Skip to main content
NCBI home page
ACS AuthorChoice logoLink to ACS AuthorChoice
. 2023 Jun 7;15(24):28954–28963. doi: 10.1021/acsami.3c03724

Magnetic Manipulation of Locomotive Liquid Electrodes for Wireless Active Cardiac Monitoring

Sumin Kim †, Young-Geun Park †,⊥, Ju-Young Kim ‡,§, Enji Kim †, Dong Ha Lee †, Jae-Hyun Lee ‡,§,*, Jinwoo Cheon ‡,§,∥,*, Jang-Ung Park †,‡,§,⊥,*
PMCID: PMC10288434  PMID: 37283562

Abstract

graphic file with name am3c03724_0007.jpg

For electrocardiogram (ECG) detection, the position of conventional patch-type electrodes based on solid-state metals are difficult to manipulate after attachment and also can lead to poor interface with stretchable, rough skin surfaces. Herein, we present a liquid form of ECG electrodes that can be magnetically reconfigured on human skin by providing its conformal interfacing. These electrodes consist of biocompatible liquid-metal droplets where magnetic particles are homogeneously dispersed, and their conformal contact with skin can yield significantly low impedance as well as high signal-to-noise ratio of ECG peaks. These electrodes are also capable of complex motions such as linear movements, splitting, and merging under external magnetic fields. Furthermore, magnetic manipulation of each electrode position on human skin enables precise monitoring of ECG signals with the change in ECG vectors. The integration of liquid-state electrodes with electronic circuitry demonstrates wireless and continuous ECG monitoring while magnetically moving this entire system on human skin.

Keywords: bioelectronics, liquid metal, electrocardiogram, wireless system, magnetic manipulation

Introduction

Detection of cardiovascular diseases (CVD) needs to be as early as possible, and the late detection of CVD is a major cause of alarmingly high death rates. Non-invasive devices for proactive diagnosis and monitoring of cardiovascular parameters enable early diagnosis of CVDs and protect patients’ lives. In particular, an electrocardiogram (ECG) is an essential diagnostic method for various abnormal heart conditions, being a cornerstone for early detection of CVD.

In standard clinical ECG systems, Ag/AgCl solid gel electrodes with a disk shape are directly attached to the patient’s skin with wet electrolyte gels as conductive adhesives. Though these electrodes can be easily found in hospital and home health monitoring systems, several drawbacks have led to erroneous interpretation, potentially leading to mistreatment and discomfort to patients.1,2 Such measurements using unstretchable and flat geometries of solid electrodes are occasionally insensitive and significantly prone to noise, due to high impedance at their poor interfaces with stretchable and rough skin surfaces.3 Therefore, they typically require electrolytic conductive gels and skin preparation to reduce the impedance and improve signals. However, in prolonged usage, the conductive gels may cause allergic reactions or skin irritation to patients, and dehydration of gels can dry out and destabilize the electrodes, reducing signal conduction and compromising the data acquired. In addition, incorrect placement of an ECG lead (i.e., a pair of ECG electrodes) may also result in signal inaccuracies of more than 20%, leading to false diagnosis of CVD.4 Although the typical 12-lead ECG system records cardiac signals at multiple different locations to exclude erroneous recordings, repeated peel-offs of these leads to find correct locations can induce a decrease in adhesion, severely degrade sensitivity, and prompt inevitable skin irritation.

To overcome these limitations of wet electrodes (which require conductive gels), the development of dry electrodes with improved signals and accuracy has been demonstrated, based on conductive5 and semiconductor materials6 fabricated by microfabrication technology. However, most of them are made of either stiff or solid substrates with a high elastic modulus causing poor interface.7−9 Besides, liquid metal (LM)-based electrodes are of rising interest for bio-interfacing electrodes. Ga-based LM and its alloys, including a eutectic gallium–indium (EGaIn) alloy, are considered to be a suitable material due to its low elastic modulus (∼200 kPa), high conductivity,10−13 stretchability,14,15 and biocompatibility.14,16,17 Unfortunately, due to the fluidity of LMs, the task of accurate positioning and fixation of the LM electrodes on the surface of a living matter for monitoring vital physiological signals still remains a major challenge. Herein, we demonstrate LM-based ECG electrodes with precise control of their positioning as well as improved skin–electrode interfaces. To do this, we introduced magnetic nanoparticles (Fe3O4 NPs) into LM to synthesize a homogeneous dispersion of Fe3O4 NPs in a liquid form of EGaIn. Although similar methods of mixing magnetic particles (e.g., iron and nickel) with LMs (e.g., galinstan) have been previously studied,18,19 our approach focuses on the application of this Fe3O4/LM composite to ECG monitoring via its magnetic movement on human skin, rather than the synthesis. Both Fe3O4 and EGaIn are biocompatible, and our study represents the first demonstration of this biocompatible composite for wireless ECG monitoring using its magnetic control on the human skin. Despite the toxicity of iron and nickel, the negligible toxicity of Fe3O4 makes our composite relatively biocompatible and safe to use without skin irritations, and this composite (as a fluid) can be magnetically manipulated along complex trajectories under external magnetic fields and located at desired positions. The human pilot experiment as well as in vivo tests conducted using live mice have verified that the use of this composite for ECG recordings does not degrade the signal quality of ECG by its magnetic manipulation and exhibits stable signal-to-noise ratio (SNR) even after multiple electrode positions are shifted on the skin surface.

Results and Discussion

Synthesis and Characterization of Fe3O4/LM Composites

The synthesis process of this Fe3O4/LM composite is schematically illustrated in Figure 1a. A binary eutectic alloy of gallium and indium (EGaIn, 75.5 wt % Ga and 24.5 wt % In, Changsha Santech Materials Co., Ltd) and Fe3O4 NPs were prepared in a beaker with a mass ratio of 1:0.4. Sodium hydroxide (NaOH, 30 wt %) was added to temporarily remove the thin oxide layer at the surface of EGaIn, allowing Fe3O4 NPs to be suspended into EGaIn.20,21 Then, this mixture was mechanically stirred with a glass rod for 20 min until Fe3O4 NPs were uniformly dispersed. Finally, the mixture was immersed in a solution of hydrochloric acid (HCl) for 10 min for pH neutralization at 7. The resulting Fe3O4/LM composite was readily oxidized in air, re-forming a thin oxide layer (thickness: ∼1 nm) to keep a solid shape while maintaining the fluidity.

