µ-wave vacuum electronics
1. Why THz ?
THz electromagnetic radiation in the region 0.1 – 1 THz [1, 2] falls between the so-called electronics and photonics areas of research as shown in Figure 1. It has gained considerable attention in the last few years due to attractive features [3] including (a) penetration through most non-metallic materials, (b) non-ionizing as compared to X-rays, (c) penetration dependent upon water concentration and tissue densities, and (d) wide bandwidth possibility at these high frequencies.
Figure 1 The THz regime 100 GHz – 1 THz and beyond lies in between the electronics and photonics area of research. The green line emphasizes the sharp drop in output power (P~f-2 scaling no longer holds) observed for both solid state and vacuum based compact microwave devices in the THz region.
2. THz Technological Gap and Substantial Applications
As shown in Figure 1, the term “THz gap” is commonly employed in current literature [1, 2] to describe the fact that despite substantial applications, there has been a lack of available source technology with reasonable power in a compact volume package.
This has generated considerable interest in the scientific community to realize the sources for a number of proposed important applications [4-8] ranging from industrial quality control, non-intrusive contra-band item detection, medical imaging/cancer diagnostics, all weather visibility systems for aviation safety, advanced telecommunication systems and radar with high wireless data transfer rates for LPI (low probability of interception), and commercial applications as described in Figure 2.
Figure 2: Millimeter Wave to THz applications
3. Vacuum Electronics: Superior Choice for High Power in compact package
At all frequencies, the vacuum electron beam device (VED) supersedes in producing higher power than a solid state power amplifier (SSPA) device as shown in the figure below[9].
There are some fundamental differences in solid state and vacuum electronic devices. In a vacuum device, the electron beam flow in vacuum is collisionless, while in a solid state device the collision-dominated stream diffuses through a semi-conducting solid. Thus, a solid state device is unable to conduct away excessive heat generated by the electron current in the interaction region. Moreover, there is a high probability of a dielectric breakdown at high microwave field strengths. These limitations do not exist at least to this high a degree as in solid state devices. Furthermore, the efficiency of a vacuum electron device can be enhanced by using an energy recovery mechanism or a multistage depressed collector[10]. Thus, for achieving reasonable power in a compact sized package at THz frequencies, vacuum integrated power amplifier (VIPA) technology is a superior choice [11-14].
Figure 3: State of the art and prospective devices in terms of Pavgf2Df/f figure of merit
Figure 4 (a) Double-vane half period staggered 0.22 THz Sheet Seam TWT (SBTWT) micro metallic structure (b) narrowest cavity width ‘x’ and cavity depth ‘y’ are shown with frequency scaling up to ~ 1 THz.
4. Enabling Technologies/Efforts at UCD:
Although Vacuum Electron Device (VED) technology is potentially an appropriate technology for providing reasonable power, band width in a compact package, it is challenging to realize the sources in the millimetre wave to THz region. There are mainly two reasons. First, the circuit dimensions scale with the frequency and thus goes beyond the capabilities of conventional fabrication schemes for holding the requisite tolerances. Furthermore, the surface roughness at near-THz frequencies become comparable to the skin depth thereby leading to losses in the circuit. Secondly, as the beam tunnel dimensions scale down with frequency, very high current density electron beams are required at mm wave – THz frequencies necessitating very high current density cathodes.
These two issues are highlighted and UCD efforts are summarized below:
4.1 MEMS Fabrication/Nano CNC Milling
To provide a perspective, Figure 4 (a, b) shows the frequency scaling of slot width and slot depth of the UCD designed double vane half-period staggered grating type SBTWTA [15-18] slow wave structure up to 1 THz.
For the fabrication of a vacuum integrated power amplifier with THz operating frequency where the grating dimensions reach the order of ~ 75 μm and the required surface roughness at least less than the skin depth (140 nm at 0.22 GHz), it is important to analyze and compare the fabrication regimes of different relevant technologies.
It is clear that at THz frequencies, conventional machining is unable to fabricate the sub-mm dimensions. Moreover, the real challenge is to fabricate structures with narrow cavity widths but shallower depths that we describe as High Aspect Ratio Structures (HARS). In the last decade, with the advent of the so-called MEMS (Micro-Electro-Mechanical-Systems) technology, it was made possible to potentially bridge the so called “THz gap” and to fabricate the microstructures [19-22] needed for micro-vacuum-electron-devices (μVEDs) with high precision. Figure 5 (a) highlights the typical equipment available at the University of California, Davis class-100 clean room. Figure 5 (b) describes the flow chart for the UV-LIGA process. UC Davis has worked on specially fabricated 4 inch Oxygen Free High Conductivity OFHC Cu wafers with highly flat (+ 0.5 μm) and polished surfaces. The process consists of spin coating, soft bake, UV exposure, post exposure bake, electroplating, lapping/polishing, and mold removal processing. The UC Davis team has successfully characterized the entire process to precisely fabricate 220 GHz Traveling-Wave Tube Amplifier TWTA copper circuits ~ 400 μm thick in a single layer process[23].
