References 1 and STEWART, w.J.: ‘Magnetism from conductors and enhanced nonlinear phenomena’, fEEE Trans. Microw. Theory Tech., 1999, 47, pp. 2075-2084 2 WILTSHIRE. M.C.K., PENDRS, J.B., YOUNG, I.R., LARKMAN, D.J., GILDERDALE, D J., and HAJNAL, J.V: ‘Microstructured magnetic materials PENDRY, J.H., HOLDEN, A.I., RoBBiNs, D.J., for RF flux guides in magnetic resonance imaging’, Science, 2001, 291, (5S05), pp. 849-851 3 LANDAU, L D., and LIFSHIZ. E.M.: ‘Electrodynamics of continuous media’ (Pergamon Press, Oxford UK, 1984) Fast electromagnetic characterisation method of thin planar materials using coplanar line up to V-band Extraction method: The extraction method of the dielectric substrate properties is based on S-parameter measurements achieved at the coplanar access planes. It requires the propagation to be the quasiTEM dominant mode and dispersion to be low ( h > W+2s). The reflection/transmission method [5] allows the first reflection (r)to be obtained at the input of the coplanar cell and the first transmission ( r ) along the coplanar cell of length d. The resolution of the reflection/ transmission methods equations leads to the terms of complex effective permittivity and permeability [ 5 , 61. However, for nonmagnetic materials to be measured, it is possible to increase the accuracy of the E:. and E: characteristics by fixing the permeability at I in the reflection/transmission method. This leads us to look for only one complex unknown instead of two. Two equations can then be determined [6, 71: t&f/ = (g)[email protected] J. Hinojosa, K. Lmimouni and G. Dambrine A very broadband method for determiningthe electromagnetic properties of isotropic thin planar materials, which uses a coplanar line as a cell, is presented. The complex permittivity is quickly computed from S-parametermeasurements of coplanar cells propagating the dominant mode by using analytical equations. Measurements between 0.05 and 75 GHz of alumina and doped silicon show good agreement between measured and predicted values. where Z, is the characteristic impedance of the test device (50 Q) and Zb is the characteristic impedance of the coplanar cell when E,. = p,. = 1, which is computed from an analytical equation [3]. In the case of (I), owing to the periodic behaviour of the coplanar cell with frequency, dimensional resonances appear on the Sll and parameters at frequencies corresponding to integer multiples of one half wavelength (Q. At these particular frequencies, the S I Iand magnitudes are small and the phase uncertainties are large [6-81. This leads to the appearance of inaccuracy peaks in the computed complex E,. values. These inaccurate peaks can be avoided by using (2) as in [7]. The exchange between complex effective permittivity and the sample desired relative permittivity (E,) is obtained from analytical equations [3,41: r (3) (4) where qtg 6,,,/1= (1 -(~;)-‘)/(l - (E:.&’) is the filling factor for the effective dielectric loss [4] and tg Sder*= E & / E ; . ~ The parameters k, kl , k’ and k; are bound to the coplanar line structure. K(k), K(k’), K(kl), K(k;) are the complete elliptical integrals of first order of modulus k and k , and complementary modulus k‘ and k ; . These integrals are also obtained by analytical equations [4]. 51.O- (0 H: I (io\ 50.5-+$ + 0 * +- m 73 a .._E 50.0- c .- I . . . : L . - iii)’ 0 0 L W+2S=175ym h=635 pm coplanar cell E2 ; 49.549.0! . L L v ti IS L h r r E , pr I 5 I I I I 10 15 20 25 I 1 I I I I I I I I 30 35 40 45, 50 55 60 65 70 75 frequency, GHz Fig. 1 Coplanar characteristic impedances against frequency and various substrates Coplanar configurations have been optimised to SO R and values have been computed from spectral domain approach numerical method [2] (i) W c,= 10 -jO.O, AI, = 1 -;O.O (ii) c,= IO-jO.1, &.=0.6-;0.006 (iii) A E,.= IO-jl.0, /1,.