Figure 1.

Figure 1

Open in a new tab

Synthesis of the Fe3O4/LM composite and its characterization. (a) Schematic illustration of preparation of the Fe3O4/LM composite. (b) Photograph of the pristine LM. Scale bar, 1 mm. (c) SEM image of the pristine LM’s surface. Scale bar, 1 μm. (d) Photograph of the Fe3O4/LM composite. Scale bar, 1 mm. (e) SEM image of the Fe3O4/LM composite’s surface. Scale bar, 1 μm. (f) Impedance spectroscopy of the pristine LM and the Fe3O4/LM composite dependent on the content of Fe3O4 NPs. (g) The impedance (at 1 kHz) of the pristine LM and the Fe3O4/LM composite with different contents of Fe3O4 NPs. (h) Magnetization–magnetic field (M–H) curves of the pristine LM and Fe3O4/LM composite dependent on the content of Fe3O4 NPs.

Optical images and scanning electron micrograph (SEM) images in Figure 1b–e show the differences in the surface morphology of the pristine LM and the Fe3O4/LM composite. As shown in Figure 1b, the pristine LM had a very smooth and glossy surface. Wrinkles in the natural oxide layer of the pristine LM surface, known as Ga2O3, were observed in the SEM images (Figure 1c). On the other hand, the surface of the Fe3O4/LM composite was roughened by the dispersion of Fe3O4 NPs, resulting in less gloss compared to the pristine LM (Figure 1d). The microstructure of the surface shown in Figure 1e indicates the relatively uniform dispersion of Fe3O4 NPs in the LM matrix, as shown in the energy-dispersive X-ray spectroscopy (EDS) (Figure S1).

Properties of Fe3O4/LM Composites

In order to develop a highly sensitive LM-based ECG electrode with good interface with the skin as well as precise motion control by magnetic fields, it is necessary to preserve the intrinsic properties of LM by maintaining its magnetic response. Although the introduction of large amounts of Fe3O4 NPs into the LM matrix can improve the magnetic response, this can degrade LM’s electrical conductivity and impedance property for ECG recording. Therefore, the amount of Fe3O4 NP was optimized to preserve both the electrical and magnetic properties of this Fe3O4/LM composite (Figure 1f–h). To measure impedance, the pristine LM and the Fe3O4/LM composites having 5, 10, 15, and 20 vol % of Fe3O4 NPs were disposed in a syringe with a diameter of 2 mm, and a rod-type Ag/AgCl electrode was used as a reference electrode. Electrochemical impedance spectra were acquired with the multichannel potentiostat in PBS 0.1 M. The pristine LM exhibited an impedance of 2.462 kΩ. Although the impedance slightly increased to 3.227 kΩ at 1 kHz on increasing the content of Fe3O4 NPs to 20 vol %, this increase in impedance was not significant (within ∼ 30%), as shown in Figure 1f,g. In particular, these impedance values of our Fe3O4/LM composites were about two times lower than the impedance of the standard Ag/AgCl gel electrode (∼105 Ω),22,23 which can be advantageous for ECG recording.

The magnetic properties of the pristine LM and Fe3O4/LM composites were characterized using vibrating sample magnetometry (VSM). Figure 1h shows the magnetization versus magnetic field (M–H) curves of the Fe3O4/LM composites with different contents of Fe3O4 NPs. The pristine LM denoted by the black line exhibited non-magnetic behavior with a saturation magnetization (Ms) value close to 0 emu g–1, whereas Ms values of the Fe3O4/LM composites increased with the content of Fe3O4 NPs. For example, the Ms values were 0.01, 3.446, 7.304, 10.232, and 12.576 emu g–1 for 0, 5, 10, 15, and 20 vol % of Fe3O4 NPs, respectively. Compared to the magnetic property of pure Fe3O4 NPs, the Fe3O4/LM composites showed lower Ms values due to the non-magnetic volume of pristine LM, while the coercivity was negligibly changed (Figures S3 and S4).

Magnetic Manipulation of Fe3O4/LM Composites

Figure 2 demonstrates the control of various motions of the Fe3O4/LM composites by magnetic manipulation. In this study, the composite with 10 vol % of Fe3O4 NPs was selected for its relatively good magnetic response and low impedance as well as its fluidic characteristics. In addition, an external magnetic field was applied using a neodymium–iron–boron magnet (NdFeB). In comparison with the absence of a magnetic field, apparently different behaviors of this composite were observed in the presence of the magnetic field (Figure 2a–d). When a droplet of this composite (volume: 0.2 mL) was placed on the untilted acrylic plate (thickness: 3 mm) without applying a magnetic field, an initial contact angle of 128° was observed. Effects of the Fe3O4 ratio on the contact angles of the composites are summarized in Table S1. The composites with higher Fe3O4 contents showed a lower contact angle. Then this droplet started rolling down due to gravity when this plate was tilted more than 35° (Figure 2a,b). When the gradient magnetic field was applied across the untilted acrylic plate (by locating the magnet under this plate), the composite was deformed into a compressed hemispherical shape so that the composite and the plate were in close contact, increasing the contact angle up to ∼138° (Figure 2c). Moreover, this close contact was maintained even when the composite was vertically positioned (by tilting the acrylic plate to 90°) without this droplet rolling down (Figure 2d). When the magnet was placed on the opposite side of this Fe3O4/LM composite across the vertically tilted acrylic plate, the magnetic force was strong enough to hold or move this composite droplet upward against gravity. When the magnetic field was not applied, the Fe3O4 NPs did not exhibit magnetic dipole moments and were homogeneously dispersed in the LM matrix. When the magnet was drawn near the composite, however, the non-uniform gradient magnetic field induced magnetic dipolar interactions, attracting the NPs toward each other to form sparse clusters within the LM matrix. The clusters then dispersed inside this composite, distorting its shape from the inside, and the composite started to slide following the motion of the magnet. The schematic illustration in Figure 2e depicts the force balance on the droplet of the Fe3O4/LM composite vertically moving along the magnet. Here, the total force balance in the horizontal and vertical directions can be written as24

graphic file with name am3c03724_m001.jpg 1
graphic file with name am3c03724_m002.jpg 2

where Fm is the magnetic attractive force applied to the composite toward the magnet, Fγ is the capillary force from the deformation of the composite, Fs is the retention force between the composite and the acrylic plate, and Fg is the gravitational force (see the Supporting Information for detailed explanation). If the magnetic force is the same as the net force associated with the deformation and contact area of the composite as well as the gravitational force, the composite can be positioned and then held at any position following the motion of magnet. Movie S3 shows the movement of the Fe3O4 NPs within the composite under a magnetic field. When the magnetic field was applied, the Fe3O4 NPs moved and pulled the EGaIn matrix along the direction of the field.

