SCIENCE CHINA Information Sciences

. RESEARCH PAPERS .

November 2011 Vol. 54 No. 11: 2435–2442 doi: 10.1007/s11432-011-4354-0

Dry electrode for the measurement of biopotential signals WANG Yu1,2 , PEI WeiHua2 ∗ , GUO Kai2 , GUI Qiang 2 , LI XiaoQian 2 , CHEN HongDa2 & YANG JianHong1 1Institute

of Microelectronics, School of Physical Science and Technology, Lanzhou University Lanzhou 730000, China; 2State Key Laboratory on Integrated Optoelectronics, Institute of Semiconductors, Chinese Academy of Sciences, Beijing 100083, China Received May 12, 2010; accepted September 14, 2010

Abstract This paper introduces a kind of silicon-based dry electrode for measuring biological signals. It uses microneedle arrays to penetrate into the stratum corneum to reduce skin impedance. The dry electrode requires neither skin preparation nor the electrolytic gel, is easy to use and causes no skin allergy. Two different technologies are chosen to manufacture microneedle arrays of dry electrode. One is deep dry etching combined with isotropic wet etching. The other is mechanical dicing combined with chemical wet etching (including isotropic wet etching and anisotropic wet etching). Microneedle arrays are coated with metal and divided into 25 mm2 as dry electrode patch. Impedance testing shows that the impedance value of dry electrode can be comparable with that of commercial electrode in the 20 Hz–10 kHz frequency range. The steady-state visual evoked potential recording and analysis prove that the dry electrode can be used to detect electroencephalography. Keywords

dry electrode, microneedle array, dry etching, wet etching, EEG

Citation Wang Y, Pei W H, Guo K, et al. Dry electrode for the measurement of biopotential signals. Sci China Inf Sci, 2011, 54: 2435–2442, doi: 10.1007/s11432-011-4354-0

1

Introduction

Biomedical detection of biopotential (e.g., electroencephalography (EEG), electrocardiogram (ECG) and electromyogram (EMG)) is now playing an increasingly important role in the medical field. Electrode, acting as a transducer, transforms the ionic current (generated by cells) into an electronic current to be processed by electronic devices [1–3]. Commercial wet electrodes have some merits, including simplicity, reliability and low weight [4]. However, in practice they have the following two shortcomings. First, their preparation is time-consuming. The skin should be cleaned or degreased by soft abrasion to reduce the thickness of the stratum corneum (SC). Electrolytic gel, which can diffuse into the SC to improve the electrical contact, must be applied between the electrodes and skin [2]. Unfortunately, the electrolytic gel may cause skin irritation and allergic reaction. Second, commercial wet electrodes must be operated by professionals and the procedures are complex, so it can hardly be used for home care [5]. Therefore, the commercial wet electrode is not ∗ Corresponding

author (email: [email protected])

c Science China Press and Springer-Verlag Berlin Heidelberg 2011 

info.scichina.com

www.springerlink.com

Wang Y, et al.

2436

Figure 1

Sci China Inf Sci

November 2011 Vol. 54 No. 11

Schematic of skin structures and dry electrodes penetration concept. The thickness of the stratum corneum

(SC) is 10–15 µm, which is the primary barrier for biopotential recording. Epidermis (50–100 µm) contains living cells and few nerves. Dermis has a large number of vessels and nerves, so microneedles penetrate the skin more than 10–15 µm but less than 50–100 µm to avoid the emergence of pain.

