France



Magnetic flux leakage (MFL) testing is a widely used method for detection of defects such as the metal loss due to the corrosion in steel bars located in reinforced concrete. The magnetic flux leak caused by the defect is then measured by a magnetic sensor, which gives information to the user of the presence of the defect [1]. For small defects, it is difficult for the sensor to determine the presence and location of the defects. Therefore, we propose a gradiometer measurement setup [2] using highly sensitive magnetic tunnel junction (MTJ) sensors, which is based on the tunnel magnetoresistance (TMR) effect, to increase the magnitude of the measured signal. In this work, 2 MTJ sensor modules were used to measure a gap of 0.1 mm between two steel bars (diameter = 20 mm, length = 500 mm). The length of the overall specimen is 1000 mm, where the gap is at the 500 mm mark (touching point of the two steel bars). The MTJ sensor modules consist of each sensor connected in a full-bridge circuit [3]. The signal is then passed through an amplifier circuit and low pass filter to further reduce noise in the measured signal. A robot is then used to move the sensor along the steel bars at a speed of 50 mm/s, measuring the magnetic flux signal. The lift-off value for the overall measurement is set at 50 mm. Before each measurement, the steel bars are magnetized using a solenoid (approximately 1.52 mT). A planar gradiometer setup was used, whereby the MTJ modules are placed in the same &lt;i&gt;x&lt;/i&gt;-axis. The effect of the baseline (&lt;i&gt;b&lt;/i&gt;), the distance between the first and second MTJ sensor module, was investigated (&lt;i&gt;b&lt;/i&gt; = 50, 100, 150, 200 mm). From the results, the overall magnitude of the gradiometer sensor signal showed an increase, almost twice as large as the single sensor module signal (Figure 1). Furthermore, the gradiometer sensor signal produced a single peak response (gap is right above the peak), in contrast to the single sensor signal which produced a sine-like wave response (gap is between the peak and trough) (Figure 2). This is helpful in terms that the location of the gap can be determined in real time. For the estimated gap location, a loss in accuracy was observed in all the signals, with the most accurate location determined for b = 100 mm. We demonstrated that a planar gradiometer setup consisting of 2 MTJ sensors could improve the overall signal value of the measurement setup resulting in an easier way to determine the location of a 0.1 mm gap at 50 mm lift-off. The results obtained from this research are expected to contribute to the further application of MTJ sensors not only in MFL testing but in other fields of magnetic-based non-destructive testing. This work was partly supported by the Center of Innovative Integrated Electronic Systems (CIES) and the Center for Spintronics Research Network (CSRN). The sensors were provided by Mr. Kosuke Fujiwara from Spin Sensing Factory (SSF).

**References:**

[1] Hirose. M, Maeda. T, Matsuda. K, Yokota. M, Hattori. A, and Miyagawa. T, J. Stru. Eng., 58A, p.867-878 (2012).&lt;br/&gt;[2] W. Dehui, S. Lingxin, W. Xiaohong, and L. Zhitian&lt;i&gt;,&lt;/i&gt; J Nondestruct Eval., 36, p.24 (2017).&lt;br/&gt;[3] J. Cao, and P.P Freitas, J. Appl. Phys., 107, 09E712 (2010).

**Images:**

![](https://assets.underline.io/uploads/markdown_image/1/image/3227474d3e22de78e908e4d2ba9a7a6b.jpg)

![](https://assets.underline.io/uploads/markdown_image/1/image/74835cefbcae5c802089d4aa85ecad44.jpg)




INTERMAG 2021

 Detection of Gap in Steel Bars by Magnetic Tunnel Junction Sensor Gradiometer based Magnetic Flux Leakage Testing Method

INTERMAG 2021 Welcome Video



	
<figure class="video_container">
 <iframe src="https://screencast-o-matic.com/player/crflqFVnUpZ?v=6&ff=1&title=0&controls=1" width=640 height=360 frameborder="0" allowfullscreen="true"> </iframe>
</figure>

Want to buy a ticket? 

Register for INTERMAG 2021 here

[![](https://assets.underline.io/uploads/markdown_image/1/image/17ce76e7960db633258024eef8d6da7f.png)](https://intermag.org/register)

Got your ticket, but still can’t access the event?

Contact us and our team will look into it.

Magnetic flux leakage (MFL) testing is a widely used method for detection of defects such as the metal loss due to the corrosion in steel bars located in reinforced concrete. The magnetic flux leak caused by the defect is then measured by a magnetic sensor, which gives information to the user of the presence of the defect [1]. For small defects, it is difficult for the sensor to determine the presence and location of the defects. Therefore, we propose a gradiometer measurement setup [2] using highly sensitive magnetic tunnel junction (MTJ) sensors, which is based on the tunnel magnetoresistance (TMR) effect, to increase the magnitude of the measured signal. In this work, 2 MTJ sensor modules were used to measure a gap of 0.1 mm between two steel bars (diameter = 20 mm, length = 500 mm). The length of the overall specimen is 1000 mm, where the gap is at the 500 mm mark (touching point of the two steel bars). The MTJ sensor modules consist of each sensor connected in a full-bridge circuit [3]. The signal is then passed through an amplifier circuit and low pass filter to further reduce noise in the measured signal. A robot is then used to move the sensor along the steel bars at a speed of 50 mm/s, measuring the magnetic flux signal. The lift-off value for the overall measurement is set at 50 mm. Before each measurement, the steel bars are magnetized using a solenoid (approximately 1.52 mT). A planar gradiometer setup was used, whereby the MTJ modules are placed in the same x-axis. The effect of the baseline (b), the distance between the first and second MTJ sensor module, was investigated (b = 50, 100, 150, 200 mm). From the results, the overall magnitude of the gradiometer sensor signal showed an increase, almost twice as large as the single sensor module signal (Figure 1). Furthermore, the gradiometer sensor signal produced a single peak response (gap is right above the peak), in contrast to the single sensor signal which produced a sine-like wave response (gap is between the peak and trough) (Figure 2). This is helpful in terms that the location of the gap can be determined in real time. For the estimated gap location, a loss in accuracy was observed in all the signals, with the most accurate location determined for b = 100 mm. We demonstrated that a planar gradiometer setup consisting of 2 MTJ sensors could improve the overall signal value of the measurement setup resulting in an easier way to determine the location of a 0.1 mm gap at 50 mm lift-off. The results obtained from this research are expected to contribute to the further application of MTJ sensors not only in MFL testing but in other fields of magnetic-based non-destructive testing. This work was partly supported by the Center of Innovative Integrated Electronic Systems (CIES) and the Center for Spintronics Research Network (CSRN). The sensors were provided by Mr. Kosuke Fujiwara from Spin Sensing Factory (SSF).

