Application of Molded Interconnect Device technology to the realization of a self-biased circulator

Application of Molded Interconnect Device technology to the realization of a self-biased circulator

Journal of Magnetism and Magnetic Materials 404 (2016) 126–132 Contents lists available at ScienceDirect Journal of Magnetism and Magnetic Materials...

2MB Sizes 0 Downloads 10 Views

Journal of Magnetism and Magnetic Materials 404 (2016) 126–132

Contents lists available at ScienceDirect

Journal of Magnetism and Magnetic Materials journal homepage: www.elsevier.com/locate/jmmm

Application of Molded Interconnect Device technology to the realization of a self-biased circulator Vincent Laur a,n, Jean-Luc Mattei a, Grégory Vérissimo a, Patrick Queffelec a, Richard Lebourgeois b, Jean-Pierre Ganne b a b

Lab-STICC, University of Brest, Brest 29238, France Thales Research & Technology, Palaiseau 91767, France

art ic l e i nf o

a b s t r a c t

Article history: Received 9 June 2015 Received in revised form 5 October 2015 Accepted 4 December 2015 Available online 7 December 2015

This paper describes the first electromagnetic characterization of a self-biased circulator in molded interconnect device (MID) technology. The circulator was designed using a 3D full-wave commercial simulator. It consists of microstrip access lines connected to a Y-junction in Substrate Integrated Waveguide (SIW) technology. Unlike classical technologies, the SIW Y-junction was not fabricated using metallic vias but by a Laser Direct Structuring (LDS) technique. A molded Cyclo-Olefin Polymer (COP) was used as a substrate and 3D metallized. The microwave properties of LDS-compatible COP are not well known so we investigated them through the use of cavity-perturbation and rectangular waveguide characterization methods. The device was then machined to insert a pre-oriented strontium hexaferrite puck doped with cobalt and lanthanum (Sr0,7La0,3Fe11,7Co0,3O19). The characteristics of the MID circulator were assessed between 28 and 32 GHz. Without magnets, insertion losses of 3.32 dB were measured at 30.7 GHz. At the same frequency, an isolation level of 13.89 dB and return losses of 19.89 dB were observed. These measurements demonstrate for the first time the high potential of MID technology for the realization of low-cost non-reciprocal devices. & 2015 Elsevier B.V. All rights reserved.

Keywords: Hexaferrite Microwave measurement Molded Interconnect Device technology Remanence Self-biased circulator

1. Introduction A Molded Interconnect Device (MID) is generally defined as an injection molded plastic substrate that incorporates a conductive circuit pattern and integrates both mechanical and electrical functions. This technology grew up in the past few years and is now capable of providing components for different markets, such as automotive electronics, telecommunications, computers and medical equipment. Extensive research was conducted on microwave passive components such as antennas in MID technology [1–3], and hundreds of millions of mobile phone antennas are today fabricated using this technology. However, fabricating RF front-ends in MID technology requires going further, so laboratories now focus their attention on the integration of filters [4] and on the improvement of flip chip assembly processes [5,6]. Realizing circulators in MID technology is also a challenge that will be investigated in this paper. Circulators and isolators are commonly used in modern telecommunications systems. These non-reciprocal devices are used as duplexers to allow the use of a single antenna in full-duplex n

Corresponding author. E-mail address: [email protected] (V. Laur).

http://dx.doi.org/10.1016/j.jmmm.2015.12.021 0304-8853/& 2015 Elsevier B.V. All rights reserved.

telecommunications systems or in monostatic radars. When one of the ports is loaded with a power load (isolator mode), these circuits make it possible to protect power amplifiers from radiation or impedance mismatch. In contrast to frequency-selective filters (diplexers), circulators enable transmission/isolation in the same frequency range. However, these devices are mainly fabricated by hybrid technologies (insertion of ferrite pucks in a triplate or microstrip structure) leading to high bulkiness and cost. Furthermore, bulkiness is exacerbated by the need for permanent magnets to polarize the ferrite pucks. Indeed, mass production of lowcost compact circulators remains a hot topic and new ideas and technologies are needed to improve the integration of these devices. Removing the magnets appears to be a way to decrease circulator size and this solution has attracted the interest of laboratories and industry, leading to intensive studies over the past twenty years. Publications on this topic are mainly focused on the use of pre-oriented polycrystalline hexagonal ferrites [7–12], especially M-type strontium hexaferrites. Several demonstrators were characterized from Ku (10.7–14.5 GHz) to Q band (33–50 GHz) and above. These studies proved that removing magnets by the use of oriented hexaferrites is a viable solution. However, these devices are mainly based on the hybrid integration

