Planar Antennas in LTCC Technology With Transceiver ... - IEEE Xplore

2830

IEEE TRANSACTIONS ON MICROWAVE THEORY AND TECHNIQUES, VOL. 54, NO. 6, JUNE 2006

Planar Antennas in LTCC Technology With Transceiver Integration Capability for Ultra-Wideband Applications Grzegorz Brzezina, Student Member, IEEE, Langis Roy, Member, IEEE, and Leonard MacEachern, Member, IEEE

Abstract—We present two novel ultra-wideband (UWB) antennas embedded in a low-temperature co-fired ceramic (LTCC) package designed to house the UWB transceiver chip. Given their planar topology, circuit integration possibilities, and compact size, a partial ground-plane triangular monopole antenna (PGP-TM) and an antipodal Vivaldi antenna (AVA) are fully characterized. The performance in both the frequency and time domain are presented. The PGP-TM employs parasitic elements for tuning of the antenna’s return loss. The PGP-TM antenna’s measured 3.5–6.5-GHz bandwidth and omnidirectional pattern with 0-dB gain is suitable for the direct-sequence UWB (DS-UWB) lower subband, while the AVA’s measured bandwidth of 3.35 GHz from 6.65 to 10 GHz and 5-dB gain make it suitable for the DS-UWB upper subband. The complete LTCC module containing the PGP-TM measures only 30 mm 25 mm 1.2 mm, while the AVA module measures 50 mm 25 mm 1.2 mm. Both LTCC modules can accommodate transceiver electronics because of a specially designed circuit feature. The effects of path loss can be canceled by combining these antennas in a transmission system. These are believed to be the first demonstrations of system-in-package technology for UWB applications. Index Terms—Antipodal Vivaldi antenna (AVA), low-temperature co-fired ceramic (LTCC), path loss, system-in-package (SIP), ultra-wideband (UWB) antennas.

I. INTRODUCTION HE development of ultra-wideband (UWB) technology for short-range high-speed wireless communication is progressing rapidly. A UWB antenna should be effective in transmitting very short and low-power pulses in the 3.1–10.6-GHz range. Ideally, the UWB antenna should be compact, planar, low-cost, and reliable. Compatibility and ease of integration with electronics is also desirable. Recently, UWB antenna designs have been presented that achieve very broad impedance bandwidths without evaluating the response in the time domain to a pulse excitation [1], [2]. Other work has shown a UWB slot antenna realized in low-temperature co-fired ceramic (LTCC) technology [3]. However, this design does not show any method for integrating transceiver electronics. The development of compact modules that intimately combine the antenna and transceiver electronics is receiving increasing interest, although the LTCC receiver presented in [4] does not contain the antenna. Advances in circuit miniaturization and packaging technology have made the goals of

T

Manuscript received October 17, 2005; revised January 26, 2006. The authors are with the Department of Electronics, Carleton University, Ottawa, ON, Canada K1S 5B6 (e-mail: [email protected]; [email protected]; [email protected]). Digital Object Identifier 10.1109/TMTT.2006.875448

Fig. 1. PGP-TM in LTCC (angled view).

smaller size and lighter weight attainable for a general packet radio service (GPRS) mobile application [5]. In this paper, we build upon the results presented in [6] and propose a second system-in-package (SIP) for UWB using LTCC technology. A novel partial ground-plane triangular monopole (PGP-TM) antenna having parasitic elements and a compact antipodal Vivaldi antenna (AVA) are employed. Both antennas have transceiver circuit integration capabilities. Time- and frequency-domain performances are investigated for a pair of identical and mixed antennas. Path loss causes a roll-off in received power with frequency [7], but the results for the latter case demonstrate effective path-loss compensation. II. LTCC PGP-TM ANTENNA DESIGN Recently, a simple antenna topology that incorporates a microstrip feed and a partial ground plane has been used to design broadband antennas [2]. This topology makes it possible to combine the antenna with integrated RF electronics and passive lumped components. Also, there is no need for a costly and performance-limiting balun or a large matching network, as in [8]. The PGP-TM employed here consists of a top metal layer, which differs from that in [1] by its triangular shape and the addition of adjacent parasitic elements. Fig. 1 shows the configuration of the proposed UWB antenna. The PGP-TM was designed to operate in the lower frequency band as defined by the direct-sequence UWB (DS-UWB) group [9]. Table I displays the performance and material specifications. The topology of the PGP-TM mimics a bow-tie shape to facilitate a wide impedance bandwidth. The microstrip feed line diverges linearly at an angle of 40 to make the bow-tie-shaped section. The partial ground plane terminates when the top layer metallization begins to diverge to form the

