Xiaobang Shang, Nick Ridler, Daniel Stokes, James Skinner, Faisal Mubarak, Uwe Arz, Gia Ngoc Phung, Karsten Kuhlmann, Alireza Kazemipour, Martin Hudlička, François Ziade
©SHUTTERSTOCK.COM/QUARDIA
The millimeter-wave (mm-wave) and terahertz (THz) regions of the electromagnetic spectrum are seeing increasing prominence, with a range of established and emerging applications—wireless backhaul for mobile networks for 5G and 6G infrastructure, automotive radar sensors, space-deployed radiometers for Earth observation, climate monitoring, weather forecasting, and more—exploiting these frequencies. Measurement techniques informed by the latest metrological research into the measurement of electrical quantities underpin the development of components and circuits for such applications by enabling the accurate and consistent measurement results to be obtained.
However, achieving accurate measurements at these frequencies is no simple task. Therefore, the European Metrology Program for Innovation and Research (EMPIR) Traceability for Electrical Measurements at Millimetre-Wave and Terahertz Frequencies for Communications and Electronics Technologies (TEMMT) project (from 2019 to 2022) [1] brought together the metrology institutes of various nations, industrial partners, and university faculties with the aim of establishing traceability to the International System of Units (SI) for three important electrical measurement quantities: 1) S-parameters (in various transmission line media), 2) power, and 3) the complex permittivity of dielectric materials at mm-wave and THz frequencies. This article summarizes the main outcomes of TEMMT, along with other notable advancements in this field.
In the international trading and supply chain, measurement traceability plays a key role, for example, by enabling suppliers to provide confidence in their products through defensible measurements and specifications. The establishment and maintenance of, as well as providing access to, this traceability is the responsibility of National Metrology Institutes (NMIs). As described in [2], for end users, traceable measurements can be obtained by calibrating (or checking) the measurement instrument using certified reference standards. These standards are usually calibrated (or checked) against other references belonging to or located at another organization (e.g., a certified standards lab), and ultimately, linked to the standard that is owned by an NMI. The national standards themselves are periodically compared with those of other nations to ensure that these standards are linked and equivalent to each other. As a result of this, measurements made anywhere in the world can be compared since they are linked to a common reference (i.e., national and/or international standards), leading to international equivalence in measurements. In this traceability process, measured quantities of all varieties are ultimately referred to standards defined by the seven base quantities (and units) used within the SI [2], [3].
The capability and functionality of test instrumentation enabling state-of-the-art high-frequency measurements up to THz frequencies have advanced dramatically over the last decade, but traceability to the SI has lagged behind these technical developments. Hence, it was timely to establish traceability for electrical measurements at mm-wave and THz frequencies by the TEMMT project partners. The consortium was formed of 19 partners, ensuring access to the array of facilities and skills demanded by such a project. The participating NMIs typically spearheaded the activities, not only due to their role as guardians of the SI but also as their remit involves a strong connection with industry, government, and academia, making them natural leaders for this collaborative endeavor. In the developments that will be described in this article, the general approach taken by the project partners to establish traceability for each measurement quantity can be summarized as follows:
With the recent increase in the use of the mm-wave and THz regions of the electromagnetic spectrum, new interconnects in different transmission lines, particularly coaxial lines and rectangular metallic waveguides, have been introduced to enable the development of systems operating at these frequencies. As new interconnects are introduced, new traceable measurement capabilities (particularly for S-parameters) are required to reliably test and verify the performance of components and systems utilizing these interconnects.
Coaxial lines are the most common transmission lines for frequencies up to around 100 GHz. Connectors for these coaxial lines are defined in standards such as IEEE Standard 287 [4]. mm-Wave connectors such as the 1.85-mm (V-band) connector, to 67 GHz, and the 1.00-mm (W-band) connector, to 110 GHz, are included in this standard, and measurement traceability for both these connector sizes has already been established (e.g., [5], [6], [7]). In recent years a new connector—the 1.35-mm (E-band) connector, to 90 GHz—has been introduced. This connector has recently been defined by a new International Electrotechnical Commission (IEC) standard [8] and has now been added to the latest edition of the IEEE 287 series of standards [4].
This 1.35-mm connector fills the gap between the 1.85- and 1.00-mm connectors and is ideal for applications such as 5G front-haul equipment (above 70 GHz) and automotive radar systems operating at around 80 GHz. It provides an attractive alternative to using the delicate 1-mm connector for applications up to 90 GHz as its relatively larger conductor dimensions, particularly for the center conductor, and higher connector coupling torque provide greater durability. This connector is already being used on test equipment such as the Rohde & Schwarz (R&S) 90 GHz thermal power sensor [9].
Traceability for S-parameter measurements in 1.35-mm coaxial line has been advanced by three European NMIs as part of the TEMMT project. Although some work on traceability for this connector has been described elsewhere [10], the work in the TEMMT project has established both dimensional and electrical traceability.
Dimensional traceability is necessary for the definition of calibration standards at higher frequencies where models based on traditional electrical circuit parameters (inductance, resistance, and capacitance) are no longer adequate. Precision dimensional measurements of conductor and connector geometries, including the diameters of the center conductor, d, and outer conductor, D, and of reference standards, such as offset short circuits (which are composed of a line section and shorting plane) and air dielectric coaxial transmission lines (air lines), enable the simulation and calculation of the standards’ characteristics. For example, the characteristic impedance, Z0, of an air line can be calculated using \[{Z}_{0} = \frac{1}{{2}{\pi}}\sqrt{\frac{\mu}{\varepsilon}}\ln\frac{D}{d} \tag{1} \] where µ is the permeability, and $\varepsilon$ is the permittivity of the dielectric of the coaxial line. For the 1.35-mm connector, the nominal values for D and d are 1.350 and 0.586 mm, respectively [4], [8]. Calibrated dimensional measurement systems such as air gauges and laser micrometers can be used to measure conductor diameters, while telecentric imaging systems can be used to profile connectors, and thus, provide traceability to the SI base unit of length—the meter—for the derived electrical characteristics of the standards. These can then in turn be used to make traceable measurements of electrical quantities such as S-parameters using a vector network analyzer (VNA) and calibration schemes such as through-reflect-line (TRL) and multiple-offset-short-through (SOOT) [10]. A photograph of a 1.35-mm air line is shown in Figure 1(a).
