A New Hall Microdevice with Minimal Complexity ^†

Abstract

A new Hall microdevice with minimal complexity and orthogonal magnetic field activation is suggested. The microsensor contains a rectangular n-type silicon substrate. On the long sides, three ohmic contacts are formed symmetrically and opposite each other. The first two opposite electrodes are connected and the second two are fed in the same way, and third ones are the outputs. The increased sensitivity constituting 40 V/AT is due to the reduced parasitic surface currents. Furthermore, output electrodes are moved out of the area where the supply currents flow. The 80 × 135 μm² size of the sensor increases the resolution and provides detailed mapping of the magnetic field’s topology.

1. Introduction

The first three-contact (3C) silicon Hall sensor, activated by the magnetic field B parallel to the substrate’s plane, known also as a vertical or in-plane Hall element, was designed in 1983 by C. Roumenin and P. Kostov. It contains an n-Si plate with only three contacts on one of its sides—one central and two symmetrical to it. The end contacts through load resistors are connected to the central electrode via a supply source. Moreover, they constitute the output. The disadvantage of this single-ended microsensor is the decreased sensitivity resulting from the parasitic surface currents flowing between the contacts. Technological implementation is complicated because of the different-in-their-essence formation processes of the ohmic electrodes and load resistors. Vertical Hall elements represent a promising approach for implementing fully integrated 3D vector probes, as they have a multitude of applications in which the orthogonal activation by field B is preferred. This paper presents a transformation of a vertical Hall element into an orthogonal sensor configuration with minimal design complexity and a promising performance.

2. Device Structure and Operation

In Figure 1, the top view of the new Hall microdevice is shown. On the long sides of the thin n-Si substrate, two groups of three heavily doped n⁺ ohmic contacts, namely C₁, C₂, and C₃ and C₄, C₅, and C₆, respectively, are formed symmetrically and opposite to each other. The first two opposite electrodes are connected and the second two are fed in the same way; the third ones are the output. A deep surrounding p-ring is also present, which reduces the spreading of surface currents and confines the transducer region in the substrate. Contacts C₃ and C₆ of the sensor are those which are moved out of the areas where supply currents I_C1,2 and I_C4,5 flow. Field B is perpendicular to the plane, and electrodes C₃ and C₆ constitute the differential output V_HC3,C6(B) ≡ V_H(B). The operation is as follows. The supply contacts C₁, C₂, C₄, and C₅ are equipotential surfaces. As a result, the current lines I_C1,2 and I_C4,5 are initially perpendicular to them and penetrate deeply into the x-y area of the substrate. The field B leads to the Lorentz lateral deflection of the current paths in the x-y plane. They generate opposite-sign charges in the zones with contacts C₃ and C₆, i.e., V_H(B). The specific two single Hall devices share the same active region. The elements with contacts C₁, C₂, and C₃ and C₄, C₅, and C₆, respectively, are controlled by the components I_C2 and I_C5. The increased sensitivity of the device results from the drastically reduced parasitic surface currents from the location of contacts C₁ and C₂ and C₄ and C₅ on the various long sides.

Figure 1. Schematically plan-view of the new Hall configuration. In the x-y plane, the length of contacts C₁–C₆ is 30 μm, their width is 10 μm, and the depth is about 2 μm. The distance between electrodes C₁, C₂, and C₄ varies: C₅ is 20 μm, while l_C2,3 and l_C5,6 are 15 μm (on the mask). The width of the p-ring on the chip surface is about 20 μm.

3. Experimental Results and Conclusions

The prototype of the new configuration was manufactured using some of the processing steps applied in bipolar IC technology by employing four masks. The resistivity of the n-Si substrate is ρ ≈ 7.5 Ω·cm (N_D = n₀ ≈ 4 × 10¹⁵ cm⁻³). The penetration depth of the current trajectories into the x-y plane is about 30 μm. The active sensor zone is 80 × 135 μm². The essential result is that the sensor is implemented within a single technological cycle. The output characteristics, V_H(B), of the configuration are linear and odd when derived from field B and supply I_S.

The sensitivity is S_RI ≈ 40 V/AT. The nonlinearity constitutes no more than NL ≤ 0.2% in field B ≤ ±0.3 T and NL ≤ 0.6% in ±0.3 T ≤ B ≤ ± 0.7 T. According to the obtained data, the sensitivity temperature coefficient TCs is 0.1%/°C. It was established that the temperature coefficient of resistance R within the range 0 °C ≤ T ≤ 80 °C is approximately TC_R ≈ 1.1%/°C. The initial offset of the new configuration is by about 25% lower than structures similar to the 3C microdevices. The offset voltage is a linear function of the current I_s, Figure 2. The noise behaviour of the device is shown in Figure 3. At low frequencies, the spectrum f ≤ 1.0 kHz is typical for usual Hall plates, i.e., 1/f type. The extracted resolution in terms of the equivalent induction Bmin = [S_n,v(f)Δf²]/S_A, [T] at current I_s = 4 mA over bandwidth Δf = 5 Hz–400 Hz at the signal-to-noise-ratio S/N = 1 is around B_min ≈ 16 µT.

Figure 2. Offset voltage versus bias of current IS at T = 20 °C.

Figure 3. Internal noise power spectral density with the current IS as a parameter.

In conclusion, the transformation of the 3C vertical Hall element into the orthogonal sensor configuration yields a high device performance, which is very promising for applications in automobiles, industrial control systems, consumer devices, and, especially, robotics and robotised surgery.