Figure 5 (a) Typical Class-100 clean room equipment available at the Northern California Nano-technology Centre, U. C Davis. (b) UV LIGA process flow chart optimized in our group for the precision fabrication of ~ 400 μm tall micro-metallic structures in single layer process.
Figure 6: UCD LIGA fabricated 220 GHz TWTA circuit in a single layer ~ 400 μm process (a) Microscope image (b) Fully metalized copper circuit (c) TWT circuit using SU-8 mold (d) TWT metalized circuit using KMPR mold
Figure 6 (a) shows the microscope image of the TWT circuit after UV-lithography and mold development process, but before electroplating. After the electroplating the samples in home-made optimized electroplating tank, the TWTA circuits were sent for lapping and polishing to the desired thickness, i.e 385 μm. After the lapping and polishing process, the fully metalized TWTA circuits which were made in two halves are shown in Figure 6 (b). For the UV-LIGA process, we used KMPR and SU-8 photo-resists. Figure 6 (c) shows the TWTA circuit made of SU-8. This circuit was made for investigating the metal coating schemes using E-beam evaporation and sputter coating. Figure 6 (d) shows the KMPR-LIGA circuit after mold removal processing that was made with KMPR inside the slow wave structure. The reason it is done for KMPR but not SU-8 is that KMPR is relatively easier to remove without destroying the delicate circuit parts.
The Nano CNC milling machine (Mori Seiki NN1000) designed by DTL was first tested on an aluminum piece for the fabrication of a 220 THz TWTA circuit. Aluminum was chosen for a first circuit prototype fabrication because it is much easier to machine than copper. The main issue when machining copper is that it is very ductile and “drags” with the tool, not only causing heavy tool wear, but also leaving a prominent surface asperity. (UC Davis researchers were able to avoid these problems in their later tests with copper.) The circuit was fabricated in two halves cut along the “zero-current” plane. A 125 micron diameter tool was used for fabrication of this circuit, and the final depth of cut was 385 microns per half. The circuit was cold tested and found to exhibit > 10 dB loss (S21) over ~ 50 GHz bandwidth that is quite high as compared to the simulation which predicted ~ 5 dB over 60 GHz bandwidth. After the initial tests, with the improvement in machining capability, availability of better tools, and using better conductivity material OFHC Cu, we obtained excellent results that are described in next section. The SEM image of the 0.22 THz TWTA circuit made in Aluminum is shown in Figure 7.
Figure. 7. 0.22 THz slow wave circuit machined by nano CNC milling in Aluminum
Figure 8: (a) commercially available tungsten scandate cathode 311X made by SpectraMat, (b) scandia doped tungsten nano composite cathode made by UC Davis, (c) I-V emission curve for scandia doped tungsten nano composite cathode
4.2 High Current Density Cathodes
mVEDs require high start-oscillation current density electron beams proportional to f02, where f0 is the frequency, in the millimeter and sub millimeter wave region. Reduction of the electron beam cross-section due to an increase of operational frequency dramatically increases the beam current density. For instance, a straightforward estimation of current density through scale-down indicates that 3 A/cm2 at 10 GHz run corresponds to 3´104 A/cm2 at 1 THz [24]. This is hardly achievable by conventional electron beam sources, particularly given the need to transport the electron beams with low interception. Recently, some electron emitters capable of producing a high emission current density have been challenged for THz mVED application and the electron gun issues have also been discussed [25-27]. The quality of these cathodes is usually evaluated by using scanning electron microscope images of the emitting surface (Figure 8, (a) and (b)) and by plotting testing data on logarithmically scaled I-V (Current-Voltage) curves (Figure 8, (c)).
5. 220 GHz Travelling Wave Tube Amplifier at UCD
The conceptual schematic for the 220 GHz travelling wave tube amplifier circuit with proposed instantaneous bandwidth > 50 GHz and output power > 50 W for an input drive of 50 mW is shown in Figure 9. The high current density cathodes are required to produce an emission current density of 50 – 100 A/cm2. After the electrostatic focusing that forms the electron beam into an elliptic sheet shape for high current transport with fixed perveance, a periodic permanent magnet (PPM) is proposed for an efficient sheet beam transport[28].
