Document Version Publisher s PDF, also known as Version of Record (includes final page, issue and volume numbers)

Characterization of the relative permittivity and homogeneity of liquid crystal polymer (LCP) in the 60 GHz band Huang, M.; Kazim, M.I.; Herben, M.H.A.J. Published in: Proc. Cost 2100 TD (10) 12031, Bologna, Italy, November 23-25, 2010 Published: 01/01/2010 Document Version Publisher s PDF, also known as Version of Record (includes final page, issue and volume numbers) Please check the document version of this publication: A submitted manuscript is the author's version of the article upon submission and before peer-review. There can be important differences between the submitted version and the official published version of record. People interested in the research are advised to contact the author for the final version of the publication, or visit the DOI to the publisher's website. The final author version and the galley proof are versions of the publication after peer review. The final published version features the final layout of the paper including the volume, issue and page numbers. Link to publication General rights Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights. Users may download and print one copy of any publication from the public portal for the purpose of private study or research. You may not further distribute the material or use it for any profit-making activity or commercial gain You may freely distribute the URL identifying the publication in the public portal? Take down policy If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim. Download date: 17. Aug. 2018

EUROPEAN COOPERATION IN THE FIELD OF SCIENTIFIC AND TECHNICAL RESEARCH EURO-COST SOURCE: Department of Electrical Engineering Eindhoven University of Technology Eindhoven, The Netherlands COST 2100 TD(10)12031 Bologna, Italy November 23-25, 2010 Characterization of the Relative Permittivity and Homogeneity of Liquid Crystal Polymer (LCP) in the 60 GHz Band Mingda Huang, M. Imran Kazim, Matti H. A. J. Herben Eindhoven University of Technology Department of Electrical Engineering P.O. Box 513 5600 MB Eindhoven The Netherlands Phone: +31 40 247 2326 Fax: +31 40 244 8375 Email: m.huang@tue.nl

Characterization of the Relative Permittivity and Homogeneity of Liquid Crystal Polymer (LCP) in the 60 GHz Band Mingda Huang, M. Imran Kazim, Matti H. A. J. Herben Department of Electrical Engineering, Eindhoven University of Technology m.huang@tue.nl, m.i.kazim@tue.nl, m.h.a.j.herben@tue.nl Abstract The relative permittivity of LCP material has been characterized within the whole 60 GHz frequency band using the microstrip ring resonator (MRR) method. Using a circuit model, the gap capacitance of the MRR has been taken into account in order to improve the accuracy of the determined relative permittivity. The results show that the relative permittivity of LCP is almost constant (ε r 3.1) within the whole 60 GHz frequency band. The homogeneity of the LCP panel has also been examined. It is found that the variation of the relative permittivity is within 1.5% across the LCP panel. I. INTRODUCTION 60-GHz millimeter wave (mmwave) communication systems are getting increasing attention in recent years, especially for low-cost consumer applications [1]. For instance, wireless uncompressed high definition video streaming and ultra-fast wireless LAN are typical indoor environment applications at 60 GHz. These applications require the antenna array to have a large scan coverage in order to operate in both line-of-sight (LOS) and nonlight-of-sight (NLOS) conditions. In order to achieve the scan coverage requirements, cylindrically bending a planar antenna array can be employed [2]. In practice this means the use of flexible PCB. Liquid crystal polymer (LCP) is a promising flexible substrate and packaging material for mmwave applications, especially for a conformal antenna array implemented on flexible PCB [3]. The electrical properties of LCP material at 60 GHz have been investigated in literature using different methods [4] [6]. In [4], microstrip ring resonators (MRR) and cavity resonators are used in order to characterize the relative permittivity (ε r ) and loss tangent (tanδ) of LCP from30 to110 GHz. In [5], measurement results of printed T-resonators and transmission lines are compared with the simulation results of 3D EM solvers to determine the electrical properties of LCP in the frequency range 60 95 GHz. In [6], the overmoded circular cavity approach is used to characterize the LCP material from 60 to 80 GHz. This method can provide accurate results over a very wide frequency range with a single frequency sweep. The measurement results for the 60 GHz frequency band are summarized in Table I. It is shown that the dissipation factors are in agreement with each other, but the variation of the relative permittivity is up to 7%. TABLE I COMPARISON OF LCP CHARACTERIZATION AT 60 GHZ Method f (GHz) ε r tanδ (10 3 ) Ref MRR 62 3.15 - [4] Cavity resonator 60-3.5 4.5 [4] T-resonator 60 95 3.25 4.5@75GHz [5] Circular cavity 60 80 2.916±0.01 4.7 [6] In this work, the MRR method will be used to confirm the relative permittivity of LCP material in the 60 GHz frequency band. One reason for choosing this method is that the MRR method is simple to realize as a planar circuit, and has higher accuracy