Figure 1. (a) A photograph of a 1.35-mm unsupported air line. The center conductor (shown on the right) is inserted inside the outer conductor (shown on the left) during connection (e.g., to VNA test ports). (b) A plot of the three comparison participants’ measurement results and uncertainties for the magnitude of the reflection (S11) and transmission (S21) parameters for a mismatched adaptor in 1.35-mm coaxial line to 95 GHz.
With these new measurement systems and calibration schemes established, the next phase of the project activity saw a comparison of measurements among the three NMIs [11]. A set of four 1.35-mm devices with a variety of electrical characteristics was measured by each laboratory: an offset short-circuit (high reflect), a matched load (low reflect), a matched adaptor (high transmission), and a mismatched adaptor (varying transmission). Some example results from this comparison, for S11 and S21 of the mismatched adaptor, is shown in Figure 1(b).
As can be seen, two of the participants’ measurement methods produced results with good agreement, with the results of the third deviating at higher frequencies. This deviation highlights one of the key motivations behind establishing SI traceable calibration schemes in this new line size as this third measurement result, shown by the black traces, was obtained by using a conventional short-open-load-reciprocal (SOLR) calibration scheme, relying on the manufacturer’s standard definitions based on traditional electrical circuit models, as opposed to the red and dark red traces, which were produced by calibrations with traceably characterized reference standards.
Rectangular metallic waveguides have been used for many years as part of microwave and mm-wave systems. Due to their relatively low loss, such waveguides are the transmission line medium of choice for frequencies above 100 GHz to at least 1 THz (see “Transmission Line Media for mm-Wave and THz Frequencies”). Although measurement traceability has already been established for some of the mm-wave and THz waveguide bands [12], [13], [14], [15], [16], there are still frequency ranges lacking in traceability—in particular, from 330 to 750 GHz and above 1.1 THz.
Transmission Line Media for mm-Wave and THz Frequencies
When it comes to mm-wave and THz applications, you may be wondering which transmission line medium is most suited to your needs. In Table S1, an overview is given of the three key media, covering the fundamental features of each.
Table S1. Key features of three conventional transmission line medium.
The TEMMT project partners developed new measurement systems for the WM-570 (330–500 GHz), WM-380 (500–750 GHz) and WM-164 (1.1–1.5-THz) bands. For the calibration of their systems, the partners generally utilized conventional calibration techniques such as short-offset short-load-through (SOLT) using off-the-shelf calibration standards and manufacturers’ standard definitions. However, the participating NMI in this activity developed measurement traceability for the WM-570 and WM-380 bands through the use of the three-quarter-wave TRL calibration technique [17], [18]. This consisted of acquiring, characterizing, and testing suitable reference standards through a rigorous process described in [19] and [20], involving thorough dimensional and electrical assessments.
Interlaboratory comparison exercises were undertaken for each of the three waveguide bands [21], [22]. For instance, the WM-380 measurement comparison involved the measurement by four of the project partners of the devices shown in Figure 2(a). Figure 2(b) shows the measurements of the S21 transmission S-parameter of the 10-dB attenuator obtained by the four participants, with the mean of the results shown in green.
Figure 2. (a) A photograph of the devices measured in the TEMMT WM-380 measurement comparison exercise: 1-in section (top left); 10-dB attenuator (top right); cross-guide line (bottom left); offset short circuit (bottom right). (b) A plot of the comparison results for the transmission (S21) measurements of the 10-dB attenuator. The sudden decrease in transmission at around 560 GHz is due to signal absorption by water vapor present in the air inside the waveguide.
The results shown in Figure 2(b) are a good demonstration of the benefits of traceability. Whereas some results exhibit rapid oscillations or deviations from the general trend, the blue trace (produced by the calibration using the standards described in [19] and [20]) processes relatively smoothly with frequency and agrees well with the mean result throughout the band, both indicative of a reliable result.
Metrological traceability for S-parameters has been extended and improved substantially due to work undertaken in the TEMMT project, specifically for the 1.35-mm (E-band) coaxial connector and for waveguides in the WM-570, WM-380, and WM-164 bands. These achievements will enable the transfer of this traceability from the participating NMIs to industry in support of applications utilizing these connectors operating in the mm-wave and THz frequency bands.
Monolithic mm-wave integrated circuit technology underpins a wide variety of applications, from communications to sensing and security. However, the development of such integrated circuits is hampered by the lack of traceability for on-wafer S-parameter measurements. Traceability has been demonstrated up to 110 GHz [23] but no higher—despite the availability of systems (i.e., VNAs, probes, and calibration substrates) operating up to 1.1 THz. The extension of traceability to these frequencies poses challenges in every element of on-wafer measurement, from the design of reference standards to calibration and verification techniques, probing methods, and measurement uncertainty.
To establish accurate traceable on-wafer S-parameter measurements up to THz frequencies, the project partners designed, manufactured, and measured custom reference substrates. Identical contact pads were included in the design of all structures as accurate definitions of the reference planes are essential in realizing consistent and accurate results [24], [25]. The substrates included a low-band set of 35 devices, designated kit 1, designed for operation between 110 and 330 GHz, and a high-band set of 35 devices, designated kit 2, designed for operation between 330 GHz and 1.1 THz. The combination of structures allows a range of calibration techniques to be implemented, i.e., TRL, multiline TRL (mTRL), SOLR, and 16-term calibrations. To account for the destructive nature of on-wafer probing, eight identical reference dies were fabricated on each 3-in semi-insulating high-resistivity (>5,000 Ω/cm) silicon wafer, shown in Figure 3. The full fabrication process is detailed in [26]. Each die included both kits, and the dies were subsequently used in an interlaboratory comparison between the TEMMT project partners, detailed in a following section.
Figure 3. SEM images of (a) a section of a fabricated reference substrate, including four CPW structures from kit 1, and (b) the thru standard from kit 2, with various device parameters labeled that were measured for uncertainty evaluation [26].