Figure 9: Conceptual Drawing of 220 GHz TWTA including electron gun section, input/output couplers, PPM focusing assembly and collector
The half-period-zigzagged grating arrays, such as depicted in Figure 10 (a), provide a strong symmetric axial electric field distribution (in-phase field variation) on the fundamental mode in the sheet beam tunnel. At the center frequency of 0.22 THz, we define the design parameters as ϕ (= kzd) = 2.5π for phase synchronism with the 20 keV electron beam. The dimensional parameters, illustrated in Fig. 10 (a), are obtained from the broad spectrum synchronous condition. The period d of 460 mm is derived from j = kzd = (2pf0/ue)d, where f0 = 0.22 THz, ue = bc = 0.27c (Ve = 20 kV), and j = 2.5p. Here, b = (1 − γ−2)0.5 is the relativistic factor, γ = 1 + Ve/Vn (Vn = mec2/e = 511 keV), and me and e are the rest mass and the charge of the electron, respectively. As shown in Figure 10 (b), in the pass band of the optimally designed circuit, the beam line is very close to the upper cutoff and crosses over the n = 1 and n = 2 space harmonic backward wave modes, which can cause instability during device operation due to parasitic cutoff oscillations or backward wave oscillations. From simulation analysis the design values, 270 and 770 mm, of L and h provide the broadest synchronism with the beam line without any noticeable cutoff oscillation. The beam tunnel dimensions of 150 mm (y) × 770 mm (x) correspond to a beam filling factor of 61% for an elliptical sheet beam size of 100 mm (y: minor axis) × 700 mm (x: major axis) with 250 mA and 454.7 A/cm2.
The overall device performance of the designed circuit is fully characterized by the numerical modeling analysis using 3D particle-in-cell (PIC) simulation (MAGIC3D) and a 1D TWT code (CHRISTINE). These results are in general agreement with preliminary studies using the CST PIC code, thereby providing further confirmation.
Figure 10 (a) Conceptual 3D drawing of the staggered double vane THz sheet beam TWT circuit with the elliptical sheet beam (b) 1st and 2nd pass bands with the beam line in the spatial harmonic dispersion curve (c) Power versus distance graph (frequency-filtered) of PIC simulation model (MAGIC3D, inset) with PML (d) power versus time graph (frequency-filtered, inset is frequency spectrum)
The circuit is analyzed by PIC simulations with a 20 keV and 250 mA elliptical sheet beam[29]. A PIC simulation circuit model (OHFC, s = 5.8 ´ 108 [W-1m-1]) with PML circuit terminations is shown in Figure 10 (c) (inset). With the 0.22 THz coherent driving signal of 50 mWand periodic noise pulses, the 4 cm long (~ 90 periods) drift tube has 97.2 % and 99.44 % beam transmission (non-thermionic electron beam) with 6.25 and 7.5 kG solenoid fields, which only correspond to 143 and 25.6 W dissipated beam power, respectively. The simulation indicates that ~ 5 – 7 kG would be the required solenoid magnetic field to focus the elliptical sheet beam for practical pulsed or CW operation.
In Figure 10 (c), the plot of power versus distance graph shows steep power growth of the 220 GHz signal along the axial distance. The power is rapidly amplified in the axial direction and gradually falls off after saturation, which occurs at ~ 34 mm for these conditions. The maximum power is ~ 180 W at ~ 34 mm, and our minimum power goal of 50 W is achieved at an axial position of 25 mm. The growing output power is monitored in the time domain at the 34 mm saturation position and Figure 10 (d) shows the time-dependent power graph and Fourier-transformed frequency spectrum (inset). The time-dependent power remains quite stable during the entire simulation, as evidenced by the monochromatic frequency spectrum (inset). Note that a small ripple appears in the time-domain power graph due to intrinsic backward circuit waves and boundary reflections, but that the power fluctuation rapidly damps away. Figure 10 (d) also exhibits the low noise level of the frequency spectrum in that the peak of the 0.22 THz driving signal is ~ 42 dB higher than the higher order modes near the upper cutoff frequency (0.27 ~ 0.275 THz). These PIC simulations verified that no noticeable cutoff oscillation and mode competition, described in Figure 9 (b), arise with the perfectly matched boundary condition assumption within the numerical convergence time frame.
