MgB2 SQUID for Magnetocardiography

The discovery of the superconductivity at a transition temperature (TC) of 39 K in magnesium diboride (MgB2) has attracted much attention from many researchers for scientific as well as technical reasons [1]. Compared with Cu-based superconductors (cuprates), MgB2 has lower anisotropy and larger coherence length, in addition to high Tc [2]. These characteristics of MgB2 give rise to new applications for superconductor devices that can operate in the temperature range between 20 and 30 K; examples of such devices are Josephson junctions and integrated circuits. This temperature range can be easily achieved by using economical and compact cryocoolers or liquid hydrogen. The use of cryocoolers may transform the superconductor from being specialized and advanced technology into common usage in consumer devices. In the future, hydrogen gas may be widely used for carbon-free power generation such as in fuel cells. Liquid hydrogen would be available for these purposes, and may be utilized for the cooling of low-temperature devices. In addition, MgB2 is considered to be a clean superconducting material, using neither toxic nor rare earth elements.


Introduction
The discovery of the superconductivity at a transition temperature (T C ) of 39 K in magnesium diboride (MgB 2 ) has attracted much attention from many researchers for scientific as well as technical reasons [1]. Compared with Cu-based superconductors (cuprates), MgB 2 has lower anisotropy and larger coherence length, in addition to high T c [2]. These characteristics of MgB 2 give rise to new applications for superconductor devices that can operate in the temperature range between 20 and 30 K; examples of such devices are Josephson junctions and integrated circuits. This temperature range can be easily achieved by using economical and compact cryocoolers or liquid hydrogen. The use of cryocoolers may transform the superconductor from being specialized and advanced technology into common usage in consumer devices. In the future, hydrogen gas may be widely used for carbon-free power generation such as in fuel cells. Liquid hydrogen would be available for these purposes, and may be utilized for the cooling of low-temperature devices. In addition, MgB 2 is considered to be a clean superconducting material, using neither toxic nor rare earth elements.
Thus far, we have developed SQUID (Superconducting Quantum Interference Device) apparatuses, such as magnetocardiography (MCG) and non-destructive evaluation (NDE) systems, by using Nb and cuprates. Since the discovery of superconductivity in MgB 2 and its promising potential, we have focused our research on developing SQUID as well as a highfrequency filter made of MgB 2 . As a first step, we developed a synthesis method for highquality films, which is the basis of fabrication of such devices. Next, since 2004, we have been working on developing a SQUID and a high-frequency filter device. Finally, we succeeded in measuring the magnetic signal from a human being (MCG signal) by using MgB 2 SQUID with a specially developed control circuit (digital FLL). In this article, we discuss our achievements and progress on the development of superconducting devices using MgB 2 films.
In contrast to HPCVD, the MBE method adopts a low synthesis temperature of about 573 K and is carried out in a high vacuum. The T C of MgB 2 films grown by the MBE method is relatively lower than that obtained from other methods and is limited to be around 34 K. The growth conditions for the MBE method adopted by the research groups other than our group are, however, high growth rate and high substrate temperature. The high growth rate compensates for the deficiency of Mg through re-evaporation due to its volatility and prevents the oxidation of Mg. This concept is very similar to that of the HPCVD method. High deposition rate and high growth temperature have been a mainstay in the fabrication of the high-quality MgB 2 films. We adopted a low substrate temperature, a low deposition rate, and an ultra-high vacuum to obtain high-quality MgB 2 films. The characteristic feature of the MBE method is lowtemperature synthesis. According to Liu et al., MgB 2 is synthesized in a wider temperature range, particularly at low temperatures, as the vacuum increases [10]. The growth at low temperature reduces the re-evaporation of Mg. The deposition rate and the supply of Mg and B have to be reduced due to the super-saturation of Mg. Under these conditions, Mg may be oxidized. However, the ultra-high vacuum in an MBE apparatus prevents the oxidation and permits the growth of high-quality films.