A white-light interferometer and a scanning electron microscope (SEM) were used to measure the characteristics of the fabricated wafer. This information was subsequently used to estimate the reference values with uncertainties corresponding to each device. 3D electromagnetic simulation tools were used to achieve these estimates for the calibration standards, an approach proven to improve on-wafer calibration accuracy significantly [27], [28]. The nominal S-parameters corresponding to each standard were estimated using electromagnetic (EM)-field simulations. Each device’s dimensional and material parameter tolerances were propagated through this simulation to estimate the corresponding variation in the S-parameters, leading to an assessment of the uncertainty contributions.
The extension of traceability up to 1.1 THz requires an evaluation of the suitability of commonly used calibration techniques. mTRL [29] calibration has been widely regarded as one of the most accurate on-wafer calibration techniques. This was also demonstrated by an interlaboratory study [30] conducted in the G-band (140–220 GHz) where three common calibration methods (i.e., SOLT, on-wafer mTRL, and off-wafer TRL) were employed by the participants.
Several guidelines have been developed to evaluate uncertainty in connectorized measurements [31] and on-wafer measurements [32] (produced from the PlanarCal project [33]). The sources of uncertainties associated with high-frequency on-wafer measurements have been extensively studied, for example, the studies on probe crosstalk and coupling effects [34], [35], [36], [37], [38] and boundary conditions [39].
A second key aspect of traceability relies on measurement comparisons to demonstrate the equivalence of measurement results across different systems. Hence, a comparison of on-wafer S-parameter measurements was held between the TEMMT partners, using the fabricated reference substrates at frequencies ranging from 10 GHz to 1.1 THz [26], with each of the participating laboratories receiving one of the custom reference wafers. Measurement setups used by the participants consisted of VNAs, frequency extender units, and probes, each from various vendors. Four laboratories participated in the comparison, measuring reference kit 1 up to 330 GHz, and three laboratories measured kit 2 from 330 GHz up to 1.1 THz. The results for a selection of the devices are shown in Figure 4.
Figure 4. (a) and (b) comparison results for reflection measurements of 50-Ω load devices corresponding to (a) kit 1 and (b) kit 2. (c) Comparison device reference values determined by the median of the measurement results for 50-Ω load, open circuit, and short circuit devices. Results with round markers correspond to devices from kit 1, and square markers correspond to kit 2 [26]. (a) Kit 1: 50-Ω load. (b) Kit 2: 50-Ω load. (c) Measurement results. FBH: Ferdinand‐Braun‐Institut.
Combined uncertainty contributions were estimated by PTB for their measurements up to 330 GHz. The Van Swinden Laboratorium (VSL) uncertainty estimates for measurements of the comparison artifacts primarily focused on calibration standards and probe crosstalk uncertainties up to 1.1 THz. The uncertainty estimates from the two laboratories show good agreement up to 330 GHz. The comparison results for kit 1 agreed with the estimated uncertainties from the EM simulations. For kit 2, results were found in moderate agreement with the estimated uncertainties. Detailed analysis of the comparison results can be found in [26].
Of the various electrical measurement domains, the on-wafer domain has proven one of the most challenging for extending traceability to higher frequencies. This effort has necessitated a truly multifaceted approach: the use of the latest in VNA instrumentation, nanometer-scale fabrication techniques, electron imaging, electromagnetic simulation, machine learning, and the application of the combined knowledge of the leaders in the field. Metrological research in this area is ongoing, but the outcomes of the TEMMT project demonstrate vital progress.
Power is one of the fundamental electrical quantities, and it is of key importance to many industrial and end-user applications. Nowadays, power meters operating up to 750 GHz are commercially available, e.g., [40]. However, there is no SI traceability accessible for industrial and other end users to benchmark measurements at mm-wave and THz frequencies; until very recently, power measurement traceability had been established only for frequencies up to 110 GHz [41], [42], [43], [44]. The European Telecommunication Standards Institute (ETSI) has stated that output power limits up to 246 GHz cannot be supported due to the lack of both precision power measurement equipment and traceability to the SI, which impacted a range of standards: for radio spectrum usage, short-range communication devices, movement detection equipment, level-probing radar, intelligent transport systems, and road transport telematics [45].
Furthermore, to verify radiation safety limits, there is also a need for the traceability of power-related free-space quantities such as power flux density and electric and magnetic field strength. These quantities are directly traced to the key quantity of high-frequency power. Based on the guidelines of the International Commission on Non-Ionizing Radiation Protection (ICNIRP), which specifies the maximum radiation ratings up to 300 GHz, the European Physical Agents Directive (EC 2013/35/EU) has set limits for personal working between 6 and 300 GHz to 50 W/m2 in controlled environments. In addition, the availability of traceable power flux density measurements will be a prerequisite for future standardization of personal scanners and spectrometers by the IEC or European Committee for Electrotechnical Standardization (CENELEC). These important safety legislations are relying on the development of power measurement traceability into the mm-wave and sub-THz frequency bands. Therefore, in the framework of the TEMMT project, several European NMIs extended their power measurement capabilities with the goal of establishing this much-needed traceability.
Microcalorimetry is commonly utilized to establish power measurement traceability [41]. In the TEMMT project, three NMIs—NPL (United Kingdom), PTB (Germany), and LNE (France)—developed microcalorimeters for the 110–170 GHz waveguide band—denoted as the WG29, WR-06, or D-band. Each laboratory implemented different methods for the characterization of these systems—multiple-offset-short (PTB) [46], [47], flush-short (NPL) [48], and VNA methods (LNE). To validate these methods, microcalorimeter measurements were carried out using both thermoelectric and thin-film bolometric power sensors. These sensors were designed and manufactured by R&S and the University of Birmingham, respectively.
Seven prototypes of the R&S thermoelectric power sensor (see Figure 5) were manufactured, with LNE, PTB, and NPL each calibrating two sensors on their microcalorimeter systems. To demonstrate the equivalence of each of these systems, a comparison of the generalized effective efficiency of the transfer standards was performed using the direct comparison transfer measurement systems (DCTMS) at PTB [49], [50].
Figure 5. The thermoelectric power sensor used for establishing power measurement traceability up to 170 GHz. (a) The sensor schematic. (b) A photograph of the device manufactured by R&S.