Figure 2.

Figure 2

Open in a new tab

Magnetic manipulation of the Fe3O4/LM composite. (a) Photograph of the Fe3O4/LM composite on an acrylic plate without applying magnetic fields. Scale bar, 5 mm. (b) Photographs of the Fe3O4/LM composite rolling down the inclined plane at a critical angle (35°), with no magnetic field. Scale bars, 5 mm. (c) Photograph of the Fe3O4/LM composite on an acrylic plate under the application of a magnetic field. Scale bar, 5 mm. (d) Photograph of the vertically positioned Fe3O4/LM droplet without its rolling down under the applied magnetic field. Scale bar, 5 mm. (e) Schematic illustration on the movement of the Fe3O4/LM droplet against gravity at the tilted angle of 90° when moving a magnet. (f–h) Time-lapse photographs of various motions using a magnet. Scale bars, 1 cm. (i) Photographs of splitting and merging by applying magnetic fields. Scale bars, 5 mm.

We further demonstrated a series of complex manipulations of this Fe3O4/LM composite. A droplet of this Fe3O4/LM composite (volume: 0.2 mL, diameter: 1 cm) was placed on an acrylic plate, and a magnet was positioned under the plate to induce a gradient magnetic field. The Fe3O4/LM composite was moved in linear, rectangular, and circular trajectories (Figure 2f–h) at a rate of 0.1 m s–1 along the trajectory of the magnet. In addition, the magnetic response was fast enough to achieve a magnet-synchronized vibration motion. Each movement of this Fe3O4/LM composite is shown in Movie S1. Furthermore, a single droplet could be separated into multiple droplets by introducing multiple magnets aligned with the same polarity (Figure 2i). This magnetically induced splitting occurred when the magnetic attractive force of Fe3O4 NPs dispersed in the LM matrix was large enough to overcome the surface tension and adhesive force of the LM to be separated in each direction. After this splitting, all droplets were also magnetically controllable due to the uniform dispersion of Fe3O4 NPs in the LM matrix during their actuation and splitting. When the droplet was withdrawn to its original location, all split droplets were merged into a single droplet again (Movie S2), indicating that multiple droplets as highly sensitive ECG electrodes can be produced from a single droplet.

Biocompatibility of Fe3O4/LM Composites

To detect and diagnose CVD, long-term monitoring of ECG such as a Holter monitor can be essential for patients.25 Hence, the attachment of ECG electrodes should be compatible with human skin without severe irritation or allergic reactions during the long-term operation for reliable collection of ECG data. Cytotoxicity of the Fe3O4/LM composite for epidermal cells and human skin was evaluated, as shown in Figure 3. Human epidermal cells (CCD-986sk) were used to test the cell viability for our composite sample. For this evaluation, the cells were cultured for 24 h in either a mixture of the culture medium and pristine LM, or a mixture of the culture medium and our Fe3O4/LM composite. In addition, the cells were cultured in the pristine culture medium as a control, and in a mixture of the culture medium and a medical ultrasound gel (1 vol %) as a negative control. Figure 3a shows fluorescent and bright-field micrographs of the cultured CCD-986sk cells. Each cultured group showed good viability comparable to the control group case in terms of the area occupied by live and dead cells. According to the bright-field micrographs, there was no significant difference in cell morphology between these groups. Figure 3b shows the MTT assay results for each group, with the viability of the ethyl alcohol-treated group (positive control), medical ultrasound gel (negative control), cell medium (positive control), the pristine LM, and the Fe3O4/LM composite being 40 ± 1.28, 86 ± 2.59, 100 ± 3.40, 97 ± 2.72, and 96 ± 1.42%, respectively. The experiments were performed three times with a sample size of 10 per each assay. The cell viability after culturing with the Fe3O4/LM composite was higher than that with the medical ultrasound gel, ensuring that our composite is negligibly toxic.

Figure 3.

Figure 3

Open in a new tab

Biocompatibility of the Fe3O4/LM composite. (a) The calcein-AM staining fluorescent and bright-field micrographs of human fibroblast cells (CCD-986sk) for control (DMEM), negative control (medical gel), the pristine LM, and the Fe3O4/LM composite after 24 h culture. Scale bars, 100 μm. (b) MTT assay of human fibroblast cells (CCD-986sk) for the positive control (ethyl alcohol), negative control (medical gel), positive control (DMEM), the pristine LM, and the Fe3O4/LM composite after 24 h culture. ***p ≤ 0.001 and comparison of the test group with the positive control (70% ethanol). A t-test was used to calculate the statistical significance (Benjamini–Hochberg adjusted p-values). *p ≤ 0.05; **p ≤ 0.01; ***p ≤ 0.001; ****p ≤ 0.0001. (c) Skin irritation test with the standard Ag/AgCl gel electrode (upper) and the Fe3O4/LM composite (bottom) after 0 day (no attachment) and 3 and 7 days, respectively. Scale bars, 1 cm.

For the long-term monitoring of electrophysiological signals, a skin irritation test was also performed by placing this Fe3O4/LM composite on the human skin for a specific period of time (Figure 3c). The standard Ag/AgCl gel electrode (control) and this Fe3O4/LM composite were attached to the skin for 3 and 7 days during the experiment. For the complete fixation of this Fe3O4/LM composite, a transparent film dressing was attached over the composite and skin area for a week (Figure S5). As shown in Figure 3c, the standard Ag/AgCl gel electrode caused urticaria and tingling along with its adhesion marks. On the other hand, our Fe3O4/LM composite sample presented negligible inflammation or irritation even after 7 days of the attachment.

