HTS SQUID gradiometer using substrate resonators operating in an unshielded environment - a portable MCG system
ABSTRACT We have demonstrated and verified the basic feasibility of performing magnetocardiographic (MCG) measurements without magnetic shielding when using a first-order electronic gradiometer with our novel dielectric substrate resonator rf SQUIDs. The setup at the operation site involved adjustment of the gradiometer's baseline length and adaptive balancing. Our experimental portable system was tested in three environments differing in the level of electromagnetic interference.
- SourceAvailable from: fz-juelich.de[show abstract] [hide abstract]
ABSTRACT: In this paper, we analyze the influence of the superconducting quantum interference device (SQUID) gradiometer baseline on the recording of magnetocardiographic measurements. The magnetometers consist of high-temperature superconducting radio-frequency SQUIDs fabricated from YBaCuO thin films, and a substrate resonator which serves as tank circuit. The gradiometers are formed using two or three such magnetometers with individual readouts in electronic difference. We have compared the measurement results using a magnetometer and first- and second-order gradiometers with different baselines. In a standard magnetically shielded room, we found not only an increasing signal-to-noise ratio in adult magnetocardiographic measurements, but also a decreasing distortion of the magnetic field map with increasing baseline of the gradiometer. Using a first-order gradiometer with an ultralong baseline of 18 cm, we have successfully measured the heart signal of a fetus in real time.IEEE Transactions on Appiled Superconductivity 01/2004; · 1.20 Impact Factor
IEEE TRANSACTIONS ON APPLIED SUPERCONDUCTIVITY, VOL. 13, NO. 2, JUNE 2003 389
Yi Zhang, Norbert Wolters, Jürgen Schubert, Dieter Lomparski, Marko Banzet, Grigory Panaitov,
Hans-Joachim Krause, Michael Mück, and Alex I. Braginski
Abstract—We have demonstrated and verified the basic feasi-
bility of performing magnetocardiographic (MCG) measurements
diometer with our noveldielectric substrate resonator rf SQUID’s.
The setup at the operation site involved adjustment of the gra-
diometer’s baseline length and adaptive balancing. Our experi-
mental portable system was tested in three environments differing
in the level of electromagnetic interference.
Index Terms—Gradiometer, MCG, resonator, SQUID.
resonator rf SQUID’s having high sensitivity . In this work,
we report on experimental verification that the system is
well suited for performing MCG measurements in unshielded
In the past ten years, a number of HTS SQUID gradiome-
shielded environment have been demonstrated. Reported were
first and second order axial electronic rf SQUID gradiometers
with mechanical balance adjustment , , and without , as
well as dc SQUID gradiometers with planar first order gradio-
metric antennas . However, in contrast to the presently tested
concept, these instruments operated with a fixed baseline length
and, in most cases, with conventional rather than adaptive bal-
In this paper, we highlight the following new aspects of our
a) A new generation of rf SQUID magnetometers with di-
electric substrate resonators suitable for operation in un-
ECENTLY, we developed a first-order electronic gra-
diometer incorporating our novel dielectric substrate
Manuscript received August 5, 2002.
H.-J. Krause are with Institute of Thin Films and Interfaces (ISG-2), Research
Center Jülich GmbH (FZJ), D-52425 Jülich, Germany (e-mail: y.zhang@fz-
M. Mück is with Institute of Applied Physics at the University of Giessen,
A. I. Braginski was with Institute of Thin Films and Interfaces (ISG-2), Re-
search Center Jülich GmbH (FZJ), D-52425 Jülich, Germany (e-mail: abra-
Digital Object Identifier 10.1109/TASC.2003.813864
b) The optimization of the first order gradiometer baseline
length for MCG measurements in conjunction with adap-
c) The new readout electronics, integrated with micropro-
cessor remote control by a serial RS 232 connection to
a laptop computer.
d) Test MCG measurements at three sites in different coun-
new portable instrument.