V. Laur et al. / Journal of Magnetism and Magnetic Materials 404 (2016) 126–132

of ferrites into an alumina substrate and are thus not suitable for low-cost mass production. Thus, we propose to investigate the realization of a self-biased circulator in MID technology, which is a technology widely used in the telecommunication industry. In this paper, the magnetic properties of doped strontium hexaferrites will first be presented. Then, the MID process employed to realize the circulator will be briefly discussed. An investigation of the microwave properties of a Laser Direct Structuring (LDS)-compatible polymer will also be presented at this point. The third part of the paper will focus on the design of the self-biased circulator. We took advantage of the potential for technical innovation offered by MID technology to propose an original topology of a Y-junction circulator in a 3D Substrate Integrated Waveguide (SIW). Performances measured without magnets and as a function of the applied field will be presented and discussed in the final part of the paper.

2. Magnetic properties of doped strontium hexaferrites Among the available materials that can be used to realize selfbiased circulators, M-type strontium hexaferrite (SrM) with magnetoplumbite structure appears to be the most promising candidate because of its high remanent to saturation magnetization ratio and its high coercive field, which make it possible to maintain a stable magnetization state without an externally applied field. SrM generally exhibits a remanent to saturation ratio of about 85%. We previously demonstrated that Lanthanum-Cobalt substitutions make it possible to increase the squareness of the hysteresis loop, and thus, the remanent magnetization level [7]. In this paper, our interest is focused on Sr0,7La0,3Fe11,7Co0,3O19 ferrite material, here called (La,Co)0.3-SrM. The studied samples of pre-oriented (La,Co)0.3-SrM took the form of a flat cylinder (diameter ¼2.48 mm, length ¼1.02 mm, length to diameter ratio ¼0.411). The remanent magnetization was oriented along the c-axis, perpendicularly to the plane of the sample (axial axis of the disk). The magnetic hysteresis loops were measured (VSM MicroSense, LotQuantum company, Lowell, Massachssets, USA) over a field range of  22 kOe to 22 kOe at room temperature (290 K). Measurements were performed with the applied field Happ directed either along the axial axis or along one of the radial axes (Fig. 1). The measured value of the coercive field along the axial axis and radial axis were HCa ¼4668 Oe and HCr ¼3900 Oe,

respectively. The hysteresis loop of the sample showed a high squareness with a remanent-to-saturation ratio of about 90%. The extrinsic remanent magnetization of the ferrite disk has a value of 304 emu/cm3 (304 kA/m). As a self-biased operation mode is considered, this value of magnetization will be used for the design → of the self-biased circulator. The internal magnetic field H was calculated with the relation:

→ → H = Happ − N .4πM

(1)

where the diagonal demagnetizing tensor N¯ is written ⎛N 0 0 ⎞ x ⎜ ⎟ ⎜ 0 Ny 0 ⎟, with Nx þNy þ Nz ¼1. ⎜ ⎟ ⎜ ⎟ ⎝ 0 0 Nz ⎠ According to Aharoni's formula [13] modified in the case of a cylinder sample by considering a constant surface in the x–y plane compared with a plane-parallel sample, a length to diameter ratio equal to 0.411 leads to Nz ¼ 0.51, Nx ¼Ny ¼0.245. The intrinsic magnetization curves thus obtained are shown in Fig. 1. The high field regions of these loops were fitted using a law of approach to saturation to extract the saturation magnetization. The law of approach to magnetic saturation widely used for the analysis of the magnetization curves of polycrystalline materials is expressed as [14]:

