment collars, also results in a significant reduction of the probe cost.
The detailed principle and performance of a resonance-based, fiber-coupled EO probe, as well as an entire photonic-down-mixing EO probing system, were reported in our previous work, where we have realized an enhanced signal-to-noise ratio with a significant reduction of beam-line polarization elements . Here, the EO interaction occurs only over a very confined area of the crystal cavity, determined by the mode-field diameter of the fiber, as well as the thickness of the crystal and the Q-factor of the resonator. Often for free-space, bulk-probe sensing, the EO interaction volume is significantly expanded so that the optical-sensing area can be increased to include a larger part, or even nearly the entire probe volume (e.g., probe C in Fig. 1(b) is 60 µm thick, but has a 10 mm diameter). While this type of sensor can facilitate rapid sensing of field distributions over a large area, it can also influence the performance of the DUT, as shown later in Sec. 3.
While the length over which the volume of the light beam interacts with the microwave electric field in the EO crystal governs the EO sensitivity per unit optical and microwave intensity, the invasiveness is primary determined by the increase in capacitance (i.e., volume and permittivity) due to the presence of the crystal. The optical-resonator cavity increases EO interaction length by keeping the photons inside of the cavity for a longer time without increasing the thickness and volume of the crystal, thus reducing the microwave-field distortion. Hence, the micro-cavity EO sensor mounted directly on a fiber facet helps to minimize invasiveness while maintaining a reasonable sensing quality.
3. Three component vector-field mapping with a minimally invasive EO probe
To compare probe invasiveness between a conventional ferrule-supported probe and a micro-scale probe, an optical-heterodyne system for down-mixing of the optical beam and extracting the microwave amplitude and phase was adopted [9,10]. Using a configuration that kept the entire optical system the same except for the different fiber-based probes, the three independent, orthogonal vector electric-field components of the 10.485-GHz signal at the resonance frequency of an X-band patch antenna were measured. In this technique, a cw laser beam is modulated within an electro-optic modulator (EOM) at a frequency that becomes the local oscillator (LO), before that beam is again modulated by the DUT electric field in the EO sensor crystal. Here the second-order harmonic of a 5.421-GHz drive signal to an EOM is used to efficiently create the photonic LO . The down-converted beating components (or intermediate frequency, IF) at 3 MHz (10.485 GHz minus 2×5.421 GHz) are used to reconstruct the original amplitude and phase of the electric fields from the DUT.
Generally, fiber probes are positioned vertically with respect to a DUT because of accessibility and invasiveness concerns. This configuration enables the sensing of only certain vector field components, based on the orientation of the optic-axis of the crystal. For instance, an x-cut LiTaO3 wafer has its optic-axis in the plane of the wafer, so the EO effect gets maximized as the electric field is applied along this axis. In this case, the x-cut LiTaO3 fiber probe (the probe A placed as in Fig. 1(b)) can be used to measure two transverse (i.e, x and y) electric-field components. If one wishes to measure the normal (i.e., z) component, the probe has to be replaced with another one fitted with a z-cut crystal.
The characteristics of EO materials vary with the orientation of an optic-axis (i.e, x or z-cut). For instance, the induced birefringences due to an applied electric field E for x-and z-cut LiTaO3 are Δn x-cut=(n 3 e r 33-n 3 o r 13)E/2 and Δn z-cut=n 3 o r 13 E/2, respectively. Using the known values for the EO coefficients at 1558 nm , the EO-induced birefringence values per unit electric field (i.e., EO figure of merit) are 97.1975 pm/V and 33.0794 pm/V for x-and z-cut, respectively. Thus, the x-cut probe has a nearly three times superior sensitivity as the z-cut probe for the same electric field, making it highly advantageous to utilize the x-cut configuration for measuring the two transverse and the normal field components. However, cumbersome re-alignment and sensitivity-calibration procedures would be necessary if one needed to switch between probes.
3.1 Minimally invasive three component field mapping
Since the miniaturized EO probe of Fig. 1 has an increased accessibility to the DUT, full three-dimensional, directional-field sensing becomes possible with a single probe. For instance, placing the micro-scale probe B laterally as in Fig. 1(b), the vector field sensitivity is directed along the optic-axis. If the optic axis of probe B in Fig. 1(b) is set normal to the plane of the DUT, the probe measures the field in the z direction. To measure the transverse field components, the probe fiber would first be simply rotated around its longitudinal axis by 90° so that the optic axis of the EO crystal was parallel to, and sensitive to, the DUT field polarization along the x direction. To capture the y component of the field, the probe fiber would be rotated by 90° so that the fiber itself was made parallel to the x axis, with the crystal optic axis remaining in the plane of the DUT, both parallel and sensitive to its y-directed polarization. If this was attempted with probe A, the ferrule diameter would limit the proximity that the probe could have to the DUT, and the parasitic capacitance presented by the probe would also increase.