In terms of synthesis conditions, our method is conceptually opposite to the conventional fabrication methods for MgB 2 films. Figure 1 shows the typical temperature dependence of the resistivity curves of MgB 2 films fabricated by our method. The growth temperature was set as 473 K, which is about 100 K lower than that in other reports. In our method, the range of the deposition temperature was between 373 K and 523 K. This is the same range as that in the phase-diagram of Lie et al. and may be obtained by extrapolating the pressure range to 10 -7 Pa and 10 -8 Pa. The Tc of the sample in Fig. 1 was 32 K, and the transition width is smaller than 1.0 K. In our method, T c can reach a value of up to 37 K by the adoption of various additional process improvements, which will be presented later.  The measurement was carried out using patterned micro-bridges, which were made by standard photo-lithography and the Ar ion beam milling method. The inset of Fig. 2 is an optical microscope image of the MgB 2 micro-bridge. The dimensions of the micro-bridge were 10 m (width)  30 m (length). The electrical contact was made by gold wire directly bonded to a Cu/MgB 2 contact pad using silver paste. The highest value of J c is 11.6 MA/cm 2 at 4.2 K on an MgO (100) substrate. The J C (0) value is 12.5 MA/cm 2 . This value is considered to be very high for MgB 2 films. We will discuss the reason why the J C of our sample synthesized on an MgO substrate is so high. From AFM and SEM measurements, the grain size of MgB 2 films on MgO substrates is 20-50 nm. This value is smaller than that on an Al 2 O 3 or Si substrate. The mismatch between the in-plane lattice constant of MgB 2 and MgO is 26.7 %. MgO (100) plane has a square surface structure, whereas MgB 2 has a hexagonal surface structure. These conditions led to a smaller grain size of MgB 2 when it is deposited on MgO. The grain shape of MgB 2 is of the columnartype, and the grain boundaries work as effective pinning centers. Therefore, the smaller grain size provides more pinning centers and leads to high values of J C in the perpendicular field.
Next, we tried to further improve the quality of the grown MgB 2 films such that they can be used for the fabrication of superconducting devices. According to the works of Shimane University [11] and NICT [12], the in-plane lattice matching between MgB 2 and the substrate plays an important role in the growth of high-quality MgB 2 films using the MBE method.
The structural quality of MgB 2 films was improved by using AlN ( ∆d = 1.9%) and TiZr ( ∆d = 3.6%) buffer layers having lattice constants close to that of MgB 2 . We initially employed ZnO (0001) as the substrate as its substrate closely matched that of MgB 2 . The in-plane lattice spacing of the hexagonal ZnO lattice (a = 0.3522 nm) is close to that of the MgB 2 lattice (∆d = 5.4%; ∆d of Al 2 O 3 , for example, is 35.2 %) [13]. However, we could not do an in-plane alignment for these films, although their T c is relatively high at 35 K. The cross-sectional transmission electron microscope (TEM) image showed a large amount of reaction products near the MgB 2 /ZnO interface. In addition, intermixing between Zn and MgO was observed near the MgB 2 /ZnO boundary, which is ascribed to the free energy difference between the substrate and the film. The free energy of MgO is -547.1 kJ/mol and is lower than the value of -303.3 kJ/mol for ZnO. Therefore, Zn in ZnO is easily replaced by Mg, and MgO is formed.
The best-quality MgB 2 film was obtained by adopting titanium (Ti) film as a buffer layer on a ZnO substrate, which was fabricated by evaporation using an electron-beam gun cell in an MBE chamber [14]. Figure 3 shows the superconducting properties of MgB 2 films when a 15 nm Ti buffer layer is adopted. When the MgB 2 film was deposited on a ZnO (0001) substrate, the T C increased from 33 K to 36K with the insertion of a Ti buffer layer. By using in situ reflection high-energy electron diffraction (RHEED), Ti buffer layers were grown epitaxially   We achieved a value of T C = 37 K by the MBE method, which is the same as that with HPCVD. Fujiyoshi et al. report that the J C of the MgB 2 films on the Ti buffer layer is higher than that of MgB 2 on the substrates [16]. Therefore, the pinning force due to the grain boundaries at the MgB 2 film on the Ti buffer layer is stronger than that of other films.