The generalized efficiency of sensor X is as follows: \[{\eta}_{eff,\text{gen},X} = \frac{{\eta}_{\text{cal},X}}{\left({{1}{-}{\left|{{\Gamma}_{TS,X}}\right|}^{2}}\right)} \tag{2} \] where ${\eta}_{\text{cal},X}$ is the calibration factor, and ${\Gamma}_{TS,X}$ are the reflection coefficients of the device under test (X).
Measurements were made at frequencies between 110 and 170 GHz for both sensors. The generalized effective efficiency results from the DCTMS and the two microcalorimeter systems, with expanded measurement uncertainties, are given in Figure 6. Overall, the results of the calorimeter measurements lie within the uncertainties of DCTMS results, demonstrating good agreement between the three NMIs.
Figure 6. (a) and (b) The results of a measurement comparison of thermoelectric power sensors between the direct comparison method utilized by PTB against calorimeter systems at (a) NPL and (b) LNE (right).
The second type of power transfer standard developed for this frequency band was a bolometric power sensor (see Figure 7). This type of sensor, as presented in [51], [52], [53], and [54], is based on a temperature-dependent element such as a thermistor. The temperature of this element varies when subjected to a microwave signal. This causes a change in its electrical resistance, which in turn gives a measure of the level of the microwave power. It has a multilayer sensor chip, consisting of a silicon substrate as a microwave absorber and a platinum thin-film layer that acts as the detector.
Figure 7. (a) A thin-film bolometric power sensor model and (b) manufactured devices supplied by the University of Birmingham, assembled with the printed circuit board (PCB) and connectors attached. (c) A graph of the power sensor/meter’s effective efficiency across D-Band frequencies measured at NPL.
To characterize the thin-film bolometric power sensors, a series of measurements was performed: return loss, short-term time response, resistance change, effective efficiency, and measurement system coefficient. Figure 7(c) shows the effective efficiency results for one of the sensors measured at NPL. The results vary between 0.93 and 0.99 across the entire frequency range. The sensors demonstrated excellent matching, with a return loss of better than 15 dB across the entire D-band. Applying levels of microwave power between 0.01 and 7 mW produced a response in the sensing element resistance of around 37 µΩ/Ω/mW, demonstrating good linearity. In addition, the sensor demonstrated a stable long-term frequency response.
The equivalence of the power measurement setups for mm-wave frequencies at the three European NMIs has been validated via a comparison of the generalized effective efficiency of thermoelectric power sensors provided by R&S. Overall, good agreement between each of the different partners was obtained—a key indicator of the success of this effort to extend power measurement traceability. Furthermore, the project has seen the successful design, fabrication, and characterization of a novel D-band bolometric power sensor. The results of the characterization of the sensor demonstrate its suitability for use as a transfer standard for traceable power calibrations at mm-wave frequencies.
The success of this development work has paved the way for the extension of traceable power measurements up to 170 GHz. Through dissemination of the project outputs, the participating NMIs are able to extend the benefits of these achievements to industry. Furthermore, the success of both the manufacturing of new standards and the application of measurement techniques in this range provides a solid foundation for future developments in power traceability at even higher frequencies.
Electromagnetic characterization of materials is a key parameter for space applications, imaging, and the high-speed electronics and telecommunications industries. In fact, “a priori” knowledge of different materials’ permittivity and permeability informs electromagnetic simulation to optimize prototyping/testing processes and thereby reduce the cost of production. A variety of techniques are available for the measurement of these quantities, each utilizing different instrumentation and physical phenomena. As such, each technique is best suited for a particular part of the frequency spectrum and for the measurement of particular materials, whether lossy or low loss or of various sizes and thicknesses [55].
Closed and open transmission cells and waveguides [56] or resonant cavities [57] are usually used in the microwave region. However, these are less effective in the mm-wave/THz domain due to mechanical limitations and dimensional restrictions. Therefore, free-space non-contact methods are often used at these frequencies [58], [59], [60]. Laser-based optical techniques have traditionally been the mainstream solution for THz spectroscopy [61], but recent progress in high-frequency semiconductor technology has enabled the development of waveguide frequency extenders that enable the characterization of materials on VNA-based setups at frequencies as high as 1.5 THz. With both optical and VNA-based methods available to them, the project partners were able to approach the challenge of characterizing materials at the crossover region of these methods, with the potential for a rich and varied comparison of results from a range of techniques.
Frequency-domain spectroscopy (FDS), time-domain spectroscopy (TDS) [62], and time-domain ellipsometry methods are established techniques for the measurements of THz signals and were used by the partners for optical-based measurements. Commercially available compact mode converters for different frequency bands, i.e., SWISSto12 material characterization kits (MCKs) [63], were used for the VNA-based measurements. The following sections provide an explanation of the different techniques and measurement setups established by the project partners and give a sample of the results of an interlaboratory comparison where each participant deployed their systems and techniques for the measurement of a range of materials under test (MUTs).
FDS systems typically use photomixers and two continuous wave (CW) lasers to generate and detect CW THz signals. The FDS system established at METAS is shown in Figure 8(a). Traceability for the measurement of complex permittivity on such systems is established by measuring the optical frequencies of the two CW lasers with a reference wavemeter.
Figure 8. (a) A THz frequency-domain spectrometer at METAS, Switzerland. The black cylinders are photomixers, and the green area indicates the THz signal path, guided via parabolic mirrors. (b) The SWISSto12 MCK connected to VNA frequency extension modules, with the MUT clamped between the two antenna apertures.
From the complex transmission data and the previously determined specimen thickness values, the material parameters are extracted via the standard method. The Fresnel transmission coefficient equation is inverted to get the complex refractive index, from which all other material properties, such as the permittivity or the absorption coefficient, can be determined.
THz TDS systems were established by the project partners using commercially available instrumentation. The THz beam path is purged with dry air to eliminate absorption from atmospheric water vapor. Samples are placed in the collimated section of the beam, and laser alignment is used to ensure the sample positioning normal to the THz beam. The typical frequency resolution achieved is 10 GHz.