**References:**

[1] Hirose. M, Maeda. T, Matsuda. K, Yokota. M, Hattori. A, and Miyagawa. T, J. Stru. Eng., 58A, p.867-878 (2012). [2] W. Dehui, S. Lingxin, W. Xiaohong, and L. Zhitian, J Nondestruct Eval., 36, p.24 (2017). [3] J. Cao, and P.P Freitas, J. Appl. Phys., 107, 09E712 (2010).

**Images:**

![](https://assets.underline.io/uploads/markdown_image/1/image/3227474d3e22de78e908e4d2ba9a7a6b.jpg)

![](https://assets.underline.io/uploads/markdown_image/1/image/74835cefbcae5c802089d4aa85ecad44.jpg)

poster

Many ferromagnetic substances are naturally magnetized under the influence of the geomagnetic field, and information about the corresponding objects can be obtained based on the analysis of monitoring the magnetic fields generated around magnetic targets. The detection and tracking of magnetic targets is the current research hotspot, and the target monitoring based on magnetic signals as well as widely used in various fields. Currently, the magnetic target localization method is commonly used to obtain the relative position of the target by inversion of the magnetic field gradient tensor matrix. However, this method has a detection blind spot, and even if the final results are corrected by mean filtering, there is still a certain amount of error. Therefore, the correction of the localization results needs to be based on the information of the magnetic field gradient matrix collected at the detection point. This paper uses the magnetic field gradient tensor target localization algorithm as the basis of single-point localization, analyzes its error in detection, and proposes a weighted route fitting method based on magnetic field gradient matrix information for the characteristics of route calculation which is vulnerable to great errors. After simulation experiments, the method can reduce the fluctuation of detection results to a certain extent, and can effectively propose the influence of the extreme error generated by the detection blind area on the positioning results. To investigate the sources of error in target localization, we designed a series of experiments for the magnetic field gradient matrix-based localization method. The location of the detection blind zone is verified by changing the relative position of the magnetic target to the detection point. When the position vector is perpendicular to the magnetic moment vector, the error rate increases significantly. The experimental results are consistent with the theoretical formula, and the higher the accuracy of the magnetic sensor, the smaller its detection blind zone. When the detection point out is in the detection blind area, its magnetic field gradient matrix is a non-full-rank matrix, so the accuracy of target detection is often inversely proportional to the dispersion of its matrix eigenvalues. Therefore, by introducing a weighted fitting method based on the gradient matrix, the detection error can be reduced to a certain extent and the interference caused by abnormal data can be eliminated. Experiments are designed to verify the feasibility of the method. The schematic diagram of the positioning detection experiment is shown in Figure 1 below, where ABCED represents the location of different detection points, and the oval dashed line is its detection range. The magnetic target is set to come out from the coordinates (-0.5m,0m) and advance along a straight line to (0.5m,0m), and the comparison of its route monitoring results is shown in Figure 2 below. Based on the results, it can be concluded that the error-corrected route is closer to the set simulation route.

**References:**

Zhao G, Han Q, Tong X, et al. Adaptive filtering method for magnetic anomaly detection[J]. Journal of Applied Remote Sensing, 2018, 12(2):1-10. Kim K Y, Jeong E, Kim S, et al. Magnetic Anomaly Detection Using Continuous Angle Alignment of Three-Axis Magnetic Signal[J]. IEEE sensors journal, 2019, 19(2):743-750. Zheng J, Fan H, Zhang Q, et al. Magnetic Anomaly Target Recognition Based on SVD and SVMs[J]. IEEE Transactions on Magnetics, 2019, 55(9):1-8. Shen Y, Wang J, Shi J, et al. Interpretation of Signature Waveform Characteristics for Magnetic Anomaly Detection Using Tunneling Magnetoresistive Sensor[J]. Journal of Magnetism and Magnetic Materials, 2019, 484(AUG.):164-171. Huang Y, Wu L H, Sun F. Underwater Continuous Localization Based on Magnetic Dipole Target Using Magnetic Gradient Tensor and Draft Depth[J]. IEEE Geoscience and Remote Sensing Letters, 2014, 11(1):178-180. Nara T, Ito W. Moore–Penrose generalized inverse of the gradient tensor in Euler's equation for locating a magnetic dipole[J]. Journal of Applied Physics, 2014, 115(17PT.3):3646-3648.

**Images:**

![](https://assets.underline.io/uploads/markdown_image/1/image/c3e0fd30c505b2a24e40c70d1a568e97.jpg)
Fig.1 Schematic diagram of magnetic target movement route monitoring

![](https://assets.underline.io/uploads/markdown_image/1/image/18c7f44678b5fffecd2245e1eb1c5fc9.jpg)
Fig.2 Comparison chart of route fitting correction results

