Journal of the Korean Physical Society, Vol. 68, No. 7, April 2016, pp. 908∼913

Design of a Loop Resonator with a Split-Ring-Resonator (SRR) for aHuman-body Coil in 3 T MRI Systems

Hyeok Woo Son and Young Ki Cho∗

School of Electronics Engineering, Kyungpook National University, Daegu 41566, Korea

Byung Mun Kim

Division of IT Cooperative Systems, Gyeongbuk Provincial College, Yecheon 36830, Korea

Hyun Man Back

Division of Magnetic Resonance Research, Korea Basic Science Institute, Daejeon 34133, Korea, andDepartment of Bio-Analytical Science, Korea University of Science and Technology, Daejeon 34113, Korea

Hyoungsuk Yoo

Department of Biomedical Engineering, University of Ulsan, Ulsan 44610, Korea

(Received 15 July 2015, in final form 28 December 2015)

A new radio-frequency (RF) resonator for Nuclear Magnetic Resonance (NMR) imaging at clinicalmagnetic resonance imaging (MRI) systems is proposed in this paper. An approach based on theeffects of the properties of metamaterials in split-ring resonators (SRRs) is used to design a newloop resonator with a SRR for NMR imaging. This loop resonator with a SRR is designed for NMRimaging at 3 T MRI systems. The 3D electromagnetic simulation was used to optimize the designof the proposed RF resonator and analyze it’s performance at 3 T MRI systems. The proposed RFresonator provides strong penetrating magnetic fields at the center of the human phantom model,approximately 10%, as compared to the traditional loop-type RF resonator used for NMR imagingat clinical MRI systems. We also designed an 8-channel body coil for human-body NMR imagingby using the proposed loop resonator with a SRR. This body coil also produces more homogeneousand highly penetrating magnetic fields into the human phantom model.

PACS numbers: 87.61.-c, 87.61.FfKeywords: MRI, Loop resonator, SRR, B+

1 , Human-body coil, 3 TDOI: 10.3938/jkps.68.908

I. INTRODUCTION

Two methods can be used to obtain a good signal-to-noise ratio (SNR) in magnetic resonance imaging (MRI)systems. One of the methods is to increase the staticmagnetic field (B0), as this increases the numbers of hy-drogen nuclear spins [1]. Another method is to increasethe radio-frequency (RF) magnetic fields (B+

1 ) that aregenerated at the RF coil [1]. In the latter case, a goodSNR can be easily obtained because the RF coil is ef-fectively designed to enhance the RF magnetic fields. Itis also lower in cost than the former method. Varioustypes of RF coils are used in clinics, and loop-type coilsare widely used for human-body imaging [2]. A surfacecoil has the advantage of producing RF magnetic fieldsthat penetrate into the human-body, and it can be easily

∗E-mail: ykcho@ee.knu.ac.kr

applied for MRI of various areas of the body, such as thehand, knee, and breast.

In this paper, we describe a new loop type resonatorwith a SRR for 3 T MRI systems. The properties of themetamaterials in the SRR have been studied in electro-magnetic fields for the last twenty years [3–9]. Electro-magnetic metamaterials are broadly defined as artificial,effectively homogeneous, electromagnetic structures withunusual properties not readily available in nature [10].The history of metamaterials started in 1967 with the vi-sionary speculation by the Russian physicist Viktor Vese-lago [11]. In his paper, Veselago called these substancesleft-handed media (LHM) to express the fact that theywould allow the propagation of electromagnetic waveswith the electric field, the magnetic field, and the phaseconstant vectors building a left-handed triad, as opposedto conventional materials where this triad is known to beright-handed. There was no further discussion or study

-908-

Design of a Loop Resonator with a Split-Ring-Resonator · · · – Hyeok Woo Son et al. -909-

Fig. 1. General unit SRR: (a) illustration and (b) dis-tributed inductance and capacitance.

of this material for 30 years until an artificial, effectivelyhomogeneous structure (i.e., a metamaterial) was pro-posed by Pendry et al. [3] at Imperial College in 1999.Pendry et al. introduced realizable structures with nega-tive permittivity and negative permeability. In 2001 and2002, Smith et al. introduced manufacturable LHM anddemonstrated the negative properties of the media in anexperiment [12–14]. In 2002, Caloz and Itoh proposed anew composite right/left-handed transmission line thatis now widely used in physics, optics, and microwaveengineering [15]. The most common and well-known ele-ment with negative permeability is the SRR. A SRR op-erates as a magnetic-field-driven inductor-capacitor (LC)circuit associated with the resonant currents, when themagnetic fields are excited perpendicular to the SRR.These resonant currents give to rise to a resonant mag-netic dipole moment; thus, a SRR can be characterized ashaving a resonant effective permeability, that is usuallynegative [16]. The magnetic fields that the SRR radiatesare homogeneous because of the properties of the SRR.Therefore, an RF coil with a SRR will have strong andhomogeneous magnetic fields due to the coupling effectand the properties of the metamaterial respectively. Inthis paper, the SRR is applied to a loop coil, which is awell-known RF coil used in MRI systems.