A Fourier transform is applied to the time-domain data to obtain the frequency-dependent field amplitude (E(f)) and phase (z(f)). The frequency-dependent refractive index (n(f)) and absorption coefficient (a(f)) of each sample are calculated using the following equations: \begin{align*}{n}\left({f}\right) & = {1}\,{+}\,\frac{\left({{\phi}_{s}\left({f}\right){-}{\phi}_{\text{ref}}\left({f}\right)}\right){c}}{{2}{\pi}{ft}} \tag{3} \\ {\alpha}\left({f}\right) & = {-}\frac{2}{t}\,\ln\left[{\frac{{\left({{n}\,{+}\,{1}}\right)}^{2}}{4n}\frac{{E}_{s}\left({f}\right)}{{E}_{\text{ref}}\left({f}\right)}}\right] \tag{4} \end{align*} where c is the speed of light, t is the sample thickness, and the subscripts s and ref refer to the sample and reference data, respectively.
Nowadays, VNA systems typically operate in their native coaxial configuration up to 110 GHz, but waveguide extension units can be used to access THz frequencies. The addition of quasi-optical components can enable the characterization of materials by converting waveguide modes to free-space propagation. Such a setup typically consists of a VNA with extender heads connected to receiving and transmitting antennas on either side of a sample holder, with mirrors and dielectric lenses used to focus the beam onto the MUT.
The SWISSto12 MCK uses a set of two long corrugated antennas with a suitable aperture feature to clamp a material slab [Figure 8(b)]. The complex permittivity of a MUT can be obtained from measurements of transmission coefficient (S21) following a normalization or calibration process [58]. The manufacturer suggests using a typical calibration method and time-domain gating (as often used in many other commercial and research-oriented setups) of the measured S-parameters to help extract the permittivity of the material.
Open or closed resonator techniques have been widely utilized for the accurate measurement of low- to medium-loss dielectric materials at microwave frequencies. The material properties can be determined from the changes in Q-factor and resonant frequencies of the resonator with or without the specimen loaded. Under TEMMT, an open resonator operating at around 144 GHz was developed at NPL. The measurement uses the fixed-frequency variable-length method. Resonator-based techniques typically have stringent requirements on specimen sizes, e.g., the specimen thickness should be approximately an integral number of half-wavelengths (in the medium of the dielectric). Therefore, a narrower range of materials can be characterized using these systems.
Having established these measurement setups and techniques for characterizing materials at mm-wave and THz frequencies, the project partners put them to the test in a measurement comparison. MCKs operating at frequencies between 50 and 750 GHz, utilizing optimized calibration and parameter-extraction techniques, were compared against two “traditional” methods—i.e., optical-based and narrow-band resonator systems. Samples with a range of characteristics were measured (i.e., lossy, low loss, thin, and thick) to compare the effectiveness of each method for measuring complex permittivity. Comparison results for the measurements of two of the samples are shown in Figure 9. In general, good agreement can be observed between the results from different methods and setups, and the full comparison report has since been published [64].
Figure 9. The comparison results for the measurement of real components of the complex permittivity of (a) a fused silica sample of 1-mm thickness and (b) a Borofloat glass sample of 0.5-mm thickness. Open resonator (OR) data at 144 GHz is included for the fused silica sample. UniFR: University of Fribourg.
As an outcome of the TEMMT project, new measurement capabilities were established to characterize material parameters at mm-wave and THz frequencies. A comprehensive comparison between different setups and different parameter-extraction techniques was achieved for the first time over a very wide frequency band—all the way up to 750 GHz. A good general agreement between different participants was observed for many samples, indicating the success of this effort to advance materials characterization capabilities into the mm-wave and THz region.
Through their achievements in this work, the TEMMT project partners have paved the way for traceability and improvement in measurement accuracy for three fundamental electrical quantities at mm-wave and THz frequencies: S-parameters, power, and complex permittivity. Improvements in measurement accuracy and the establishment of measurement traceability provide manufacturers and industrial laboratories with confidence in their measurement capabilities, enabling reliable specifications of their products and devices to be obtained.
New measurement systems are now in place at three NMIs—LNE, NPL, and PTB—to enable both dimensional and electrical measurements of coaxial devices in the 1.35-mm line size. The work done is completely in line with industrial needs; for instance, R&S, a partner in the project, has already implemented the new 1.35-mm E-band coaxial connector in the design of two new thermal power sensors, the R&S NRP90T and NRP90TN. This project also saw a significant extension of the measurement capabilities of the partners: to more than 1 THz for S-parameters, 170 GHz for power, and 750 GHz for material characterization. These capability extensions will provide measurement confidence to industry for developments in electronics and communications technology, which are climbing higher up the frequency spectrum year after year, not to mention the many developments taking place in the space industry, which will require the highest level of measurement accuracy.
As the old saying of the microwave community goes, “The higher the frequency, the more challenging the measurement.” In pushing the boundaries of electrical measurement capabilities, the TEMMT project partners have boldly taken on the challenges presented by the mm-wave and THz bands of the electromagnetic spectrum, and the capabilities that have been established will facilitate the advancement of next-generation electronics and communications technologies.
This work was supported by the European Metrology Program for Innovation and Research (EMPIR) Project 18SIB09 TEMMT. The EMPIR Program is cofinanced by the participating states and by the European Union’s Horizon 2020 Research and Innovation Program. The authors would like to thank all TEMMT consortium members for their contributions to the work described in this article.
[1] “Traceability for electrical measurements at millimetre-wave and terahertz frequencies for communications and electronics technologies,” LNE, Paris, France, 2019–2022. Accessed: Dec. 23, 2022. [Online] . Available: https://projects.lne.eu/jrp-temmt/
[2] N. Ridler, R. Clarke, M. Salter, and A. Wilson, “The Trace is on measurements: Developing traceability for s-parameter measurements at millimeter and submillimeter wavelengths,” IEEE Microw. Mag., vol. 14, no. 7, pp. 67–74, Nov./Dec. 2013, doi: 10.1109/MMM.2013.2280311.
[3] A. Rumiantsev and N. Ridler, “VNA calibration,” IEEE Microw. Mag., vol. 9, no. 3, pp. 86–99, Jun. 2008, doi: 10.1109/MMM.2008.919925.
