Fabrication method of the DLC-UME
A quartz capillary (Outer diameter: 1.0 mm and inner diameter: 0.7 mm, Sutter Instrument Co., Novato, CA, USA) was pulled into two nanopipettes assisted by CO2 laser-puller (P2000, Sutter Instrument Co., Novato, CA, USA) with appropriate pull parameters (Heat: 950, Velocity: 50, Delay: 168, Filament: 2, Pull: 120) (Fig. 1a-I). Cr and Au were sputtered onto the nanopipette surface at the power of 105 W and the chamber pressure of 1 Pa for both 1 min, and an enameled Cu wire was fixed on the Au layer with conductive silver paste and further baked in a vacuum drying oven at 120°C for 30 mins until the paste solidifies (Figs. 1a-II, III, IV). An Ir layer was sputtered onto the Au layer as the seed layer of iridium oxide (IrOx) at the power of 105 W with an Ar flux of 13 sccm and an IrOx layer was further sputtered onto the Ir layer at the power of 105 W with an Ar flux of 13 sccm and an O2 flux of 13 sccm (Fig. 1a-V). Then, a DLC layer with a thickness of 400 nm was deposited onto the IrOx layer by pure ion coating technology (Anhui Chunyuan Coating Technology Co., Ltd, China) (Fig. 1a-VI) and removed by microplasma jet to expose the IrOx layer on the UME tip (Fig. 1a-VII). The surface coating distribution of the processed UME is shown in Fig. 1a-VIII.
Processing of microplasma jet treatment for UMEs
Under the monitoring of the positioning camera and the observation camera in real-time, the angle of the nozzle was controlled through the robotic arm, and the position of the UME tip was adjusted through the five-axis motion platform to be about 100 µm below the nozzle. Then, A 2 kHz sinusoidal voltage output by the waveform generator was amplified to peak-to-peak voltage of 12 kV to 14 kV by the high voltage amplifier and applied to the ring electrode (Fig. 1b).
Principle for the microplasma jet removal of the DLC coating on the UME tip
The DLC coating contains diamond structure and graphite structure, with carbon atoms mainly bound by sp3 and sp2 hybrid bonds. Therefore, it will be damaged by the physical interaction of microscopic particles in the plasma and will also undergo chemical reactions with reactive species such as oxygen plasma, producing volatile products (Fig. 1c). Electrons and ions (such as He) in microplasma jet can be accelerated by the electric field to be directly bombarded on the DLC surface, leading to the breakdown of covalent bonds on the DLC-UME surface and activation of the DLC. Then, the activated surface easily reacts with oxygen free radicals to introduce oxygen-containing functional groups. After microplasma treatment43. These two processes together lead to the removal of the DLC coating and the changes in the surface morphology of the UME tip.
Selection of the reaction gas and the horizontal machining distance
The species generated in the microplasma jets acting on UME tips under different reaction gases was detected by the optical emission spectrum using the spectrometer with a fiber optic probe (Fig. 2a-c). The spectrum of He microplasma jet under the He flow rate of 20 sccm was dominated by neutral helium atoms lying in the wavelength range of 550–750 nm. Reactive oxygen atoms were also found at 777.7 nm, which indicated that the air in the jet path also undergoes a certain degree of ionization to generate reactive oxygen species during the process of He microplasma jet generation (Fig. 2a). After mixing a certain proportion of oxygen into helium, significantly reactive O atoms at 777.7 nm dominated the spectrum and reduced the production of neutral helium atoms (Fig. 2b). These reactive O atoms mainly come from the collisions between oxygen molecule and He* or the inelastic electron-impact collisions with oxygen molecule43, and play a significant role in the removal and surface modification of DLC. As a comparison to the He microplasma jet, the spectrum of Ar microplasma jet under the Ar flow rate of 20 sccm was dominated by neutral argon atoms lying in the wavelength range of 700–850 nm with higher emission intensity and less reactive O atoms (at 777.7 nm) (Fig. 2c).
Due to the smaller atomic mass and lower glow ignition voltage of He compared to Ar, the reaction between the He microplasma jet and the UME is less intense and easier to control. Moreover, some oxygen-containing functional groups from the oxygen microplasma jet undergo thermal decomposition and react with DLC to generate small molecule gases, such as CO, CO2, and water vapor during activation processes53, thereby reducing the residue on the surface of UME. Therefore, in this work, we chose a mixed microplasma of He and O2. Due to O2 being a negatively charged gas, excess O2 can absorb electrons from microplasma jets, leading to a decrease in the amount of reactive oxygen atoms and ultimately reducing the DLC removal rate and effect. So, the O2 flow rate was chosen to be 2 sccm here.