than the linear resonator method due to its higher quality factor. Furthermore, there is about 9 GHz spectrum allocated around 60 GHz (57.24 65.88 GHz) in the draft standard of IEEE 802.15.3c Task Group [7]. But in the literature, the measurement results can not sufficiently cover that whole bandwidth since generally the resonant structure method is very accurate but can only measure the relative permittivity at the resonant frequency. Therefore, 5 different MRRs with different resonant peaks, i.e. at 58, 60, 61.5, 63, and 65 GHz, are designed in order to examine the electric properties over the whole 60 GHz frequency band. In addition, the homogeneity of the materials can also cause the variation of the relative permittivity. Therefore, the MRRs are distributed in a periodic way on the LCP panel in order to investigate the homogeneity of the LCP panel in two dimensions. Due to the capacitance effect of the gap between the microstrip transmission line and the ring resonator, the observed resonant frequency will be lower than that of the unloaded ring. This will result in overestimating the relative permittivity of the substrate. To the author s knowledge, no literature which uses the MRR method to characterize LCP material in the 60 GHz band takes this effect into account. In this paper, a circuit model of the ring resonator structure [8] is adapted to be used to

correct the frequency pushing effects of the gap in order to characterize LCP material accurately. From the measurement results, it is found that the relative permittivity of LCP materials is almost constant (ε r 3.1) within the whole 60 GHz frequency band. The variation of the relative permittivity is found to be within 1.5% across the LCP panel. Therefore, the homogeneity of the LCP panel is suitable for mass production of bent antenna arrays operating in the 60 GHz frequency band. II. MRR DESIGN The layout of the designed microstrip ring resonator is shown in Fig. 1. R is the mean radius of the microstrip ring, S is the spacing of the coupling gap, and W is the width of the microstrip line. via probe pitch Fig. 1. ref. plane R W S ref. plane Layout of a two-port microstrip ring resonator. The parallel resonant frequency of the unloaded MRR is given by cn f 0,N = 2πR, (1) ε eff where ε eff is the effective permittivity, N is the order of resonance, and c is the speed of light in vacuum [8]. Therefore, with the physical dimensions of the microstrip, the relative permittivity of LCP can be obtained with the relation with u eff = u+ 1.25t πh and ε eff = ε r +1 2 + ε r 1 2 ( ( )) 2h 1+ln, t 1 1+ 12, (2) u eff ( for u > 1 ), 2π (3) u = W h, (4) where t is the thickness of the microstrip and h is the height of the dielectric substrate [9]. The use of the effective width of the microstrip u eff is because the thickness of the microstriptis not negligible in this case. However, the unloaded MRR has to couple with microstrip transmission lines in order to be measured. Therefore, parasitic capacitances are introduced by the gap between the MRR and the feeding line. This causes the resonant frequency of the loaded MRR to become lower than that of the unloaded MRR. As a result, the relative permittivity will be overestimated if the resonant frequency of the loaded MRR is used. As shown in Fig. 2, a circuit model of the loaded MRR can be used to take this effect into account [8], [10]. C p Fig. 2. C g Z r C g Circuit model of loaded MRR. Z r presents the impedance of the MRR, which can be calculated by Z r = Z 0 2 C p coth(γπr) (5) where Z 0 is the characteristic impedance of the transmission line and γ is the complex propagation constant. C p and C g represent the parasitic capacitances, which can be determined by a planar simulation of a T-gap configuration. The resonant peak of the circuit model can be used to compare with that of the loaded MRR in order to determine the effective permittivity ε eff. The MRRs have been designed on Rogers ULTRA- LAM 3850 LCP substrate. The LCP panel has the dimension of 457mm 610mm with the thickness h of 101 µm (4 mil). The design layout is shown in Fig. 3. The subblock is shown in the right side of the figure, it contains the de-embedding structures, a small ring, and a big ring. The small and big rings have the 4th and 8th resonant peaks respectively around the design frequency, which is given by the numeric numbers with the unit of GHz. It is seen that there are 5 different resonant frequencies which are sampled within the whole 60 GHz frequency band in order to determine the electric properties of LCP material. The radii of the designed MRRs for these 5 different resonant frequencies are listed in Table II. The width of the microstrip line W is 227 µm and the metal thickness t is 18 µm in order to obtain a characteristic impedance of 50 Ω. The spacing of the gap S is 100 µm, which is the minimum achievable spacing of the manufacturer. The probe pitch is designed in order to land the ground-signal-ground (GSG) probes with 250 µm probe tip spacing. Using a through-reflectline (TRL) calibration, the two-port measurement results are de-embeded to the reference plane as shown in Fig. 1 in order to remove the effects of the transition from the GSG probe to microstrip. Therefore, the measurement results after de-embedding can be used to obtain the resonant frequency of the loaded MRR.