The proposed loop-type resonator with a SRR [17] isbased on transverse electromagnetic (TEM) principles tolimit the RF radiation loss and preserve the RF magneticfields’ inhomogeneity at high fields (> 3 T). The TEMcoil has the advantage of B+

1 optimization in the regionof interest (ROI), which is achieved by driving the cur-rents of the individual coil elements [18]. The proposedloop resonator with a SRR consists of distributed ele-ments and lumped elements with discrete capacitors forthe desired resonance frequencies. It is effective to varythe resonance frequency (i.e., Larmor frequency) from1.5 T to 7 T.

Unlike conventional surface coils, RF shielding is notneeded for the loop resonator with a SRR because it’sdesign is based on the TEM resonator. The proper-ties of the magnetic field B+

1 of the loop resonator witha SRR for a human-body model were determined and

Fig. 2. (Color online) Effective permeability of the unitSRR.

demonstrated via 3D electromagnetic simulations. Thestrength of the magnetic field B+

1 at the center of thehuman-body model of the proposed loop resonator witha SRR was approximately 10% higher than the mag-netic field strength of the general loop resonator withouta SRR. The proposed loop resonator with a SRR caneffectively be used to obtain high-quality NMR imageswith the 3 T MRI systems that are commonly availablein clinics and hospitals.

II. MATERIAL AND METHODS

1. Split-ring Resonator

Pendry et al. introduced a metamaterial composed ofperiodic SRRs and demonstrated the property of neg-ative effective permeability in 1999 [3]. In that paper,the negative effective permeability of the unit SRR isreported to enhance the B+

1 of the RF coil. Figure 1shows an illustration of the unit SRR, along with its dis-tributed components. The resonance of the unit SRRoccurs with distributed components such as inductorsand capacitors. The inductances L1 and L2 are definedby the inner conductor and the outer conductor, respec-tively, and a mutual inductance Lm also occurs betweenthe two conductors. The gap between the inner con-ductor and the outer conductor, which are used for theinductance component, can be represented as inductanceC1 and inductance C2, respectively. The conductance forthe adjacent conductors is the inductance C3. The com-ponents mentioned above, the unit SRR can operate anLC resonator. The resonance frequency is optimized touse distributed components, with the exception of the 15pF lumped capacitors, which are used instead of C1 andC2.

-910- Journal of the Korean Physical Society, Vol. 68, No. 7, April 2016

Fig. 3. Structure of the RF resonator: (a) general loopresonator and (b) proposed loop resonator with a SRR.

The effective permeability of a unit SRR can beobtained by using the retrieval method [19] with S-parameters, as shown in Fig. 2. We note that the unitSRR has negative permeability at the Larmor frequency(128 MHz). The resonance frequency can be easily con-trolled by adjusting the lumped capacitor in the unitSRR.

2. Loop Resonator with a SRR

The structure and the parameters of the general loopresonator and the proposed loop resonator with a SRRare shown in Fig. 3. Figure 3(b) shows the proposed loopresonator based on a SRR, which has negative perme-ability as shown in Fig. 2. Figure 3(a) depicts a generalloop resonator. For the general planar loop resonator,the low-loss dielectric material Teflon (εr = 2.08, losstan = 0.004) with a height and width of 5 mm and 90mm, respectively, is used as the substrate. A conduc-tor plate lies on the underside of the dielectric substratefor RF shielding and a grounding effect because the loopresonator is based on a microstrip line. For impedancematching at the Larmor frequency, parallel capacitorswith capacitances of 155 pF and 12.3 pF are connectedto the input and the output terminals of the general loopresonator. In the case of the proposed loop resonator,the series capacitor and parallel inductor are connectedto the input terminal. The values of the capacitance andthe inductance are 12 pF and 600 nH, respectively. Atthe output terminal, a 300pF parallel capacitor is con-nected.

Figure 4 shows the frequency response with return lossfor the 1.5 T and the 3 T MRI systems. The resonancefrequency (Larmor frequency) for the 1.5 T and the 3T systems is easily adjusted to control the lumped com-ponents in the input and the output terminals. In thisresearch, 7 T system had no resonant frequency, but webelieve that the proposed loop resonator will be able tobe used in 7 T MRI systems.

Fig. 4. (Color online) Frequency response of the proposedloop resonator (1.5 T and 3 T).

Fig. 5. (Color online) Frequency response of the proposedloop resonator according to the series capacitance in the inputterminal.