0018-9480/$20.00 © 2006 IEEE

BRZEZINA et al.: PLANAR ANTENNAS IN LTCC TECHNOLOGY WITH TRANSCEIVER INTEGRATION CAPABILITY FOR UWB APPLICATIONS

2831

TABLE I LTCC-BASED PGP-TM SPECIFICATIONS

TABLE II PGP-TM DIMENSIONS

Fig. 2. PGP-TM antenna topology (without circuit feature).

triangular-monopole-shaped radiating element. A rectangular metal area forms the rest of the main section of the antenna. Two novel triangular parasitic elements on either side of the bow-tie section serve as capacitive loads. Capacitive loading reduces the input impedance variation with frequency of the antenna while maintaining its efficiency. The capacitance can be adjusted by varying the separation between the parasitic elements and the main part of the antenna. This feature provides another important parameter that can be used to change the performance of the antenna. The effect of parasitic element separation was explored using HFSS and is shown in Section V. The microstrip feed section was designed for a characteristic of impedance of 50 , resulting in a microstrip width 1.8 mm. A parametric analysis in HFSS was used to vary the other dimensions of the antenna. The final optimized dimensions of the LTCC-based PGP-TM antenna are presented in Table II. The area requirements for this design are less than 25% of those for the LTCC UWB antenna in [3]. The descriptions of the variables in Table II are as follows: and represent the lengths of the substrate, partial ground plane, and monopole section, respectively. and represent the widths of the substrate, microstrip feed, and monopole section, respectively. Fig. 2 shows a frontal view of the PGP-TM antenna with all the dimensions labeled accordingly.

Fig. 3. LTCC circuit feature for UWB transceiver chip (top).

The circuit feature is placed 3.25 mm away from the adjacent microstrip feed line of the antenna to minimize interference. The circuit feature consists of a cavity with a grounded floor. The cavity has a footprint of 3 mm 4 mm to ensure that even large microchip dies can be used. Metal traces were drawn along one edge of the cavity. In this way, a daughter board could be easily attached to the LTCC substrate to feed ground and power to the microchip. A second 50- microstrip line was added to serve as an alternate output port. The microchip itself would be placed inside the cavity and bonded to the ground plane with conductive epoxy. Bond wires would be used to connect the microchip to the output of the antenna. Bond wires add inductance (sometimes advantageously) to the circuit, and any chip insertion or accompanying measurements must take them into account.

III. PLANAR CIRCUIT INTEGRATION The LTCC substrate was not only selected for its superior microwave properties (low loss and dispersion) that are not necessarily found in “standard technology,” such as FR-4, but to take advantage of its integration capabilities. As a demonstration of SIP design, the PGP-TM antenna required the multilayer aspects of LTCC to integrate a circuit feature that could accommodate a microchip die. The choice of 12 layers was made to obtain the correct substrate height and to permit the realization of the circuit cavity, which is eight layers deep, without compromising structural integrity. Importantly, the antenna can operate with or without a functional microchip die. Fig. 3 shows the circuit of the chip in detail.

