Alexander Arsenovic

A Micromachined Terahertz Waveguide 90

Waveguide twists are often necessary to provide polarization rotation between waveguide-based components. At terahertz frequencies, it is desirable to use a twist design that is compact in order to reduce loss; however, these designs are difficult if not impossible to realize using standard machining. This paper presents a micromachined compact waveguide twist for terahertz frequencies. The Rud-Kirilenko twist geometry is ideally suited to the micromachining processes developed at the University of Virginia. Measurements of a WR-1.5 micromachined twist exhibit a return loss near 20 dB and a median insertion loss of 0.5 dB from 600 to 750 GHz.

^{\circ}

An improved micromachined on-wafer probe covering frequencies 500-750 GHz is demonstrated in this paper to address sub-millimeter-wave integrated-circuit testing. Measurements of a prototype WR-1.5 micromachined on-wafer probe exhibit a return loss better than 12 dB and a mean insertion loss of 6.5 dB from 500 to 750 GHz. The repeatability of on-wafer measurements with the micromachined probe is investigated. Monte Carlo simulations are used to identify the dominant error source of on-wafer measurement and to estimate the measurement accuracy. The dominant error source is positioning error, which results in phase uncertainty. Reliability tests show the probe is robust and can sustain over 20 000 contacts.

Twist

Although progress has been made in the development of submillimeter-wave monolithic integrated circuits, the evaluation of these circuits still relies on test fixtures, which makes testing expensive and time consuming. Based on a W-band prototype, a micromachined on-wafer probe covering frequencies 500–750 GHz is built to simplify submillimeter-wave integrated circuits testing. This paper demonstrates the repeatability and the robustness of this terahertz micromachined on-wafer probe.

Terahertz Micromachined On-Wafer Probes: Repeatability and Reliability

An experimentally based technique for characterizing calibration uncertainty is presented. The approach described calculates uncertainty metrics at the output of the calibration processing chain as opposed to the input. In doing so, this method replaces the complexities of error propagation with the computational effort associated with performing numerous calibrations. Practical applications are demonstrated for a variety of scenarios focused on one-port calibration, illustrating the versatility of the technique. An open-source implementation has been made publicly available as part of the Python module scikit-rf.

Terahertz micromachined on-wafer probes: Repeatability and robustness

This paper presents a study characterizing the connection repeatability and reflection coefficient of submillimeter waveguide flanges in the 500-750 GHz band (WR-1.5 or WM-380). The connection repeatability of four types of flange was measured using one-port measurements and a “load-reference” technique with a vertically mounted system to mitigate gravitational bias. To measure the error-corrected complex reflection coefficients of pairs of waveguide flanges, a calibration procedure insensitive to flange misalignment was used. This SDD(RO) calibration method employs four standards: a flush short, two delay shorts with different but unspecified offsets, and a radiating open-ended waveguide. The uncertainty associated with this calibration method is investigated and it is used to estimate the reflection coefficient resulting from flange misalignment.

An Experimental Technique for Calibration Uncertainty Analysis

We accurately calculate the reflection coefficient and normalized admittance of radiating open-ended rectangular waveguides and fit our results with a linear combination of Legendre polynomials. We verify the expression to an accuracy of 0.005 with other calculations and examine the impact of flanges and burrs on the accuracy to which the reflection coefficient can be approximated in practice.

Repeatability and Mismatch of Waveguide Flanges in the 500–750 GHz Band

A micromachined on-wafer probe has been designed to facilitate the development of integrated circuits in the 600-900 GHz frequency range. The probe tip is fabricated on a 5-micrometer thick high-resistivity silicon substrate using a silicon-on- insulator fabrication process. This letter updates previous work on WR-1.2 wafer probes and presents for the first time the full RF characterization of the probe. These are the first reported on-wafer measurements above 750 GHz.

Legendre Fit to the Reflection Coefficient of a Radiating Rectangular Waveguide Aperture

A prototype imaging reflectometer based on the coded aperture technique and operating in the WR-1.5 (500–750 GHz) frequency band is described. Masks for the coded aperture system are realized through optical modulation of the conductivity of a high-resistivity silicon wafer. A network model representation of the imaging system is developed and applied to measuring beam maps of a submillimeter-wave diagonal horn antenna.

Characterization of Micromachined On-Wafer Probes for the 600–900 GHz Waveguide Band

As waveguide measurements continue to push upwards in frequency, waveguide misalignment becomes a severe problem. In this paper we combine two known calibration algorithms; the SDDL and Usuggest that inaccuracies in the predicted response of the radiating open limits the perfornknown Thru, to create a two-port calibration insensitive to flange misalignment. The resultant algorithm, termed Misalignment Resistant Calibration (MRC), is numerically verified and tested experimentally. While the numerical simulation proves that MRC works as intended, the experimental results suggest that inaccuracies in the predicted response of the radiating open limits the performance of the MRC algorithm in practice.