M (H ) = MS (1 − a/H − b/H2) + χhf H

(2)

where MS is the magnetization at saturation. The term χhf is the forced magnetization coefficient that describes the linear increase in spontaneous magnetization at high fields. The term a/H is attributed to structural defects and non-magnetic inclusions. It is also well-known that the term b/H2 is caused by uniform magnetocrystalline anisotropy. Fig. 2 shows the magnetization, measured along the axial axis, plotted as a function of 1/H2. A linear variation between M and 1/H2 was observed in the field range 14.7–19.8 kOe. According to Eq. (2), the following value was obtained: MS ¼ 345 emu/cm3 (345 kA/m). These results (for the coercive field and saturation magnetization) are consistent with extrapolated values of other experimental data [15]. It has been shown in theory and experiment that investigation of the magnetization curve of a polycrystalline material, through the Singular Point Detection (SPD) method, can lead to the determination of the anisotropy field [16,17]. Actually, according to the SPD approach, the saturation magnetization is obtained at a finite field value, that is the anisotropy field Hk. As a consequence, a cusp in the experimental behavior of the second derivative

Fig. 1. Dependence of magnetization as a function of magnetic field of the cylindrical sample. Open symbols: Happ parallel to the axial axis (circles) and Happ perpendicular to the axial axis (triangles). The continuous curves show the dependence of magnetization as a function of the internal magnetic field H (blue curve: H parallel to the axial axis; red curve: H perpendicular to the axial axis). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

127

∂ 2M ∂ 2H

Fig. 2. Magnetization along the axial axis as a function of 1/H2. The dashed curve indicates a linear relation between the magnetization and 1/H2 in the field range 14.7–19.8 kOe.

128

V. Laur et al. / Journal of Magnetism and Magnetic Materials 404 (2016) 126–132

Fig. 3. First and second derivative of the magnetization as a function of applied field measured on a (La,Co)0.3-SrM sample.

along the magnetization curve in a polycrystalline material is 2

predicted to occur. The second derivative ( ∂ 2M ) curve measured ∂ H

parallel to the direction of easy magnetization (Fig. 3) shows an SPD signal for an internal field equal to HK ¼ 1439 kA/m. in good agreement with previous results [16].

3. MID-LDS: materials and processes MIDs can be produced in a variety of ways (single-shot injection, two-shot molding and film stamping) to suit a wide variety of applications. In this study, we used a single-shot injection technique combined with a Laser Direct Structuring technology, also known as MID-LDS technology. At first, a LDS-compatible material (thermoplastic with organic metallic complex particles) is shaped using a classical molding process. Then, the surface of the material is selectively activated by a laser in order to separate the metal atoms from the organic ligands and roughen the surface, enabling copper coating (Fig. 4). The final step consists in building up the metallization on the activated area by the use of electroless plating. MID-LDS technology makes it possible to produce 3D-metallized devices and is thus revolutionizing the traditional design of microwave circuits. A wide range of materials is now available for the MID-LDS process. However, their microwave properties are generally not well known and only a few of them have been used to realize microwave devices. The dielectric loss tangent is a key point in the realization of microwave devices. Among the LDS-compatible

Fig. 5. (a) Characterization cell in a rectangular waveguide and (b) permittivity and permeability spectra of Zeonex RS-420 LDS material in the Ka band (26–40 GHz).

polymers, liquid crystal polymers (LCP Vectras E820i [4] and E840i [1]), polyethylene terephthalate / polybutylene terephthalate (Pocans DP T7140 LDS [3,4]) and acrylonitrile butadiene styrene / polycarbonate (Xantars LDS 3710 [4]) have proved to be usable at microwave frequencies and led to the realization of demonstrators. However, dielectric losses of these materials remain higher than 3.10  3, which is quite a high value, especially for millimeter-wavelength applications. In this study, our interest focused on cyclo-olefin polymers (COP), especially Zeonexs RS-420 LDS. These easy-to-mold thermoplastic resins offer a low dielectric constant, low loss tangent and reliable performances over a wide range of environmental conditions due to the hydrophobic nature of COP. According to the datasheet [18], a dielectric constant of εr ¼2.1 and loss tangent of tan δ ¼5.10  4 are found at 1 GHz. In order to assess the microwave properties of this material, we performed rectangular waveguide and cavity-perturbation

Fig. 4. MID-LDS technology principle: single-step injection, laser direct structuring and electroless plating.