58 60 61.5 63 65 63 63 63 63 63 58 60 61.5 63 65 61.5 61.5 61.5 61.5 61.5 58 60 61.5 63 65 60 60 60 60 60 58 60 61.5 63 65 Fig. 3. Layout of LCP panel. TABLE II RADIUS OF MRRS f 0,N (GHz) 58 60 61.5 63 65 R (µm), N = 4 2070 2010 1950 1909 1845 R (µm), N = 8 4072 3936 3842 3750 3651 III. MRR MEASUREMENT RESULTS AND DISCUSSIONS There are two fabricated LCP panels since there were too many defected sub-blcoks in the first panel to analyses the homogeneity of the LCP panel. The measurements were done firstly over 0.1 67 GHz band in order to observe the number of resonant peaks of the MRR. As shown in Fig. 4, the MRR from the first fabricated LCP panel has 4 resonant peaks in the whole frequency sweep range, and the 4th resonant peak is around 58 GHz as designed after de-embedding. S21 (db) -30-40 -50-60 -70-80 -90-100 -110 0 10 20 30 40 50 60 LCP panel as shown in Fig. 3. The same naming rule of the position is used in the following part of this paper. Applying a low pass filter (LPF) to the measurement result, it is found that the 4th resonant frequency of this MRR is at 60.42 GHz, which is close to the designed resonant frequency 60 GHz. S21 (db) -35-40 -45-50 -55-60 -65-70 -75 Measurement After LPF 52 54 56 58 60 62 64 66 Fig. 5. S 21 measurement of a small MRR. It is also observed that the amplitude of the transmission coefficient in Fig. 5 is about 5 db lower than that in Fig. 4. This is mainly because the spacings of the gap S of the second-run MRR is larger than that of the firstrun MRR. It is seen that the spacing S of the second-run MRR is still small enough to allow adequate coupling of power, otherwise the resonant peak will be difficult to be recognized. Two MRRs from different fabrication runs are inspected using a microscope with the same scale factor, as shown in Fig. 6. It is seen that the spacing of the gap of the second-run MRR is obviously larger than that of the first-run MRR. This leads to less coupling between the microstrip and the ring resonator in the second-run MRR. Thus the amplitude of the transmission coefficient of the second-run MRR becomes lower. In order to determine the fabricated dimensions of the gap and the microstrip, a 200 µm coplanar line on the calibration substrate is used as the reference. As shown in Table III, the fabricated dimensions of the first-run Fig. 4. S 21 measurement of a small MRR. After the second-run LCP panel was available, the s-parameters measurements were all carried out using the sub-blocks in the second panel with a narrower frequency band sweep (51-67 GHz). Fig. 5 shows the measurement results after de-embedding of the small ring at the position C1R2. C1R2 is located at column 1 from the left, row 2 from the bottom of the second (a) the first-run MRR (b) the second-run MRR Fig. 6. The comparison of MRRs from two fabrication runs.