A He/O2 microplasma jet formed on the silicon substrate deposited with the DLC coating after dielectric barrier discharge. As the UME gradually approached the jet, a weak branch of the microplasma jet began to be attracted to the UME tip when the horizontal spacing was around 150 µm (Fig. 2d). This is because the motion of conductive particles causes the microplasma jet to focus on the position with the closest distance and the highest relative conductivity. At this moment, the jet intensity was not sufficient to remove the dense DLC coating on the UME surface. When the horizontal distance was around 80 to 120 µm, the jet intensity was just enough to slowly remove the DLC coating on the UME surface, resulting in better machining accuracy and controllability of the exposed tip length, which helps to process the UMEs with finer tips and the coatings with a wider range of available materials (Fig. 2e). When the horizontal distance was less than 50 µm, the entire jet with a greater intensity was directly focused on the UME tip, resulting in higher processing rate and lower processing accuracy, which is more suitable for processing the UMEs with the larger tip size and the protective coatings of larger thickness and harder-to-machining materials (Fig. 2f). Therefore, in this work, we placed the UME at a horizontal distance of 100 µm and used the branch of microplasma jet to process the UME tip.
Characterization of the DLC-UME surface coatings
The fabricated DLC-UME is presented in Fig. 3a. IrOx was chosen as the conductive layer for the UME because it is a typical material with Faraday pseudocapacitive properties and is compatible with MEMS processes, with good detection repeatability, electrochemical stability, biocompatibility, and targeted detection ability for intracellular pH15,47. The IrOx layer on the UME tip presented an earthworm-like nanowire structure, increasing the surface area of the UME detection point and reducing impedance (Figs. 3b and c). Due to the simultaneous reversible reactions between Ir3+ and Ir4+ states48, the capacitance of the UME increases and the ability of UME to transfer charges within cells is enhanced, thereby increasing detection sensitivity and signal-to-noise ratio. From the cross-section of the UME tip, the thickness of the DLC layer covering the IrOx was consistent with the expected thickness of 400 nm (Fig. 3d). The DLC layer had a dense nanoparticle structure (Figs. 3e and f), which contributes to the electrochemical stability and effective protection of the UME.
Selection of processing time and control of processing effect for UME tip
In this experiment, the controllable fixed-point processing of UME tips was achieved by controlling the processing time of the He/O2 microplasma jet branch with a diameter of approximately 3 µm. The processing time was kept at 1 second, 5 seconds, and 10 seconds, with the material removal length at the UME tip being 0.5 µm, 8.5 µm and 21.2 µm, respectively (Fig. 4a). The approximate etching rate can be estimated by measuring the relationship between the removal length and the processing time, which was approximately 2.12 µm/s. The material removal length at the UME tip increased with increasing processing time. The DLC layer at the UME tip began to be removed after 1 second, exposing the inside IrOx layer, and the surface of the UME was slightly roughened (Figs. 4a-I, II). After a processing time of 5 seconds, the exposed tip length of the IrOx layer has exceeded 5 µm (Fig. 4a-III). At this time, the IrOx coating has also been subjected to surface modification and roughness to a certain extent by microplasma jet (Fig. 4b), and the surface area continued to increase, which helps to reduce the impedance of the UME and improve detection sensitivity. Furthermore, the excessive processing time of more than 10 seconds caused significant damage to the conductive layer material and the exposed length to be too large of the UME tip (Fig. 4a-IV). Therefore, the processing time should be controlled within 5 seconds. Here, we controlled the processing time of the microplasma jet to 2 seconds, and the exposed tip length of the processed UME was about 900 nm (Fig. 4c), achieving controllable processing of the UME tip with submicron resolution, which is ahead of the work of other researchers (Table. S1).
Composition, structure, and biocompatibility testing of processed DLC-UME
EDS showed that the DLC-UME was composed of C, O, Ir, Au, Cr and Si elements, with the weight percentage of 26.35, 15.27, 32.47, 24.14, 0.39, 1.38, respectively (Fig. 5c). Moreover, the atomic ratio of Ir to O is 4.83: 27.30. The relative content of each element is influenced by the thickness and distribution of each coating from the outside to the inside of the UME. By comparing the distribution and density of C atoms in the red box area of Fig. 5a with other elements, it can be concluded that the DLC coating at the UME tip has been successfully removed, and other coatings were still intact (Fig. 5a). The surface of the untreated UME tip was coated with DLC film with typical mixed structure of sp2 and sp3 carbon49, and its Raman spectrum consisted of the typical D-peak of 1350 cm− 1 and G-peak of 1580 cm− 1, while the presence of D-peak or G-peak was not detected at the UME tip after microplasma jet processing due to the complete removal of the DLC coating on the UME surface (Fig. 5b).