IV. MEASUREMENT SETUP Radiation pattern measurements were performed inside an anechoic chamber. An Agilent 8720ES vector network analyzer and a computer workstation running FR959 far-field measurement software was used. Pattern cuts in the two principal planes (i.e., azimuth and elevation) were taken at 4, 6, and 8 GHz. The azimuthal plane measurement constitutes rotation in the plane, which is the plane of the antenna. An elevation plane meaplane. surement is performed in the orthogonal Scattering parameter measurements were performed within a Faraday cage to minimize environmental interference. First, the return loss of each antenna was obtained to verify broad-

2832


Fig. 5. LTCC-based PGP-TM S

measurement.

Fig. 4. Effect of parasitic element separation.

band impedance matching. Then, the transmission coefficient of a two-antenna system (using identical antennas under or test) was measured to assess linear phase and flat amplitude response. Measurements were performed at 20 and 75 cm to confirm the results. Dotted lines of best fit are included to capture the essential behavior and remove effects due to the nonideal environment. These lines are generated by MATLAB’s curve fitting function and show the best prediction for the variation of and group delay with frequency. Time-domain measurements were performed with a burst CW UWB pulse generator, Agilent 86100A wide-bandwidth oscilloscope, and a computer workstation. An ideal base line for the measurements was created by directly connecting two ports of the oscilloscope. V. PGP-TM SIMULATED AND MEASURED RESULTS As discussed earlier, the parasitic elements on either side of the main body of the PGP-TM have a large impact on its performance. Fig. 4 shows the effect of changing the separation distance between the elements and the antenna body on the return loss. Increasing the distance of the elements has the effect of reducing the of the antenna. At a distance of 1.5 mm, the match is very good, while at 3.0 mm the match is poorer but the impedance bandwidth has increased by almost 1 GHz. A distance of 2.0 mm was chosen to be optimum. It provides a good compromise between a high and broad bandwidth. Frequency domain -parameter and radiation pattern measurements along with time-domain pulse responses were obtained to fully characterize the antenna. As shown in Fig. 5, a total impedance bandwidth of 2.96 GHz from 3.57 to 6.53 GHz was measured. The agreement with the simulated results is good. , The transmission coefficient of a two-antenna system, or is an important frequency-domain indicator of the time-domain performance of a UWB antenna [10], [11]. For good perforshould be flat over the same bandmance, the magnitude of width that the voltage standing wave ratio (VSWR) is low. All of the frequency components in the frequency range where

Fig. 6. Measured S

for LTCC-based PGP-TM.

is near constant will be transmitted equally. Verifying that the is linear over the bandwidth should not be omitted. phase of A linear phase represents the application of a constant time delay to all of the frequency components of the transmitted signal. Achieving these two aspects of the transmission coefficient over the bandwidth of the UWB signal guarantees that distortion will be minimized. along with Fig. 6 displays the magnitude and phase of group delay. These measurements were obtained from a two-antenna system using identical PGP-TM antennas in a Faraday cage. The measurements were performed at two different distances to average out variations. The dashed lines show the best and group delay with freprediction for the variation of quency. shows In the antenna’s operating band, the magnitude of a variation of 7 dB when a line of best fit, with the equation , is applied. The trend with frequency is decreasing at is linear beyond 4.5 GHz. 18 dB per decade. The phase of Even with some initial instability before 4 GHz, the group delay


Fig. 7. Measured azimuthal patterns for LTCC PGP-TM at 4 GHz.

2833

Fig. 8. Measured elevation patterns for LTCC PGP-TM at 4 GHz.

stays within a variation of 1.8 ns in the operating band of the antenna when a seventh-degree polynomial is fitted to the data. can be The decreasing trend of 20 dB per decade for readily explained by using the Friis transmission equation [12]

(1) The requirement to maintain a flat system is dependent on product that is the ability to achieve a constant close to the regulatory limit. However, a property of antennas called path loss has a frequency dependence that is described as and is given by

path loss

(2) Fig. 9. Measured azimuthal patterns for LTCC PGP-TM at 6 GHz.