II. SUBSTRATE RESONATOR CONCEPT
The first order electronic gradiometer consists of two SQUID
magnetometers. In order to achieve the required gradiometer
sensitivity, one has to use magnetometers sufficiently sensitive
over the signal frequencies of interest, down to below 1 Hz in
the case of MCG. We have already shown in  that the insta-
bility of the operating point of an rf SQUID significantly in-
noise. The rf SQUID operating point depends
not only on the value of the rf current flowing in the tank cir-
cuit, but also on the coupling between SQUID and tank circuit,
and thus on the mutual inductance
different in different environments. Nearby metal or dielectric
objects can reflect or absorb rf energy and thus change the rf
field distribution. Any motion or displacement of such objects
changes then the value of the mutual inductance
affects the SQUID’s operating point. As the environment is un-
controllable, confiningtherffielddistribution tothespace close
to the SQUID can substantially reduce the influence of the en-
, improve the stability of operating point and
thus reduce the
noise. Kornev et al. were first to show that
a dielectric resonator having high permittivity
For confining the rf field to a small spatial volume, the SrTO is
at 77 K.
We have developed a very practical concept – the single sub-
strate dielectric resonator . A standard SrTO substrate of 10
ventional tank circuit. A thin YBCO film is deposited onto it
and the flux focuser patterned. A small rf washer SQUID with
a step-edge junction is pressed against the focuser in flip-chip
geometry to attain suitable coupling. The resonance frequency
of the dielectric resonator with the YBCO film flux focuser
structure is about 650 MHz.
. The value ofcan be
can be used to
1051-8223/03$17.00 © 2003 IEEE
390IEEE TRANSACTIONS ON APPLIED SUPERCONDUCTIVITY, VOL. 13, NO. 2, JUNE 2003
flip-chip configuration, the readout loop and heater, which are all contained in
sealed fiberglass capsule.
Schematic drawing of the dielectric substrate resonator and SQUID in
In the flip-chip configuration, the planar thin film SQUID is
placed between two substrates with high , and the rf energy is
confined towithin these substrates. Because only little rf energy
is leaking out of the resonator-SQUID stack, the mutual induc-
is much less dependent on the environment than, for
example,inthe coplanar resonator configuration. Hence, the
strated in the 1970s , .
Fig. 1 shows schematically the flip-chip configuration used
and the arrangement of components inside a sealed fiberglass
capsule. A readout (coupling) loop etched in a printed circuit
board is positioned on the uncoated side of the resonator sub-
strate to couple the SQUID to the readout electronics via a 50
coaxial cable. A planar heater is placed on the SQUID chip as
well. The encapsulation not only protects the YBCO thin film
from moisture, but also reduces the low frequency noise due
to temperature fluctuations caused by bubbling nitrogen. Such
SQUID magnetometersachievea verylownoiselevel.We mea-
sured white noise levels of about 24 fT
the noise is lower than 40 and 100 fT
Hz. At 10 and 1 Hz,
III. CHOOSING THE OPTIMUM BASELINE LENGTH
The goal of practical MCG measurements is to attain the
highest possible signal-to-noise ratio (SNR). Using sensitive
SQUID magnetometers, one must also minimize the magnetic
noise (external spurious signals) from the environment. In our
approach, we adaptively adjust the balance, select thefilters and
optimize protection against the high-frequency electromagnetic
interference. In this paper, we focus on the optimization of the
gradiometer baseline length for a given signal source type, i.e.,
the heart signal of a given human subject.
Fig. 2 shows our SQUID holder. The two encapsulated
SQUID magnetometers are mounted in an axial configuration.
The bottom SQUID (close to the measured signal source) is
fixed, whereas the position of the top SQUID can be changed.
The baseline length, , is adjustable from 2 to 20 cm. The
can be selected from a series of measurements with
varying , when setting up at a given location. In the present
position of the top SQUID can be adjusted.