[4] IEEE Standard for Precision Coaxial Connectors at RF, Microwave, and Millimeter-Wave Frequencies—Part 1: General Requirements, Definitions, and Detailed Specifications, IEEE Standard 287.1-2021, Sep. 16, 2022.
[5] N. Ridler, “Traceability to national standards for S-parameter measurements of devices fitted with precision 1.85 mm coaxial connectors,” in Proc. 68th Autom. RF Techn. Group (ARFTG) Conf. Dig., Broomfield, CO, USA, Nov./Dec. 2006, pp. 1–10, doi: 10.1109/ARFTG.2006.8361648.
[6] M. Horibe, M. Shida, and K. Komiyama, “Calibration for precision coaxial air lines in the frequency range up to 110 GHz,” in Proc. Conf. Precis. Electromagn. Meas. (CPEM), Broomfield, CO, USA, Jun. 2008, pp. 478–479, doi: 10.1109/CPEM.2008.4574861.
[7] J. Hoffmann, M. Wollensack, J. Ruefenacht, D. Stalder, and M. Zeier, “Traceable calibration with 1.0mm coaxial standards,” in Proc. 87th Microw. Meas. Conf. (ARFTG), San Francisco, CA, USA, Jun. 2016, pp. 1–4, doi: 10.1109/ARFTG.2016.7501943.
[8] Radio-Frequency Connectors – Part 65: Sectional Specification for RF Coaxial Connectors, 1,35 mm Inner Diameter of Outer Conductor, With Screw-Coupling, Characteristic Impedance 50 Ω, IEC Standard 61169-65:2021.
[9] “Rohde & Schwarz first to equip test instruments with the new 1.35 mm E-band coaxial connector,” Microw. J., May 18, 2020. Accessed: Oct. 12, 2023. [Online] . Available: https://www.microwavejournal.com/articles/33994-rohde-schwarz-first-to-equiptest-instruments-with-new-135-mm-e-band-coaxial-connector
[10] M. Horibe and R. Kishikawa, “Precision offset short calibration standards for 1.35 mm coaxial line sizes,” in Proc. 97th ARFTG Microw. Meas. Conf., Atlanta, GA, USA, Dec. 2021, pp. 1–4, doi: 10.1109/ARFTG52261.2021.9640175.
[11] D. Stokes, F. Gellersen, D. Allal, J. Skinner, G. N. Phung, and K. Kuhlmann, “Traceable S-parameter measurements up to 90 GHz in 1.35 mm coaxial,” Meas. Sci. Technol., vol. 34, no. 6, pp. 1–8, Mar. 2023, doi: 10.1088/1361-6501/acc04c.
[12] R. Clarke, R. Pollard, N. Ridler, M. Salter, and A. Wilson, “Traceability to national standards for S-parameter measurements of waveguide devices from 110 GHz to 170 GHz,” in Proc. 73rd ARFTG Microw. Meas. Conf. Dig., Boston, MA, USA, Jun. 2009, pp. 127–136, doi: 10.1109/ARFTG.2009.5278079.
[13] N. Ridler, R. Clarke, M. Salter, and A. Wilson, “Traceability to national standards for s-parameter measurements in waveguide at frequencies from 140 GHz to 220 GHz,” in Proc. 76th ARFTG Microw. Meas. Conf. Dig., Clearwater Beach, FL, USA, Dec. 2010, pp. 8–14, doi: 10.1109/ARFTG76.2010.5700046.
[14] N. Ridler, R. Clarke, M. Salter, and A. Wilson, “Traceability to national standards for S-parameter measurements in waveguide at frequencies from 220 GHz to 330 GHz,” in Proc. 78th ARFTG Microw. Meas. Conf. Dig., Tempe, AZ, USA, Dec. 2011, pp. 103–108, doi: 10.1109/ARFTG78.2011.6183882.
[15] N. Ridler and R. Clarke, “Establishing traceability to the international system of units for scattering parameter measurements from 750 GHz to 1.1 THz,” IEEE Trans. THz Sci. Technol., vol. 6, no. 1, pp. 2–11, Jan. 2016, doi: 10.1109/TTHZ.2015.2502068.
[16] M. Horibe and R. Kishikawa, “Metrological traceability in waveguide S-parameter measurements at 1.0 THz band,” IEEE Trans. Instrum. Meas., vol. 62, no. 6, pp. 1814–1820, Jun. 2013, doi: 10.1109/TIM.2013.2240953.
[17] N. Ridler, “Choosing line lengths for calibrating waveguide vector network analysers at millimetre and sub-millimetre wavelengths,” Nat. Physical Lab., Teddington, U.K., NPL Rep. TQE 5, Mar. 2009.
[18] N. Ridler, R. Clarke, C. Li, and M. Salter, “Strategies for traceable submillimeter-wave vector network analyzer measurements,” IEEE Trans THz Sci. Technol., vol. 9, no. 4, pp. 392–398, Jul. 2019, doi: 10.1109/TTHZ.2019.2911870.
[19] X. Shang, N. Ridler, W. Sun, P. Cooper, and A. Wilson, “Preliminary study on WM-380 waveguide TRL calibration line standards at the U.K.’s National Physical Laboratory,” in Proc. 92nd ARFTG Microw. Meas. Conf., Orlando, FL, USA, 2019, pp. 1–4, doi: 10.1109/ARFTG.2019.8637214.
[20] N. Ridler, S. Johny, M. Salter, X. Shang, W. Sun, and A. Wilson, “Establishing waveguide lines as primary standards for scattering parameter measurements at submillimetre wavelengths,” Metrologia, vol. 58, no. 1, 2021, Art. no. 015015, doi: 10.1088/1681-7575/abd371.
[21] J. Skinner et al., “Interlaboratory comparison of S-parameter measurements in WM-570 waveguide at frequencies from 325 GHz to 500 GHz,” Nat. Physical Lab., Teddington, U.K., NPL Rep. TQE 24, Jul. 2022.
[22] J. Skinner et al., “Interlaboratory comparison of S-parameter measurements in WM-380 waveguide at frequencies from 500 GHz to 750 GHz,” Nat. Physical Lab., Teddington, U.K., NPL Rep. TQE 23, Jul. 2022.