V. Laur et al. / Journal of Magnetism and Magnetic Materials 404 (2016) 126–132

characterizations. A rectangular waveguide characterization is a transmission/reflection method based on the measurement of the scattering parameters (S-parameters) of a loaded rectangular waveguide (Fig. 5). After a Through-Reflect-Line (TRL) calibration procedure, the 5-cm long WR-28 waveguide section was measured, without inserting the Material Under Test (MUT), using an Agilent E8364A network analyzer in the Ka band (26.5–40 GHz). The S-parameters of the empty waveguide enabled us to take into account metallic losses in the waveguide section. Then, a Zeonex RS-420 LDS sample was inserted into the waveguide. The S-parameters of the loaded waveguide were analyzed by a NicolsonRoss-Weir [19,20] (NRW) procedure modified by Baker-Jarvis et al. [21] previously used at Lab-STICC to characterize foam materials [22]. As expected, this polymer does not show magnetic properties and its permeability remains near unity (m′ E1) over the whole frequency band (Fig. 5). No dielectric dispersion is observed between 26 and 40 GHz. The microwave permittivity is kept constant around a mean value of 2.46. One should note that this value is slightly higher than the one given by the manufacturer and is in good agreement with measurements at lower frequencies (1– 18 GHz) performed in the laboratory using a coaxial method. Dielectric losses appear to be below the sensitivity threshold of this method, and we thus used a cavity-perturbation method to evaluate them more accurately. The cavity-perturbation method is based on the measurement of a cavity into which a sample is introduced. If the perturbation induced by the sample on the characteristics of the cavity (field pattern, resonance frequency, quality factor) is small, the dielectric properties of the material under test can be extracted analytically from the resonance frequency and quality factor shifts of the cavity. In this study, the cavity consisted of a 5-cm long WR-90 waveguide section coupled by a circular iris. The dielectric sample was inserted through a non-radiating rectangular slot placed at the center of the cavity (Fig. 6). This method was previously successfully used for the characterization of ferroelectric thin films [23]. Moreover, we proposed a numerical calibration process to enlarge the validity range of this technique [24]. A set of wellknown samples were thus used to calibrate the perturbationcavity method. The empty cavity shows a resonance frequency of fr0 ¼7.20364 GHz and a quality factor of Q0 ¼3389 (Fig. 6). When a 0.92  4.94  10.16 mm3 Zeonex RS-420 LDS sample is inserted in the cavity, the resonance frequency shifts to fr1 ¼ 7.12705 GHz and the quality factor decreases to Q1 ¼3226. By applying the cavityperturbation analysis, a dielectric constant of εr ¼2.48 and dielectric losses of tan δ ¼7.6.10  4 were obtained for this polymer at 7 GHz. This measurement is in good agreement with the rectangular waveguide measurement and confirms that the dielectric constant is slightly higher than the one given by the manufacturer. Dielectric losses are also slightly higher than those of the datasheet but remain quite low and compatible with microwave applications. One should note that the dielectric properties of polymers are closely linked to their density, and thus, to the molding conditions. These dielectric properties were then used for the design of the circulator.

4. Self-biased circulator design The MID-LDS process offers new possibilities for the design of microwave devices as it opens up a new dimension. Classical circulators are generally realized in a rectangular waveguide where high power handling capability is needed, or in stripline or microstrip technologies for which standard Printed Circuit Board (PCB) processes are sufficient. We chose Substrate Integrated Waveguide (SIW) technology to investigate the potential of the MID-LDS process. Indeed, if SIW is classically realized using

129

Fig. 6. (a) Characterization cell showing the cavity and (b) amplitude of transmission parameters S21(dB) of the empty and COP-loaded cavity.