suitable for daily use at home. Dry electrode can overcome these disadvantages. There are many types of dry electrodes. One is composed by microneedle arrays. The microneedles can penetrate the SC barrier of the skin and create pores to decrease the impedance of skin, so electrolytic gel is not required. The design of microneedle is based on the structure of skin. Skin consists of three layers: epidermis, dermis and hypodermis (Figure 1). The thickness of epidermis is varying from 50–100 µm. The outermost layer of epidermis is SC which has a thickness of 10–15 µm. It acts as a barrier to fluid entering the body, and has high electrical resistivity. Other parts of epidermis contain living cells, but no blood vessels and few nerves. The dermis contains abundant blood vessels and nerves [6, 7]. The length of microneedles is greater than 50 µm but less than 100 µm; they can penetrate the SC but can hardly reach the nerves in the dermis. Over the past few years, several research groups have investigated the micromachining processes of microneedles. To date, several materials have been tried, including silicon [8–11], metal [12, 13] and polymer [14–17]. Manufacture methods are also various. In terms of silicon-based microneedle array, dry etching is a common method [18]. Griss et al. [2] manufactured microneedle arrays by threestep-deep reactive ion etching process. This method is difficult to control. In our work, one-step dry etching was used to define microneedle-length. In order to reduce costs, Wilke et al. [10] replaced dry etching with KOH wet etching. But the density of microneedles was low with this method. In our work, mechanical dicing combined with chemical wet etching was used. This method is cost saving. In this paper, dry electrode fabricated by two methods was introduced. One was dry etching (deep reactive ion etching, DRIE) combined with isotropic wet etching, and the other was mechanical dicing combined with chemical wet etching. In section 3, the impedance of the dry electrode was measured and compared with that of commercial wet electrode. In section 4, the EEG signals were measured to assess the quality of the dry electrode.

2

Fabrication

Production process of dry electrode goes in three steps. The first step is to produce microneedle arrays. The second step is to sputter the metal. The third step is to divide the wafer into the pieces of appropriate size (5×5 mm2 ). 2.1

Combination of dry etching and isotropic wet etching

Four-inch double-sided polished, n-type silicon wafer, with 100 crystallographic orientation, 500 µm thickness and 0.001 Ωcm resistance was used in this study. Figure 2 shows the fabrication process. (1) All wafers were immersed in Piranha solution (H2 SO4 :H2 O2 =2:1) at 125 ◦ C for about 30 min followed by thoroughly rinsing with DI water. (2) 0.3 µm-thick SiO2 was grown on the wafers in a thermal oxidation furnace at 1050 ◦ C. (3) Then the SiO2 layer was patterned by hydrofluoric acid buffer. (4) Pillars were formed by deep reactive ion etching (DRIE) process. To obtain higher pillars with less underetch,

Wang Y, et al.

Sci China Inf Sci

Figure 2

Figure 3

November 2011 Vol. 54 No. 11

2437

Flow of fabrication process of dry electrode.

SEM images. (a) Side view of pillars; (b) side view of microneedles.

the inducting coupling plasma (ICP) etching and the “BOSCH” process were used [2]. The selectivity between silicon and silicon oxide can be as high as 300:1. The height and diameter of the completed pillars were 80 µm and 50 µm respectively. The SEM image of pillars is shown in Figure 3(a). (5) After the dry etching, a isotopic etching solution (HF:HNO3 :CH3 COOH =1:8:3) [19] was adopted to sharpen the tops of the pillars. Etching of silicon in the solution proceeds by sequential oxidation and followed by dissolution process. The overall reaction is Si+HNO3 +6HF−→H2 SiF6 +HNO2 +H2 O+H2 . The procedure of wet etching must be carried out in a vibration-free environment. The etch rate is greater at the top of the pillars because the activity of the etching solution at the base of the pillars is reduced, and little fluid motion is present to replenish it [20]. The result is shown in Figure 3(b). The tips of the microneedles were sharp enough to penetrate the skin, and no fracture was present after removal from the skin. Even if any microneedles were broken and remained in the skin, it would not cause infection, because skin self-renews and broken microneedles would be eliminated after a period of time. Finally, 50/300 nanometer Ti/Au was sputtered on the microneedles and the backside of the wafers. Then the silicon wafer was divided into several pieces 5×5 mm2 in size. Figure 4 is the photo of the dry electrode. 2.2

Combination of mechanical dicing and chemical wet etching

Fabrication of microneedle by KOH wet etching has been reported in [10]. The microneedles had low density. In our method, the density depends on the interval of dicing, so it can have high-density of microneedle arrays. The fabrication processes are as follows: (1) prepare and clean the wafer (reported in subsection 2.1); (2) thermally oxidize the silicon wafer to form the chemical wet etching mask; (3) transform the wafer into rectangular columns by making deep orthogonal cuts using an automatic dicing

Wang Y, et al.