[23] U. Arz et al., “Traceable coplanar waveguide calibrations on fused silica substrates up to 110 GHz,” IEEE Trans. Microw. Theory Techn., vol. 67, no. 6, pp. 2423–2432, Jun. 2019, doi: 10.1109/TMTT.2019.2908857.
[24] L. Galatro, F. Mubarak, and M. Spirito, “On the definition of reference planes in probe-level calibrations,” in Proc. 87th ARFTG Microw. Meas. Conf., San Francisco, CA, USA, 2016, pp. 1–4, doi: 10.1109/ARFTG.2016.7501968.
[25] A. Rumiantsev, R. Doerner, and G. N. Phung, “Calibration substrate design for accurate mm-wave probe-tip calibration,” in Proc. 94th ARFTG Microw. Meas. Symp., San Antonio, TX, USA, 2020, pp. 1–4, doi: 10.1109/ARFTG47584.2020.9071757.
[26] F. Mubarak et al., “An interlaboratory comparison of on-wafer S-parameter measurements up to 1.1 THz,” IEEE Trans. Thz Sci. Technol., to be published.
[27] L. Galatro et al., “Towards commercially available quartz calibration substrates,” in Proc. 95th ARFTG Microw. Meas. Conf., Los Angeles, CA, USA, 2020, pp. 1–5, doi: 10.1109/ARFTG47271.2020.9241374.
[28] L. Galatro and M. Spirito, “Millimeter-wave on-wafer TRL calibration employing 3-D EM simulation-based characteristic impedance extraction,” IEEE Trans. Microw. Theory Techn., vol. 65, no. 4, pp. 1315–1323, Apr. 2017, doi: 10.1109/TMTT.2016.2609413.
[29] R. B. Marks, “A multiline method of network analyzer calibration,” IEEE Trans. Microw. Theory Techn., vol. 39, no. 7, pp. 1205–1215, Jul. 1991, doi: 10.1109/22.85388.
[30] R. Clarke, X. Shang, N. Ridler, R. Lozar, T. Probst, and U. Arz, “An interlaboratory study of the reproducibility of on-wafer S-parameter measurements from 140 GHz to 220 GHz,” in Proc. 94th ARFTG Microw. Meas. Symp., San Antonio, TX, USA, 2020, pp. 1–4, doi: 10.1109/ARFTG47584.2020.9071783.
[31] “Guidelines on the evaluation of vector network analysers (VNA): Calibration guide no. 12, version 3.0,” EURAMET, Braunschweig, Germany, Mar. 2018. [Online] . Available: https://www.euramet.org/Media/news/I-CAL-GUI-012_Calibration_Guide_No._12.web.pdf
[32] U. Arz et al., “EMPIR — 14IND02 PlanarCal: Best practice guide for planar s-parameter measurements using vector network analysers,” Physikalisch-Technische Bundesanstalt, Braunschweig, Germany, 2018. [Online] . Available: https://planarcal.ptb.de/fileadmin/documents/empir/14Ind02/documents/Open%20Access/PlanarCal_BestPractice_Guide_D8_final_OAR.pdf
[33] “EMPIR project 14IND02 PlanarCal: Microwave measurements for planar circuits and components,” Physikalisch-Technische Bundesanstalt, Braunschweig, Germany, 2017. Accessed: May 25, 2023. [Online] . Available: https://planarcal.ptb.de/fileadmin/documents/empir/14Ind02/documents/Meetings/20171214_Bern/0_-_20171214_PlanarCal_Introduction.pdf
[34] R. Lozar et al., “A comparative study of on-wafer and waveguide module S-parameter measurements at D-band frequencies,” IEEE Trans. Microw. Theory Techn., vol. 67, no. 8, pp. 3475–3484, Aug. 2019, doi: 10.1109/TMTT.2019.2919538.
[35] D. F. Williams, F. Schmückle, R. Doerner, G. N. Phung, U. Arz, and W. Heinrich, “Crosstalk corrections for coplanar-waveguide scattering-parameter calibrations,” IEEE Trans. Microw. Theory Techn., vol. 62, no. 8, pp. 1748–1761, Aug. 2014, doi: 10.1109/TMTT.2014.2331623.
[36] G. N. Phung and U. Arz, “Anomalies in multiline-TRL-corrected measurements of short CPW lines,” in Proc. 96th ARFTG Microw. Meas. Conf., San Diego, CA, USA, 2021, pp. 1–4, doi: 10.1109/ARFTG49670.2021.9425345.
[37] G. N. Phung and U. Arz, “Parasitic probe effects in measurements of conductor-backed coplanar waveguides,” in Proc. Kleinheubach Conf., Miltengerg, Germany, 2021, pp. 1–4, doi: 10.23919/IEEECONF54431.2021.9598396.
[38] G. N. Phung and U. Arz, “On the influence of thru- and line-length-related effects in CPW- based multiline TRL calibrations,” in Proc. 97th ARFTG Microw. Meas. Conf., Atlanta, GA, USA, 2021, pp. 1–4, doi: 10.1109/ARFTG52261.2021.9639909.
[39] G. N. Phung and U. Arz, “Impact of chuck boundary conditions on wideband on-wafer measurements,” in Proc. IEEE 25th Workshop Signal Power Integrity (SPI), Siegen, Germany, 2021, pp. 1–4, doi: 10.1109/SPI52361.2021.9505192.
[40] D. Stokes, Y. Liu, X. Shang, and N. Ridler, “Millimeter-wave and terahertz power meter characterisation at W-band frequencies,” J. Infrared Millimeter THz Waves, vol. 44, pp. 134–150, Jan. 2023, doi: 10.1007/s10762-022-00898-8.
[41] R. F. Clark, “The microcalorimeter as a national microwave power standard,” Proc. IEEE, vol. 74, no. 1, pp. 102–104, Jan. 1986, doi: 10.1109/PROC.1986.13411.
[42] L. Brunetti, L. Oberto, M. Sellone, and E. T. Vremera, “Comparison between thermoelectric and bolometric microwave power standards,” IEEE Trans. Instrum. Meas., vol. 62, no. 6, pp. 1710–1715, Jun. 2013, doi: 10.1109/TIM.2013.2253914.