6. RF Measurements (cold test) of 220 GHz Traveling Wave Tube Amplifier
UCD has two experimental setups to make measurements in the millimeter wave range:
(a) Figure 11 shows the scalar Network Analyzer system employing a Backward Wave Oscillator (BWO) operating in the frequency range of 165 – 270 GHz. Virginia Diodes, Inc. detectors were used for the sensitive detection of RF signals in this frequency range. The system was fully calibrated and tested for full operation before actual experimentation on UV-LIGA and Nano-machined fabricated 0.22 THz SBTWT circuits
Figure 11. Scalar Network Analyzer (RF measurement setup) for LIGA and Nano-machined TWT circuits
(b) For confirmation measurements, the DMRC Lab is equipped with an Agilent state-of-the-art PNA-X vector network analyzer with T/R modules available in the frequency rnage 67 – 110 GHz, 140 – 220 GHz and 220 – 330 GHz. Figure 12 shows the TWTA circuit under test with the PNA-X T/R module WR5.1 in frequency range 140 – 220 GHz.
Figure 12. 220 GHz TWTA under test using PNA-X with WR5.1 T/R modules in frequency range 140 – 220 GHz.
Figure 13. Comparison between RF measurements (cold test) of the TWT LIGA circuits with simulation.
6.1 RF Measurements (cold test): LIGA TWT circuit
The cold test results for the UV KMPR LIGA fabricated samples were quite impressive[30]. The circuit exhibited transmission in the frequency range 214 GHz – 266 GHz with insertion loss varying between 5 to 10 dB in general. Figure 23(a) shows the comparison of the RF measurements and simulations for two different thicknesses of circuit halves, i.e 360 μm and 370 μm. As the figure demonstrates, the simulation prediction and measurement result match very well except that we observe a drop in S21 beyond 268 GHz which is attributed to the sharp reduction in the BWO output power at this frequency range.
6.2 RF Measurements (cold test): Nano-machined circuit
With the same RF measurement setup described in the earlier section, nano machined SBTWT circuits were analyzed for RF transmission and reflection[31]. Figure 24 shows the comparison of simulation of the broadband couplers and measurement results. The broadband couplers have a gradual depth variation for impedance matching over a wide range of frequencies; this design was crucial for the efficient operation of the ultra wide-band SBTWTA designed at UCD. The simulation predicted a 1 dB bandwidth of ~ 70 GHz. The S21 measurement matched very well with simulation up to around 265 GHz beyond which the measurement setup was unable to take the data at the same power settings. The return loss remained a little higher than the simulation predicted around -10 dB in the pass band. This is attributed to the imperfections of the fixture assembly that holds the two halves of the SBTWT circuit together.
Figure 14 (left) SEM image of nano-machined broadband coupler and S-parameter comparison of simulation and RF measurements (right) SEM image of full TWT structure with couplers and S-parameters comparison of measured and simulated values.
Figure 14 shows the SEM image of the full model of 40 mm long nano-machined TWT circuit with the integrated broad-band couplers. The simulation of transmission (S21) as shown in figure predicted a 1 dB band width greater than 60 GHz. The RF measurement agrees excellently with the 3D electromagnetic computational analysis. The return loss was seen to be matched with the computational predictions very well (< -10 dB in the passband). This was due to the fact that surface finish was improved by chemical cleaning (called bright dip) and an improved fixture setup to better hold the circuit.
LINKS
REFERENCES
1. Williams, G.P., Filling the THz gap - high power sources and applications. Rep. Prog. Phys., 2006. 69: p. 301-326.
2. Siegel, P.H., Terahertz Technology. IEEE Transactions on Microwave theory and Techniques, March 2002. 50(3).
3. Xiaodong, C. and L. Daohui. High power Terahertz generation by using vacuum electronic technology. in Broadband Network & Multimedia Technology, 2009. IC-BNMT '09. 2nd IEEE International Conference on. 2009.
4. Chattopadhyay, G., et al. A 600 GHz Imaging Radar for Contraband Detection. in 19th International Symposium on Space Terahertz Technology. 28 - 30 April 2008. Groningen.
5. Appleby, R. and H.B. Wallace, Standoff Detection of Weapons and Contraband in the 100 GHz to 1 THz Region. IEEE Transactions on Antennas and Propagation, 2007. 55(11): p. 2944.
6. Davies, A.G., et al., Terahertz spectroscopy of explosives and drugs. Materials Today, 2008. 11(3): p. 18-26.
7. Tonouchi, M., Cutting-edge terahertz technology. Nat Photon, 2007. 1(2): p. 97-105.
8. Yujiri, L., M. Shoucri, and P. Moffa, Passive Millimeter-Wave Imaging. IEEE Microwave Magazine, 2003: p. 39 - 50.
9. Borsuk, G.M. and B. Levush. PL-2: Vacuum electronics research perspective at the naval research laboratory. in Vacuum Electronics Conference (IVEC), 2010 IEEE International. 2010.