A photograph of the SQUID holder. The bottom SQUID is fixed, the
room temperature, while a practical clinical system could be
designed to allow adjustment from outside the cold cryostat.
tude of the measured output signal of the gradiometer, as com-
pared to the signal of the bottom magnetometer. We modeled
the magnetic activity of human heart by currents in a coil of 8.3
cm diameter, generating magnetic field signals at a distance of
14 cm to the bottom SQUID. At this distance, the peak signal
SQUID of about 50 pT. For baseline lengths of about 3, 5, 10
and 15 cm, the respective output signal of our gradiometer is
about 50, 60, 80 and 90% of the magnetometer signal, respec-
tively. Indeed, it is known that, in order to obtain a large MCG
signal, a gradiometer baseline length should not be too short.
Subsequently, to find out how to eliminate the noise from the
environment, we have performed measurements with our MCG
simulator coil in the unshielded environment of our laboratory,
settings of our gradiometer and adaptive bal-
ancing to maximize SNR. The time series filtered by a low-pass
filter to a bandwidth of 30 Hz are shown in Fig. 3(a). In the
plot of Fig. 3(b), we show that the SNR vs. baseline length is
not the expected curve with only one maximum. The dominant
noise was not the intrinsic noise of our system, but rather the
16 2/3 Hz interference from the German railway. This distur-
bance cannot easily be filtered because it lies within the major
MCG signal frequency range. The use of a notch filter would
either cause signal distortions or result in signal loss . This
spurious signal may have a large spatial phase change. There-
fore, we observed more than one maximum in the SNR as the
baseline length was varied. In our environment we obtained the
SNRwas observed for baselinelengths of 5to 6cm, which until
now were typically used in HTS MCG gradiometers –.
In most cases, the overall noise of the first order gradiometer
system is not given by the theoretical level,
is the magnetic field resolution of one magnetometer. In our
work, the noise was often found to be determined by strong
interference from either the railway or nearby electric power
is adjusted after warming up to the
ZHANG et al.: HTS SQUID GRADIOMETER USING SUBSTRATE RESONATORS 391
an MCG), recorded in unshielded environment with different baseline lengths.
The video bandwidth was about 30 Hz. (b) Signal-to-noise ratio vs. baseline
length. With a baseline length of 14 cm, we obtained the best SNR for MCG
measurements in our environment.
(a) Real time trace (source: coil driven by current source simulating
plants. Indeed, a simple noise measurement of the SQUID gra-
diometer in a quiet environment has no practical relevance. The
optimum baseline length for the best SNR depends on the mea-
sured subject and also on the environment.
IV. READOUT ELECTRONICS
Fig. 4 shows our portable SQUID gradiometer demonstrator
system for MCG feasibility testing. The readout electronics is
mounted on top of the liquid nitrogen cryostat.
The tripod holding the cryostat can be dismantled for trans-
portation. The complete system fits into a traveling suitcase
(measuring 0.75 m
0.5 m0.28 m).
The readout electronics is designed for two SQUID magne-
tometer channels and also contains the summation unit and fil-
ters. It is integrated with a microprocessor remote control via a
serial RS 232 connection toa laptopcomputer. Evennonexperts
can operate the system. The user program interface, shown in
Fig. 5, permits the operator to automatically or manually adjust
demonstrator). The tripod holding the cryostat may be dismantled for easy
transportation. The complete system fits into a travel suitcase.
A portable SQUID gradiometer system for MCG measurements (our
the SQUID’s working points (upper part), the balance (summation), as well as
on the top display. The averaged cardiac cycle is shown on the lower display.
the SQUID’s operating points (upper part of Fig. 5), the adap-
tive balance (summation), as well as the selection of the filters.
The measured real-time MCG time series is shown on the top
display. The averaged cardiac cycle is presented on the lower
display. The measured MCG signals can be saved and printed
out on paper.
V. MCG TEST MEASUREMENTS
Using the instrument described above, we performed MCG
measurements at different sites in different countries. Examples
392IEEE TRANSACTIONS ON APPLIED SUPERCONDUCTIVITY, VOL. 13, NO. 2, JUNE 2003
and China. The video bandwidth was about 30 Hz.