[43] J. Kwon, T. Kang, and N. Kang, “V-band waveguide microcalorimeter for millimeter-wave power standards,” IEEE Trans. Instrum. Meas., vol. 66, no. 6, pp. 1598–1604, Jun. 2017, doi: 10.1109/TIM.2017.2662358.
[44] L. Brunetti, L. Oberto, M. Sellone, and E. T. Vremera, “Comparison among coaxial microcalorimeter models,” IEEE Trans. Instrum. Meas., vol. 58, no. 4, pp. 1141–1145, Apr. 2009, doi: 10.1109/TIM.2008.2011091.
[45] European Telecommunications Standards Institute. Accessed: May 25, 2023. [Online] . Available: https://www.etsi.org/standards
[46] R. H. Judaschke, K. Kuhlmann, T. M. Reichel, and W. Perndl, “Millimeter-wave thermoelectric power transfer standard,” IEEE Trans. Instrum. Meas., vol. 64, no. 12, pp. 3444–3450, Dec. 2015, doi: 10.1109/TIM.2015.2454632.
[47] R. Judaschke and J. Ruhaak, “Determination of the correction factor of waveguide microcalorimeters in the millimeter-wave range,” IEEE Trans. Instrum. Meas., vol. 58, no. 4, pp. 1104–1108, Apr. 2009, doi: 10.1109/TIM.2008.2012381.
[48] M. Celep, M. Salek, D. Stokes, J. Skinner, and Y. Wang, “Power sensor characterization from 110 GHz to 170 GHz using a waveguide calorimeter,” IEEE Trans. Instrum. Meas., vol. 71, pp. 1–7, 2022, doi: 10.1109/TIM.2022.3144209.
[49] J. R. Juroshek, “A direct calibration method for measuring equivalent source mismatch,” Microw. J., vol. 40, no. 10, pp. 106–113, Oct. 1997.
[50] X. Cui, Y. S. Meng, Y. Shan, W. Yuan, C. Ma, and Y. Li, “Evaluation and validation of a national WR-15 (50 to 75 GHz) power measurement system,” in Proc. 84th ARFTG Microw. Meas. Conf., Boulder, CO, USA, 2014, pp. 1–3, doi: 10.1109/ARFTG.2014.7013418.
[51] K. Sakurai and T. Nemoto, “A thin-film bolometer unit,” IEEE Trans. Instrum. Meas., vol. 16, no. 3, pp. 206–211, Sep. 1967, doi: 10.1109/TIM.1967.4313623.
[52] L. Brunetti, “Thin-film bolometer for high-frequency metrology,” Sensors Actuators A, Phys., vol. 32, nos. 1–3, pp. 423–427, Apr. 1992, doi: 10.1016/0924-4247(92)80024-W.
[53] X. Wang, Q. Zhong, J. Li, W. Yuan, Y. Li, and X. Cui, “WR-06 power standard devices,” in Proc. Conf. Precis. Electromagn. Meas. (CPEM), Paris, France, 2018, pp. 1–2, doi: 10.1109/CPEM.2018.8500962.
[54] D. D. Dinh and M. J. Lancaster, “Microwave power sensors with integrated filtering function for transfer power standards,” IEEE Microw. Wireless Compon. Lett., vol. 30, no. 3, pp. 308–311, Mar. 2020, doi: 10.1109/LMWC.2020.2969570.
[55] R. N. Clarke et al., A Guide to the Characterisation of Dielectric Materials at RF and Microwave Frequencies. Teddington, U.K.: National Physical Laboratory/ Institute of Measurement and Control, Sep. 2003.
[56] S.-K. Yee et al., “Design and calibration of a wideband TEM-cell for material characterization,” in Proc. Asia-Pacific Symp. Electromagn. Compat., Singapore, 2012, pp. 749–752, doi: 10.1109/APEMC.2012.6237852.
[57] B. Komiyama, M. Kiyokawa, and T. Matsui, “Open resonator for precision dielectric measurements in the 100 GHz band,” IEEE Trans. Microw. Theory Techn., vol. 39, no. 10, pp. 1792–1796, Oct. 1991, doi: 10.1109/22.88556.
[58] A. Kazemipour et al., “Standard load method: A new calibration technique for material characterization at terahertz frequencies,” IEEE Trans. Instrum. Meas., vol. 70, pp. 1–10, May 2021, doi: 10.1109/TIM.2021.3077660.
[59] T. Tosaka, K. Fujii, K. Fukunaga, and A. Kasamatsu, “Development of complex relative permittivity measurement system based on free-space in 220–330 GHz range,” IEEE Trans. THz Sci. Technol., vol. 5, no. 1, pp. 102–109, Jan. 2015, doi: 10.1109/TTHZ.2014.2362013.
[60] X. Liu, L. Gan, and B. Yang, “Millimeter-wave free-space dielectric characterization,” Measurement, vol. 179, Jul. 2021, Art. no. 109472, doi: 10.1016/j.measurement.2021.109472.
[61] G. Mourou, C. V. Stancampiano, A. Antonetti, and A. Orszag, “Picosecond microwave pulses generated with a subpicosecond laser-driven semiconductor switch,” Appl. Phys. Lett., vol. 39, no. 4, pp. 295–296, 1981, doi: 10.1063/1.92719.
[62] J. L. Coutaz, F. Garet, and V. Wallace, Principles of Terahertz Time-Domain Spectroscopy. New York, NY, USA: Jenny Stanford Publishing, 2018.
[63] Y. Wang et al., “Material measurements using VNA-based material characterization kits subject to thru-reflect-line calibration,” IEEE Trans. THz Sci. Technol., vol. 10, no. 5, pp. 466–473, Sep. 2020, doi: 10.1109/TTHZ.2020.2999631.
[64] D. Allal et al., “Report describing the comparison of material parameter measurements at millimetre-wave and THz frequencies,” EURAMET, Braunschweig, Germany, 2022. [Online] . Available: https://zenodo.org/record/7442687
Digital Object Identifier 10.1109/MMM.2023.3321516