ECG Monitoring with Good Interface to Skin

We conducted an in vivo experiment in a mouse model to demonstrate high SNR of this Fe3O4/LM composite electrode for ECG recording after the change in its position following Einthoven’s triangle. In Figure 4a, three empty circles represent the position of ECG-recording electrodes, and R circled on the right hind paw indicates the position of a reference electrode. This experiment was conducted by applying the Fe3O4/LM composite (100 mg) to the skin surface of Hos:HR-1 hairless mice. Figure 4b–d illustrates photographs in which the Fe3O4/LM composite electrode was magnetically split into the positions of lead I, II, and III. The negative, positive, and reference electrodes of the Fe3O4/LM composite were connected to blue, red, and yellow wires, respectively. These electrodes were moved and relocated on the skin surface using a magnet moving from lead I to lead II (Figure 4e), and the wires were adjusted for each electrode to connect with the ECG measurement system. To compare the ECG waves recorded using a conventional electrode, a patch-type standard Ag/AgCl gel electrode with a diameter of 25 mm was used as a control. Figure 4f shows ECG waves recorded at three different locations using both Ag/AgCl gel electrodes and our Fe3O4/LM composite electrodes. Compared to signals from Ag/AgCl gel electrodes, higher amplitude with distinct QRS complexes were clearly observed in signals from the Fe3O4/LM composite electrodes. Especially in lead III, according to Einthoven’s triangle, the sum of ECG amplitudes in positive and negative directions should be close to zero, which was not found in the Ag/AgCl gel electrode cases. The quality of the ECG measured on Ag/AgCl electrodes with a large contact area (491 mm2) was insufficient to distinguish the ECG signal for each lead in the small animal. In addition, the flat and rigid structure of this solid Ag/AgCl gel electrode caused its unstable contact on the nonplanar and stretchable skin surface in the presence of wrinkles or bumps, whereas the liquid and soft characteristics of our Fe3O4/LM composite electrode can reconfigure its shape to enhance its contact properties with the skin significantly (Figure 4g). For a more quantitative analysis, SNR was calculated using the following equation

graphic file with name am3c03724_m003.jpg 3

Here, each SNR value was calculated as 25.59 ± 1.26, 28.47 ± 1.18, and 26.27 ± 2.01 dB for the three lead locations of our Fe3O4/LM composite electrode, and 21.81 ± 1.29, 23.53 ± 1.71, and 20.85 ± 1.36 dB for the Ag/AgCl gel electrode, respectively (Figure 4h). The SNR values measured using our Fe3O4/LM composite electrode were higher in all leads and maintained relatively constant as the location of the lead was changed. However, despite the large contact area, the SNR values of the Ag/AgCl gel electrode cases were lower than those of the Fe3O4/LM composite. This is not solely due to unstable contact but also to the distance between the measuring electrodes. When the distance between the measuring electrodes is large enough, a potential difference is clearly observed. In the case of the Ag/AgCl electrode with a large surface area, the distance between the measuring electrodes was too close, and as a result, the overall ECG amplitudes were reduced since the potential difference was small.

Figure 4.

Figure 4

Open in a new tab

In vivo mouse experiment. (a) Schematic illustration of bipolar limb leads and each location in a mouse. (b–d) Photographs of the lead location during ECG monitoring in lead I, II, and III, respectively. Blue, red, and yellow interconnect wires represent negative, positive, and reference electrodes. Scale bars, 2 cm. (e) Photographs demonstrating magnetic manipulation of the Fe3O4/LM electrode on the surface of the mouse skin. Scale bars, 3 cm. (f) Representative ECG amplitude traces at each lead measured from the standard Ag/AgCl gel electrode (black) and the Fe3O4/LM composite (blue). (g) Photographs of the standard Ag/AgCl gel electrode (top) and the Fe3O4/LM composite (bottom) on the surface of the mouse skin. Scale bars, 1 cm. (h) The comparison in SNR of the standard Ag/AgCl gel electrode case (black) and the Fe3O4/LM composite case (blue) at lead I, II, and III. Data are presented as the value ± s.d. (i) The comparison in SNR per contact area of the Ag/AgCl gel electrode case (black) and the Fe3O4/LM composite case (blue) at lead I, II, and III.

The impedance decreases with an increase in the contact area of the electrode to skin, which can lead to an increase in the SNR.7 In our experiment, the Ag/AgCl gel electrode with a larger size (area: 491 mm2) than the Fe3O4/LM composite electrode case (area: 19.6 mm2) was used. Figure 4i shows the SNR values of these two electrodes normalized with their sizes. These normalized SNR values recorded using the Ag/AgCl gel electrode in lead I, II, and III were 0.04, 0.05, and 0.04 dB/mm2, whereas the normalized values of the Fe3O4/LM electrode cases were 1.31, 1.44, and 1.34 dB/mm2, respectively. It is noteworthy that the SNR of the liquid-electrode cases was at least 28 times larger for all leads. Based on these results, using the liquid-type magnetic composite as the electrode for ECG monitoring can be a promising alternative to the conventional solid-type electrode for targeting various sizes and sites of nonplanar and stretchable skin surfaces.

Human ECG Monitoring with Fe3O4/LM Composites

The Fe3O4/LM composite electrode was further demonstrated for human ECG monitoring, and the signals recorded at different locations of the human skin were compared. Due to the presence of an oxide layer of the LM, the composite can be adhered to various surfaces.26Movie S4 shows the adhesion of the composite to human skin, as it maintained its position on skin when the arm flipped. This adhesion of the composite to human skin enables the stable measurement of ECG on human skin.

The first location of the positive electrode was the human wrist (Figure 5a). After the first ECG monitoring on the wrist, the Fe3O4/LM composite electrode was magnetically moved and relocated to the upper arm for recording the second ECG (Figure 5b). Since the two different locations of positive electrode were classified as lead I, similar ECG waveforms were observed. Each negative electrode and reference electrode were placed on the left wrist and the right ankle, respectively. ECG from two different locations were recorded, as shown in Figure 5c. Both ECG waveforms of the potential distributions were oriented in the same direction, denoted as lead I. The P wave, QRS complex, and T wave used in cardiac condition analysis were clearly distinguished in both locations, as shown in Figure 5d. Although the overall waveform was similar, it is worth pointing out that the ECG signal on the upper arm showed a specific elevation between the QRS complex and the T wave. This peak was only observed on the upper arm, which could be caused by muscle contraction when bending the arm, indicating that the Fe3O4/LM composite electrode was very sensitive to detect subtle physiological changes that were difficult to be detected by commercially available ECG electrodes. Moreover, these three representative waves remained within approximately 9% of the variation in amplitude after repositioning the Fe3O4/LM composite electrode on human skin, ensuring that the signal accuracy was preserved after its magnetic manipulation on long trajectories.