MRR are close to the design value. But in the secondrun, both the spacing of the gap S and the width of the microstrip line W have about 30 µm tolerance compared with the design value. TABLE III COMPARISON OF THE DIMENSIONS OF MRRS Parameters W (µm) S (µm) Design 227 100 1st-run 214 106 2nd-run 196 129 Fig. 7 shows the amplitude difference between the normalized S 21 of the two different run MRRs using the circuit model presented in Fig. 2 with ε r = 3.1. It is seen that the S 21 resonant peak of the first-run MRR with S = 106 µm is about 5.3 db higher than that of the second-run MRR with S = 129 µm. This is in a good agreement with the measurement results shown in Fig. 4 and Fig. 5. Normalized S21 (db) Fig. 7. 0-5 -10-15 -20-25 -30 1st-run 2nd-run -35 56 57 58 59 60 61 62 63 64 The normalized S 21 with different MRR dimensions. Fig. 8 shows all of the measurement results in the corresponding position, in which the numeric numbers present the 4th resonant frequency of the small MRR in that sub-block with the unit of GHz. The X presents defected sub-block. With the use of the circuit model presented in Fig. 2, the relation between the relative permittivity of the LCP materials and the resonant frequencies of the MRRs can be obtained with the second-run fabricated dimension values which are listed in Table III. Fig. 9 shows this relation for the small MRRs which are designed to have 4th resonant peaks at 60 GHz. It is found that the resonant frequencies varies between 60.3 60.64 GHz. As a result, the corresponding ε r is in the range of 3.073 3.113. In Fig. 9 it is also observed that, compared with the relative permittivity determined by the fabricated dimension, the relative permittivity is about 1.3% lower 58.75 63.63 63.67 X 63.65 63.9 58.79 58.73 60.42 60.46 60.45 60.64 58.56 60.33 60.48 62.34 62.18 62.22 62.37 62.38 60.5 60.3 62.31 62.23 62.28 62.39 63.73 63.49 63.82 63.71 65.99 65.97 65.98 60.48 65.77 Fig. 8. The 4th resonant peaks of S 21 measurement of small MRR on LCP panel. εr 3.16 3.14 3.12 3.1 3.08 3.06 3.04 (60.3, 3.113) Fabricated value Designed value w/o circuit model (60.64, 3.073) 3.02 59.8 60 60.2 60.4 60.6 60.8 61 61.2 The relative permittivity of LCP materials determina- Fig. 9. tion. if the design dimensions are used. It is also seen that the relative permittivity will be overestimated about 0.45% if the circuit model is not applied. The overestimation is a bit less than that presented in [8]. One reason is that the realized gap spacing is larger than the design value. Therefore, the effects of the parasitic capacitances are reduced. As shown in Fig. 10, the relative permittivity is overestimated about 0.68% when the design values are used (h = 4 mil, S = 100 µm). Another reason is that the parasitic capacitances are related with the height of the substrate. It is seen in Fig. 10 that, when the thickness of the substrate becomes larger (h = 4.72 mil), the relative permittivity is overestimated about 0.9%. Fig. 11 shows the relative permittivity of the LCP materials at 5 different resonant frequencies, as obtained using the circuit model. It is observed that the relative permittivities at 5 different frequencies are almost constant. The average and the sample standard deviation of the relative permittivities of the measured LCP

ǫr 3.18 3.16 3.14 3.12 3.1 3.08 3.06 3.04 4.72 mil 4.72 mil (w/o) 4 mil 4 mil (w/o) 3.02 59.6 59.8 60 60.2 60.4 60.6 Fig. 10. LCP. Overestimation analysis of the relative permittivity of samples are given in Table IV. It can be seen that ε r = 3.093 ± 0.035 in the whole 60 GHz frequency band. εr 3.16 3.14 3.12 3.1 3.08 3.06 3.04 58 GHz 60 GHz 61.5 GHz 63 GHz 65 GHz 3.02 58 59 60 61 62 63 64 65 66 67 Fig. 11. The relative permittivity of the LCP panel. TABLE IV THE RELATIVE PERMITTIVITY OF LCP IN THE 60 GHZ FREQUENCY BAND Resonant Freq. (GHz) ε r ε r,av σ 58.56 58.79 3.084 3.113 3.094 0.0128 60.3 60.64 3.073 3.113 3.095 0.0118 62.18 62.39 3.086 3.11 3.096 0.0088 63.49 63.9 3.067 3.112 3.089 0.0136 65.77 65.99 3.079 3.103 3.086 0.0115 Total 3.067 3.113 3.093 0.0115 Fig. 12 shows the obtained relative permittivity ε r in the corresponding position. It is found that the variation of the relative permittivities at different positions of the LCP panel is less than 1% in both horizontal and vertical directions. This variation can be caused by the fabrication tolerance of the MRR radius, the measurement errors, and the homogeneity of the LCP panel. Fig. 12. panel. 