 Magnetic Target Motion Monitoring Based on Weighted Route Fitting

With the development of the Internet of things, the demand for wearable flexible sensors is increasing. In order to research the new type of flexible magnetic sensors, in this paper, aiming at the vacancy of the application surface acoustic wave(SAW) devices in the field of wearable flexible magnetic sensors, the SAW magnetic sensors with ZnO/FeGa/PI structure are modeled and simulated[1,2], and the resonant characteristics and wave modes are studied. The frequency domain analysis of four kinds of devices with different thickness of 1.2 μm, 2 μm, 3.4 μm and 4 μm is carried out respectively. It is found that the device has multiple wave modes in each case, and the resonance peak of the device shifts to high frequency with the increase of the thickness of the piezoelectric layer. By comparing the modal deformation figure, it is found that the main modes of piezoelectric film with different thickness are various. For example, for 1.2 μm, the main mode is mode 2, while for 2 μm, the main mode is mode 4. The reason is that the device is composed of multi-layer thin films, in which the acoustic wave is not only affected by the piezoelectric medium, but also by the magnetostrictive layer and flexible substrate. Therefore, the optimal working mode is different for different thickness devices. In this paper, the flexible magnetic sensor is designed according to different thickness structure, and its optimal working mode is studied. When the maximum resonant mode is used as the working mode, the linear region of the magnetic sensor is basically unchanged, which is between 2 and 10 Oe. When the thickness of ZnO is 2 μm, the sensitivity in the linear region is the largest, about 253.3 kHz/Oe, and the working frequency is about 70MHz. When the thickness of ZnO is 3.4 μm, the sensitivity is only 30.5 kHz/Oe, but the working frequency is the highest, about 106MHz. This provides a theoretical basis for the development and application of flexible magnetic sensors. It is also found that when the thickness of ZnO is 1.2 μm, the maximum absolute value of admittance is only 0.0564, while for the structure with 4 μm, the maximum absolute value of admittance is 10.3413, which is of great significance for the subsequent design of impedance matching of wireless SAW magnetic sensor.

**References:**

[1] Datta S, Atulasimha J, Mudivarthi C, et al. Stress and magnetic field-dependent Young's modulus in single crystal iron-gallium alloys[J]. Journal of Magnetism and Magnetic Materials, 2010, 322(15): 2135-2144. [2] Chung G S, Phan D T. Finite Element Modeling of Surface Acoustic Waves in Piezoelectric Thin Films[J]. Journal of the Korean Physical Society, 2010, 57(3): 446-450.

**Images:**

![](https://assets.underline.io/uploads/markdown_image/1/image/7b7faa45b7f12d0b6a8f402c537530e5.jpg)
FIG. 1 Absolute value of admittance and deformed shape plot of piezoelectric layers with different thickness (a) 1.2 μm (b) 2 μm (c) 3.4μm (d) 4 μm.

![](https://assets.underline.io/uploads/markdown_image/1/image/d20a6bf2d577a547df1976e022bde57b.jpg)
FIG. 2 (a) Relationship between resonant frequency of main mode and magnetic field intensity when the thickness of ZnO is 1.2 μm. (b) Sensitivity of main modes corresponding to different ZnO thickness.

 Simulation of wave mode of flexible SAW magnetic sensor based on ZnO/FeGa/PI structure

Magnetic field sensors using the magnetoelectric (ME) effect in the composite multiferroic structures have attracted intensive attention due to their convenient operating at room temperature and high sensitivity. Large works have focused on the improvement of sensitivity and linearity by designing novel device structures but put rare attention on the optimization of equivalent noise that is also an important parameter for actual device application. In this work, we present a DC magnetic field sensor consisting of PZT/Metglas/PZT composites. With applying a voltage bias on the bottom PZT piezoelectric layer, the ME coefficient of the top PZT/Metglas bilayers can be increased obviously, as shown in Fig.1(b). This effective increase of ME coefficient will benefit the improvement of the magnetic field-to-voltage transfer coefficient and the reduction of the noise in the magnetoelectric sensor that have been shown in Fig.1(b). By increasing the electric field from 0 V/mm to 360 V/mm, the sensitivity of the sensor can increase from 13.04 mV/Oe to 32.85 mV/Oe at the resonant frequency of 131.6 kHz, and the equivalent magnetic noise of the sensor will decrease from 20 nT/√Hz to 9.8 nT/√Hz, as shown Fig.1(c). This three times increase of sensitivity and two times reduction of equivalent noise in the DC magnetic field detection by using ME effect may offer a new route for the development of novel magnetic field sensors or arrays.

**References:**

1. J. L. Hockel, T. Wu, and G. P. Carman, Journal of Applied Physics, Vol. 109, p. 064106 (2011). 2. V. N. Serov, Y. K. Fetisov, and A. A. Berzin, IEEE Sensors Journal, Vol. 18, p. 8256 (2018). 3. Y. Shen, D. Ma, and J. Gao, IEEE Electron Device Letters, Vol. 39, p. 1417 (2018).

**Images:**

![](https://assets.underline.io/uploads/markdown_image/1/image/4f90ba99bf370ba612a719e50bcc089f.jpg)
(a) Dependence of the output voltage with the DC magnetic field at different bias voltages; (b) Dependence of the sensitivity with the bias electric field for the proposed magnetic field sensor; (c) Frequency dependent equivalent noise density spectra at different bias electric fields of the proposed sensor.

 Optimization of Equivalent Noise in the DC Magnetic Field Sensor Based on Magnetoelectric Effect

Nowadays，the detection of the ultra-weak magnetic fields (&lt; 1 mOe) in micro- or nano-scope (&lt; 100 nm) is a hot research field. The magnetic sensor based on the principle of anisotropic magnetoresistance effect is attractive in the application of detecting the ultra-weak magnetic field in a contactless manner [1]. In this paper, a sensitive, 2-axis magnetic sensor based on anisotropic thin-film structure is proposed and fabricated. As the sensor’s induction layer, the anisotropic films of permalloy Ni (80%) Fe (20%) with different thicknesses (20 nm and 50 nm) are deposited by means of magnetron sputtering [2]. Seed layer under NiFe layer is used to control the grain size in the NiFe layer, which is an important parameter for magnetoresistance. In order to obtain the NiFe layer with good performance, tantalum is deposited as the seed layer, making the NiFe layer with &lt;111&gt; crystallographic texture with suitable grain size [3]. Figure 1 shows the schematic structure of the designed sensor for the ultra-weak magnetic field detection. The characterizations are conducted on the prepared layer, including surface roughness by atomic force microscope (AFM) and the crystalline structure by X-ray diffraction (XRD). According to the test results of AFM, it is found that the film with a thickness of 20 nm has a smoother surface with an average roughness of 1.985 nm, which is less than that of the film with a thickness of 50 nm with an average roughness of 3.005 nm. With the difference of surface roughness, the coercivity of the two thickness films is also different, and it ultimately has an effect on the film properties. Figure 2 shows the characterization results of the surface roughness of the deposited NiFe thin film. Based on the characterized results, the magnetoresistance related to the anisotropic effect is demonstrated and analyzed. The influence of the thickness of the anisotropic film on the characteristics of the magnetization and the dynamic range of the sensors, as well as the range of the magnetic field sensitivity is studied and analyzed in the paper [4]. The deposited film with 20 nm thickness shows obvious anisotropy with low coercivity, resulting in high sensitivity and outstanding detecting ability in the weak magnetic field.