Then, five diffraction peaks were found at 2θ values of 26.2°, 44.1°, 51.6°, 75.4°and 89.1° in the XRD spectra of the processed UME (Fig. 4d). It can be observed that the XRD spectra was dominated by three intense peaks located at 2θ ∼ 44.1°, 75.4° and 89.1°, which can be identified by the reflection of diamond (111), (220) and (311) planes49. The peaks located at 2θ ∼ 26.2° and 51.6° were corresponding to the graphite (002) and (102), respectively50. In a word, the DLC coating on the surface of the processed UME was composed of a certain proportion of diamond graphite with high crystallinity. Furthermore, the growth status of the cells with the UME placed within seven days was as good as that of the control group within the seven days of testing, which indicated the DLC-UME has good biocompatibility (Fig. S1).
Electrochemical characterization of DLC-UME
The electrochemical performance was tested using a micromanipulator to clamp the DLC-UME as the working electrode (WE), an Ag/AgCl electrode as the reference electrode (RE), and a Pt electrode as the counter electrode (CE) (Fig. 6a). Accelerated aging testing has been used to evaluate the intracellular lifespan of UMEs51. Placing the UME in a high-temperature environment will accelerate the chemical reaction of the UME surface coatings to accelerate its degradation, thus evaluating the long-term stability of the UME in a short period of time52. A commonly used formula is Eq. S153. In our work, the reference temperature was set to 37°C suitable for HT22 cells growth and the ambient temperature was kept at 60 ℃ for an accelerated aging factor of 4.9 according to Eq. S1 by placing the beaker with electrodes on a hot plate. The impedance frequency response curves and CV curves of the DLC-UME were recorded with the accelerated aging time of 0 hours, 34 hours, 68 hours, 147 hours, and 294 hours, corresponding to approximately 0 days, 7 days, 14 days, 30 days, and 60 days, respectively (Figs. 6b and 6c). As the UME ages, the impedance slowly decreased at the frequency of 1 kHz, and the limit current slowly increased at the applied potential of 0.8 V. The changes in impedance and current within 60 days were very small (Fig. 6d).
The electrochemical performance of the UME was still stable after 60 days, which may be attributed to the good electrochemical stability of the DLC coating and the modification of the DLC coating by microplasma jet. With the introduction of oxygen-containing functional groups in the microplasma jet, sp2 bonding was broken, reducing the degree of graphitization of DLC54,55. Oxygen atoms adsorbed at defect sites in DLC material to form various oxygen-containing functional groups and more sp3 carbon appears, resulting in a decrease in conductivity and an increase in internal Ohmic resistance of the DLC coating56, which may improve the compactness and the electrochemical shielding ability of the DLC coating and contribute to the effective protection of the UME.
Electrophysiological recordings of intracellular signals of neurons
Electromotive force (EMF), which represents the potential difference between the RE and the WE at zero current, which can be used to characterize the Intracellular potential response57. The DLC-UME was inserted into the cultured HT22 cells, and the cell membrane underwent slight deformation (Fig. 6e). Then, the UME was sequentially implanted into three different cells for the signal recording (Fig. 6f). Equivalent circuit diagram of the UME inserted into the cytoplasm in a dual electrode system is shown in Fig. 6g, where Re represents electrode resistor, Rm represents cell membrane resistor, Rb represents bulk solution resistor, Ce represents electrode capacitor, and Cm represents cell membrane capacitor, which is used to simplify complex behaviors inside and outside cells15.
Intracellular pH value is closely related to cellular metabolism, carcinogenesis, and apoptosis, which is crucial for a deeper understanding and diagnosing diseases. The corresponding relationship between pH and EMF, named the E-pH response sensitivity, was measured to be approximately 60.92 mV per pH by placing the UME in solutions of different pH values (Fig. S2), which was closed to the Nernst response. The EMF curves and the corresponding pH curves of the DLC-UME were tested when the UME was sequentially inserted from extracellular to intracellular of the three cells in Fig. 6f in the same culture medium at 7.4 pH (Fig. 6h), which indicates that the DLC-UME has the ability to continuously and stably detect pH within different cells, which can be used to diagnose the health status of different cells.
The EMF difference between intracellular and extracellular cytoplasm in the cell 1 in Fig. 6f was approximately 4.26 mV. When a small amount of NaOH solution was dropped into the culture medium to increase the pH to 7.6, the potential difference increased to about 13.4 mV and the intracellular pH value within 60 s was still about 7.33, basically consistent with the value before adding NaOH solution (Fig. 6i). The implantation of the DLC-UME caused minimal damage and the cell membrane was not damaged, so the extracellular fluid did not penetrate into the cell, and the cell was not affected by changes in the pH of the culture medium in a short period of time. Results above demonstrated that the DLC-UME was sensitive to pH changes and had good reversibility and stability in intracellular recording.