Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://www.cplire.ru/html/lab234/pubs/2011_13.pdf Äàòà èçìåíåíèÿ: Sat Sep 17 17:30:52 2011 Äàòà èíäåêñèðîâàíèÿ: Tue Oct 2 08:49:12 2012 Êîäèðîâêà: Ïîèñêîâûå ñëîâà: þæíàÿ àòëàíòè÷åñêàÿ àíîìàëèÿ

Available online at www.sciencedirect.com

Physics Procedia
www.elsevier.com/locate/procedia

Physics Procedia 00 (2011) 000­000

Superconductivity Centennial Conference

Multilayer high-Tc DC SQUID magnetometers for magnetoencephalography
M.I.Faleya*, U.Poppea, R.E.Dunin-Borkowskia, M.Schiekb, F.Boersc, H.Chocholacsc, J.Dammersc, E.Eichc, N.J.Shahc, A.B.Ermakovd, V.Yu.Slobodchikovd, Yu.V.Maslennikovd, and V.P.Kosheletsd
d

PGI-5, bZEL, cINM-4, Forschungszentrum JÝlich GmbH, D-52425 JÝlich, Germany The Kotel'nikov Institute of Radio Engineering & Electronics RAS, 125009, Moscow, Russia

a

Abstract We describe tests of multilayer high-Tc DC SQUID magnetometers for magnetoencephalography (MEG) and compare our measurements with results obtained using low-Tc SQUID sensors. The integration of bias reversal readout electronics for high-Tc DC SQUID magnetometers into a commercial MEG data acquisition system is demonstrated. Results of measurements performed on a saline-filled head phantom, as well as the detection of an auditory evoked magnetic response of the human cortex elicited by a stimulus, are shown. Future modifications of high-Tc DC SQUID sensors, for applications in MEG in order to reach a resolution of 1 fT/Hz at 77.4 K over a wide frequency band, are outlined.

© 2011 Published by Elsevier Ltd. Selection and/or peer-review under responsibility of Horst Rogalla and Peter Kes.
Keywords: High-Tc superconducting heterostructures; quantum interferometers; SQUID; magnetometer; magnetoencephalography

1. Introduction Magnetoencephalography (MEG) is a non-invasive imaging modalities, providing valuable information ab superb (millisecond) temporal resolution. Some of the MEG from brain tissue originate from evoked cortical technique out neural largest ma activity an that complements other functional brain dynamics of the living human brain with gnetic fields that have been measured by d have typical amplitudes in the range of

* Corresponding author. E-mail address: m.faley@fz-juelich.de.