**References:**

[1] N. A. Djuzhev, N. S. Mazurkin, and V. S. Pozdnyakov, Semiconductors, Vol. 49, no. 13, pp. 1739-1742 (2015) [2] M. Y. Chinenkov et al., NANOCON 2010 - 2nd International Conference, Conference Proceedings, pp. 270-273 (2010) [3] S. V. Luong, A. T. Nguyen, T. L. Nguyen, and A. T. Nguyen, IEEE Transactions on Magnetics, Vol. 54, no. 99, pp. 1-4 (2018) [4] D. Murzin, V. Belyaev, and F. Groß, Japanese Journal of Applied Physics, Vol. 59 (2019)

**Images:**

![](https://assets.underline.io/uploads/markdown_image/1/image/7d4e34ba9f63b52244bdc965e672fa73.jpg)
The model of the magnetic sensor. Internal figure (a) Local top view of the model. (b) Enlarged view of permalloy stripes and parallelograms which are biased

![](https://assets.underline.io/uploads/markdown_image/1/image/ebdc03eef8f2668ca164a2570b87437a.png)
Results of Ni80Fe20 films' surface roughness with different thicknesses tested by atomic force microscope (AFM). (a) The thickness is 20nm. (b) The thickness is 50nm.

 A Sensitive, 2-Axis Magnetic Sensor Based on Anisotropic Magnetoresistance Effect

There are many different ways and methods for measuring displacement. The influence of the magnetic material on the inductance of the inductor in its vicinity has been well investigated and used for the realization of various sensors, as well as displacement sensors [1, 2]. However, it has been shown that an improvement in the inductance of the inductor can be achieved if the grooves are realized on the surface of the magnetic film [3], which gives the great potential for the realization of magnetic sensors with better characteristics. With advances in development of materials in recent years, 3D printing technology allows a very fast fabrication of prototypes. It can be used for development of electronic components, sensors and actuators. It is a very simple, high-quality and precise way to create complex geometric structures using different materials which enable the realization of sensors of different characteristics. This paper presents a wireless magnetic displacement sensor realized in heterogeneous integration technology. The sensor is realized using additive 3D printing technology and magnetic disc with improved design. The wireless measurement method enables simpler and more precise measurement and data collection compared to the classic wired method. The exploded 3D model of the sensor is shown in the Fig. 1. The proposed sensor consists of three parts: the inductor, the sensor’s casing with spacer and the ferrite disk with grooves. The inductor windings are turned downwards towards the ferrite. The inductor is realized in Printed Circuit Board Technology, in the form of a spiral with an outer diameter of 25 mm and 22 turns (the width of the conducting lines is 500 μm and the conductor spacing is 730 μm). Using the Mn0.5Zn0.5Fe2O4 powder, the ferrite disc with grooves is pressed and sintered at a temperature of 1275 °C for 2 hours in a chamber furnace. The ferrite disc has a diameter of 25 mm. It has three concentric rings and a circle in the center on its surface. The part of the sensor that allows the sensor to be sensitive to displacement is the spacer. The spacer is printed using a 3D technology. It was made of Ninjaflex, one of the leading flexible 3D filaments. It is in the shape of the ring with the thickness of 500 µm. Using a Polylactic acid (PLA) filament, the sensor’s housing is 3D printed in which the sensor parts are connected and fixed. The applicator of displacement that acts on the surface of the sensor is also 3D printed. Displacement measurement is performed wirelessly using an impedance analyzer HP4194A. A coil (antenna) is placed around the sensor and the impedance and the phase of the antenna-sensor system are measured, from which the resonant frequency is determined. The displacement operation on the sensor is performed using a manual translation stage which enables the precise movement of the sensor surface in the steps of 50 μm. The characteristic of the antenna-sensor is shown in the Fig. 2. During the action of the displacement, the spacer is compressed and the ferrite approaches the inductor, which causes an increase in inductance, i.e. a decrease in the resonant frequency of the antenna-sensor system.

**References:**

1. M. Kisic, N. Blaz, Lj. Zivanov and M. Damnjanovic, AIP Advances, Vol. 10, paper no. 015017, p. 1-4 (2020) 2. M. G. Kisic, N. V. Blaz, Lj. D. Zivanov and M. S. Damnjanovic, IEEE Transactions on Magnetics, Vol. 53, no. 11, Art no. 4004104, p. 1-4 (2017) 3. X. Zhou, Y. Wen, Y. Mu and P. Li, 64th Annual Conference on Magnetism and Magnetic Materials, MMM2019, November 4-8. 2019, Las Vegas, Nevada, Book of abstracts, paper no. DV-05, p. 437-438 4. Available online at: https://ninjatek.com/ninjaflex/, accessed 05.01.2021.

**Images:**

![](https://assets.underline.io/uploads/markdown_image/1/image/eb2563e4c8b8f330d15ac98f997429ff.jpg)
Fig.1. The exploded 3D view of the sensor.

![](https://assets.underline.io/uploads/markdown_image/1/image/4dd0398f6f53d10c2624bcdc7dd3ddf0.jpg)
Fig. 2. The dependence of the resonant frequency on the displacement.