To better understand the concept of field distortion due to dielectric probes having different sizes, examples of how electric-field flux changes due to the capacitance of EO probes are illustrated in Fig. 2. Finite-element simulations [e.g., in 12] show that one result of substituting air with the volume and permittivity of dielectric probes is that the electric-field flux lines in the vicinity of the probe may deviate significantly, causing a field pattern to actually “flatten” or take on a more distributed characteristic with respect to the pattern in air. When comparing the effect of the small, but still relatively large volume of probe A (as compared to the micro-scale probe), one would expect to observe a more widely distributed field measured by the former as compared to the latter.
This is what is observed in a comparison of the measurement fidelity of the fields above the patch antenna for probes A and B (Figs. 3(a) and 3(b), respectively). That is, probe B extracts an apparently more localized field distribution as compared to probe A, a result of the fact that the increased area of high-permittivity material presented in probe A compresses the field under the probe. All three orthogonal components of the electric-field vector from the patch antenna, as measured with the micro-scale probe, are shown in Figs. 3(b)–3(d). The x-y-z patterns are measured at a uniform height ~200 µm above the DUT, and thus the sensitivity for each of the measured components is the same, and the amplitude field plots in Figs. 3(b)–3(d) have the same scale and can be directly compared. This is a departure from all previous work with fiber-based field probes, in which either different probe materials, different crystal axes, or different probe positions (that presented different invasiveness profiles) were employed. The peak modulated signal level at the IF as measured on a lock-in amplifier read-out instrument is -57 dBm±0.5 dB for each of the three field components.
The combined amplitude and phase information of the three orthogonal-field components measured in the reactive near field of the patch antenna provide valuable information on the signal propagation directions and radiated polarizations from the antenna, as shown in Media 1–3. Two predominant phase angles are observed in the Media, with the strong purple and red colors indicating a 180° phase difference. For the x- and z- components of electric field, the adjacent signals are seen to be out-of-phase, and thus their polarizations are expected to destructively combine in the far field, while for the y- component of electric field, the spatial field components are in-phase, constructively combining to yield the far-field polarization. Such significance of the field patterns for the patch antenna, which is not an objective of this paper, is discussed detail in Refs. 2,9,13. From the current field characterization, however, we can conclude that the micro-scale probe yields a substantially more detailed and accurate microwave near-field pattern.
3.2 Invasiveness analysis versus probe size
The invasiveness of a radiation source is heavily influenced by the distance from the DUT to an EO sensor probe. The near-field regions may be categorized into two sub-regions based on invasiveness, distance and radiation patterns of the DUT. One is the reactive near-field region, within a wavelength of the DUT. The other one is the radiating near-field region, located before the conventional border of the Fresnel/Fraunhofer regions . The fields within the reactive region are important in that they eventually determine the Fresnel to Fraunhofer distributions.
Although EO sensing, due to its photonically assisted, all-dielectric configuration, is the most promising experimental way to measure the fields that serve as the origin of the far-field pattern, the reactive near-field regime is highly susceptible and invasive to any material other than air. For instance, the X-band patch antenna used in Figs. 1 and 2 has a 10.485 GHz resonance frequency with a -11-dB return loss as measured with a vector network analyzer. The resonance characteristics of the DUT remain virtually the same when probes A and B are brought into the near field of the patch as shown in Fig. 1(b) (the antenna return loss for the case when air or either of the fiber-based probes is represented by the black curve in Fig. 4). However, the resonance dip frequency and degree of return loss are modified noticeably as the surface or volume of the probe increases. For instance, using a 150 µm thick glass slide, with a relatively low permittivity compared to most EO materials, to cover the entire DUT shifts the resonance frequency 63 MHz shorter (to the red line in Fig. 4). For more invasiveness materials like the 60 µm LiTaO3 wafer probe C of Fig. 1(b), the resonance frequency decreases by 755 MHz (the blue line in Fig. 4). Covering the DUT with heavy dielectric wafer ruins the desired resonance frequency and radiation efficiency by degrading the quality of the match of the characteristic impedance of the antenna input. To mitigate the device invasiveness, one may tolerate a decrease in EO-sensor sensitivity by offsetting the wafer probe above the DUT and leaving an air gap between the probe and DUT. The field intensity decreases significantly (~1/(offset distance)3) from the DUT , but of potentially greater concern is that the invasiveness of the probe C, even with an offset of 200 µm, still shifts the antenna resonance to a frequency 133 MHz shorter than that of the unloaded or fiber-probe-loaded antenna (green line in Fig. 4).