2

M. I. Faley et al. / Physics Procedia 00 (2011) 000­000

a few hundred femtotesla [1-3]. MEG is a technique that is complementary to electroencephalography in the sense that the two techniques have different sensitivities to different orientations of neuronal currents. MEG measurement systems presently rely on low-Tc direct current superconducting quantum interference devices (low-Tc DC SQUIDs) that are based on Nb films operate at 4.2 K and are cooled by liquid helium. Generally, SQUIDs demonstrate superior sensitivity for measuring the vector components and spatial gradients of magnetic fields, as well as an ability to resolve small changes in large signals. However, the cost of liquid helium and the inconvenience of its use for cooling the sensors are significant obstacles that prevent general acceptance of MEG systems in clinical practice. In addition, world reserves of helium are running out and prices for helium are expected to increase by up to a factor of 30 [4]. These issues highlight the need to search for magnetic field sensors that provide an alternative to the helium cooled low-Tc DC SQUIDs that are used exclusively in MEG systems. Sensors for MEG applications should have a magnetic field resolution of better than 10 fT/Hz and a size of sensor's pickup coil of below 20 mm. Direct coupled high-Tc SQUIDs [5, 6] with an intrinsic white noise that greatly exceeds 10 fT/Hz have previously been tested for the recording of MEG signals. As observed in [6], about 2000 epochs were required to achieve a signal-to-noise ratio of 3 in the measurement of an evoked MEG signal. The resulting times for MEG measurements with such sensors were unacceptably long. High-Tc DC SQUID multilayer magnetometers with 16-mm pickup coils and 12-turn input coils of high-Tc superconducting thin-film flux transformers have demonstrated magnetic field resolutions down to 4 fT/Hz at 77.4 K [7]. Such sensors offer great potential to serve as replacements for low-Tc SQUIDs in MEG systems. Such an upgrade using high-Tc SQUIDs would make MEG systems independent of helium, more user-friendly and would save about 100,000 Euros/year in the operating costs of each system. In this paper, we report on a test of the applicability of multilayer high-Tc DC SQUID magnetometers for MEG measurements in the form of a direct comparison with low-Tc DC SQUIDs. 2. Experimental results A high-Tc DC SQUID magnetometer for MEG measurements was prepared using high oxygen pressure magnetron sputtering and deep-UV (200 nm) photolithography. The sensor consisted of a DC SQUID and a 16-mm multilayer flux transformer made on a separate substrate that were combined together in a flip-chip configuration [8, 9]. A bicrystal (100) SrTiO3 (STO) substrate with a symmetric 24o in-plane misorientation angle was used for preparation of the DC SQUID. The flux transformer was prepared using YBa2Cu3O7-x (YBCO), PrBa2Cu3O7-x (PrBCO) and STO films on a single crystal 30 mm STO wafer. The multifunctional computer program "IRTECON" [10] was used for automatic measurements of the electron transport properties of the high-Tc films, Josephson junctions and DC SQUIDs. The high-Tc magnetometer was encapsulated in vacuum-tight sealed fiber-glass epoxy capsulation together with a modulation coil and a Pt heater. AC-bias SQUID control electronics and a 1.5 liter LN2 cryostat, both from Cryoton Ltd, were used for operation of the high-Tc magnetometer. Both high-Tc and low-Tc SQUID systems had magnetic field resolutions of 5 fT/Hz when operating at 77.4 K and 4.2 K, respectively. Operation of the DC SQUID control electronics in bias reversal mode led to an approximately 3-fold reduction in the intrinsic low-frequency noise originating from fluctuations of critical currents of the Josephson junctions in the high-Tc DC SQUIDs. Spectral density of the background signal of the high-Tc system measured with the high-Tc DC SQUID magnetometer placed in superconducting shield at 77.4 K is shown in Fig.1a. The distance between the pickup coil of the signal magnetometer and the bottom of the cryostat was 19 mm in the low-Tc system and 15 mm in the high-Tc system. Measurements with a saline-filled head phantom and MEG measurements were performed in a magnetically shielded room using both a one-channel high-Tc DC SQUID measurement system and a


M. I. Faley et al. / Physics Procedia 00 (2011) 000­000

3

commercial 248-channel MEG system (9-mm magnetometers; "Magnes© 3600 WH" 4D-Neuroimaging). The analog output of the high-Tc system was connected to a 16-bit analog-to-digital converter port available on the low-Tc system and processed together with the signals from the low-Tc magnetometers. Photograph taken during the test measurement with the head phantom by the high-Tc system is shown in Fig. 1b.

a.

b.

Fig. 1. (a) Noise spectral density of the 16-mm high-Tc magnetometer; (b) photograph taken during the test measurement with head phantom by the high-Tc system.

In Fig. 2a, blue lines show the results of measurements obtained from the head phantom from one channel of the low-Tc measurement system, while red lines in Fig.2b show data from the high-Tc system measured at the same point above the head phantom surface.

a.

b.

Fig. 2. Measurements of fields obtained from a saline-filled head phantom using: (a) a low-Tc SQUID and (b) a high-Tc SQUID.