Figure 5 shows the frequency response of the proposedloop resonator with a spherical phantom (radius = 100mm, εr = 58.1, σ = 0.539 S/m) and with series capaci-tor in the input terminal. The spherical phantom is po-sitioned 10 mm above the proposed loop resonator andis used as a human model. The capacitance is 23.5 pFat the resonance frequency (128 MHz), which is differ-ent from the capacitance without a spherical phantombecause of the effect of the dielectric material. The res-onance frequency can change the frequency due to thephantom, but that can be effectively controlled by usingthe matching components, as shown in Fig. 5.

Design of a Loop Resonator with a Split-Ring-Resonator · · · – Hyeok Woo Son et al. -911-

Fig. 6. (Color online) Surface current density distributionin μA/m2 for the loop resonators: (a) general loop resonatorand (b) proposed loop resonator with a SRR.

Fig. 7. (Color online) Distribution of the penetrating mag-netic field in nA/m at the center of the spherical phantom:(a) general loop resonator and (b) proposed loop resonatorwith a SRR.

III. RESULTS FOR THE LOOPRESONATORS

Figure 6 shows the surface current density distributionof the loop resonators. The proposed loop resonator op-erates with a SRR, as shown in Fig. 6(b). We note thatthe proposed loop resonator has strong magnetic fieldsas a result of the effect of the SRR. However, the innerconductor of the SRR is not coupled well with the gen-eral loop resonator compared to the coupling betweenthe outer conductor of the SRR and the general loopresonator.

The proposed loop resonator with a SRR has magneticfields that strongly penetrate into the spherical phantom,as shown in Fig. 7. Figure 8 shows the normalized inten-sity of the penetrating magnetic field at the center of thespherical phantom. The peak values of magnetic-fieldintensity of the general loop resonator and the proposedloop resonator at the center of the phantom are 0.34 A/mand 0.375 A/m, respectively.

The proposed loop resonator has a strong peak value ofthe magnetic-field intensity at the center of the sphericalphantom, approximately 10%, as compared to the gen-eral loop resonator. The proposed loop resonator can beeffectively used for parallel imaging and localization ofmagnetic fields because of the peak value.

The radiation fields without a human object can becalculated [20], but they are not sufficient for a MRI

Fig. 8. (Color online) Normalized intensity of the pene-trating magnetic field at the center of a spherical phantom.

Fig. 9. (Color online) Human model with the loop res-onators: (a) an 8-channel body coil, and (b) axial view of arealistic body model including many different tissue types.

study. In this paper, the computational analysis (SEM-CAD X, SPEAG, Zurich [21]) was based on the finite-difference time-domain method and was used to obtainthe B+

1 fields with a human model. As seen in Fig. 9, theDuke model from the Virtual Family Models was used tosimulate the 8-channel resonators independently. Thishuman model is a realistic and heterogeneous model in-cluding 20 different tissue types (e.g., skin, blood, fat,muscle, gray matter, white matter, cerebrospinal fluidetc.). In Fig. 9(a), the rectangular box is the region ofinterest in the human model. The box size is 301 × 301× 233 mm3. The distance between the resonators is 60mm in light of the coupling. Thus, for the region of in-terest, the 8-channel is suitable in consideration of thedistance and the coupling.

Figures 10 and 11 show the B+1 field distribution of the

loop resonators on the x-y plane and x-z plane with thehuman model in Fig. 9(a). In Fig. 10, the total B+

1 fieldsof the general loop resonator and the loop resonator withSRR on x-y plane are 414 mT/W0.5 and 468 mT/W0.5,respectively. The fields of the proposed loop resonatorstrongly penetrate in wider regions, and the proposedloop resonator has a strong peak value of the B+

1 field at

-912- Journal of the Korean Physical Society, Vol. 68, No. 7, April 2016

Fig. 10. (Color online) B+1 field distribution in μT/W0.5

on the x-y plane in the human-body model: (a) general loopresonator and (b) proposed loop resonator with a SRR.

Fig. 11. (Color online) B+1 field distribution in μT/W0.5

on the x-z plane in the human-body model: (a) general loopresonator and (b) proposed loop resonator with a SRR.

the center of x-z plane as compared to the general loopresonator, as shown in Fig. 11. The peak values of thetwo types loop resonators are 0.36 μT/W0.5 and 0.414μT/W0.5, respectively. We note that the SRR increasesthe penetration of the magnetic field, enhances the B+

1

fields, and makes good homogenization.