Fig. 7. Comparison of classical SIW structure with metallic vias and 3D metallized MID-LDS SIW.

metallized via holes, the MID-LDS process makes it possible to metallize the substrate in 3D (Fig. 7). In order to ensure a continuous metallization of the substrate, we designed the substrate integrated waveguide with beveled edges. SIW makes it possible to propagate Transverse Electric (TEmn) or Transverse Magnetic

130

V. Laur et al. / Journal of Magnetism and Magnetic Materials 404 (2016) 126–132

Fig. 8. Dispersion diagram (a) imaginary part β and (b) real part α of a MID-LDS SIW propagation constant with a¼ 5.63 mm, aext ¼ 6.63 mm and b ¼ 1 mm (Zeonex RS-420 LDS substrate with εr ¼2.46 and tan δ ¼8.10  4).

(TMmn) modes for which the lower cutoff frequency is given by the following equation:

fc (m, n) =

1 2π εμ

×

⎛ mπ ⎞2 ⎛ nπ ⎞2 ⎜ ⎟ ⎟ + ⎜ ⎝ a ⎠ ⎝ b ⎠

(3)

where fc(m,n) is the cutoff of the TEmn or TMmn mode, ε the permittivity of the substrate in the case of a SIW, a the width of the waveguide and b its height. Eq. (3) is ambiguous in the case of a waveguide with beveled edges, where it appears difficult to determine the effective width aeff of the SIW. Thus, we performed electromagnetic simulations of the MID-LDS SIW in order to calculate the dispersion diagram of the waveguide. We considered a MID-LDS waveguide integrated in a Zeonex RS-420 LDS substrate whose widths are a ¼5.63 mm and aext ¼6.63 mm and height is b¼ 1 mm (Fig. 8). A unique TE10 mode is required to excite properly the ferrite resonator, which will be placed at the center of the Y-junction circulator. The dispersion diagram shows that a monomodal frequency band exists between 15.6 GHz (fc TE10) and 31.2 GHz (fc TE20). These values enable us to evaluate the effective width of the SIW to aeff ¼6.14 mm, which is naturally between a and aext. One should note that TEn0 modes (n ¼2 to 5) are the first higher-order modes due to the low shape factors of the waveguide (b/aeff ¼0.162). Fig. 8 also shows that the real part of the propagation constant α remains low (less than 0.42 Np/m) between 18.5 GHz and 31.2 GHz, which is the cutoff frequency of the TE20 mode. The range of 18.5–31.2 GHz will thus be considered as the operating band in which the Y-junction circulator can be designed.

Fig. 9. (a) Illustration of the Y-junction circulator in MID-LDS technology and (b) its main dimensions.

The self-biased circulator is presented in Fig. 9. The overall dimensions of the device are 42.5  49 mm2. It should be noted that these dimensions were chosen to be convenient for the measurement, although they could be strongly reduced if necessary. A Sr0,7La0,3Fe11,7Co0,3O19 puck (radius 2.48 mm) is placed at the center of the Y-junction in SIW technology. Microstrip access lines are connected to the Y-junction by tapered lines which act as microstrip-to-SIW transitions. One should note that the height of the device is not the same for microstrip access lines and the Y-junction. Indeed, high thickness is beneficial for minimizing losses in the Y-junction. However, if the thickness had been kept constant for the access lines, this would have made the microstrip lines very wide, causing high radiation losses. One advantage offered by, MID-LDS technology is that it enables such changes of width, which are difficult to realize with classical PCB technology. The measured properties of a (La,Co)0.3-SrM hexaferrite cylinder (diameter¼2.48 mm, length ¼1.02 mm) are given in Table 1. Anisotropy field, ferromagnetic resonance linewidth and permittivity are issued from the retro-simulation of a previous (La,Co)0.3-SrM-based self-biased circulator in rectangular Table 1 Sr0,7La0,3Fe11,7Co0,3O19 hexaferrite properties. 4πMs (G)

Mr/Ms

Hk (Oe)

ΔH (Oe)