=3-jO.3 Measurement bench: To implement measurements with coplanar cells up to 75 GHz, we have chosen two commercially available onwafer systems covering DC-40 GHz and 50-75 GHz, which use coplanar lines as probes. They allow different sizes of coplanar line and easy and fast measurements to be implemented without thermocompression between the on-wafer system and the coplanar cell. The S-parameters measurements are repeatable and accurate. The calibration procedure uses a line-reflect-match (LRM) covering 0.05-75 GHz, since the quasi-TEM mode dispersion of the 50 !2 characteristic impedance coplanar standard lines on alumina substrate is very low. The two obtained reference planes are the outputs of the two probes. Retum losses, insertion losses and repeatability were better than - 15, - 1 and -50 dB, respectively, in the 0:05 to 75 GHz frequency range. Results: Alumina (~$=9.85, ~:<0.001 at 10 GHz, pv= 1) (Fig. 2) and doped silicon (E: = 1 1.7, 40 Q c m < p < 60 R cm) (Fig. 3) with well-known dielectric properties were measured in the 0.05 to 40 GHz and 50 to 75 GHz frequency ranges at room temperature. The coplanar cells were made from thin-film technology (gold conductor) ~41. In this Letter we present an easy and fast complex permittivity extraction method of the coplanar substrates from S-parameter measurements, which takes into account the quasi-TEM mode. The S-parameter measurement bench employs vector network analysers and coinmercially available high-quality on-coplanar test fixtures covering 0 . 0 5 4 0 and 50-75 GHz without the necessity to make a transition between the network analyser and the sample-cell as in [l]. ELECTRONICS LETTERS 11th April 2002 The measured E: values for alumina (Fig. 2) and doped silicon (Fig. 3) correspond to those anticipated. The error is lower than 3% with regard to the manufacturer values at frequencies above 5 GHz. At frequencies below 5 GHz, the large errors on E: are due to large uncertainties on the S-parameters. Various integer multiples of one half wavelength are required to increase the accuracy of the complex permittivity with this method [6, 71. As can be seen from Figs. 2 and 3, Vol. 38 No. 8 373 c: values do not exactly correspond at 40 and 50 GHz. These errors are linked to the position errors of the on-wafer system probes at the access planes of the coplanar cells. The small resonances in E: and E: values in the 0.05 to 40 GHz frequency range for alumina are due to the fabrication imperfections of the coplanar line, which do not disturb the result interpretation. Thus, they were avoided in the 50 to 75 GHz frequency range by omitting some measurement points. This same process was carricd out for doped silicon (0.05-75 GHz). The small increment of the E: and E:’ values against frequency for alumina and doped silicon is due to both the low dispersion of the 50 R characteristic impedance coplanar load and the low quasi-TEM mode dispersion of the measured coplanar cells. These factors were not taken into account during the LRM calibration procedures (0.0540 and 5075 GHz) nor by the analytical equations in the complex permittivity extraction method. In the case of dielectric losses, large errors are shown for alumina (Fig. 2). These errors are mainly due to the network analyser, the test fixture performance, and the whole coplanar cell (dielectric, metallic and radiation) losses. A detailed error bound analysis of this method from [7] is presented in [6]. The measurement of low-loss samples is difficult with this technique. To obtain reasonable accuracy, E:! must be greater than 0.1 as the case of measured doped silicon (Fig. 3), where the values are in good agreement with the manufacturer data. J. Hinojosa (Universidad Politbcnica de Cartagena, Departamento de Electrbnica, Tecnologia de Computadoras y Proyectos. C / . Doctor Fleming s / n , 30202 Cartagena (Murciu), Spain) E-mail: [email protected] K. Lmimouni and G. Dambrine (DGpartement HyperfrC.qzcences et Semiconducteurs, lnstitut d’Electronique et de Microblectronique du Nord, Cite Scienti$que, Avenue PoincarG, B.P 69, 59652 Villeneuve d’Ascq Cedex, France) References HINOJOSA, J., KRuCK, J.F., and DAMBRINE, G.: ‘Ridged waveguide to microstrip transition for electromagnetic characterisation of materials in V-band’, Electron. Lett., 2000, 36, (17), pp. 1468-1469 ITOH, T., and MITTRA, R.: ‘Spectral-domain approach for calculating the dispersion