Implementation of Josephson junction and SQUID by lithography technique
There are three types of Josephson junction fabrication methods for SQUID and they are as follows: SIS tunneling junctions, nano-bridge junctions and SNS junctions. The first tunneling MgB 2 junctions were fabricated by the NTT group [17]. We have developed and used an SIS junction [18]. Subsequently, both the Twente University and European groups succeeded in the fabrication of nano-bridge MgB 2 junctions [19]. We have developed nano-bridge MgB 2 SQUIDs devices that are optimized for MCG measurement, which are based on the design of the Twente University group. Nano-bridge junctions were made using focused ion beam (FIB) milling apparatus. Detailed fabrication conditions are reported elsewhere [20].    We encountered two problems during the development of SQUID. One is the reduction of the yield ratio, due to the small size of the nano-bridge, which is close to the size limitation of the FIB fabrication process. Another is the hysteretic behavior of the I-V curve at the transition edge, which appears at low temperatures. This causes instability in the bias current and makes measurements very difficult. The first problem was solved by adopting thinner films. We are able to obtain I-V and -V properties and even expand the width of the nano-bridge by decreasing film thickness. At present, we retain a high yield of about 90 % for the fabrication of 65 nm thick film and the 400 nm wide nano-bridges. The second problem was resolved by placing a resistor in parallel to the MgB 2 nano-bridges. We were able to eliminate the effect of hysteresis by inserting the resistor.

SQUID electronics
The input-output (-V) characteristic of SQUID is nonlinear as shown in Fig. 8. A sinusoidal output voltage with the appropriate periodicity is generated, when a magnetic flux is applied in SQUID. A control circuit, known as the FLL (flux locked loop) circuit, is necessary to linearize the input-output characteristic so that the SQUID can be utilized as a magnetic sensor. An example of such an FLL circuit is shown in Fig. 9. The FLL circuit provides the feedback necessary to fix the output at the defined value (the lock point) in a sinusoidal SQUID output. Then, by using only the small area in the vicinity of the lock point, we get the desired linear output. The external magnetic field applied to a pick-up coil produces the voltage in the SQUID. The EMF-voltage is amplified, integrated and then fed back to SQUID in order to maintain to the lock point. Then the applied magnetic field is nullified. This integrated value becomes the output signal of the FLL circuit, and it is equal to the input magnetic signal applied to the pick-up coil. In recent years, it has been shown by several groups that the electronics within the digital SQUID, where the flux-locked loop system is controlled by a digital circuit, can achieve a wide dynamic range [21][22][23]. These FLL systems can measure biomagnetic signals in a magnetically unshielded environment. We call a digitally controlled flux-locked loop system (D-FLL).
A D-FLL system satisfies unshielded conditions: its dynamic range is expanded by jumping to a lock point on the SQUID's periodic -V characteristic and by keeping a count of the number of such events. This method was named flux-quanta counting (FQC) [24,25]. The

www.intechopen.com
integrator of the D-FLL is reset every time the feedback flux exceeds ±1 0 and the corresponding  0 steps are counted. Usually, the original input flux is reconstructed on a digital signal processor that is a part of the FLL circuit. The dynamic range depends on the architecture of the digital signal processor, which is typically 32-bit. Furthermore, highspeed signal processing makes a digital feedback loop delay shorter. A high-performance digital signal processor is required to increase the dynamic range and slew rate. So far, it has p r o v e n d i f f i c u l t t o m a k e a h i g h p e r f o r m a nce system with the typical single-chip microcontroller that is generally used in such circuits.