10. Lin, M.C., et al. Modeling of multistage depressed collectors using a 3D conformal finite-difference time-domain particle-in-cell code. in Plasma Science, 2010 Abstracts IEEE International Conference on. 2010.
11. Qiu, J.X., et al., Vacuum Tube Amplifiers. IEEE Microwave Magazine, 2009: p. 38 - 51.
12. Abrams, R.H., et al., Vacuum Electronics for the 21st century. IEEE Microwave Magazine, 2001: p. 61 - 72.
13. Coaker, B. and T. Challis, Traveling Wave Tubes: Modern Devices and Contemporary Applications. Microwave Journal, 2008: p. 32 - 46.
14. Booske, J.H., et al., Vacuum Electronic High Power Terahertz Sources. Terahertz Science and Technology, IEEE Transactions on, 2011. 1(1): p. 54-75.
15. Young-Min, S., L.R. Barnett, and N.C. Luhmann, Phase-Shifted Traveling-Wave-Tube Circuit for Ultrawideband High-Power Submillimeter-Wave Generation. Electron Devices, IEEE Transactions on, 2009. 56(5): p. 706-712.
16. Shin, Y.-M. and L.R. Barnett, Intense wideband terahertz amplification using phase shifted periodic electron-plasmon coupling. Applied Physics Letters, 2008. 92(9): p. 091501-3.
17. Shin, Y.-M., L.R. Barnett, and J.N.C. Luhmann, Strongly confined plasmonic wave propagation through an ultrawideband staggered double grating waveguide. Applied Physics Letters, 2008. 93(22): p. 221504-3.
18. Baig, A., et al., Design, Fabrication and RF Testing of Near-THz Sheet Beam TWTA (Invited). Terahertz Science and Technology, Dec 2011. 4(4).
19. Seong-Tae, H., Investigations on a Microfabricated FWTWT Oscillator. IEEE transactions on Electron Devices, 2005. 52(5).
20. Young-Min, S., et al., Novel Coupled Cavity TWT Structure Using Two-Step LIGA Fabrication. IEEE Transactions on Plasma Science, 2003. 31(6).
21. Ives, R.L., Microfabrication of High Frequency Vacuum Electron Devices. IEEE Transactions on Plasma Science, 2004. 32(3): p. 1277 - 1291.
22. Lueck, M.R., et al., Microfabrication of diamonf-based slow wave circuits for mm-wave and THz vacuum electronics sources. J. Micromech. Microeng. , 2011. 21: p. 1 - 7.
23. Young-Min, S., et al., UV Lithography and Molding Fabrication of Ultrathick Micrometallic Structures Using a KMPR Photoresist. Microelectromechanical Systems, Journal of, 2010. 19(3): p. 683-689.
24. Shin, Y.-M., et al., Investigation of terahertz sheet beam traveling wave tube amplifier with nanocomposite cathode. Physics of Plasmas, 2010. 17(12): p. 123105.
25. Wang, S., Scandate cathode for TWT. Applied Surface Science, 2005. 251(1-4): p. 114-119.
26. Jinfeng, Z., et al., High Current Density and Long-Life Nanocomposite Scandate Dispenser Cathode Fabrication. Electron Devices, IEEE Transactions on, 2011. 58(4): p. 1221-1228.
27. R. L. Ives, et al., Electron Guns for Terahertz Vacuum Electron Sources (Invited Paper). Special Issue on Microfabricated THz Vacuum Electron Devices, International Journal THz Science and Technology., In Press, 2011.
28. Baig, A., et al. Beam transport modeling of PPM focused THz sheet beam TWT circuit. in Vacuum Electronics Conference (IVEC), 2011 IEEE International. 2011.
29. Young-Min, S., et al., Modeling Investigation of an Ultrawideband Terahertz Sheet Beam Traveling-Wave Tube Amplifier Circuit. Electron Devices, IEEE Transactions on, 2011. 58(9): p. 3213-3218.
30. Baig, A., et al. Experimental characterization of LIGA fabricated 0.22 THz TWT circuits. in Vacuum Electronics Conference (IVEC), 2011 IEEE International. 2011.
31. Robert, B., et al., THz sheet beam traveling wave tube amplifier for microwave power module (MPM) application: MEMS-fabrications and characteristic analysis. Ieee International Conference On Plasma Science. Proceedings, 2011: p. 1-1.