characteristics of microstrip lines’, IEEE Trans. Microw Theory Tech., 1973, 21, (7), pp. 496-499 CHIONE, G., and NALDI, C.: ‘Analytical formulas for coplanar lines in hybrid and monolithic MICs’, Electron. Lett., 1984,20, (4), pp. 179-181 HOFFMANN, R.K.: ‘Handbook of microwave integrated circuits’ (Artech House Inc., 1987) WEIR, W.B.: ‘Automatic measurement of complex dielectric constant and pemieability at microwave frequencies’, Proc. IEEE, 1974, 62, (l), pp. 33-36 HINOJOSA, J.: ‘Contribution I’klaboration d’une nouvelle methode de caracterisation electromagnttique de matkriaux a partir de lignes plaqukes - Applications a I’etude de nouveaux matkriaux’. Thkse d’universite en Electronique, Lille, France, May 1995 BOUGHRIET, A., LEGRAND, C., and CHAPOTON. A : ‘Noniterative stable transmission/reflection method for low-loss material complex permittivity determination’, ZEEE Trans. Microw. Theory Tech., 1997, 45, (I), pp. 52-57 DONEKER, B.: ‘Accuracy predictions for new generation network analyzer’, Microwave 1,1984, 6 , pp. 127-141 Millimetre-wave wideband reflection-type 41 0 I I I I I I I I I I I I I 1 0 10 15 20 25 30 35 40 45 50 55 60 65 70 75 I 5 frequency, GHz 2 Measured E, data ,for alumina Cell dimension: W = 50 pm, W+ 2S= 175 pm, h = 635 pm,f = 5 pm, length d = l cm , 10 20F . I CPW MMIC phase shifter Hong-Teuk Kim, Dae-Hyun Kim, Youngwoo Kwon and Kwang-Seok Se0 A reflection-type coplanar waveguide (CPW) phase shifter fabricated using a standard monolithic microwave integrated circuit (MMIC) process is presented. Air-gap overlay CPW couplcrs were employed for wideband 3 dB coupling and low loss at millimetre wave. The twostage cascaded analogue phase shifter showed insertion losses of 6.9 1.6 dB, return losses > 10 dB, and maximum r m s phase error of 15.5” for the relative phase shift from -20” to 135”, over a wideband 27 to 47 GHz. + 5 4 0 , 5 , , , , , , , , 1 , , , , ! 0 10 15 20 25 30 35 40 45 50 55 60 65 70 75 frequency, GHz Fig. 3 Measured E,. data for doped silicon Cell dimension: W=70 pm, W+ 2 S = 175 pm, h = 230 pm, t = 3 pm, length d = l em Conclusion: The experimental results have demonstrated the validity of the method. This method can be conveniently applied to the study of isotropic dielectric film-shaped materials in the 0.05 to 75 GHz frequency range. Acknowledgments: The authors thank S. Lepilliet for his contributions to the measurements with the network analyser. 0 IEE 2002 Electronics Letters Online No: 20020272 Dol: IO.1049/el:200202 72 37 4 8 November 2001 Introduction: Various reflection-type microstrip monolithic microwave integrated circuit (MMIC) phase shifters have been reported using Lange couplers [l-31. However, coplanar waveguide (CPW)type reflective MMIC phase shifters have not been common since the low-loss broadband 3 dB CPW coupler compatible with the standard MMIC process have not been available. Previously, analogue reflection-type CPW MMIC phase shifters with narrow bandwidth ( < I GHz) were realised using multilayer CPW couplers in which the overlapped conductors were separated by an extra dielectric layer [4, 51. However, the phase shifters showed high mean losses of 31 dB/36Oo with eight couplers at 24 GHz [4], 16 dB/12Oo with two couplers at 31.5 GHz [SI due to lossy polyimide dielectrics and strong field intensities between the coupler conductors. Recently, the authors proposed a standard MMIC-compatible Lange coupler using air-gap overlay CPW structures [6]. The coupler showed very low loss (0.6 dB at 30 GHz) over a wide frequency range 20-39 GHz. In the overlay CPW coupler, the air-gap offset broadside coupling between two lines offers tight coupling and reduces the conductor loss by redistributing currents over broad surfaces. In this work, a millimetrewave two-stage cascaded reflection-type analogue phase shifter was realised using a standard uniplanar MMIC process. There are no additional dielectric layers for the couplers. The air-gap overlay CPW couplers and GaAs Schottky varactor diodes were used as key components of the phase shifter. ELECTRONICS LETTERS 11th April 2002 Vol. 38 No. 8

1/--страниц