In decibels, the free-space path-loss dependency from GHz to GHz is given by

dB/dec

(3)

Antenna gain in general also has a frequency dependence. However, in this case, it is virtually constant due to the monopole being considerably less than over all frequencies. Therefore, omnidirectional or constant gain antennas (or 20 dB per cause the received power to decrease as decade) because of the path loss[7]. Radiation pattern cuts at 4 and 6 GHz in the two principal planes are shown in Figs. 7–10. The azimuthal cut corresponds to the plane of the antenna, while the elevation cut corresponds to an orthogonal plane. They indicate monopole-like patterns with omnidirectional radiation in elevation and gain of approximately 0 dB. At 6 GHz, the cross-polarized levels are, in general, below 15 dB, but are noticeably higher at 4 GHz due to

inadequate shielding of the antenna test fixture. The radiation performance of this antenna is stable over the entire band with decreased ripples at higher frequencies. Time-domain measurements were performed with identical PGP-TM antennas mounted on plastic columns and placed 50 cm apart. To create a near-ideal benchmark case, a cable was used to directly connect the pulse generator to the oscilloscope. A burst CW UWB pulse centered at 6 GHz and having a bandwidth of 2 GHz was applied to the antenna to obtain the time-domain response [13]. The spectral content (normalized to peak power) of the burst CW pulse satisfies the UWB spectrum mask as defined by the FCC. Fig. 11 shows the response in the time domain for this antenna system superimposed on the pulse of the benchmark case with normalized amplitudes. The antenna produces a pulse shape that is very well maintained with little dispersion or distortion evident in the received pulse. The correlation coefficient is calculated to be 94.56%.

2834


TABLE III LTCC AVA SPECIFICATIONS

Fig. 10. Measured elevation patterns for LTCC PGP-TM at 6 GHz.

Fig. 11. LTCC-based PGP-TM response in the time domain.

VI. COMPACT ANTIPODAL VIVALDI DESIGN utilizing two PGP-TM antennas To achieve a flat system or, in general, any two omnidirectional radiators with sizes less than would require a method to increase the gain of the system at higher frequencies. As shown in Section V, in the band from 1 to 10 GHz, path loss increases linearly with a slope of , we would need an 20 dB/dec. Therefore, to maintain a flat amplifier, the gain of which increases at the same rate to cancel the path loss. However, this adds to system complexity and cost. Another more elegant solution exists that takes advantage of the fact that the gain of most directional antennas increases at the same rate with frequency as path loss so that they naturally cancel each other. This can be seen when antenna gain is defined in terms of antenna effective aperture as

(4)

Conversely to omnidirectional antennas, the effective aperture of most directional antennas increases as , thus causing the radiation pattern to narrow and for gain to increase. Therefore, a link that includes one omnidirectional and one directional antenna would result in a flatter system while keeping complexity unchanged [7]. A directional antenna was designed based on a Vivaldi topology. Planar Vivaldi antennas have been used in radar-like communications since 1979 [14]. They are a form of tapered slot radiators that support traveling waves. Ground-penetrating radar applications [15] and UWB communications studies [16] have shown that this antenna can preserve the shape of transmitted UWB pulses. The performance of this antenna is limited by the need for a wideband balun. Considerable effort is required to give broadband performance since the feed structure is so complicated. Traditionally, a Marchand-type stripline to slot feed transition is used [17]. More recently, a type of Vivaldi antenna (known as the antipodal Vivaldi antenna) has been developed that overcame these problems by enabling a simple microstrip feed method [18], [19]. The AVA is a two-layer design that uses a microstrip feed. While making the feed simpler, the AVA also inherits the good time-domain performance of the original tapered slot antenna. The AVA is a dual-layer structure composed primarily of three different features: a microstrip feed, a paired-strip middle section, and a radiating section. The smooth transition from microstrip to radiating section allows for broadband performance. The transition region is responsible for connecting the highly capacitive feed structure to the inductive radiating section [20]. The radiating section is formed by the metallizations on either side of the substrate that flare in opposite directions, forming a tapered slot. Given the promising characteristics of the AVA, a compact version of this antenna, measuring only 50 mm 25 mm, was designed to operate in the upper frequency band as defined by the DS-UWB group. The upper band occupies the spectrum from 6.2 to 9.7 GHz. The complete design specifications are provided in Table III. The physical characteristics of the Vivaldi antenna that affect its operation are the mathematical description of the radiating tapers, length, and width of the transition region and the groundplane patterning. Fig. 12 labels these critical dimensions. The first step in the design process is to design the microstrip feed for a - characteristic impedance. If represents the width of the microstrip line and represents the height of the dielectric substrate, then the required ratio is [21] (5)