3.089 3.097 3.084 3.091 3.091 3.099 3.11 3.11 3.113 3.113 3.086 3.092 X 3.094 3.091 3.089 3.094 3.104 3.105 3.098 3.094 3.095 3.086 3.112 3.088 3.076 3.073 3.088 3.079 3.067 3.081 3.087 3.08 3.091 3.103 The homogeneity of the relative permittivity of LCP In order to examine the errors introduced by the fabrication tolerance of the MRR radius, the microscope and a 6600 µm coplanar line are used to measure the diameter of the rings. Three small rings are examined as shown in Table V. It is seen that the diameters of the fabricated MRRs are slightly smaller (0.05 0.17%) than the design values. This difference can also be partly introduced by the measurement accuracy. It is found that the relative permittivities shift upward about 0.1 0.4%. Therefore, from these three samples, it can be verified that the relative permittivity of LCP materials ε r is within 3.093±0.035, and the homogeneity of the LCP panel is within 1%. If the average shifting value 0.3% is used, the relative permittivity ε r 3.1. In order to obtain more accurate results of ε r and the homogeneity of the LCP panel, all of the fabricated dimensions of the ring diameters need to be examined with a more accurate method. TABLE V THE FABRICATED DIAMETER OF MRRS AND THE CORRESPONDING ε r Position Design (µm) Fabricated (µm) ε r C2R1 4020 4018 3.116 C3R2 4020 4013 3.108 C4R2 4020 4013 3.086 IV. CONCLUSIONS In this paper, the relative permittivities of the LCP material has been examined over the whole 60 GHz frequency band using the method of microstrip ring resonators (MRR). A circuit model is used to correct the overestimation of the relative permittivities due to the frequency pushing effects of the gap between the microstrip transmission line and the ring resonator. It

is found that the relative permittivities at 5 different frequencies within 60 GHz band are almost constant (ε r 3.1). The homogeneity of the LCP panel is examined by distributing the same MRRs at different locations on the LCP panel. It is found that the variation of the relative permittivities is within 1.5% across the LCP panel. As a result, the homogeneity of the LCP panel is suitable for mass production of bent antenna arrays operating in the 60 GHz frequency band. ACKNOWLEDGMENT This work has been carried out within the European Medea+ project QStream Ultra-high data-rate wireless communication. The authors would like to thank A.C.F. Reniers and A.R. van Dommele from the Electromagnetics group at TU Eindhoven for their valuable supports with the design and measurements and the Mixed-Signal Microelectronics group at TU Eindhoven for the use of their measurement equipment. REFERENCES [1] P. F. M. Smulders, H. Yang, and J. A. G. Akkermans, On the design of low-cost 60-GHz radios for multigigabit-per-second transmission over short distances, IEEE Commun. Mag., vol. 45, no. 12, pp. 44 51, December 2007. [2] M. D. Huang and M. H. A. J. Herben, Effects of bending a planar antenna array on its scan performance, in European Conference on and Propagation (EuCAP 2010), Barcelona, Spain, 12 16 April 2010, pp. 1 5. [3] N. Kingsley, Liquid crystal polymer: Enabling next generation conformal and multilayer electronics, Microwave Journal, vol. 51, no. 5, pp. 188 200, May 2008. [4] D. C. Thompson, O. Tantot, H. Jallageas, G. E. Ponchak, M. M. Tentzeris, and J. Papapolymerou, Characterization of liquid crystal polymer (LCP) material and transmission lines on LCP substrates from 30 to 110 GHz, IEEE Trans. Microwave Theory Tech., vol. 52, no. 4, pp. 1343 1352, April 2004. [5] S. Smith and V. Dyadyuk, Measurement of the dielectric properties of Rogers R/flex 3850 liquid crystalline polymer substrate in V and W band, in and Propagation Society International Symposium, 2005 IEEE, vol. 4B, 3 8 July 2005, pp. 435 438. [6] Y. Lu, Y. Huang, K. Teo, N. Sankara, W. Lee, and B. Pan, Characterization of dielectric constants and dissipation factors of liquid crystal polymer in 60 80 GHz band, in and Propagation Society International Symposium, 2008. AP-S 2008. IEEE, 5 11 July 2008, pp. 1 4. [7] [Online]. Available: http://www.ieee802.org/15/pub/tg3c.html. [8] J. Bray and L. Roy, Microwave characterization of a microstrip line using a two-port ring resonator with an improved lumpedelement model, Microwave Theory and Techniques, IEEE Transactions on, vol. 51, no. 5, pp. 1540 1547, may 2003. [9] L. F. Chen, C. K. Ong, C. P. Neo, V. V. Varadan, and V. K. Varadan, Microwave Electronics: Measurement and Materials Characterization. Chichester: Wiley, 2004, pp. 60 61. [10] J. A. G. Akkermans, Planar beam-forming antenna array for 60- GHz broadband communication, Ph.D. dissertation, Eindhoven University of Technology, 2009.