Figure 5.

Figure 5

Open in a new tab

Human pilot experiment and wireless ECG monitoring. (a,b) Photographs of the Fe3O4/LM composite for wireless ECG monitoring of a human male subject, placed on the wrist (a) and magnetically relocated on the upper arm region (b). Scale bars, 5 cm. (c) Representative amplitude traces of ECG detected on the wrist (red) and the upper arm (blue). (d) Magnification of an individual ECG wave on the wrist (red) and the upper arm (blue) to indicate each P wave, QRS complex, and T wave. (e) Comparison of the ECG amplitudes of P wave, QRS complex, and T wave recorded on the wrist (red) and the upper arm (blue). (f) Difference in the ECG amplitude of each wave (P wave, QRS complex, and T wave) between the wrist and the upper arm. (g) Entire schematic illustration (top) and side-view illustration (bottom) of the wireless ECG monitoring system composed of two liquid electrodes, integrated circuits, connector, and a battery. (h,i) Photographs of the entirely integrated system for wireless ECG monitoring of a human male subject, placed near the elbow (h) and magnetically relocated on the wrist (i). Scale bars, 3 cm. (j) Representative ECG amplitude traces monitored on the initial position (black line) and the magnetically relocated position (green line).

Wireless ECG Monitoring during Magnetic Manipulation

Wireless ECG monitoring would pave the way for achieving long-term applicability in medical interventions. As the magnetic manipulation of our liquid electrodes using the Fe3O4/LM composite enables untethered locomotion, we directly interconnected these liquid electrodes with wireless ECG circuits and a battery as an untethered robotic ECG platform. Figure 5g shows a schematic layout on the integration of a wireless Wi-Fi circuitry (JAGA Penny, JAGA Systems, Inc.), a battery, and two liquid electrodes (i.e. two droplets of the Fe3O4/LM composite) for ECG recording and a reference. Two liquid drops of this composite were separately positioned on skin (volume of each droplet: 0.2 mL, the spacing between two droplets: 2.24 cm), and then a flat plate (3 cm × 1 cm) of the wireless communication module (including a battery and a connector) was directly placed on these two drops through their electrical connection using two metal wires (length of shorter wire: 1 cm, length of longer wire: 3 cm) from the connector. More detailed schematic illustrations are shown in Figure S6. This entirely integrated system (including two liquid electrodes) for wireless ECG monitoring was lightweight (∼30.9 g in total) and narrower than approximately 3 cm in width. Also, this integrated system enabled the continuous monitoring of ECG signals while moving easily on human skin by applying an external magnetic field, as shown in Movie S3. For example, Figure 5h,i presents photographs of this magnetically manipulable system on a human male subject. Figure 5j exhibits the ECG waves detected wirelessly at two different locations during movement of this integrated system on the skin. In this measurement, we put this system near the elbow area initially (Figure 5h), and the measured ECG peak is shown on the left (black line) in Figure 5j. In the presence of an applied magnetic field, the liquid electrodes of the Fe3O4/LM composite dragged this entire system to the wrist area (Figure 5i), and the detected peak is plotted on the right (green line) in Figure 5j. There was a slight increase in overall noise and the generated noise was caused by vibration that occurred during magnetic manipulation (Figure S7). In addition, the amplitudes of all ECG peaks as well as the noise were increased, indicating an improvement in signal quality as the composite adhered to the skin through magnetic manipulation. As demonstrated in Figure S8, the composite can be magnetically squished to the skin surface, thereby increasing the contact area and ultimately improving the quality of the ECG signals. As shown in Movie S5, ECG signals were reliably monitored, and no significant increase in noise levels was observed during overall magnetic manipulation of this wireless system. Especially, P wave, QRS complex, and T wave were also clearly distinguished even during this magnetic manipulation. Additionally, no residue of the composite was left on the skin after these human experiments (Movie S6).

Conclusions

In this study, we developed a liquid state of ECG electrodes using the Fe3O4/LM composite that enables accurate position control by an external magnetic field while maintaining high sensitivity for ECG recording and showing outstanding skin–electrode interface. The liquid electrode was synthesized by dispersing Fe3O4 NPs in the pristine LM matrix. Although blending magnetic particles with LMs has been previously reported elsewhere, these previous studies have focused on the movement of the material itself or magnetic actuation techniques, resulting in a lack of biomedical applications.27−31 In contrast, our ECG electrode exhibited good electrical conductivity, magnetic responsibility, biocompatibility, and reliable wireless ECG monitoring even when continuously moving over human skin. Various modes of motion control, including nonlinear movement, vertical fixation, vibration, magnetically induced splitting, and merging, have been demonstrated under the gradient magnetic fields. Moreover, the biocompatibility of our ECG electrode was outstanding without causing skin irritation or allergic reactions, compared to the cases of conventional ECG electrodes. Remarkably, we have illustrated its capabilities as a promising alternative to the standard Ag/AgCl gel electrode by performing 28 times higher signal intensities in SNR for ECG recording. This performance demonstrates significant potential for a bench-to-bedside translation to further use in assisting physicians with more accurate diagnosis of CVD and treatment of patients. Furthermore, its long-term ECG monitoring (over 7 days), its integrity with electronic circuitry for wireless recording, and the magnetic movement of this entirely integrated system (i.e., the Fe3O4/LM electrode + wireless circuits) on human skin represent substantial progress toward its functionality as a soft robot and suggest the future promise of various electrophysiological recordings such as electromyogram, electroencephalogram, and electrocorticogram.

Materials and Methods

Materials

EGaIn (75.5% gallium, 24.5% indium alloy by weight; Changsha Santech Materials Co. Ltd.), and Fe3O4 NPs (70 nm diameter, 637106, Sigma-Aldrich) were used to fabricate the Fe3O4/LM composite. Also, 30 wt % NaOH aqueous solution was prepared by mixing NaOH pellets and distilled water. The N35 grade commercial NdFeB magnet (5 mm × 10 mm × 100 mm, Alibaba) with a magnetic flux density (B) on the surface of 200 mT was used for magnetic actuation of the Fe3O4/LM composite.

Electrochemical Characterization of the Fe3O4/LM Composite

For electrochemical property measurement, impedance measurements of each pristine LM and Fe3O4/LM composite were conducted in PBS solution (Sigma-Aldrich). All impedance measurements were performed over a frequency range of 0.1–100 kHz using the multichannel potentiostat (PMC-1000, AMETEK).