 Heterogonous Integrated Displacement Sensor with Grooved Ferrite

1. INTRODUCTION The non-destructive testing technology is appropriate to inspect high-priced product because it can inspect while maintaining the shape of the structure. And as the structure of the inspection body becomes complex and diversified with the development of the industrial society, the demand for non-destructive inspection is increasing mainly in developed countries. These techniques investigate structures of inspection by detecting energy deformation caused by the defects or constitution of products using specific physical energy to the target. Depending on the type of energy used for investigation, non-destructive testing is divided into technologies such as magnetic-particle testing, eddy-current testing, ultrasonic testing, radiation transmission testing, liquid-penetration testing and is applied to a wide range of application such as automobiles, nuclear power, steel structures, and pipelines [1]. Oil and gas pipes are mostly buried underground, so it is necessary to regularly manage corrosion or internal defects in pipes. Non-destructive testing is mainly used for safe long-distance transportation of oil and gas, among which Eddy-Current Testing (ECT), which does not have direct contact with the test object, is widely used. In ECT, operating frequencies have a significant impact on detecting faults. Traditional ECT mostly operates within the frequency range of a few kHz to MHz [2]. When the following high frequencies applied at the structure of the inspection, there is some difficulty to detect defects under pipe because eddy current does not flow to the bottom of the pipe due to the skin effect. In order to compensate for these limitations, the industry recently started to use the Magnetic Eddy Current technique that combines the principle of eddy current testing and magnetic flux leakage testing. The principle of this technology is based on the change in the pipe permeability due to defects by applying a certain amount of magnetic flux to the pipe using a permanent magnet or an electromagnet. In this technique, the magnetic flux used here does not need to go to the saturation point like MFL. In general, the magnetic field size required for MFL is at least 10 KA/m, but in MEC, a magnetic field of 3 KA/m is sufficient [3]. In the MEC technique, the value of magnetization level applied to the structure of the inspection has a significant influence on the defect determination performance. If the magnetic field is too small, it is impossible to determine the defects on the outer part of the pipe, and if the magnetic field is too large, defects of various sizes cannot be distinguished. Therefore, it is important to properly apply magnetic flux to the pipe so that defects can be found with high resolution in a given driving condition and the defect position can be distinguished. However, it has not been long since MEC technology has been introduced and researched, so there are insufficient papers on what driving conditions the sensor should operate under for optimal defect detection performance. For this reason, in this paper, a study was conducted to compare the defect detection performance according to the applied magnetic field point and to find the optimal applied magnetic field under a given driving condition. In order to analyze the distribution of the magnetic field around the defect according to the applied magnetic field value, a 3D model is set as shown in Fig.1-(a) and analyzed using 3D FEM. The defect underneath the pipe increases the density of magnetic flux in the pipe, thereby the permeability of pipe decreases and the impedance of the coil changes accordingly to generate a defect signal. The thickness of the defect used in the simulation is fixed at 1.5mm, which is 30% of the pipe thickness, and the diameter of the defect is fixed at 20mm. The value of magnetic field applied within the pipe was changed by changing the current value using an electromagnet. Fig.1-(b) show the respective magnetic field point of the pipe when 1A, 2A, 3A, 4A current is applied to the electromagnet coil. When a current of 1A is passed, it is difficult to determine the defects outside the pipe due to the large permeability value, and at 3A and 4A, the magnetic field applied to the pipe reaches almost saturation and the magnetic permeability is close to 1, making it inappropriate for detection of various size of defects. However, at 2A, it has a permeability, which is the middle value of the previous values, so it can detect various size of defects. Thus, using MEC techniques allows optimal defect detection performance at lower magnetic flux than MFL. In order to compare the defect detection performance in a real situation according to the applied magnetic field value, a prototype was made as shown in Fig.2-(a) and the experiments were conducted to verify compare to the simulated values. As shown in Fig.2-(b_Above), a specimen with artificial defects was produced and an actual MEC driving experiment was conducted. The specimens consist of defects of 10%, 30%, and 50% depth of the pipe thickness, where the solid line represents the external defect of the pipe and the dotted line represents the internal defect of the pipe. And defects werer detected by using coils as shown in Fig.2-(b_Below). The setting and procedures of experiments and the discussion of the results will be covered in the manuscript.

**References:**

[1] Seung-Seok Lee, Ki-Bok Kim, Journal of the korean society for nondestructive testing, vol. 24, pp. 64-85 (2015) [2] K.Reber, A.Boenisch, Stefanie Asher, PPSA Seminar in Scotland, pp.1-12 (2015) [3] Mohd Aufa Hadi Putera Zaini, Mohd Mawradi Saari, Nurul Ain Nadzri, 2020 11th IEEE Control and System Graduate Research Colloquium(ICSGRC 2020), Malaysia, pp.1-4 (2020)

**Images:**

![](https://assets.underline.io/uploads/markdown_image/1/image/77bba74437caf3ebf9387a3d20c9fd21.jpg)
Fig. 1. (a) MEC simulation models. (b) Magnetic field points according to the applied magnetic flux.

![](https://assets.underline.io/uploads/markdown_image/1/image/f9a621e0e02b8adcb115167ec3f92c7a.jpg)
Fig.2. (a) MEC prototype. (b) specimen with artificial defects(Above) and MEC coil(Below).

 Selection of Applied Magnetic Field for Optimal Detectability in Eddy Current Testing with DC bias