We have developed a double-counter equipped D-FLL system in Fig. 10. The doublecounter method has two counters on a single-chip microcontroller and a host computer. The digital integrator unit calculates feedback flux data. If feedback flux exceeds ±0.5 0 , the integrator unit subtracts 1  0 from the digital integrator data or adds 1  0 to it, instead of resetting the FLL. The feedback loop process works more quickly than the data transmission from the one-chip microcontroller into a host computer. The integrator data and Couner1 data is sent at periodic time intervals defined by a timer. Counter1 is reset every time after this transmission is complete and commences the count of  0 steps during the next interval. Counter2 integrates the data transmitted from Counter1. The host computer reconstructs the measured flux data using Counter2 as well as integrator data. Because a host computer reconstructs input flux, the workload of the single-chip microcontroller is decreased, leading to a quicker operation of the feedback loop than that of the system without a double-counter method. The high-performance D-FLL system using a single-chip microcontroller is achieved by a method that employs two counters. Our latest D-FLL system is composed of a commonly used 8-bit single-chip microcontroller, a 12-bit A/D converter, two 16-bit D/A converters and some other inexpensive and easily obtained devices [26]. The measured magnetic flux data is recorded at the rate of 1 kHz via an RS-232 interface. The time delay of the digital feedback loop is 8.4 s. The first-order gradiometer is made of a low-T c dc SQUID (niobium:Nb), has a pickup coil with a diameter of 17.8 mm and a baseline of 50 mm, and was used for the characteristic evaluations of the D-FLL system. The sensitivity which is calculated using the relation B/  = 0.6 nT/ 0 , where B is the flux density in the pickup coil, gives a measure of the flux resolution of 11.35 fT/digit (19  0 /digit) for this system. Figure 11 shows an active area of the D-FLL system. The terminology "active area" means that the system can measure magnetic signals just in this area. The right vertical axis is shown as a dynamic range when a 16-bit D/A converter is used for about 1  0 feedback flux. The measured dynamic range is 218  0 (141 dB) at 1 Hz, 22  0 (121 dB) at 10 Hz, and 2.1  0 (100 dB) at 100 Hz. In addition, the measured slew rate is 1.3 k 0 /s. The D-FLL system can operate in a stable manner if power line noise that is usually the main problem encountered with unshielded operations is less than 4.1  0 in amplitude at 50 Hz. Figure 12 shows the measured D-FLL output noise spectra both inside and outside a magnetically shielded room (MSR). The lower line is measured in the MSR and the upper line is measured in an unshielded environment. The white noise level used was 36.1 fT/Hz 1/2 and the 1/f corner frequency was 2 Hz in an MSR. The noise spectrum measured outside an MSR contained some unidentified noises beside power line noise that exhibited 465 pT/Hz 1/2 (0.78  0 /Hz 1/2 ) at 50 Hz. Dynamic range (dB) Fig. 11. Active area of the D-FLL system. An "active area" implies that the system can measure magnetic signal just in this area. The right vertical axis is shown as a dynamic range when 16 bit D/A converter is used for 1  0 feedback flux.

Magnetocardiogram measurements
SQUIDs have been utilized for the detection of weak magnetic fields in our daily life from non-distraction evaluation to bio-magnetic field detection.
Recently, the magnetocardiography (MCG) has attracted much attention as a potential application of SQUID, because MCG systems with SQUID provide higher sensitivity than other systems and because cardiac diagnosis plays an important role in the prevention of heart disease. Thus far, we have developed a multi-channel MCG system by using Nb SQUID [27]. Following that effort, we developed a conventional MCG system made of MgB 2 . Figure 13 shows the block diagram of a cryocooler system and the close-up of a cold cylinder. A commercially available pulse-tube cooler (Sumitomo RP-052D) was used. The cold head of the cryocooler is connected to the compressor via a valve unit. Because the valve unit produces magnetic noise, the cold head and the refrigerator component are installed in an MSR, and the valve unit is set outside of it. They are connected by a long copper tube.
The mechanical vibration of the refrigerator is ±3 m at 1.2 Hz. The cold stage is connected to the refrigerator by a copper column to enable refrigeration. The distance between them is 608 mm.