2438

Figure 4

Figure 5

Figure 6

Sci China Inf Sci

November 2011 Vol. 54 No. 11

(a) Photo of dry electrode; (b) details of dry electrode.

(a)Top view of underetching process in theory; (b) top view of underetching process in experiment.

SEM image of microneedle arrays fabricated by combination of mechanical dicing and chemical wet etching.

saw. (The abrasive disc was 100 µm in thickness. The cutting width was 110 µm when the above disc was used. The rectangular column was 100 µm in width. The cutting depth was 80 µm); (4) after mechanical dicing, use 34.0 wt% KOH to etch the columns [21]. Figure 5(a) shows the top view of the columns shape during KOH wet etching process [22]. And the corresponding experimental micrographs are shown in Figure 5(b). At the beginning, the undercutting of convex corners takes place, which reduces the diagonal but not the lateral of the polygon [23]. After about 20 minutes the lateral width starts to decrease. When the diagonal width is close to 70 µm, the etching is stopped. KOH wet etching makes the columns smoother. Finally, isotropic wet etching is used to sharpen the columns. The solution in this experiment is an isotopic etching solution (HF:HNO3 :CH3 COOH =1:8:3). The result is shown in Figure 6. The next steps are metal sputtering and silicon cleavage. 2.3

Discussion

The first method includes deep reactive ion etching (DRIE) and HNA(HF+HNO3 +CH3 COOH)wet etch-

Wang Y, et al.

Figure 7

Sci China Inf Sci

November 2011 Vol. 54 No. 11

2439

Measurement set-up for recording electrode-skin-electrode impedance.

ing. The parameters of DRIE are easy to accurately control, and are repeatable and reliable. The equipment of DRIE is expensive, so the production cost of dry electrode is high in this way. The second method consists of mechanical dicing and two-step wet etching process. Mechanical dicing is easy to operate. The two-step wet etching processes include KOH wet etching and HNA wet etching. In the process of KOH wet etching, the temperature of solution and etching time need to be strictly controlled. This method does not require expensive equipment, and thereby greatly reduce production costs. ICP process makes the shape of microneedle look like a pin: the top is very sharp and other parts have small difference. Mechanical dicing produces mountain-shaped microneedles.

3

Testing

Two tests have been performed to verify the efficiency of the dry electrode for extraction of biosignals. The first test is impedance measurement. The second test is extraction of EEG. In the tests, the dry electrodes required no skin preparation and electrolytic gel. The commercial electrode need skin preparation and electrolytic gel. 3.1

Impedance measurement

The measurement setup for evaluating electrode-skin-electrode impedance is shown in Figure 7. Two electrodes are required. They are adjacent to each other on the voltmeter’s forearm at a separation of 10 cm, center to center. The electrodes are connected with an LCR meter (Agilent 4284A). Measurement frequency range is 20 Hz–10 kHz. The result is shown in Figure 8. In order to compare the impedances of dry electrode and commercial ECG electrode, the same test has been carried out with commercial ECG electrode. The commercial ECG electrode is a kind of silver/silver chloride electrode. The commercial ECG electrode surface is covered with electrolytic gel. Electrolytic gel contains conductive ions. Conductive ions can penetrate into the stratum corneum to increase its conductivity. But the electrolytic gel may cause skin allergy. The effective area of commercial ECG electrode is 1 cm2 . The dry electrodes require no skin preparation and electrolytic gel. The effective area of dry electrode is 0.25 cm2 . The result (Figure 8) shows that the impedances of the dry electrodes are lower than those of the commercial ECG electrode. To ensure the reliability of data, the experiments lasted for 30 minutes. 3.2