Tunneling magnetoresistive (TMR) sensors can easily be used for magnetic sensing due to high sensitivity, small size, and low power consumption. [1] They also have extremely short response times and are suitable for precise sensing applications. The magnetic hysteresis in the sensing layer of TMR sensors usually has a negative effect on the linearity and reversibility of the sensors. Therefore, reduced magnetic hysteresis, as well as low power and low noise, is favored in low magnetic field measurement. Previous research has shown that the magnetic hysteresis in TMR sensors can be reduced using AC modulation magnetic fields. [2] In this study, a much smaller AC modulation amplitude is used for lower overall power consumption of TMR sensors and the linear response range of the sensors under a small modulation field is investigated. The magnetic tunnel junction (MTJ) stack in the studied TMR sensor was designed as Si/150 SiO2/ 1.5 Ta/25 Ru/7 Ir20Mn80 / 2.2 Co30Fe70 /0.85 Ru/ 2.8 Co40Fe40B20/ 2 MgO /1.3 Co40Fe40B20 / 0.6 MgO/ 50 Co40Fe40B20 /2 Ta/15 Ru (in nm). The TMR ratio is about 180 % and the coercivity of the free layer is about 0.5 mT after magnetic annealing at 340 oC for 2 hour. The MTJs were connected into a Wheatstone bridge to form a sensor. An alternating current driven modulation coil was added surrounding the TMR sensor during magnetic field measurement with the AC modulation field applied along the sensing direction. The DC voltage output of the TMR sensor with different measurement conditions is shown in Fig. 1. The sensor was powered by a 1.5 Volt battery and a sinusoidal modulation field was applied with a frequency of 1 kHz and an amplitude of 0.3 mT. The output voltage of the sensor was measured in these DC magnetic field ranges: (a) -3 mT to 3 mT, marked in black in the figure, (b) -1.8 mT to 1.8 mT, marked in green, (c) -0.6 mT to 0.6 mT, marked in blue, and (d) -0.1 mT to 0.1 mT, marked in red. The figure shows that magnetic hysteresis in TMR sensor was effectively reduced if measured in a small range of DC magnetic field. With the amplitude of AC modulation field being as small as 0.3 mT, the sensor exhibits a linear response in the range of -0.6 mT to 0.6 mT, where the sensitivity is 20.6%/mT. It’s demonstrated that based on a simple construction, TMR sensors with small AC modulation fields can be used for low field sensing with good linearity. ACKNOWLEDGMENT This work was supported by Ministry of Science and Technology of Taiwan under grant number: 108-2314-B-003 -001 and 109-2314-B-003 -004. The MTJ samples were prepared in the Nebraska Nanoscale Facility: National Nanotechnology Coordinated Infrastructure and the Nebraska Center for Materials and Nanoscience, which are supported by the National Science Foundation under Award ECCS: 2025298, and the Nebraska Research Initiative.

**References:**

[1] Zheng, C. et al. Magnetoresistive Sensor Development Roadmap (Non-Recording Applications). IEEE Trans. Magn. 2019, 55, 1–30. [2] S. Liao et al., "Hysteresis Reduction in Tunneling Magnetoresistive Sensor with AC Modulation Magnetic Field," IEEE Transactions on Magnetics,(2020) doi: 10.1109/TMAG.2020.3009900.

**Images:**

![](https://assets.underline.io/uploads/markdown_image/1/image/064c1f75dc42a064ec6e3200b38de2c4.jpg)
Fig.1 The output voltage of the sensor as a function of the applied DC magnetic field in the range of four conditions: -0.1 mT to 0.1 mT, -0.6 mT to 0.6 mT, -1.8 mT to 1.8 mT, and -3 mT to 3 mT. The sensor was powered by a 1.5 Volt battery and a AC modulation field was applied with a frequency of 1 kHz and an amplitude of 0.3 mT for all conditions of DC field range.

 A study of hysteresis reduction of small AC magnetic field modulated tunneling magnetoresistive sensor

Introduction Induction hardening treatment is used for a variety of steels because of its short hardening time and ability to hardened steel with complex shapes. In this research, the non-destructive inspection method for the hardened depth in surface of the induction hardened steel plate by an electromagnetic sensor is investigated. The hardened depth in surface of the steel plate is measured by detecting the difference in the permeability and the conductivity between the hardened and the non-hardened layers inside the steel plate by the electromagnetic sensor [1,2]. On the other hand, the detection signal in the electromagnetic sensor is also affected by the changing in the distance (Lo: Lift-off) between the surface of the induction hardened steel plate and the sensor. Therefore, the method for simultaneously measuring the surface hardened depth in the hardened steel plate and the lift-off is proposed. The model and the evaluation method Fig.1 shows the 1/2 region of the proposal inspection model for detecting the surface hardened depth in the surface hardened steel plate of SCM440 steel material. The proposal electromagnetic sensor is consisted of an electromagnetic yoke made of laminated silicon steel plate, an alternating excitation coil, and two types of detection coils. The z-direction of the flux density (Bz) inside the magnetic yoke is detected by the z-direction detection coil on one leg of the yoke, and the x-direction magnetic field distributed (Bx) between the two legs of the yoke is detected by the x-direction detection coil. Since the permeability and conductivity in the hardened layer are lower than that of the non-hardened layer, the flux density in the x- and z- direction of these detection coils is changed when the hardened depth is changed. However, even if the depth of the hardened layer is constant, their detected magnetic flux in these detection coils is changed when the lift-off is changed. Therefore, both the lift-off and the surface hardened depth are estimated by the Bx and Bz. 3-D electromagnetic FEM considering the nonlinearity of magnetic properties The flux density and eddy current inside the surface hardened steel plate are calculated by 3-D nonlinear finite element method (FEM) taking account of the initial magnetization curve and conductivity of the hardened and non-hardened layers in the steel. In an induction hardened steel plate, there is an intermediate layer called a heat-affected zone between the hardened and non-hardened layers. Therefore, in this analysis, the initial magnetization curves and conductivity inside the intermediate layer are calculated by linear interpolation of the initial magnetization curves and conductivity of the hardened and non-hardened layers. Measurements results and estimates Fig.2 shows the obtained values of Bx and Bz in the two detection coils when the hardened depth (D) and Lo are changed, respectively. This figure denotes that if Lo is constant, Bx is increased and Bz is decreased when D is increased. This is, because the flux density Bz inside the magnetic yoke is decreased, and the Bx distributed between both feet of the yoke is increased, since the permeability and conductivity in surface domain of the steel plate is decreased when the D is increased. On the other hand, if D is constant, Bx is increased and Bz is decreased when Lo is increased. This is, because the flux density Bz in the closed magnetic path between the sensor and the steel plate is reduced when the Lo is increased, and the magnetic flux Bx distributed between both feet of the yoke is increased. The figure shows that the signals of Bx and Bz due to the changes of D and Lo are distinguished. Therefore, Both D and Lo are inspected by the linear interpolation using measured Bx and Bz. Table I shows the interpolated results of D and Lo using the Bx - Bz plane of Fig.2. The measured values are almost in agreement with the obtained ones. The table illustrates that both D and Lo can be detected from the Bx and Bz in two detection coils. Conclusions The results obtained are summarized as follows: (1)It is possible to estimate the depth (D) of surface hardened steel plate by detecting the change of flux density in two detection coils due to the difference of the permeability and the conductivity in the steel with and without hardened layers. However, the flux density detected by the search coil is influenced by the change of both lift-off and depth of surface hardened steel. (2)The flux density (Bx) in the detection coil of x-direction is increased and the flux density (Bz) in the detection coil of z-direction is decreased when the surface hardened depth (D) of the induction hardened steel or the lift-off (Lo) are increased. However, the signals of Bx and Bz in two detection coils due to the changes of D and Lo are distinguished. Therefore, it is possible to inspect the D of the surface hardened steel without the influence of Lo by detecting both the Bx and Bz of two detection coils in the proposed inspection sensor.