IV. CONCLUSION

The 3 T MRI system has been used as an importantmethod for clinical diagnosis and treatment, but its SNRis lower than that exhibited by the 7 T MRI system,which is used only for research. Therefore, developmentof a RF resonator is needed for effective and strong B+

1

except for increasing static magnetic fields. The RF res-onator is also required for the multi-channel coil that isused for parallel imaging. In the case of a multi-channelcoil, the RF resonator can be made to have the high peakvalue of B+

1 for shimming and optimization by adjust-ing the phase and the amplitude of the RF resonator foreach channel. In this paper we proposed a loop resonatorwith a SRR that uses the properties of metamaterials toenhance B+

1 in 3 T MRI systems. The SRR had neg-ative permeability at the Larmor frequency (128 MHz).The proposed loop resonator had a strong peak valueof the magnetic field intensity at the center of a spher-ical phantom, approximately 10%, as compared to thegeneral loop resonator. This type of loop resonator caneasily be designed up to 7 T or more when a dielectric

material with a permittivity and low loss is used for coilswith the desired diameters according to the purpose ofthe MRI systems.

The proposed loop resonator can be effectively usedas a multi-channel coil because the peak value of B+

1 isat the center of the phantom. For shimming and opti-mization of B+

1 in the future, an 8-channel body coil willbe operated with the proposed loop resonator and thephase and the amplitude of each loop resonator will beoptimized. The 8-channel body coil with the proposedloop resonator had a more homogeneous distribution anda higher peak value of the B+

1 field compared to an 8-channel body coil with a general loop resonator. Afteroptimization, the multi-channel coil with the proposedloop resonator will be expected to provide an improvedB+

1 , and the coil should be able to be widely used forparallel imaging with parallel excitations to increase andhomogenize the B+

1 field in 3 T clinical MRI system.

ACKNOWLEDGMENTS

This study was supported by BK21 Plus funded by theMinistry of Education, Korea (21A20131600011) and inpart by KBSI No. E35600, and T35426.

REFERENCES

[1] J. Vaughan et al., Magn. Reson. Med. 46, 24 (2001).[2] G. Adriany et al., Magn. Reson. Med. 53, 434 (2005).[3] J. B. Pendry, A. J. Holden, D. J. Robbins and W. J.

Stewart, IEEE Trans. Microw. Theory Tech. 47, 2075(1999).

[4] F. Martin, F. Falcone, J. Bonache, R. Marques and M.Sorolla, IEEE Microw. Wirel. Componen. Lett. 13, 511(2003).

[5] J. Garcia-Garcia et al., IEEE Trans. Microw. TheoryTech. 53, 1997 (2005).

[6] K. Y. Liu, C. Li and F. Li, Microw. Opt. Technol. Lett.49, 1613 (2007).

[7] R. Marques, F. Mesa, J. Martel and F. Medina, IEEETrans. Antennas Propag. 51, 2572 (2003).

[8] E. Rajo-Iglesias, O. Quevedo-Teruel and M. N. MouKehn, IEEE Trans. Antennas Propag. 57, 1570 (2009).

[9] G. zamora, S. Zuffanelli, F. Paredes, F. J. Herraiz-Martinez, F. Martin and J. Bonache, IEEE AntennasWirel. Propag. Lett. 12, 1424 (2013).

[10] C. Caloz and T. Itoh, Electromagnetic metamateri-als: transmission line theory and microwave applications(Wiley & Sons Inc., New Jersey, 2006).

[11] V. Veselago, Soviet Physics Uspekhi. 10, 509 (1968).[12] D. R. Smith, W. J. Padilla, D. C. Vier, S. C. Nemat-

Nasser and S. Schultz, Phys. Rev. Lett. 84, 4184 (2000).[13] R. A. Shelby, D. R. Smith and S. Schultz, Science 292,

77 (2001).[14] R. A. Shelby, D. R. Smith, S. C. Nemat-Nasser and S.

Schultz, Appl. Phy. Lett. 78, 489 (2001).

Design of a Loop Resonator with a Split-Ring-Resonator · · · – Hyeok Woo Son et al. -913-

[15] C. Caloz and T. Itoh, IEEE T. Antenn. Propag. 52, 1159(2004).

[16] R. S. Penciu, K. Aydin, M. Kafesaki, Th. Koschny, E.Ozbay. E. N. Economou and C. M. Soukoulis, Opt. Ex-press 16, 18131 (2008).

[17] H.-J. Lee and J.-G. Yook, J. Korean Phys. Soc. 58, 826(2011).

[18] H.-W. Son, Y.-K. Cho and H. Yoo, J. Korean Phys. Soc.64, 813 (2014).

[19] X. Chen, T. M. Grzegorczyk, B. I. Wu, J. Pacheco Jr.and J. A. Kong, Phy. Rev. E 70, 016608 (2004).

[20] T. H. Hubing, J. Electromag. Engin. Sci. 13, 4 (2013).[21] SEMCAD X by SPEAG, www.speag.com.