2835

TABLE IV LTCC AVA DESIGN PARAMETERS (mm)

Fig. 12. AVA schematic.

where

Fig. 13. LTCC-based AVA.

The variable substitutions were made based on the material information provided in Table III, resulting in a microstrip width of 1.8 mm. The ground plane slowly tapers along the AVA to transform the microstrip structure to a paired strip transmission line. The equation of this taper is circular with a radius calculated from

is circular in shape and has the same radius defined by (6). The exterior edge of the flare is elliptical in shape and is defined by (9) (10)

mm

(6)

The center of the circle is made to be on the edge of the substrate and a distance defined by (7) away from the feed point

mm

(7)

Using [22], the width of the paired strip line to maintain is

(8) where is the width of the paired line, is the thickness of the dielectric substrate, and This resulted in a paired strip width of 1.92 mm. The paired strip width deviates more from the width of a simple microstrip line when the ratio of the microstrip width to substrate thickness is larger. An empirical analysis established the optimum length of the paired strip to be 1 mm. At the end of the transition region, the paired strip metallization on either side of the substrate widens in opposite directions to become the radiating section of the antenna. The shape of this flare is critical to the performance of the AVA and should be as smooth as possible to improve high-frequency performance. The interior of the flare

is the location of the center of the ellipse along where the edge of the AVA and, after optimizations, was selected to be 15 mm away from the feed plane, and and are the length and width of the AVA, respectively. The length of the Vivaldi antenna is approximately one wavelength at the lowest frequency of operation, while the width is approximately half of one wavelength. The elliptical equations were developed to improve low-frequency performance by maximizing the width of the slot aperture. Here, the lowest frequency of operation is 6.2 GHz, which corresponds to a length and width of 48 and 24 mm, respectively. After optimizations, the final dimensions of the AVA were found to be 50 mm in length and 25 mm in width. Table IV lists all of the dimensions of the AVA calculated from (5)–(10). VII. COMPACT ANTIPODAL VIVALDI SIMULATED AND MEASURED RESULTS Fig. 13 shows the AVA antenna discussed in the previous section. As with the PGP-TM antenna, the AVA includes a circuit feature that can accommodate a microchip die. The return loss results of Fig. 14 show an impedance bandwidth from 6.65 to 10 GHz with good agreement between HFSS and the measured values. results, which are shown in Fig. 15, of a The measured two-AVA system show a very stable magnitude variation. The variation is within 5 dB in the operating band of the antenna , is when a line of best fit, whose equation is used. As expected, the magnitude shows an increasing trend with frequency of almost 20 dB per decade in the range from

2836

Fig. 14. Measured jS


j

of LTCC-based AVA.

Fig. 16. Measured azimuthal patterns for LTCC AVA at 6 GHz.

Fig. 17. Measured elevation patterns for LTCC AVA at 6 GHz. Fig. 15. Measured S

of LTCC-based AVA.