εr

tan δ

4240

0.9

19,750

400

21

1.10  4

V. Laur et al. / Journal of Magnetism and Magnetic Materials 404 (2016) 126–132

131

Fig. 12. Circulator realized in MID-LDS technology. Fig. 10. Evolution of the internal field along z Hz as a function of the sample aspect ratio for different remanent magnetization level (Mz ranging from 3392 to 4240 G) of a (La,Co)0.3-SrM cylinder.

waveguide technology which presented insertion losses of 0.86 dB at 41 GHz without an applied field [25]. These properties need to be corrected during the design of the circulator by taking into account the squareness of the hysteresis loop and the aspect ratio of the ferrite puck inserted in the circulator. Indeed, the hysteresis loop reveals that the remanent magnetization level, here named Mz, reaches 90% of the saturation level and can thus be evaluated at 3820 G. Aspect ratio, and thus demagnetizing field, can be taken into account by using Eq. (1). Fig. 10 shows the evolution of the internal magnetic field in the hexaferrite cylinder as a function of the remanent magnetization along z Mz . One should note that, for a constant Hk ¼19,750 Oe, a low level of remanent magnetization makes it possible to keep a high value of internal field along z Hz because of demagnetizing fields. However, although a lanthanum- cobalt-substitution tends to increase the squareness of the hysteresis loop, it also causes an increase in the saturation magnetization [16]. In our case, the ferrite puck shows an aspect ratio ( h ) equal to 0.82 leading to a r

demagnetizing factor along z Nz equal to 0.51, the internal field being then equal to 17,800 Oe. These properties were used to model the self-biased circulator through the use of Ansys HFSS. Fig. 11 presents the simulated S-parameters of this MID-LDS self-biased circulator. The device shows minimum insertion losses of 1.69 dB at 28.79 GHz. Isolation and return losses remain below  15 dB between 28.56 GHz and 28.96 GHz leading to a bandwidth of 400 MHz. In the bandwidth, insertion losses remain lower than 2.20 dB.

Fig. 11. Simulated S-parameters (Insertion losses: IL, Isolation: Isol and Return Losses: RL) of the MID-LDS self-biased circulator in the 28.3–29.2 GHz frequency band.

5. Microwave characterization The circulator was realized using the MID-LDS process described in section III. Zeonex RS-420 LDS was molded and 3D metallized using the LDS process. The resulting device is shown in Fig. 12. The change in thickness between the microstrip access lines (300 mm) and the SIW Y-junction (1.02 mm) should be noted (see inset), which was made possible by the MID-LDS process. A hole was then machined at the center of the Y-junction in order to insert the ferrite puck. Then, the device was electrically closed using thin metallic films. The circulator was measured using an Agilent PNA E8364A network analyzer. After a SOLT calibration, the circulator was connected to the analyzer using End Launch connectors from Southwest Microwave, Tempe, Arizona, USA. The circulator was measured in isolator mode by using a matched load connected to one of its ports. Fig. 13 presents the S-parameters of the circulator measured in the 29.5–31.5 GHz frequency band. Compared to the simulation, the working frequency band is slightly increased but kept within the working frequency band of the MID-LDS SIW defined in section IV. Minimum insertion losses of 3.32 dB are observed at 30.7 GHz without magnets (self-biased mode). At this frequency, the isolation level is 13.9 dB and return losses are near 20 dB. A -10 dB bandwidth can be defined between 30.28 GHz and 30.84 GHz (560 MHz). However, in this bandwidth, insertion losses reach 4.57 dB. Experimental insertion losses are significantly higher than the one predicted by the simulation. However, End Launch connectors and coaxial adapters (2.4 mm V connectors to 2.92 mm K

Fig. 13. Comparison between measured S-parameters (solid lines) and retro-simulated ones (dashed lines) of the self-biased circulator in MID-LDS technology.