LiTaO3 has a relatively higher invasiveness due to its higher microwave permittivity, so using the bulk configuration of this crystal affects the original field significantly. To investigate the field penetrating through the LiTaO3 wafer, probe C was placed 200 µm above the DUT, and we measured the field distribution 200 µm above wafer C using the micro-scale probe B. Figures 5(a)–5(c) are the x, y and z field components at 10.485 GHz of the original resonance frequency, respectively. Because of the resonance frequency shift and induced invasiveness by the wafer probe C, the measured fields are distorted with lower sensitivity. Even at the new shifted resonance frequency of 10.319 GHz, the fields are still distorted as observed in Figs. 5(d)–5(f).
The main advantage of using a bulk wafer probe is to decrease the scanning and data acquisition time rather than to enhance the sensitivity. The wafer probe that is larger than the DUT scale enables one to either scan a probe beam over the DUT without moving the probe, or to use a large, relatively high-power probe beam to capture a large area of the RF electric field at the same time, processing the EO detection in parallel with multi-channel image processing . This technique allows even virtually live field images to be sensed, while avoiding the single probe-channel field mapping used in this paper.
The instantaneous live imaging is a true breakthrough where real-time monitoring of a DUT is essential. For broadband, non-resonant radiation devices such as microstrip line or co-axial cables, invasive is less a concern , although changes in impedance could still be significant. However, for most resonant devices, such as antennas, filters and resonators, it is crucial to keep the original radiation performance even for sensors that are placed in the reactive near-field region. Hence, sensing speed and invasiveness is a trade-off between live imaging with a bulk wafer probe and slow raster mapping with a micro-scale fiber probe.
Although slow raster-scanning does not provide live-imaging, it is possible to see the animated flux flow in a scale time-scale slow enough to visualize after the scanned data is acquired. Because the photonic-down-mixing EO measurements with the harmonic LO-sidebands are also able to capture the microwave-signal phase from the IF signal phase, one may utilize the expression for electric field, E(ω,t)=Asin(ωt+ϕ), where A is the measured signal amplitude and ϕ is the measured phase, with a numerically-iterated ωt term in order to animate the field maps. In essence, the spatial distribution of X-band microwave signals is transformed into a map of sub-hertz signals that can be observed, with subtle color changes corresponding to small field amplitudes, and dramatic changes to large field amplitudes. Although the animated image is not real-time, it reconstructs the recorded actual cycles of the field flows from the DUT with a re-scaled, sub-hertz periodic iteration as shown in Figs. 3(b)–3(d) (Media 1–3).
3.3 Comparison of spatial accessibility
Besides its advantages in regards to low invasiveness, the micro-scale probe also offers greater potential accessibility to DUTs such as integrated-circuit packages (measurements between pins), circuits that possess air bridges or vias, and so forth. Thus, due to the miniaturized size of the fiber probe, sensing within gaps as narrow as 150 or 200 mm is feasible. For an example measurement, we probed the field in a plane underneath a wire-over-ground-plane structure, in which a 2.5-mm-diameter copper wire was suspended 0.8 mm over a ground plane. As seen in Fig. 6(a) the probe B can easily be guided to the middle of the gap, where it is raster-scanned, sensing three standing-wave modes (Fig. 6(b)). The fiber probe is set to sense the normal field and the DUT was fed at 10 GHz with an open termination. As expected for a 10-GHz standing-wave pattern in free-space, a 4.5 cm separation between three strong standing modes are observed. Probe A, also seen positioned above the copper wire in Fig. 6(a), had a diameter more than twice as large as the wire/ground-plane separation, and thus was useless for extracting the fields from the small space within the DUT.
3.4 Invasiveness and spatial resolution advantages for millimeter-wave applications
The micro-scale probe becomes more valuable for higher frequency resonant devices. Generally, the size of an antenna decreases as the resonance wavelengths become smaller. For instance, Fig. 7 shows an array of 35.5-GHz patch antennas, such as those that might be found in a phased-array radar. Considering the sub-centimeter radiation wavelength and the half-wavelength of the unit antenna scale, the size of probe A is now comparable to an entire unit antenna, whereas the micro-scale probe B is not. Antenna performance is primarily evaluated by how efficiently it broadcasts the guided microwave power to the air. To realize such a radiation condition, the reflected power portion has to be minimized, and this can be achieved typically by implementing a quarter-wavelength impedance transformer into the signal’s path in the same fashion as a quarter-wavelength anti-reflection layer in used in an optical coating. Such impedance matching networks are typically very narrow, as seen in Fig. 7 at the sections where the small line parallel to the y-axis enters the different patches. Thus, perturbations during probe scanning could alter the impedance (particularly the reactive part) of the matching network, causing serious resonances and a general degradation of the quality of operation of the device.
The measured field comparisons using probes A and B are presented in Fig. 7(b). Probe A clearly disrupts the field pattern, especially at the narrow matching lines as the probe captures data from the lines. However, fortunately the radiation pattern and performance is basically maintained except for in the matching-layer regions. This is because as the probe moves onto the radiation terminal edges where the radiation fields emerge, the matching layers are not greatly affected by the probe, for instance as compared to when the whole-wafer probe is used. However, it is still clear that the patterns from probe B are much more distinct and well-confined than the patterns from probe A.