1 to 10 GHz. The phase is very linear in this band. This leads to a very stable group delay that has a 0.5-ns variation when a seventh-degree polynomial is fitted to the data. Radiation pattern cuts at 6 and 8 GHz in the two principal planes are shown in Figs. 16–19. In all cases, the agreement between the simulated and measured radiation patterns is very good. The radiation pattern is stable and shows no degradation at 8 GHz. The plots indicate a boresight gain of 2.76 dB at 6 GHz and 5.1 dB at 8 GHz. At 6 GHz, the cross-polarized levels are 20 dB below the copolarized levels in both principle planes. These results show that, along with the small increase in gain, the LTCC substrate provides good high-frequency radiation performance. A summary of the simulated and measured radiation parameters is provided in Table V. results for this antenna would indicate good The excellent performance in the time domain. Fig. 20 confirms this supposition. The response in the time domain is very well matched to the ideal case and has a correlation coefficient of 93.54%. This confirms that frequency-domain parameters such as linear phase

Fig. 18. Measured azimuthal patterns for LTCC AVA at 8 GHz.

and a small variation in the group delay indicate good time-domain performance. This design shows that compact Vivaldi antennas can be made to operate at relatively low frequencies and be combined with planar integrated circuits.


2837

TABLE VI GAIN AND LINK PERFORMANCE FOR ELECTRICALLY SMALL ANTENNAS

Fig. 19. Measured elevation patterns for LTCC AVA at 8 GHz.

TABLE V LTCC-BASED AVA RADIATION PATTERN RESULTS COMPARISON

Fig. 21. Measured plied).

Fig. 20. Measured LTCC-based AVA time-domain response.

VIII. MEASURED RESULTS FOR A MIXED ANTENNA SYSTEM As discussed in Section VI, a communications link employing an omnidirectional and a directional antenna will . For comparison, four minimize the effects of path loss on distinct combinations of electrically small transmit and receive antennas are listed in Table VI along with the relationship between gain and received power.

j

S

j

for mixed-antenna system (with lines of best fit ap-

Case 1 refers to a communications link consisting of omnidirectional transmit and receive antennas; case 2 consists of an omnidirectional transmitter and directional receiver; cases 3 and 4 both consist of directional transmit and receive antennas but have decreasing or flat transmit power spectrums, respectively. In case 1, the constant gain antennas cause the received power to (or 20 dB per decade) because of path loss. decrease as This shortcoming must be compensated for by the transmitter designer. Case 2 does not have this shortcoming since the increasing gain of the receive antenna cancels the increasing path loss with respect to frequency. This is also true in case 3, howis used to achieve a ever, a transmit power that decreases as spectrum. In case 4, a flat transmit power results in flat spectrum that increases as (or 20 dB per decade). a In cases 2 and 3, the received power will be higher relative to case 1 [7]. The combination measured here corresponds to case 2. A link consisting of a PGP-TM and an AVA was tested, and the results curve for this configuration is are shown in Fig. 21. The compared to those previously obtained for two PGP-TMs and response in the range from 1 to two AVAs. As expected, the 10 GHz is flatter. The trend is increasing with a variation of 8 dB , per decade when a line of best fit, with equation curve would be even flatter if the operating is used. The frequencies for these antennas were to overlap more. A comparison of the time-domain response with the ideal case is shown in Fig. 22. The plot indicates very little distorcurve leads to a better timetion or dispersion. The flatter

2838


Near distortion-less transmission can be achieved by combining these antennas in a transmit–receive system. The response in the time domain showed the smallest amount of distortion of the three cases studied here. This study shows that a compact SIP for UWB communications can be achieved using LTCC. ACKNOWLEDGMENT The authors would like to thank K. Kautio along with the micromodules group at VTT Technical Research Center of Finland that helped make this work possible. REFERENCES

Fig. 22. Measured time-domain response for a mixed-antenna system.