Magnetic Characterization

For quantitative analysis, the magnetic hysteresis loops are measured with a vibrating sample magnetometer (VSM, 7404-S, Lake Shore Cryotronics) for all samples. Since each Fe3O4/LM composite was in a liquid state, an appropriate amount (∼100 mg) was trapped in the cartridge container and fixed to a sample holder to prevent unintended vibration.

Cell Viability Test

For the cell viability test (n = 10), human fibroblast cells (CCD-986sk, ATCC) were cultured in 88% DMEM/F-12 (Gibco) containing 10% certified fetal bovine serum (Biological Industries) and 2% Penicillin–Streptomycin Solution (Corning). Then, the cells were cultured and plated at a concentration of 5 × 104 mL–1 in each well of 96-well plates. The CCD-986sk cells were immersed and cultivated by DMEM (control), medical ultrasound gel (negative control), pristine LM, and Fe3O4/LM composite for 24 h and stained by calcein AM (LIVE/DEAD Viability/Cytotoxicity Kit, Invitrogen) to detect the activity of the enzyme and the integrity of the plasma membrane. The culture medium was mixed with 1 vol % of medical ultrasound gel, pristine LM, and the composite, respectively, incubated for 24 h, and extracted for the test. The live cells were green under a fluorescence microscope excitation of 495 nm (Nikon ECLIPSE TS100). Likewise, the MTT assay was prepared with the CCD-986sk cells, which were seeded on 96-well plates (cell concentration: 5 × 104 mL–1) with 100 μL well–1 medium. After culturing the cells for 24 h, the original culture medium was removed, and 200 μL well–1 control eluate and sample eluate were added. The CCD-986sk cells were cultured for additional 24 h. After all, each well was treated with MTT reagent diluted tenfold with DMEM, for 2 h in a 37 °C CO2 incubator. The cytotoxicity was analyzed after measuring fluorescence in a 96-well plate reader (Victor X5, PerkinElmer) using a 570 nm filter. All experiments were performed at least three times.

In Vivo Experiment with Mouse

In vivo studies were conducted based on the guidelines of the National Institutes of Health for the care and use of laboratory animals and with the approval of the Institute of Animal Care and Use Committee of Yonsei University (IACUC-A-202103-1228-02). The Institute of Animal Care and Use Committee of Yonsei University was the ethics review committee.

In Vivo Experiment with Human

The protocol for this part of the study was approved by the Institutional Review Board of Yonsei University (7001988-202110-HR-1390-02). The Fe3O4/LM composite and commercial electrode were attached to both arms and ankles of the subject, respectively. Within 10 min, the composite was moved using a magnet, and ECG measurement was conducted. After the experiment, the skin was wiped with an alcohol swab (CareSens), and the immune response was monitored for 1 h.

ECG Measurements

The ECG waves of the mouse and human were recorded using a three-wire lead ECG system (PSL-DAQ RMSW, PhysioLab, Co., Korea). These three-lead wires were connected to Fe3O4/LM composites and commercial Ag/AgCl gel electrodes (Myoware, Advancer Technologies, U.S.A.) that were placed on the left, right arms (positive and ground), and femoral of the left leg (negative) at the skin of mouse and human, respectively. A separate ECG module with a signal filter and amplifier was connected to each wire. The signal amplification was set at 330 magnifications for mouse and 750 magnifications for human. Also, to acquire real-time ECG waves, the ECG module was connected to a data acquisition (DAQ) module, which was displayed through a LabVIEW-based customized software program. The mouse was sedated with isoflurane to prevent inadvertent vital sign measurement during the experiment.

Wireless ECG Monitoring

The wireless system consists of wireless circuits (JAGA Systems, Sunnyvale, CA), a rechargeable Li poly-ion battery, and two liquid electrodes of the Fe3O4/LM composite (i.e., a recording electrode and a reference). The high-pass filter was set at 0.5 Hz, and the low-pass filter was set at 150 Hz. The sampling rate was 1000 Hz. The wireless ECG system transmitted the data to the router of the external equipment (e.g., computer and laptop).

Acknowledgments

This work was supported by the Ministry of Science and ICT (MSIT), the Ministry of Trade, Industry and Energy (MOTIE), the Ministry of Health & Welfare, and the Ministry of Food and Drug Safety of Korea through the National Research Foundation (2023R1A2C2006257), Nano Material Technology Development Program (2021M3D1A2049914), ERC Program (2022R1A5A6000846, 2020R1A5A1019131), the Technology Innovation Program (20013621, Center for Super Critical Material Industrial Technology), and the Korea Medical Device Development Fund grant (RMS 2022-11-1209/KMDF RS-2022-00141392). The authors also thank the financial support by the Samsung Research Funding & Incubation Center of Samsung Electronics (SRFC-TC2003-03) and the Institute for Basic Science (IBS-R026-D1). In addition, Y.-G. P. thanks the Sejong Science Fellowship (2021R1C1C2008657).

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsami.3c03724.

  • Mechanical analysis for magnetic response of the Fe3O4/LM composite, changes in the contact angle of composites on glass substrate according to the ratio of Fe3O4 NPs, EDS analysis, photographs of the experimental setup for impedance measurement of the Fe3O4/LM composite, magnetization versus magnetic field (M–H) curves of pristine LM and Fe3O4/LM composites with different contents of Fe3O4 NPs, magnetization versus magnetic field (M–H) curve of Fe3O4 NPs, attachment of the Fe3O4/LM composite with transparent medical film for skin irritation test, a schematic layout of the wireless ECG platform using two liquid drops of the Fe3O4/LM composite, single ECG signal before and during magnetic manipulation, and photographs of the Fe3O4/LM composite (volume of the composite: 0.2 mL) on the skin before and after magnetic manipulation and ECG signals at that time (PDF)

  • Magnetic manipulation of the Fe3O4/LM composite (AVI)

  • Splitting and merging mode of the Fe3O4/LM composite (AVI)

  • Movement of Fe3O4 NPs within the Fe3O4/LM composite under a magnetic field (AVI)

  • Adhesion of the Fe3O4/LM composite on skin (AVI)

  • Real-time movement and ECG measurement of the wireless ECG platform (AVI)

  • Magnetic manipulation of the Fe3O4/LM composite on the skin (AVI)

Author Contributions

S. K., Y.-G.P., and J.-Y.K. authors contributed equally to this work. S.K. and Y.-G.P. contributed to the experiments, analyzed the data, and wrote the manuscript. J.-Y.K. characterized the magnetic properties and wrote the manuscript. E.K. and D.H.L. contributed to the wireless ECG experiments. J.-H.L. and J.C. contributed to the analysis of the Fe3O4/LM composite and revised the manuscript. J.-U.P. oversaw all the research phases and revised the manuscript. All the authors discussed and commented on the manuscript.

