This document summarizes a research paper that proposes a multiple-input multiple-output (MIMO) dielectric resonator antenna (DRA) design with improved isolation for 5G millimeter-wave applications. The design uses two rectangular DRs excited by microstrip slots underneath, with a metal strip printed on each DR to move the strongest part of the coupling field away from the slot, improving isolation between elements. Simulations show the design achieves over 12 dB improved isolation across 27.5-28.35 GHz while maintaining impedance bandwidth. A prototype was fabricated and measured, validating over 24 dB isolation in the 28 GHz band.
2. 748 IEEE ANTENNAS AND WIRELESS PROPAGATION LETTERS, VOL. 18, NO. 4, APRIL 2019
Fig. 1. Geometry structure of the proposed MIMO DRA. (a) Exploded 3-D
view. (b) Top view.
TABLE I
GEOMETRIC PARAMETERS OF THE MIMO DRA. UNITS: mm
A. MIMO DRA Design
Two identical DRs with the dimension a × b × d are mounted
on the metal ground plane with the dimension of 20 mm ×
20 mm. The TEy
311 mode is excited by a microstrip-fed rectan-
gular slot. The resonant frequency of a rectangular DR for the
TEy
pqr mode can be derived from the wavenumbers kx, ky , and
kz inside the DR, and the subscripts p, q, and r are numbers of
the standing waves along the x-axis, y-axis, and z-axis, respec-
tively. The wavenumbers can be calculated using the Marcatili’s
approximation method [12]. The dimension of the DR can be
set to a = 9.5 mm, b = 7.5 mm, and d = 2.54 mm, which
makes the DR resonate at 28 GHz for the TEy
311 mode cal-
culated using the Marcatili’s approximation method mentioned
above.
Fig. 2. Simulated S-parameters of the DRA with and without strips.
Fig. 3. Amplitude distribution of the coupling field (only Port 1 is excited).
(a) DRs are without metal strips. (b) DRs are with metal strips.
The MIMO DRA without metal strips is simulated by us-
ing the Ansys HFSS software. The simulated S-parameters are
shown in Fig. 2. It can be seen that the S11 is better than –10 dB
over 27.00–28.83 GHz, and that the S12 is around –17 dB in a
28 GHz band. The isolation should be improved further to meet
the requirement of a higher diversity gain.
B. High-Isolation Design
A metal strip is introduced and printed on the upper surface of
each DR to improve the isolation between two antenna elements.
To investigate why the metal strips can significantly improve the
isolation, the distribution of the coupling electrical field above
the exciting slot of DR 2 with and without the metal strips are
simulated and compared in Fig. 3 when only Port 1 is fed.
It can be observed in Fig. 3 that the strongest part of the cou-
pling field is directly above the microstrip-fed rectangular slot.
Therefore, the energy of the coupling field can be transmitted to
Port 2 via the slot resulting in strong coupling.
3. ZHANG et al.: MIMO DIELECTRIC RESONATOR ANTENNA WITH IMPROVED ISOLATION FOR 5G mm-WAVE APPLICATIONS 749
Fig. 4. Equivalent circuit model of the proposed MIMO DRA. (a) DRs are
without metal strips. (b) DRs are with metal strips.
The S-parameters of antennas with and without metal strips
are shown in Fig. 2. It can be seen that the MIMO DRA
with metal strips has a minimum improvement of 6 dB and
a maximum improvement of 12 dB on the isolation over 27.5–
28.35 GHz, and the metal strips do not observably influence the
impedance matching.
Fig. 3(b) shows the distribution of the coupling field in DR 2
with a metal strip printed on its upper surface. The figure shows
that the strongest part of the coupling field is no more directly
above the slot and moves to the position directly under the metal
strip, and the isolation between two elements is improved in this
way.
The reason why the metal strips can move the strongest part
of the coupling field is that the combination of the metal ground
plane and the metal strip can be considered as a plate capacitor
storing most energy of the coupling field between two elements.
The equivalent circuit of the proposed MIMO DRA is shown
in Fig. 4 to make a better understanding for the decoupling
mechanism. A series RLC resonator is adopted to model the DR.
Z0 is the characteristic impedance of the microstrip feedline.
Transformers T1, T2, and T3 represent the coupling between
microstrip feedline and the slot, the coupling between the slot
and the DR, and the mutual coupling between antenna elements,
respectively. Cp is the capacitor formed by decoupling metal
strip and metal ground. When Port 1 is excited and Port 2 is
terminated by a matched load, the introduction of capacitor Cp
directly reduces the voltages across the transformers T3 and T2,
and then reduces the voltage across the transformer T1 near Port
2, making the voltage across Port 2 lower than that without Cp.