domain response that has a correlation coefficient of 96.59%, which is 2.03% better than the case with two PGP-TMs and 3.05% better than the case with two AVAs. IX. CONCLUSION The feasibility of designing and fabricating small, versatile, UWB-compliant antennas has been investigated. The DS-UWB implementation was selected as the target application. Requirements and practical design guidelines for UWB antennas were formulated before the design and full characterization of two novel antennas was undertaken. The most important requirements are that a candidate UWB antenna have sufficient that has a flat magniimpedance bandwidth and a system tude and linear phase. These requirements are most likely met by antennas that support traveling waves or that have a low . Most often, these antennas incorporate tapers or rounded edges to give surface currents a smooth path to follow. A practical LTCC UWB antenna with a simple feed structure and special circuit feature useful for UWB transceiver chip integration has been demonstrated. A return loss below 10 dB from 3.57 to 6.53 GHz was obtained for the PGP-TM. This antenna has the advantages of compact size, omnidirectional radiation pattern and simple implementation. Good agreement was obtained between the measurements and simulations generated by HFSS. Radiation patterns were measured in an anechoic chamber at 4 and 6 GHz. The response in the time domain of the antenna to a burst CW signal was measured and exhibits low distortion. A compact AVA that also incorporates a novel circuit feature has been demonstrated. The antenna shows an increasing system slope of 19 dB per decade—making this antenna effective at path-loss cancellation. The impedance bandwidth was 3.35 GHz, from 6.65 to 10 GHz. This makes it suitable for use in the upper band of the DS-UWB implementation. Gain was measured to be 5 dB at 8 GHz. A UWB pulse was transmitted and received with very little added distortion in an antenna system utilizing these AVAs.

[1] S. H. Choi, J. K. Park, S. K. Kim, and J. Y. Park, “A new ultra-wideband antenna for UWB applications,” Microw. Opt. Technol. Lett., vol. 40, no. 5, pp. 399–401, 2004. [2] J. Liang, C. Chiau, X. Chen, and C. Parini, “Printed circular disc monopole antenna for ultra-wideband applications,” Electron. Lett., vol. 40, pp. 1246–1247, Sept. 2005. [3] C. Ying and Y. Zhang, “Integration of ultra-wideband slot antenna on LTCC substrate,” Electron. Lett., vol. 40, pp. 645–646, May 2004. [4] L. Pergola, R. Vahldieck, U. Gobel, and P. Nuchter, “An LTCC-based 5–6 GHz receiver with integrated antenna,” in Proc. 7th Eur. Conf. Wireless Technol., Oct. 2002, pp. 165–168. [5] Y. Lin, C. Liu, K. Li, and C. Chen, “Design of an LTCC tri-band transceiver module for GPRS mobile applications,” IEEE Trans. Microw. Theory Tech., vol. 52, no. 12, pp. 2718–2724, Dec. 2004. [6] G. Brzezina, L. Roy, and L. MacEchern, “LTCC ultra-wideband antenna with transceiver integration capability,” in Proc. 35th Eur. Microw. Conf., Oct. 2005, pp. 2011–2014. [7] H. G. Schantz, “Introduction to ultra-wideband antennas,” in Proc. IEEE Conf. Ultra Wideband Syst. Technol., Nov. 16–19, 2003, pp. 1–9. [8] A. Saitou, T. Iwaki, K. Honjo, K. Sato, T. Koyama, and K. Watanabe, “Practical realization of self-complementary broadband antenna on low-loss resin substrate for UWB applications,” in IEEE MTT-S Int. Microw. Symp. Dig., Jun. 2004, vol. 2, pp. 1265–1268. [9] R. Fisher, R. Kohno, M. McLaughlin, and M. Welborn, DS-UWB Physical Layer Submission to 802.15 Task Group 3a Jul. 2004. [10] Q. Ye, “Time domain response of ultra wideband dipole antennas,” in Proc. Antem/URSI Conf. Antenna Technol. Appl. Electromagn., Ottawa, ON, Canada, Jul. 2004, pp. 661–664. [11] G. M. Brzezina, “Planar antennas in LTCC technology for ultra-wideband applications,” M.S. thesis, Dept. Electron., Carleton Univ., Ottawa, ON, Canada, 2005. [12] D. M. Pozar, Microwave Engineering, 3rd ed. New York: Wiley, 2005, pp. 636–665. [13] D. Ball, P. Charlebois, and W. Lauber, “Ultra-wideband signal sources for interference measurements,” in Proc. IEEE Int. Conf. Ultra-Wideband, Sep. 2005, pp. 621–626. [14] P. Gibson, “The Vivaldi aerial,” in Proc. 9th Eur. Microw. Conf., 1979, pp. 101–105. [15] V. Mikhnev and P. Vainikainen, “Wideband tapered-slot antenna with corrugated edges for GPR applications,” in Proc. 33rd Eur. Microw. Conf., 2003, vol. 2, pp. 727–729. [16] W. Sorgel, C. Waldschmidt, and W. Wiesbeck, “Transient responses of a vivaldi antenna and logarithmic periodic dipole array for UWB communication,” in Proc. IEEE Antennas Propag. Soc. Int. Symp., 2003, vol. 3, pp. 592–595. [17] J. Shin and D. H. Schaubert, “A parameter study of stripline-fed vivaldi notch-antenna arrays,” IEEE Trans. Antennas Propag., vol. 47, no. 5, pp. 879–886, May 1999. [18] E. Gazit, “Improved design of the Vivaldi antenna,” Proc. IEE Microw., Antennas Propag., vol. 135, no. 2, pp. 89–92, 1988. [19] J. Langley, P. Hall, and P. Newman, “Balanced antipodal Vivaldi antenna for wide bandwidth phased arrays,” in Proc. IEE Microw., Antennas Propag., 1996, vol. 143, pp. 97–102. [20] J. Noronha, T. Bielwa, C. Anderson, D. Sweeney, S. Licul, and W. Davis, “Designing antennas for UWB systems,” Microw. RF Mag., pp. 53–61, 2003. [21] A. Petosa, Antennas and Arrays Course Notes. Ottawa, ON, Canada: Carleton Univ., 2002, pp. 1–30. [22] B. Wadell, Transmission Line Design Handbook. Norwell, MA: Artech House, 1991.