132

V. Laur et al. / Journal of Magnetism and Magnetic Materials 404 (2016) 126–132

connectors), required during the measurement, were not taken into account during the simulation. Insertion losses of these transitions were experimentally evaluated to 0.8 dB and thus explained, in part, the difference observed between the measured and simulated performances. Retro-simulations of the circulator were performed using Ansys HFSS software. Quite a good agreement was found between simulation and measurement (Fig. 13), especially in the working frequency band, for a permittivity of 17, an anisotropy field of 20,500 Oe and magnetic losses (ΔH) of 2000 Oe for the ferrite, which is not totally consistent with our previous results [24]. However, this can be explained in part by a dispersion of the properties of (La,Co)0.3-SrM ceramics and also by the anisotropy of dielectric properties of these pre-oriented materials, which is not taken into account during the simulations. As it was previously specified, differences of insertion losses between measurement and simulation is also in part due to connectors and coaxial adapters required during the measurement and not taken into account in the simulation.

6. Conclusion Laser Direct Structuring-Molded Interconnect Device MID-LDS technology was applied to the realization of a self-biased circulator. A pre-oriented Sr0,7La0,3Fe11,7Co0,3O19 (La,Co)-SrM hexaferrite was used to get a non-reciprocal effect without using permanent magnets. At first, we investigated the static properties (M-H cycles) of this material. This (La,Co)0.3-SrM material exhibits a very high remanent to saturation magnetization ratio of about 90% and can thus be used to realize a self-biased circulator. Dielectric properties of a LDS-compatible cyclo-olefin polymer were then investigated. A dielectric permittivity of 2.48 and dielectric losses of 7.6.10  4 were obtained at 7 GHz for this material, which appears to be compatible with microwave applications. This material was used to design a MID-LDS compatible Y-junction SIW circulator. Due to the specific shape of the substrate integrated waveguide in MID-LDS technology, an electromagnetic study was performed to determine the monomodal frequency band of the waveguide (15.6–31.2 GHz in this case). Finally, the integration of a pre-oriented Sr0,7La0,3Fe11,7Co0,3O19 puck into a MID-LDS SIW Y-junction was studied. Taking into account the aspect ratio of the ferrite puck, we optimized the design of the circulator in order to get a good isolation level in the 28.6 to 29 GHz frequency band. This structure was realized and measured. We observed a slight increase in the working frequency band and insertion losses. However, we demonstrated an isolation level of 13.9 dB at 30.7 GHz without magnets. This self-biased circulator is, to our knowledge, the first demonstration of a non-reciprocal microwave device in MID technology, which appears to be a promising technology for the realization of low-cost mass-production integrated microwave devices.