Thus, the isolation between two ports is improved.
III. RESULTS AND DISCUSSIONS
A. Impedance Performance
A prototype of the proposed MIMO DRA is fabricated by us-
ing a printed circuit board manufacturing process. The antenna
was measured using the Agilent PNA E8363C vector network
analyzer. The S-parameters are shown in Fig. 5. The measured
reflection coefficient of the proposed MIMO DRA with metal
strips is better than –10 dB over 27.19–28.48 GHz, which covers
the 28 GHz band for the future 5G applications. The measured
isolation is better than 24 dB in a 28 GHz band. It can be
verified that the isolation has been improved by introducing the
metal strips. The deviation of the scattering coefficients between
simulation and measurement is mainly due to the connector sol-
dering and practical dielectric constant of the medium varying
with frequency.
Fig. 5. Simulated and measured S-parameters of the proposed MIMO DRA
with and without metal strips.
Fig. 6. Radiation patterns of the proposed MIMO DRA with and without
strips.
Fig. 7. Measured and simulated radiation pattern of the proposed MIMO
DRA.
B. Radiation Performance
Metal strips do not significantly affect the radiation pattern.
This is due to the location of metal strips where the electric field
of DR is weak. The radiation patterns with and without metal
strips are shown in Fig. 6, which shows there is just a slight
influence of metal strips on radiation patterns.
The simulated and measured radiation patterns of MIMO
DRA with metal strips at 28 GHz are shown in Fig. 7. The
realized gain of the proposed MIMO DRA is shown in Fig. 8.
The discrepancy between simulated and measured results is due
to the connectors whose size is comparable to that of the DRA.
The testing cable also affects the measured radiation pattern.
C. Diversity Performance
The envelope correlation coefficient (ECC) is a critical fig-
ure in the MIMO communication system. For a two-element
4. 750 IEEE ANTENNAS AND WIRELESS PROPAGATION LETTERS, VOL. 18, NO. 4, APRIL 2019
Fig. 8. Realized gain and ECC of the proposed MIMO DRA.
Fig. 9. Channel capacity of the proposed MIMO DRA.
MIMO antenna, the ECC can be calculated according to radia-
tion patterns. The ECC of the proposed MIMO DRA is lower
than 0.013 in a 28 GHz band as shown in Fig. 8, which con-
tributes to a large channel capacity and diversity gain of the
MIMO communication system.
The diversity gain is another critical parameter for evaluating
the MIMO system performance. It can be obtained by [13]
DG = 10 1 − ECC2 (1)
where 10 is the maximum apparent diversity gain at the 1%
probability level for selection combining. Since ECC of the
proposed MIMO DRA is less than 0.013, the diversity gain is
more than 9.9 dB.
The channel capacity can be calculated by the following equa-
tion [14]:
C = log2 I − SH
R SR I − SH
T ST γ /2 − 1.18 (2)
where { }H
denotes the Hermitian transposition, γ is the
signal-to-noise ratio at the receiving antenna, I is an identity ma-
trix, and SR and SH are S-matrices of transmitting and receiving
antennas, respectively. The channel capacity of the proposed
MIMO DRA is shown in Fig. 9.
Fig. 10. TARC of the proposed MIMO DRA with θ varying.
TABLE II
COMPARISON WITH OTHER WORKS
A total active reflection coefficient (TARC) is presented for
multiport systems [15], which takes into consideration the effect
when ports of the multiple-antenna system are fed with signals
of different phases and can be obtained by
Γt
a = |S11 + S12ejθ |
2
+ |S21 + S22ejθ |
2
/2 (3)
where θ is the phase difference between two feeding ports. The
TARC of the proposed MIMO DRA is shown in Fig. 10. It
can be noticed that the TARC always covers a 28 GHz band in
the variation of the θ and follows the original behavior of the
antenna characteristics.
The comparisons of the proposed MIMO DRA with other
related works are listed in Table II. It can be seen that the de-
coupling structure of the proposed MIMO DRA has the small-
est size and that a lower ECC and a higher isolation are ob-
tained.
IV. CONCLUSION
In this letter, a MIMO DRA with enhanced isolation for the
future 5G mm-wave applications is designed, analyzed, and
measured. The isolation of the proposed antenna is improved
to 24 dB by introducing a metal strip printed on the upper
surface of each DR moving the strongest part of the coupling
field away from the exciting slot. A maximum improvement of
12 dB on the isolation over 27.5–28.35 GHz is achieved in this
way. The MIMO DRA is simulated, fabricated, and measured.