Grzegorz Brzezina (S’05) received the B.Eng. and M.A.Sc. degrees in electrical engineering from Carleton University, Ottawa, ON, Canada, in 2002 and 2005, respectively, and is currently working toward the Ph.D. degree at Carleton University. He has been involved with the design of ultra-wideband antennas and electromagnetic simulators at the Communications Research Center, Ottawa, Canada. His research interests include compact antennas for ultra-wideband, low-temperature co-fired ceramic packaging, miniaturized microwave filters, electromagnetics modeling, and high-performance RFICs.

Langis Roy (M’93) received the B.A.Sc. degree in electrical engineering from the University of Waterloo, Waterloo, ON, Canada, in 1987, and the M.Eng and Ph.D. degrees from Carleton University, Ottawa, ON, Canada, in 1989 and 1993, respectively. After a research fellowship with Matra Marconi Space France in 1993, he joined the Department of Electrical Engineering, University of Ottawa. Ottawa, ON, Canada, as an Assistant Professor. Since 1999, he has been an Associate Professor with the Department of Electronics, Carleton University,

2839

and, since 2003, the Department Chair. His research interests are in GaAs monolithic integrated circuits, high-performance microwave and optoelectronic circuit packaging, integrated active antennas, and numerical techniques in electromagnetics. Dr. Roy is a licensed Professional Engineer in the Province of Ontario.

Leonard MacEachern (S’92–M’98) received the B.Sc. degree from Acadia University, Wolfville, NS, Canada, in 1990, the B.Eng. and M.A.Sc. degrees from the Technical University of Nova Scotia, Halifax, NS, Canada, in 1993 and 1996, respectively, and the Ph.D. degree from the University of Waterloo, Waterloo, ON, Canada. He is currently an Assistant Professor with the Department of Electronics, Carleton University, Ottawa, ON, Canada. His current research is focused on electrooptical interfaces, laser modeling, laser predistortion techniques, and RFICs.