References [1] A. Friedrich, B. Geck, O. Klemp, H. Kellermann, in: Proceedings of the 43rd European Microwave Conference, Nuremberg, Germany, 7–10 October 2013, pp. 640–643. [2] D. Kaddour, S. Tedjini, A. Djamel, in: Proceedings of IEEE Antennas and Propagation Society International Symposium, Orlando, USA, 7–13 July 2013, pp. 1502–1503. [3] C. Orlob, Q.H. Dao, D. Komek, in: Proceedings of IEEE European Conference on Antennas and Propagation, Rome, Italy, 11–15 April 2011, pp. 361–364. [4] D. Unnikrishnan, D. Kaddour, S. Tedjini, in: Proceedings of International Symposium on Signals, Systems and Electronics, Postdam, Germany, 3–5 October 2012, pp. 1–5. [5] M. Dressler, B. Wunderle, K.F. Becker, H. Reichl, in: Proceedings of Electronic System-Integration Technology Conference, Berlin, Germany, 13-16 September 2010, pp. 1-6. [6] U. Kessler, H. Kück, W. Eberhardt, in: Proceedings of High Density Microsystem Design and Packaging and Component Failure Analysis, Shanghai, China, 30 June–3 July 2004, pp. 242–247. [7] V. Laur, G. Vérissimo, P. Queffelec, L.A. Farhat, H. Alaaeddine, J.C. Reihs, E. Laroche, G. Martin, R. Lebourgeois, J.P. Ganne, in: Proceedings of IEEE International Microwave Symposium, Phoenix, USA, 17–22 May 2015, http://dx.doi. org/10.1109/MWSYM.2015.7166760. [8] M.A. Tsankov, L.G. Milenova, Design of self-biased hexaferrite waveguide circulators, J. Appl. Phys. 73 (1993) 7018–7020. [9] J. Wang, A. Yang, Y. Chen, Z. Chen, A. Geiler, S.M. Gilette, V.G. Harris, C. Vittoria, Self-biased Y-junction circulator at Ku band, IEEE Microw. Wirel. Compon. Lett. 21 (2011) 292–294. [10] X. Zuo, H. How, S. Somu, C. Vittoria, Self-biased circulator/isolator at millimeter wavelengths using magnetically oriented polycrystalline strontium M-type hexaferrite, IEEE Trans. Magn. 39 (2003) 3160–3162. [11] B.K. O’Neil, J.L. Young, Experimental investigation of a self-biased microstrip circulator, IEEE Trans. Microw. Theory Tech. 57 (2009) 1669–1674. [12] S.A. Oliver, P. Shi, W. Hu, H. How, S.W. McKnight, N.E. Gruer, P.M. Zavracky, C. Vittoria, Integrated self-biased hexaferrite microstrip circulators for millimeter-wavelength applications, IEEE Trans. Micrw. Theory Tech. 49 (2001) 385–387. [13] A. Aharoni, Demagnetizing factors for rectangular ferromagnetic prisms, J. Appl. Phys. 83 (1998) 3432–3434. [14] S. Chikazumi, Physics of Magnetism, Wiley, New-York, 1964. [15] X. Liu, P. Hernandez-Gomez, K. Huang, S. Zhou, Y. Wang, X. Cai, H. Sun, B. Ma, Research on La3 þ -Co2 þ -substituted strontium ferrite magnets for high intrinsic coercive force, J. Magn. Magn. Mater. 305 (2006) 524–528. [16] G. Asti, F. Bolzoni, J.M. Le Breton, M. Ghidini, A. Morel, M. Solzi, F. kools, P. Tenaud, Magnetic anisotropy of LaCo substituted SrFe12O19 ferrites, J. Magn. Magn. Mater. 272 (276) (2004) E1845–E1846. [17] F. Fiorillo, Characterization and Measurement of Magnetic Materials, Elsevier Series in Electromagnetism, 2004. [18] 〈http://www.zeonex.com/electronics.aspx〉. [19] A.M. Nicolson, G.F. Ross, Measurement of the intrinsic properties of materials by time domain techniques, IEEE Trans. Instrum. Meas. 1 (1970) 377–382. [20] W.B. Weir, Automatic measurement of complex dielectric constant and permeability at microwave frequencies, Proc. IEEE 62 (1974) 33–36. [21] J. Baker-Jarvis, M.D. Janezic, J.H. Grosvenor, R.G. Geyer, Transmission/reflection and short-circuit line methods for measuring permittivity and permeability, Natl. Inst. Stand. Tech. Tech. Note, 1355-R, 1993. [22] R. Benzerga, V. Laur, R. Lebullenger, L. Le Gendre, S. Genty, A. Sharaiha, P. Queffelec, Waste-glass recycling: a step toward microwave applications, Mater. Res. Bull. 67 (2015) 261–265. [23] P. Queffelec, V. Laur, A. Chevalier, J.M. Le Floch, D. Passerieux, D. Cros, V. Madrangeas, A. Le Febvrier, S. Députier, M. Guilloux-Viry, G. Houzet, T. Lacrevaz, C. Bermond, B. Fléchet, Intercomparison of permittivity measurement techniques for ferroelectric thin layers, J. Appl. Phys. 115 (2014) 024103. [24] V. Laur, P. Queffelec, Cavity-perturbation method improvement by a numerical calibration process, Microw. Opt. Technol. Lett. 56 (2014) 904–906. [25] V. Laur, G. Vérissimo, P. Queffelec, L.A. Farhat, H. Alaaeddine, J.C. Reihs, E. Laroche, G. Martin, R. Lebourgeois, J.P. Ganne, Self-biased Y-junction circulators using lanthanum- and cobalt-substituted strontium hexaferrites, IEEE Trans. Microw. Theory Tech. 63 (2015) 4376–4381.