This refrigerator can achieve a temperature of about 5 K without a load. Our cryocooler has two modes. The first mode is the SQUIDs characterization mode. This is used for characterization of MgB 2 SQUIDs. The cold stage and the cold cylinder are covered by a radiation shield, which is made of copper and has a thickness of 3.4 mm. The MgB 2 SQUID is set underneath the cold stage. The lowest temperature in this mode was 5.6 K.  Fig. 13. Block diagram of a cryocooler system and the close-up of a cold cylinder. The MgB 2 SQUID is set under the cold stage or on the bottom plate of the cold cylinder when the system is used in either the SQUIDs characterization mode or the measurement mode.
The second mode is the measurement mode. This mode is used for measurement of an external magnetic field, as in the case of MCG. The cold stage and the cold cylinder are covered by super insulation films instead of a radiation shield. The MgB 2 SQUID is set on the bottom plate of the cold cylinder and has a thickness of 2.0 mm. The lowest temperature achieved in this mode was 11.9 K. A temperature sensor and a heater are placed on the cold stage and the column, respectively. They are connected to a PID controller, with which the temperature of the cold stage is controlled. Both the SQUID characterization mode and the measurement mode are controlled and operated by a computer through the D-FLL electronics. Figure 14 shows the magnetic flux noise spectrum measured inside a single layer MSR. The white magnetic flux noise was 6.8 pT/Hz 1/2 . The largest peak was caused by power lines of 50 Hz. Other peaks were observed at 1.2 Hz and its harmonics of that frequency. They were caused by the refrigerator and had a magnitude of 13.8 pT at 1.2 Hz.
www.intechopen.com  Figure 15 shows the photograph of the actual measurement and the MgB 2 magnetometer that was used in the experiments. The magnetocardiogram was measured by our MgB 2 SQUID system inside the MSR. This measuring apparatus was cooled by a pulse tube refrigerator instead of liquid helium.
(a) (b) Fig. 15. Photographs of the MgB 2 -SQUID device and magnetocardiogram measurements. Figure 16 shows a magnetocardiogram waveform of a healthy volunteer. This waveform was obtained by applying a digital low-pass filter whose cutoff frequency was 40 Hz and the waveform was averaged over 536 measurements [28]. We have succeeded in measuring the QRS complex and the T wave which is characterized in the activity of the human heart. This result shows a possible practical use of the MgB 2 -SQUID MCG system cooled by refrigeration above 20 K without liquid cryogen, although the system is still under development. www.intechopen.com

Summary
We built a prototype of a nano-bridge type MgB 2 SQUID device and optimized it for the measurement of magnetocardiogram. The results indicate that our method for the fabrication of SQUID of an MgB 2 magnetometer optimized for MCG measurement is effective. We are currently conducting research to reduce the noise from the refrigerator and to optimize the MgB 2 SQUID MCG system. A multi-channel type MCG system is also under development. The crystalline characteristics of MgB 2 films should be further improved to obtain better device performance. In the near future, the MgB 2 SQUID device is expected to have excellent performance. In addition to the SQUID device, we developed a highfrequency filter device. The detailed results of the filter device are shown in the papers of coworkers at Yamagata University [29]. Figure 17 shows the photographs of the filter device that was developed and studied.
(a) (b) Fig. 17. Photographs of the high-frequency filter device. The filter device was mounted on a Cu cavity. www.intechopen.com

Acknowledgments
The authors thank H. Yamaguchi, Y. Fujine, D. Oyama, K. Ikeda, S. Goto, T. Nakajima, T. Takahashi, H. Iriuda, T. Oba, A. Okubo, J. Araaki, K. Meguro, T. Abe, H. Endo, Y. Uchikawa and K. Fujisawa for the cooperation of MgB 2 SQUID project, and M. Iitake for his technical assistance of FIB process. This work was supported by the Japan Science and Technology Agency. A part of this work was conducted at the AIST Nano-Processing Facility, supported by "Nanotechnology Network Japan" of the Ministry of Education, Culture, Sports, Science and Technology (MEXT), Japan.