The excitations of the head phantom current dipole were the same for the measurements made with both systems and reduced by a factor of 2 at each step from top to bottom. The peak-to-peak magnetic fields measured by the low-Tc SQUID were 3.2 pT, 1.6 pT, 800 fT, 400 fT and 200 fT, respectively. In order to obtain comparable results, no noise compensation (weights or reference channels) were applied to the data obtained by low-Tc and high-Tc systems. Both the high-Tc and the low-Tc data are therefore


4

M. I. Faley et al. / Physics Procedia 00 (2011) 000­000

purely magnetometer data. Both data sets show averages over 100 epochs and were band-pass filtered from 3 to 30 Hz.

a.

b.

Fig. 3. (a) Photograph taken during the measurement of an auditory evoked field; (b) Results of measurements of an auditory evoked field obtained using a high-Tc SQUID sensor (red and green curves) and low-Tc SQUID sensors (blue curves).

A photograph taken during the measurement of an auditory evoked response using the high-Tc measurement system is shown in Fig.3a. Fig.3b shows a comparison of data acquired on the same day using the high-Tc SQUID (red and green curves) and two of the low-Tc channels (blue curves). The red curve shows a high-Tc curve measurement recorded at the first position of the SQUID. The green curve shows a result obtained after moving the recording position by a few centimeters, near to the position of the low-Tc sensor during the measurement of the upper blue curve. All of the data sets are acquired with a sample rate of 678 Hz and a bandwidth 200 Hz, and averaged over 300 epochs.

3. Discussion The measurements shown in Fig. 2 and Fig.3 demonstrate sufficient sensitivit y of the 16-mm high-Tc magnetometers operating at 77.4 K for MEG measurement of auditory evoked magnetic fields with a relatively small number of epochs and a short acquisition time. The high-Tc magnetometers are significantly smaller than the 21-mm magnetometers used in commercial low-Tc (4.2 K) MEG systems from Elekta Neuromag©, while achieving [7] magnetic field resolution similar to one of low-Tc system. Fewer radiation shields required in LN2 cryostats produce less magnetic field noise, which is often dominant over the intrinsic noise of low-Tc SQUID sensors [11]. Thermal noise of magnetically shielded room [11] can also increase background noise of the measuring systems additionall y reducing difference between their magnetic field resolutions. Thinner thermal insulation in LN2 cryostats allow high-Tc sensors to be placed closer to the object additionally increasing signal-to-noise ratio of the MEG measurements. Nevertheless, it is desirable to improve the magnetic field resolution of high-Tc sensors further without significant degradation of their spatial resolution and cross-talk of the sensors in multichannel systems. Magnetic field resolution of the inductively coupled magnetometers with multiturn input coil can be estimated from the following expression:


M. I. Faley et al. / Physics Procedia 00 (2011) 000­000

5

BN =

L pu + Li kA
pu

Li LS

S

1/ 2

,

(1)

where the white flux noise of the SQUIDs S is determined mainly by the thermal fluctuations in the Josephson junctions, by the maximum voltage response to the magnetic flux V/ and by the noise of the preamplifier of the control electronics SVe 0.2 nV:
2 V 12k BT S = SV / RN 2 V R N L2S V 2 + + SVe / 4 2 2

.

(2)

The white noise of the sensors ma y be reduced further, for example by the implementation of serial SQUID arrays, by a factor of N where N is the number of SQUIDs in the array. For sufficiently large values of N, the magnetic field resolution of the high-Tc DC SQUID magnetometers, using sufficiently large input coils, can potentially reach values of below 1 fT/Hz at 77.4 K. Our estimations according the equations (1) and (2) shows field resolution 1 fT/Hz at 77.4 K already for serial connection of two DC SQUIDs (N = 2), diameter of pickup loop of multila yer flux transformer 30 mm, and coupled inductances of the DC SQUID loops from about 40 pH up to about 80 pH (see Fig.4).

Fig. 4. Expected magnetic field resolution BN(LS) of high-Tc DC SQUID magnetometer consisted of a dual-SQUID inductively coupled to 30-mm superconducting flux transformer (see insert) having two multiturn input coils.