The authors declare no competing financial interest.

Supplementary Material

am3c03724_si_001.pdf (1.1MB, pdf)
am3c03724_si_002.avi (864.4KB, avi)
am3c03724_si_003.avi (486.8KB, avi)
am3c03724_si_004.avi (365.9KB, avi)
am3c03724_si_005.avi (483.3KB, avi)
am3c03724_si_006.avi (1MB, avi)
am3c03724_si_007.avi (268.2KB, avi)

References

  1. Chung H. U.; Kim B. H.; Lee J. Y.; Lee J.; Xie Z.; Ibler E. M.; Lee K.; Banks A.; Jeong J. Y.; Kim J.; Ogle C.; Grande D.; Yu Y.; Jang H.; Assem P.; Ryu D.; Kwak J. W.; Namkoong M.; Park J. B.; Lee Y.; et al. Binodal, Wireless Epidermal Electronic Systems with in Sensor Analytics for Neonatal Intensive Care. Science 2019, 363, eaau0780 10.1126/science.aau0780. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Sim K.; Ershad F.; Zhang Y.; Yang P.; Shim H.; Rao Z.; Lu Y.; Thukral A.; Elgalad A.; Xi Y.; Tian B.; Taylor D. A.; Yu C. An Epicardial Bioelectronic Patch Made from Soft Rubbery Materials and Capable of Spatiotemporal Mapping of Electrophysiological Activity. Nat. Electron. 2020, 3, 775–784. 10.1038/s41928-020-00493-6. [DOI] [Google Scholar]
  3. Cho K. W.; Sunwoo S.-H.; Hong Y. J.; Koo J. H.; Kim J. H.; Baik S.; Hyeon T.; Kim D.-H. Soft Bioelectronics Based on Nanomaterials. Chem. Rev. 2022, 122, 5068–5143. 10.1021/acs.chemrev.1c00531. [DOI] [PubMed] [Google Scholar]
  4. Rjoob K.; Bond R.; Finlay D.; McGilligan V.; J Leslie S.; Rababah A.; Iftikhar A.; Guldenring D.; Knoery C.; McShane A.; Peace A. Reliable Deep Learning–Based Detection of Misplaced Chest Electrodes During Electrocardiogram Recording: Algorithm Development and Validation. JMIR Med. Inform. 2021, 9, e25347 10.2196/25347. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Ai Y.; Hsu T. H.; Wu D. C.; Lee L.; Chen J.-H.; Chen Y.-Z.; Wu S.-C.; Wu C.; Wang Z. M.; Chueh Y.-L. An Ultrasensitive Flexible Pressure Sensor for Multimodal Wearable Electronic Skins Based on Large-Scale Polystyrene Ball@reduced Graphene Oxide Core-shell Nanoparticles. J. Mater. Chem. C 2018, 6, 5514–5520. 10.1039/c8tc01153b. [DOI] [Google Scholar]
  6. Chen K.; Ren L.; Chen Z.; Pan C.; Zhou W.; Jiang L. Fabrication of Micro-Needle Electrodes for Bio-Signal Recording by a Magnetization-Induced Self-Assembly Method. Sensors 2016, 16, 1533. 10.3390/s16091533. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Phillips J. W.; Prominski A.; Tian B. Recent Advances in Materials and Applications for Bioelectronic and Biorobotic Systems. View 2022, 3, 20200157. 10.1002/viw.20200157. [DOI] [Google Scholar]
  8. Fu Y.; Zhao J.; Dong Y.; Wang X. Dry Electrodes for Human Bioelectrical Signal Monitoring. Sensors 2020, 20, 3651. 10.3390/s20133651. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Zhao S.; Li G.; Tong C.; Chen W.; Wang P.; Dai J.; Fu X.; Xu Z.; Liu X.; Lu L.; Liang Z.; Duan X. Full Activation Pattern Mapping by Simultaneous Deep Brain Stimulation and fMRI with Graphene Fiber Electrodes. Nat. Commun. 2020, 11, 1788. 10.1038/s41467-020-15570-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Jung W.; Koirala G. R.; Lee J. S.; Kim J. U.; Park B.; Jo Y. J.; Jeong C.; Hong H.; Kwon K.; Ye Y.-s.; Kim J.; Lee K.; Kim T.-i. Solvent-assisted Filling of Liquid Metal and Its Selective Dewetting for the Multilayered 3D Interconnect in Stretchable Electronics. ACS Nano 2022, 16, 21471–21481. 10.1021/acsnano.2c09994. [DOI] [PubMed] [Google Scholar]
  11. Park Y.-G.; Min H.; Kim H.; Zhexembekova A.; Lee C. Y.; Park J.-U. Three-Dimensional, High-Resolution Printing of Carbon Nanotube/Liquid Metal Composites with Mechanical and Electrical Reinforcement. Nano Lett. 2019, 19, 4866–4872. 10.1021/acs.nanolett.9b00150. [DOI] [PubMed] [Google Scholar]
  12. Cole T.; Khoshmanesh K.; Tang S.-Y. Liquid Metal Enabled Biodevices. Adv. Intell. Syst. 2021, 3, 2000275. 10.1002/aisy.202000275. [DOI] [Google Scholar]
  13. Park Y.-G.; An H. S.; Kim J.-Y.; Park J.-U. High-resolution, Reconfigurable Printing of Liquid Metals with Three-dimensional Structures. Sci. Adv. 2019, 5, eaaw2844 10.1126/sciadv.aaw2844. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Li M.; Wu Y.; Zhang L.; Wo H.; Huang S.; Li W.; Zeng X.; Ye Q.; Xu T.; Luo J.; Dong S.; Li Y.; Jin H.; et al. Liquid Metal based Electrical Interconnects and Interfaces with Excellent Stability and Reliability for Flexible Electronics. Nanoscale 2019, 11, 5441–5449. 10.1039/c8nr09503e. [DOI] [PubMed] [Google Scholar]