MCG measurements at different sites in Germany, England, Scotland,
of recorded data are shown in Fig. 6. The top time series was
recorded in our laboratory. The one immediately below it was
recorded at NPL in London, England, and the third below at
a hospital in Glasgow, Scotland. The bottom one, most recent,
was recorded in the exhibition hall of the International Trade
Center in Beijing, during a trade fair. The first three measure-
ments were made for healthy human subjects, the fourth is for
a subject with a known heart disease, a so-called atrial septal
defect. The environment in the exhibition hall in Beijing was
extremely noisy. The 50 Hz signal was about 0.8
peak), about 20 times larger than in a typical laboratory envi-
ronment. During the one-week exhibition, we did not have to
perform any readjustment of our system. These measurements
demonstrate the feasibilityof performing MCGrecordings even
in very noisy environments.
T (peak to
VI. CONCLUSION AND OUTLOOK
We have demonstrated that a first-order portable HTS gra-
diometer utilizing substrate resonator rf SQUID’s can be a suit-
able base for future development of commercial multichannel
MCG systems. Such future systems should incorporate (i) an
externally adjustable baseline length optimized during the in-
strument’s setup at the site of operation, and (ii) adaptive, auto-
mated gradiometer balancing. We have also demonstrated that
HTS rf SQUID magnetometers using dielectric substrate res-
onators may be used in an unshielded and very noisy environ-
While the technology of LTS SQUID systems for MCG is
relatively mature, that of HTS systems is not. The fabrication
technology of thin-film YBCO Josephson junctions is not suffi-
ciently well developed, and YBCO wire suitable for gradiome-
ters is not available. With electronic gradiometers it is difficult
to construct systems having a very large number of measure-
ment channels. However, we believe that fully automated HTS
rf SQUID gradiometer systems with up to 9 measurement chan-
to have such system miniaturized to demonstrate the additional
advantages of HTS SQUID systems – portability and the oper-
The authors thank L. Hao, J. Macfarlane, J. Gallop, C. Carr,
G.B. Donaldson and the Peking University for the cooperation
and the help in performing the MCG measurements.
 Y. Zhang, J. Schubert, N. Wolters, M. Banzet, W. Zander, and H.-J.
Krause, “Substrate resonator for HTS rf SQUID operation,” Physica C,
2002, to be published.
 Y. Tavrin, Y. Zhang, M. Mück, A. I. Braginski, and C. Heiden, Appl.
Phys. Lett., vol. 62, pp. 1824–6, 1993.
 J. Borgmann, P. David, G. Ockenfuss, R. Otto, J. Schubert, W. Zander,
and A. I. Braginski, Rev. Sci. Instrum., vol. 68, pp. 2730–4, 1997.
 Y. Zhang, G. Panaitov, S. G. Wang, N. Wolters, R. Otto, J. Schubert, W.
Zander, H.-J. Krause, H. Soltner, H. Bousack, and A. I. Braginski, Appl.
Phys. Lett., vol. 76, pp. 906–8, 2000.
 K. A. Kouznetsov, J. Borgmann,and J.Clarke, Appl. Phys. Lett., vol. 75,
pp. 1979–81, 1999.
 J. Vrba, “SQUID gradiometers in real environments,” in SQUID Sen-
sors, H. Weinstock, Ed.Dordrecht: Kluwer, 1996, pp. 63–116.
 V. K. Kornev, K. K. Likharev, O. V. Snigirev, Y. S. Soldatov, and V. V.
Khanin, Radio Eng. Electronic Phys., vol. 25, pp. 122–4, 1980.
 Y. Zhang, W. Zander, J. Schubert, F. Rüders, H. Soltner, M. Banzet, N.
Wolters, X. H. Zeng, and A. I. Braginski, Appl. Phys. Lett., vol. 71, pp.
IEEE, vol. 61, p. 20, 1973.
 R. Rifkin and B. S. Deaver, Phys. Rev., vol. B13, p. 3894, 1976.