6

M. I. Faley et al. / Physics Procedia 00 (2011) 000­000

The serial connection of two DC SQUIDs in a dual-SQUID geometry is the first step in the application of high-Tc DC SQUID arrays [12]. By testing a dual-SQUID circuit separately, we have observed a doubling of SQUID voltage swings and a reduction in noise, when compared with a single SQUID sensor used with a similar SQUID washer and parameters of the Josephson junctions. Two washers in the dualSQUID system can be inductively coupled to two multiturn input coils of the large area multilayer flux transformer (see insert in Fig.4), providing improved sensitivity of the sensor to the magnetic field to be measured [8]. We expect that the application of a modulation signal to the directly coupled loop of the dual-SQUID magnetometer will result in lower noise in the sensor and less cross-talk between the sensors, when compared to the application of the modulation signal to the pick-up coil of a multilayer flux transformer. Further experiments with multichannel high-Tc MEG systems are required to prove if sufficiently low cross-talk between the sensors could be realized for the case of relatively large pickup loops of high-Tc DC SQUID sensors.

Acknowledgements The authors gratefully acknowledge the technical assistance of R. Speen.

References
[1] Fagaly RL. Superconducting quantum interference device instruments and applications Rev Sci Instr 2006;77:101101. Fagaly R.L. 1990 "Neuromagnetic instrumentation", In: Sato S, editor. In: "Magnetoencephalography. Advances in n eurology", New York: Raven Press, 11-32. [2] The SQUID Handbook Vol.2: Applications of SQUIDs and SQUID systems,: Clarke J, Braginski AI editors. Weinheim: WILEY-VCH Verlag GmbH&Co. KgaA; 2006 ISBN 3-52740408-2. [3] Magnetism in medicine. A handbook AndrÄ W, Novak H editors. Weinheim: WILEY-VCH Verlag Gmb H&Co. KgaA; 2007 ISBN 978-3-527-40558-9. [4] Witchalls C. One minute with... Robert Richardson. The New Scientist. 2010;207, Issue 2773, 14 August, Page 29; Witchalls C. Nobel prizewinner: We are running out of helium. New Scientist. 18 August 2010. [5] Zhang Y, Tavrin Y, MÝck M, Braginski AI, Heid en C, Hampson S et al. Magnetoencephalography Using High Temperature rf SQUIDs. Brain Topography, 1993;5(4):379-382. [6] Barthelmess H, Halversch eid M, SchiefenhÆvel B, Heim E, Schilling M, Zimmerman R. HTS-Multichannel-System for Magnetocardiography and Magnetoencephalography. BIOMAG 2000 Conference 2001;927­930. [7] Faley MI, Jia CL, Poppe U, Houben L, and Urban K. Meanderin g of th e grain boundary and d-wave effects in high-Tc bicrystal Josephson junctions. Superconductor Science and Technology 2006;19:S195-S199. [8] Faley MI. Epitaxial oxide heterostructures for ultimate high-Tc quantum interferometers. In: Adir Luiz editor. Applications of High-Tc Superconductivity, Rijeka: InTech; 2011, p. 147-176. ISBN 978-953-307-308-8. [9] Faley MI, Poppe U, Urban K, Paulson DN, Starr T, Fagaly RL. Low noise HTS dc-SQUID flip-chip magnetometers and gradiometers. IEEE Trans. Appl. Supercond 2001;11(1):1383-1386. [10] The PC program "IRTECON" was developed by A. B. Ermakov (ermakov@hitech.cplire.ru). [11] Nenon en J, Montonen J, and Katila T. Thermal noise in biomagnetic measurements, Rev.Sci.Instrum., 1996; 67(6): 23972 405. [12] Chen KL, Yang HC, Ko PC, Horng HE. Characterization of dual high transition temperature superconducting quantum interference device first-order planar gradiometers on a chip, J Appl Phys, 2010;108:064503(4).