The simulated and measured results are with a good agreement,
showing the validity of the proposed decoupling technology.
5. ZHANG et al.: MIMO DIELECTRIC RESONATOR ANTENNA WITH IMPROVED ISOLATION FOR 5G mm-WAVE APPLICATIONS 751
REFERENCES
[1] J. Kornprobst, K. Wang, G. Hamberger, and T. F. Eibert, “A mm-wave
patch antenna with broad bandwidth and a wide angular range,” IEEE
Trans. Antennas Propag., vol. 65, no. 8, pp. 4293–4298, Aug. 2017.
[2] C. Mao, S. Gao, and Y. Wang, “Broadband high-gain beam-scanning
antenna array for millimeter-wave applications,” IEEE Trans. Antennas
Propag., vol. 65, no. 9, pp. 4864–4868, Sep. 2017.
[3] S. Long, M. McAllister, and S. Liang, “The resonant cylindrical dielec-
tric cavity antenna,” IEEE Trans. Antennas Propag., vol. AP-31, no. 3,
pp. 406–412, May 1983.
[4] A. J. Paulraj, D. A. Gore, R. U. Nabar, and H. Bolcskei, “An overview of
MIMO communications - A key to gigabit wireless,” Proc. IEEE, vol. 92,
no. 2, pp. 198–218, Feb. 2004.
[5] J. Yan and J. T. Bernhard, “Design of a MIMO dielectric resonator antenna
for LTE femtocell base stations,” IEEE Trans. Antennas Propag., vol. 60,
no. 2, pp. 438–444, Feb. 2012.
[6] L. Zou, D. Abbott, and C. Fumeaux, “Omnidirectional cylindrical dielec-
tric resonator antenna with dual polarization,” IEEE Antennas Wireless
Propag. Lett., vol. 11, pp. 515–518, 2012.
[7] H. Yan, L. Taolin, D. Xiangyu, and W. Weiwei, “The design of a tripo-
larization rectangle dielectric resonator antenna,” in Proc. 10th Eur. Conf.
Antennas Propag., 2016, pp. 1–3.
[8] A. Abdalrazik, A. S. A. El-Hameed, and A. B. Abdel-Rahman, “A three-
port MIMO dielectric resonator antenna using decoupled modes,” IEEE
Antennas Wireless Propag. Lett., vol. 16, pp. 3104–3107, 2017.
[9] M. S. Sharawi, S. K. Podilchak, M. U. Khan, and Y. M. Antar, “Dual-
frequency DRA-based MIMO antenna system for wireless access points,”
Microw., Antennas Propag., vol. 11, no. 8, pp. 1174–1182, 2017.
[10] A. Dadgarpour, B. Zarghooni, B. S. Virdee, T. A. Denidni, and A. A.
Kishk, “Mutual coupling reduction in dielectric resonator antennas using
metasurface shield for 60 GHz MIMO systems,” IEEE Antennas Wireless
Propag. Lett., vol. 16, pp. 477–480, 2017.
[11] R. Karimian, A. Kesavan, M. Nedil, and T. A. Denidni, “Low-mutual-
coupling 60 GHz MIMO antenna system with frequency selective sur-
face wall,” IEEE Antennas Wireless Propag. Lett., vol. 16, pp. 373–376,
2017.
[12] R. K. Mongia, “Theoretical and experimental resonant frequencies of rect-
angular dielectric resonators,” IEE Proc. H, – Microw. Antennas Propag.,
vol. 139, no. 1, pp. 98–104, 1992.
[13] Y. Gao, X. Chen, Z. Ying, and C. Parini, “Design and performance in-
vestigation of a dual-element PIFA array at 2.5 GHz for MIMO termi-
nal,” IEEE Trans. Antennas Propag., vol. 55, no. 12, pp. 3433–3441,
Dec. 2007.
[14] N. Honma, H. Sato, K. Ogawa, and Y. Tsunekawa, “Accuracy of MIMO
Channel capacity equation based only on S-parameters of MIMO an-
tenna,” IEEE Antennas Wireless Propag. Lett., vol. 14, pp. 1250–1253,
2015.
[15] M. Manteghi and Y. Rahmat-Samii, “Multiport characteristics of a wide-
band cavity backed annular patch antenna for multipolarization opera-
tions,” IEEE Trans. Antennas Propag., vol. 53, no. 1, pp. 466–474, Jan.
2005.