EEG recording

Raw EEG: The EEG signals were measured using dry electrode and compared with a commercial EEG electrode. The dry electrode was positioned at the O1 position of the international 10/20 system. The commercial EEG electrode was placed alongside the dry electrode for simultaneous measurement. The electrodes were referenced to the left and right ear lobes. The result is illustrated in Figure 9. The upper trace was measured with the dry electrode and the lower trace was measured with the commercial EEG electrode. The commercial EEG electrode is a kind of silver/silver chloride electrode. Electrolytic

Wang Y, et al.

2440

Figure 8

Figure 9

Sci China Inf Sci

November 2011 Vol. 54 No. 11

Impedance of the dry electrode and commercial ECG electrode according to the frequency changes.

EEG signals measured by dry electrode and commercial EEG electrode. The upper trace was measured with

the dry electrode and the lower trace was measured with the commercial EEG electrode. The two traces were essentially the same.

Figure 10

Temporal wave (a) and frequency spectrum of SSVEP (b) induced by a 15 Hz flash. The fundamental and

second harmonics can be clearly identified at 15 Hz and 30 Hz.

gel was injected into the electrode from the hole before the test. The effective area of commercial EEG electrode is 1 cm2 . The two recordings are very similar, showing that the dry electrode is suitable for EEG measurement. Steady-state visual evoked potential (SSVEP): When a person focuses on a stimulus that is continuously flashing at a frequency of 6 Hz or above, an SSVEP can be evoked at the visual cortex in most people [24]. The dry electrode was placed at the O1 position for detection of SSVEP. A flashing lightemitting diode (LED) modulated by a square wave was used as the stimulus. The volunteer stared at

Wang Y, et al.

Sci China Inf Sci

November 2011 Vol. 54 No. 11

2441

the LED, which was flashing at a given frequency for 30 seconds and the SSVEP were recorded during this time. Signals were pre-processed with a 50 Hz notch filter and a 4–50 Hz band-pass filter. Figure 10 shows the SSVEP induced by 15 Hz stimulation as a typical example of the temporal wave and frequency spectrum. The fundamental and second harmonics can be clearly identified at 15 Hz and 30 Hz.

4

Conclusions

In this work, the high aspect ratio needle structure was tested on (100) single crystal silicon by two methods. Dry etching combined with isotropic wet etching was easy to control. Mechanical dicing combined with chemical etching was cost saving. The impedance test and EEG experiment demonstrate the efficiency of the dry electrode. Steady-state visual evoked potential was measured. It is proved that dry electrode can record high quality EEG signals. Thus, the feature component can be resolved out effectively. The signal is useful in the field of brain-computer interface. Dry electrode fabrication and biological experiments were performed successfully, but more work is still needed on biocompatibility and mechanical property testing.

Acknowledgements This work was supported by the National Natural Science Foundation of China (Grant Nos. 60776024, 60877035, 90820002), and the National High Technology Research and Development Program of China (Grant Nos. 2007AA03Z427, 2007AA04Z329, 2007AA04Z254).