We have proposed and demonstrated a micro-cavity-resonator EO probe that has minimal intrusiveness in a fiber-based sensing scheme. The proposed probe yields significantly less invasiveness with respect to conventional fiber and wafer-based sensors and an improved scanning accessibility to the radiation devices, due to the dramatic reduction in volume of the dielectric material comprising the probe that must be inserted into the near field of a device to be characterized. This accessibility enables one to reconstruct the full three dimensions of near-electric fields with a single probe. Finally, the accessibility and invasiveness advantages of the micro-scale probe were demonstrated with X- and Ka-band radiation measurements, respectively.
This work was supported through the ARO STTR program, contract number W911NF-06-C-0178. The authors also wish to thank Dr. Jeong-Jin Kang in Dong Seoul College, Korea for supplying the Ka-band antenna arrays and for his encouragement.
References and links
1. K. Yang, G. David, S. Robertson, J. F. Whitaker, and L.P.B. Katehi, “Electro-optic Mapping of Near-field Distributions in Integrated Microwave Circuits,” IEEE Trans. Microwave Theory Tech. 46, 2338–2343 (1998). [CrossRef]
2. K. Yang, J. G. Yook, L.P.B. Katehi, and J. F. Whitaker, “Electrooptic Mapping and Finite-Element Modeling of the Near-Field Pattern of a Microstrip Patch Antenna,” IEEE Trans. Microwave Theory Tech. 48, 228–294 (2000).
3. K. Yang, T. Marshall, M. Forman, J. Hubert, L. Mirth, Z. Popovic, L.P.B. Katehi, and J. F. Whitaker, “Active-amplifier-array diagnostics using high-resolution electrooptic field mapping,” IEEE Trans. Microwave Theory Tech. 49, 849–857 (2001). [CrossRef]
4. K. Yang, L.P.B. Katehi, and J. F. Whitaker, “Electro-optic field mapping system utilizing external gallium arsenide probes,” Appl. Phys. Lett. 77, 486–488 (2000). [CrossRef]
5. H. Togo, N. Shimizu, and T. Nagatsuma, “Near-Field Mapping System Using Fiber-Based Electro-Optic Probe for Specific Absorption Rate Measurement,” IEICE Trans. Electron. E90-C, 436–442 (2007). [CrossRef]
6. S. Wakana, E. Yamazaki, S. Mitani, H. Park, M. Iwanami, S. Hoshino, M. Kishi, and M. Tsuchiya, “Performance evaluation of fiber-edge magnetooptic probe,” J. Lightwave Technol. 21, 3292–3299 (2003). [CrossRef]
7. M. Sameer and Chandani, “Fiber-Based Probe for Electrooptic Sampling,” IEEE Photon. Technol. Lett. 18, 1290–1292 (2006). [CrossRef]
8. D. J. Lee and J. F. Whitaker, “A Simplified Fabry-Pérot Electrooptic-Modulation Sensor,” IEEE Photon. Technol. Lett. 20, 866–868 (2008). [CrossRef]
9. D. J. Lee, M. H. Crites, and J. F. Whitaker, “Electro-Optic Probing of Microwave Fields Using a Wavelength-Tunable Modulation Depth,” Meas. Sci. Technol. 19, 115301–115310 (2008). [CrossRef]
11. J. L. Casson, K. T. Gahagan, D. A. Scrymgeour, R. K. Jain, J. M. Robinson, V. Gopalan, and R. K. Sander, “Electro-optic coefficients of lithium tantalite at near-infrared wavelengths,” J. Opt. Soc. Am. B 21, 1948–1952 (2004). [CrossRef]
12. J. A. Deibel and J. F. Whitaker, “A fiber-mounted polymer electro-optic-sampling field sensor,” in 2003 IEEE LEOS Annual Meeting Conference Proceedings (IEEE, 2003), pp. 786–787.
13. D.J. Lee, J. J Kang, and J. F. Whitaker, “Vector Near-Field Measurements Using Optimized Electrical and Photonic Down-Conversion,” IEEE Trans. Microwave Theory Tech. 56, 3231–3238 (2008). [CrossRef]
14. A.G. Yaghjian, “An Overview of Near-Field Antenna Measurements,” IEEE Trans. Antennas Propagat. AP-34, 30–45 (1986). [CrossRef]
15. K. Sasagawa, A Kanno, T. Kawanishi, and M. Tsuchiya, “Live Electrooptic Imaging System Based on Ultraparallel Photonic Heterodyne for Microwave Near-Fields,” IEEE Trans. Microwave Theory Tech. 55, 2782–2791 (2007). [CrossRef]