  15. Cho Y. H.; Park Y.-G.; Kim S.; Park J.-U. 3D Electrodes for Bioelectronics. Adv. Mater. 2021, 33, 2005805. 10.1002/adma.202005805. [DOI] [PubMed] [Google Scholar]
  16. Park Y.-G.; Lee G.-Y.; Jang J.; Yun S. M.; Kim E.; Park J.-U. Liquid Metal-Based Soft Electronics for Wearable Healthcare. Adv. Healthcare Mater. 2021, 10, 2002280. 10.1002/adhm.202002280. [DOI] [PubMed] [Google Scholar]
  17. Lu Y.; Hu Q.; Lin Y.; Pacardo D. B.; Wang C.; Sun W.; Ligler F. S.; Dickey M. D.; Gu Z. Transformable Liquid metal Nanomedicine. Nat. Commun. 2015, 6, 10066. 10.1038/ncomms10066. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Zhou W.; Liang Q.; Chen T. 3D Manipulation of Magnetic Liquid Metals. Adv. Intell. Syst. 2020, 2, 1900170. 10.1002/aisy.201900170. [DOI] [Google Scholar]
  19. Ma B.; Xu C.; Chi J.; Chen J.; Zhao C.; Liu H. A Versatile Approach for Direct Patterning of Liquid Metal Using Magnetic Field. Adv. Funct. Mater. 2019, 29, 1901370. 10.1002/adfm.201901370. [DOI] [Google Scholar]
  20. Gough R. C.; Dang J. H.; Moorefield M. R.; Zhang G. B.; Hihara L. H.; Shiroma W. A.; Ohta A. T. Self-Actuation of Liquid Metal via Redox Reaction. ACS Appl. Mater. Interfaces 2016, 8, 6–10. 10.1021/acsami.5b09466. [DOI] [PubMed] [Google Scholar]
  21. Dickey M. D.; Chiechi R. C.; Larsen R. J.; Weiss E. A.; Weitz D. A.; Whitesides G. M. Eutectic Gallium-Indium (EGaIn): A Liquid Metal Alloy for the Formation of Stable Structures in Microchannels at Room Temperature. Adv. Funct. Mater. 2008, 18, 1097–1104. 10.1002/adfm.200701216. [DOI] [Google Scholar]
  22. Lin K.; Wang X.; Zhang X.; Wang B.; Huang J.; Huang F. An FPC based flexible dry electrode with stacked double-micro-domes array for wearable biopotential recording system. Microsyst. Technol. 2017, 23, 1443–1451. 10.1007/s00542-016-2893-4. [DOI] [Google Scholar]
  23. S Rodrigues M.; Fiedler P.; Küchler N.; P Domingues R.; Lopes C.; Borges J.; Haueisen J.; Vaz F. Dry Electrodes for Surface Electromyography Based on Architectured Titanium Thin Films. Materials 2020, 13, 2135. 10.3390/ma13092135. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Huang G.; Li M.; Yang Q.; Li Y.; Liu H.; Yang H.; Xu F. Magnetically Actuated Droplet Manipulation and Its Potential Biomedical Applications. ACS Appl. Mater. Interfaces 2017, 9, 1155–1166. 10.1021/acsami.6b09017. [DOI] [PubMed] [Google Scholar]
  25. Chua S.-K.; Chen L.-C.; Lien L.-M.; Lo H.-M.; Liao Z.-Y.; Chao S.-P.; Chuang C.-Y.; Chiu C.-Z. Comparison of Arrhythmia Detection by 24-Hour Holter and 14-Day Continuous Electrocardiography Patch Monitoring. Acta Cardiol. Sin. 2020, 36, 251–259. 10.6515/ACS.202005_36(3).20190903A. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Cao J.; Li X.; Liu Y.; Zhu G.; Li R.-W. Liquid Metal-Based Electronics for On-Skin Healthcare. Biosensors 2023, 13, 84. 10.3390/bios13010084. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Li X.; Li S.; Lu Y.; Liu M.; Li F.; Yang H.; Tang S.-Y.; Zhang S.; Li W.; Sun L. Programmable Digital Liquid Metal Droplets in Reconfigurable Magnetic Fields. ACS Appl. Mater. Interfaces 2020, 12, 37670–37679. 10.1021/acsami.0c08179. [DOI] [PubMed] [Google Scholar]
  28. Hu L.; Wang H.; Wang X.; Liu X.; Guo J.; Liu J. Magnetic Liquid Metals Manipulated in the Three-Dimensional Free Space. ACS Appl. Mater. Interfaces 2019, 11, 8685–8692. 10.1021/acsami.8b22699. [DOI] [PubMed] [Google Scholar]
  29. Shu J.; Tang S.-Y.; Feng Z.; Li W.; Li X.; Zhang S. Unconventional Locomotion of Liquid Metal Droplets Driven by Magnetic Fields. Soft Matter 2018, 14, 7113–7118. 10.1039/c8sm01281d. [DOI] [PubMed] [Google Scholar]
  30. Carle F.; Bai K.; Casara J.; Vanderlick K.; Brown E. Development of Magnetic Liquid Metal Suspensions for Magnetohydrodynamics. Phys. Rev. Fluids 2017, 2, 013301. 10.1103/physrevfluids.2.013301. [DOI] [Google Scholar]
  31. Zhang Y.; Jiang S.; Hu Y.; Wu T.; Zhang Y.; Li H.; Li A.; Zhang Y.; Wu H.; Ding Y.; Li E.; Li J.; Wu D.; Song Y.; Chu J. Reconfigurable Magnetic Liquid Metal Robot for High-Performance Droplet Manipulation. Nano Lett. 2022, 22, 2923–2933. 10.1021/acs.nanolett.2c00100. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

am3c03724_si_001.pdf (1.1MB, pdf)
am3c03724_si_002.avi (864.4KB, avi)
am3c03724_si_003.avi (486.8KB, avi)
am3c03724_si_004.avi (365.9KB, avi)
am3c03724_si_005.avi (483.3KB, avi)
am3c03724_si_006.avi (1MB, avi)
am3c03724_si_007.avi (268.2KB, avi)

Articles from ACS Applied Materials & Interfaces are provided here courtesy of American Chemical Society

RESOURCES