References 1 Matteucci M, Carabalona R, Casella M, et al. Micropatterned dry electrodes for brain-computer interface. Microelectron Eng, 2007, 84: 1737–1740 2 Griss P, Enoksson P, Tolvanen-Laakso H K, et al. Micromachined electrodes for biopotential measurements. J Microelectromech S, 2001, 10: 10–16 3 Griss P, Tolvanen-Laakso H K, Merilainen P, et al. Characterization of micromachined spiked biopotential electrodes. IEEE Trans Bio-med Eng, 2002, 49: 597–604 4 Baek J, An J, Choi J, et al. Flexible polymeric dry electrodes for the long-term monitoring of ECG. Sensor Actuat A: Phys, 2008, 143: 423–429 5 Yu L M, Tay F, Guo D G, et al. A microfabricated electrode with hollow microneedles for ECG measurement. Sensor Actuat A: Phys, 2009, 151: 17–22 6 Henry S, McAllister D V, Allen M G, et al. Micromachined needles for the transdermal delivery of drugs. In: Annual of International Conference on IEEE Microelectromech System. Germany: Heidelberg, 1998. 494-498 7 Ruffini G, Dunne S, Farres E, et al. A dry electrophysiology electrode using CNT arrays. Sensor Actuat A: Phys, 2006, 132: 34–41 8 Mukerjee E, Collins S, Isseroff R, et al. Microneedle array for transdermal biological fluid extraction and in situ analysis. Sensor Actuat A: Phys, 2004, 114: 267–275 9 Rajaraman S, Henderson H T. A unique fabrication approach for microneedles using coherent porous silicon technology. Sensor Actuat B: Chem, 2005, 105: 443–448 10 Wilke N, Hibert C, O’Brien J, et al. Silicon microneedle electrode array with temperature monitoring for electroporation. Sensor Actuat A: Phys, 2005, 123-124: 319–325 11 Wilke N, Mulcahy A, Ye S R, et al. Process optimization and characterization of silicon microneedles fabricated by wet etch technology. Microelectron J, 2005, 36: 650–656 12 Chandrasekaran S, Frazier A B. Characterization of surface micromachined metallic microneedles. J Microelectromech S, 2003, 12: 289–295 13 Parker E R, Rao M P, Turner K L, et al. Bulk micromachined titanium microneedles. J Microelectromech S, 2007, 16: 289–295 14 Aoyagi S, Izumi H, Isono Y, et al. Laser fabrication of high aspect ratio thin holes on biodegradable polymer and its application to a microneedle. Sensor Actuat A: Phys, 2007, 139: 293–302 15 Han M, Kim D K, Kang S H, et al. Improvement inantigen-delivery using fabrication of a grooves-embedded microneedle array. Sensor Actuat B: Chem, 2009, 137: 274–280

2442

Wang Y, et al.

Sci China Inf Sci

November 2011 Vol. 54 No. 11

16 Moon S J, Lee S S, Lee H S, et al. Fabrication of microneedle array using LIGA and hot embossing process. Microsyst Technol, 2005, 11: 311–318 17 Park J H, Yoon Y K, Choi S O, et al. Tapered conical polymer microneedles fabricated using an integrated lens technique for transdermal drug delivery. IEEE Trans Bio-med Eng, 2007, 54: 903–913 18 Ladenburger A, Reiser A, Konle J, et al. Regular silicon pillars and dichroic filters produced via particle-imprinted membranes. J Appl Phys, 2007, 101: 034302–034302-5 19 Norazreen A A, Muhamad R B, Burhanuddin Y M. Process characterization of wet etching for high aspect ration microneedles development. Adv Mater Res, 2009, 74: 341–344 20 Campbell P K, Jones K E, Huber R J, et al. A silicon-based, 3-dimensional neural interface: manufacturing processes for an intracortical electrode array. IEEE Trans Bio-med Eng, 1991, 38: 758–768 21 Shikida M, Hasada T, Sato K. Fabrication of a hollow needle structure by dicing, wet etching and metal deposition. J Micromech Microeng, 2006, 16: 2230–2239 22 Resnik D, Vrtacnik D, Aljancic U, et al. Different aspect ratio pyramidal tips obtained by wet etching of (100) and (111) silicon. Microchem J, 2003, 34: 591–593 23 Wilke N, Morrissey A. Silicon microneedle formation using modified mask designs based on convex corner undercut. J Micromech Microeng, 2007, 17: 238–244 24 Cheng X, Gao X, Gao S K, et al. Design and implementation of a brain-computer interface with high transfer rates. IEEE Trans Bio-med Eng, 200, 49: 1181–1186