**References:**

[1] Y.Gotoh, S.Onita, T.Horino, and Y. Misaka,” Detecting Method of Hardened Depth in Surface Hardened Steel by Magnetic Field on Steel”, IEEE Transactions on Magnetics, vol.54, no.11, Article Sequence Number:6202404, 2018. [2] Y.Gotoh, K.Sakurai, N.Takaoka, Y.Misaka, K.Kawasaki, and N.Takahashi,” Proposal of Electromagnetic Inspection Method of Hardened Depth of Steel using 3-D Nonlinear FEM Taking Account of Hysteresis”, IEEE Transactions on Magnetics, vol.47, no.10, pp.3546-3549, 2011.

**Images:**

![](https://assets.underline.io/uploads/markdown_image/1/image/e862db6def1a56d0093f74acb2e22a85.jpg)
Fig.1. Proposal inspection model (x-z plane 1/2 region).

![](https://assets.underline.io/uploads/markdown_image/1/image/c3974039805ce679ad07cfffd241dc00.jpg)
Fig.2. Experimental results for magnetic flux density (Bx,Bz) at each hardened depth (D) and lift-off (Lo). 
 
Table 1. The interpolated results of D and Lo.

 Proposal of Inspection Method for Lift-Off and Surface Hardened Depth of Induction Hardened Steel Plate Using by Alternating Magnetic Field

In recent years, the non-destructive inspection (NDI) method based on the eddy-current (EC) effect is increasingly playing an important role in many applications. The advancement in the EC NDI techniques with the miniature magnetic field sensors enables the size estimation of small magnetic particles and the detection of deep-lying defects, such as cracks, metal loss, and other mechanical damages, in highly conductive metals. Several solutions of the field-detection based EC NDI have been developed based on the combination of an excitation coil and magnetic sensors such as Hall, giant-magnetoresistance (GMR), or superconducting quantum interference device (SQUID) sensors. The high-sensitivity and cost-effect solutions are the technique with GMR sensors, which are suitable for many different applications, including the defect evaluation in the related industries [1]-[3] and the electronic compass in consumer electronics [4]. The reliability and capability of GMR sensors for defect detection have been demonstrated by using the eddy-current probe with packaged GMR sensor arrays [2] or a single chip having double-coil excitation [3]. However, the usage of the GMR sensor with a big excitation coil is not suitable for the applications for which the space for scanning the probe is small. Furthermore, the linear arrangement of elements in the array [3] makes it suitable only to the specimen with a flat surface. Further improvement in the design of the GMR EC probe should be made to enhance the sensitivity and accuracy in flaw characterization. In this work, we proposed a novel design of miniature differential EC probes consist of a GMR bare-die sensor and a tiny coil. The schematic diagram for the high spatial resolution GMR EC probe is shown in Fig. 1. This probe consists of a tiny excitation coil with a small feature size of less than 3 mm. The GMR sensor is the GF708 bare die from Sensitec GmbH. The chip layout is a Wheatstone bridge circuit with four spin-valve magnetoresistance elements, as shown in the inset of Fig. 1. The sensitivity of the two active elements, R1 and R2, are enhanced by the soft magnetic flux concentrator films, while the two reference elements, R0, are shielded by the same soft magnetic films. The output voltage of the GMR bridge sensor can be expressed as: Vab = Vcc{R0/[R(H)+R0]-R(H)/[R0+R(H)]} (1) where R(H) is the transfer curve modeled as a hyper-tangent function of the magnetic field H on R1 and R2 [4]. The GMR chip is 1 mm in width and 1.4 mm in length, and the sensing axis is along the width of the chip. It is mounted on the edge of a printed circuit board (PCB). A special aluminum wire bonding technique is developed to connect the terminals of the GMR sensor onto both sides of the PCB. With this technique, the distance from the chip surface to the sample surface can be reduced to within 0.2 mm. The sensor surface is perpendicular to the PCB and parallel to the sample surface. The sensing direction of the GMR chip is parallel to the specimen surface and perpendicular to the vertical excitation field. With this arrangement, the GMR chip acts like a gradiometer sensitive to the gradient of the defect field. The miniature chip-and-coil assembly and the short lift-off distance effectively enhance the spatial resolution in flaw detection with the GMR EC probe. To test the spatial resolution in the detection of the defect signal, the sample was placed on the translation stage controlled by the 08TMC-2 motor controller from Unice E-O Services Inc. To scan for the simulated defect in the sample. The GMR EC probe is driven by a 100-kHz sine wave voltage provided by a function generator. The differential output of the probe is measured by a low noise preamplifier and transmitted to the lock-in amplifier to analyze the amplitude and phase of the EC signal. The spatial distribution of the signals was digitized in real-time by a data acquisition (DAQ) module USB-6216 from National Instruments and recorded by the computer. An in-house developed C# program is used to set the scanning range, velocity, and step size as well as to receive the eddy-current signal taken by the DAQ device. The experiment is performed on an aluminum plate with machined slots to simulate the cracks of different depths of 0.1, 0.3, 0.5, 1.0, 1.5 and 1.8 mm and the same width of 0.2 mm. The B-scan image with a lift-off distance of 0.2 mm between the probe and the specimen is shown in Fig. 2. It can be seen that the cracks on the aluminum plate are detected and the position of cracks coincides with the change in the in-phase (Re) and quadrature (Im) of the output voltage (Vout). The amplitude for the integration of Vout over the spatial coordinate x is roughly proportional to the depth of cracks. The full width at half maximum of each peak near the crack is less than 2 mm, which has achieved the physical limit set by the lateral size of the probe. The phase plot shows that all of the surface cracks deeper than 0.1 mm can be clearly identified. The result shows that the developed EC probe is suitable for the quantitative evaluation of the geometrical characteristics, such as shapes, sizes and positions of the machined defects on the metallic sample, with a high-spatial resolution. This work is supported by the Ministry of Science and Technology of Taiwan under Grant Nos. MOST108-2221-E992-083MY2 and MOST 107-2221-E-992-094.

**References:**

[1] D. Rifai, A. N. Abdalla, K. Ali, & R. Razali, “Giant magnetoresistance sensors: A review on structures and non-destructive eddy current testing applications”, Sensors, 16(3), 298 (2016). [2] O. Postolache, A. L. Ribeiro, and H. G. Ramos, “GMR array uniform eddy current probe for defect detection in conductive specimens”, Measurement, 46(10), 4369-4378 (2013). [3] Andrea Bernieri, Luigi Ferrigno, Marco Laracca, and Antonio Rasile, “Eddy Current Testing Probe Based on Double-Coil Excitation and GMR Sensor”, IEEE Transactions on Instrumentation and measurement, vol. 68, no. 5, pp. 1533 – 1542(2019). [4] V. S. Luong, Y.-H. Su, C.-C. Lu, J.-T. Jeng, J.-H. Hsu, M.-H. Liao, J.-C. Wu, M.-H. Lai, C.-R. Chang, “Planarization, Fabrication and Characterization of Three-dimensional Magnetic Field Sensors”, IEEE Transactions on Nanotechnology, Vol. 17, no. 1, pp.11 - 25 (2018).

**Images:**

![](https://assets.underline.io/uploads/markdown_image/1/image/8d751ab27f3f29e4e77bb5ac9747b6a8.jpg)
Fig. 1. High spatial resolution GMR EC probe.

![](https://assets.underline.io/uploads/markdown_image/1/image/695ad452874753af7b4e1e084e179870.png)
Fig. 2. The integral phase for the output voltage of EC probe when scanning on the aluminum plate.

 High spatial resolution flaw detector based on the GMR eddy-current probe

Energy sources used in industries, such as gas and oil, are mainly transported through buried pipes. If an energy source leaks out from the pipe due to corrosion or cracks, an accident such as an explosion may occur, so it is very important to periodically inspect the buried pipes. Since cracks have the risk of the pipe being torn, it is important to look for cracks in pipe inspection. There are several methods for inspecting buried pipe, such as ultrasonic testing(UT), and magnetic flux leakage testing(MFL). UT method is often used for various inspections including pipe inspection due to their very high accuracy, but has the disadvantage of slow inspection speed and high battery consumption. Buried pipes are connected up to tens to hundreds of kilometers at a time, so long-distance inspections are required for pipe inspection. UT, which has a slow inspection speed and high battery consumption, has limitations in the inspection of buried pipes. On the other hand, the MFL method using MFL sensors has little battery consumption except for the power used for the sensor, the inspection speed is relatively fast, and the defect detection accuracy is high. However, because the width of the crack is very narrow, the crack in the same direction as the magnetic field direction hardly affects the flow of magnetic flux, and there is almost no leaked magnetic flux due to the crack in the same direction as the magnetic field direction. That is, conventional MFL inspection devices such as axial MFL(AMFL) and circumferential MFL(CMFL), which allow magnetic flux to flow in only one direction in the pipe, hardly find cracks in the same direction as the magnetic flux direction[1]-[2]. Fig.1 is here Conversely, if the MFL inspection device is used to allow magnetic flux to flow in various directions in the pipe, it has the advantage of MFL method and can find all cracks regardless of direction. This paper proposes an MFL inspection device that allows magnetic flux to flow through a pipe in various directions and can find all cracks regardless of the direction. Fig. 2 shows the comparison of the magnet placement and magnetic flux direction of the AMFL, CMFL, and the proposed MFL inspection device. Fig.2 is here As shown in Fig. 2(a) and (b), the magnetic flux flows only in the axial and circumferential directions of the conventional MFL inspection devices, AMFL and CMFL, respectively. In contrast, the proposed MFL inspection device arranges the magnets differently from the conventional MFL inspection device, so that the magnetic flux flows in the diagonal and circumferential directions. When the magnetic flux flows through the pipe in two directions using the proposed MFL inspection device, the cracks in the circumferential direction can be detected at the section where the magnetic flux flows in the diagonal direction, and the cracks in the diagonal direction can be detected at the section where the magnetic flux flows in the circumferential direction. In other words, even if there are cracks that can’t be found in the section where the magnetic flux flows in a specific direction, the cracks can be found in the section where the magnetic flux flows in the other direction. As a result, all cracks can be found regardless of the direction using the proposed MFL inspection device. The reason why the MFL method can’t find cracks in the same direction as the magnetic flux direction and details of the proposed MFL inspection device will be covered in full paper.

**References:**

[1] M. Rostami Kandroodi, B. Nadjar Araabi, and M. Nili Ahmadabadi, in IEEE Transactions on Magnetics, Vol. 53, No. 3, pp.1-10(2017) [2] H. M. Kim, C. G. Heo, and G. S. Park, in IEEE Transactions on Magnetics, Vol. 54, No. 11, pp.1-5(2018)

**Images:**

![](https://assets.underline.io/uploads/markdown_image/1/image/6f08173f80a8dda7a8b08465cc9ce367.jpg)
Fig. 1. The magnetic flux leakage pipe inspection devices. (a) AMFL[1] (b) CMFL[2].

![](https://assets.underline.io/uploads/markdown_image/1/image/6b2c2db6cf2e1625096560ac4a8f79c2.jpg)
Fig. 2. The magnet placement and the magnetic flux direction of (a) AMFL. (b) CMFL. (c) the proposed MFL inspection device.

Detection of Gap in Steel Bars by Magnetic Tunnel Junction Sensor Gradiometer based Magnetic Flux Leakage Testing Method

Next from INTERMAG 2021

Magnetic Target Motion Monitoring Based on Weighted Route Fitting

Similar lecture

Selection of Applied Magnetic Field for Optimal Detectability in Eddy Current Testing with DC bias