We report an electrically conductive carbon film with controllable hydrophilic properties that offers a covalent binding surface containing radicals for biomolecule attachment without using chemical linkers. Films were deposited from an acetylene-containing plasma using plasma immersion ion implantation during growth and subsequently annealed under vacuum. Electrical conductivity, spin density, contact angle, surface energy, surface composition, and covalent binding capability were studied as a function of annealing temperature, revealing three distinct regions. In the first region, surface energy is dominated by polar groups. In the second region, the polar groups are expelled, creating unpaired electrons that dominate the polar component of the surface energy. In the third region, the electrical conductivity rises and the polar component of surface energy falls as the unpaired electrons recombine, leading to an optimum combination of surface energy, spin density, and electrical conductivity for biological applications. It is proposed that persistent radicals are responsible for both high wettability and covalent binding properties. Covalently attached enzyme molecules on the C film can resist stringent washing with detergents. The C films offer the functions of conducting polymers, but with the added features of controllable wettability and a covalent binding capability.

Applications such as biosensing and neural prosthetics require electrodes that interface well with biomolecules and tissues. For implanted bionic eye and bionic ear applications, good tissue integration is important to maintain performance in the long term.1 Insertion of implantable devices into brain tissue triggers an inflammatory tissue response, inducing astroglial scar formation with consequent device malfunction.2 Providing the device with a surface covered with covalently bound biomolecules has been suggested as an approach for improving integration with nerve tissue.3 Coatings with “neural friendly” molecules such as proteins4 and peptides5 have been found to reduce scar formation in vivo. In electrochemical biosensors, direct surface attachment of bio-recognition molecules such as enzymes6 or deoxyribonucleic acid7 on electrode surfaces is required to generate an electrical signal from a biological detection. Apart from a good conductivity, electrode materials therefore need both high wettability and surface binding sites for biomolecule attachment.

Conducting materials, including conductive polymers, have been widely investigated for both biosensing8,9 and neural interfaces.1,10 However, the coupling of biomolecules to a conducting surface remains challenging because of the need for both strong binding and the preservation of molecular function. Biomolecules can bind to a surface either via physical adsorption (van der Waals, hydrophobic, or electrical interaction) or chemical bonds (covalent coupling). Chemical bonds are stronger and are preferable to avoid leaching and to ensure reasonable lifetime.8 Linker chemistry techniques for covalent binding require multiple steps to modify the electrode surface and, in some cases, the biomolecule as well before binding to the surface can take place.8,11 A further drawback of linker chemistry is the potential for perturbation of the native conformation and restrictions on the orientation of recognition molecules,11 both of which could affect their sensing capability in a biosensor and the interaction with tissues in a neural contacting electrode. Alternatively, covalent bonds can be created from reactions between biomolecules and free radicals on a surface that has been modified by plasma immersion ion implantation (PIII).12 This one-step immobilization method takes a time only of order of 1 h and takes place in mild aqueous conditions without using any toxic chemicals,13 making the method simple and effective. In addition, while linker chemistry requires different chemical groups for conjugating different molecules, radical-rich surfaces can react with a wide variety of biomolecules including enzymes,14 antibodies,15 and oligonucleotides.16 Carbon coatings prepared by plasma deposition are rich in free radicals that react with protein side chains to facilitate covalent binding of proteins.17 However, until now carbon coatings prepared in this way have insufficient electrical conductivity for applications as an electrode material. Hence, there is a clear need for conducting biosensor and neural contacting electrodes fabricated from a material that contains a high density of free radicals.

It has been reported that plasma polymers, also known as hydrogenated amorphous carbons, show an increase in free radical concentration after annealing.18,19 Fabisiak et al.20 reported a conductivity increase in plasma deposited amorphous carbon after annealing. These authors concluded that carbon annealed at higher than 400 °C transitions to a graphite-like material. However, none of these works link the generation of free radicals by heat treatment with the capacity to covalently bind biomolecules.

In this paper, we deposit hydrogenated amorphous carbon films by plasma chemical vapour deposition and subject them to post-deposition annealing to develop good conductivity, while retaining the capacity for radical-based covalent binding of biomolecules without any wet chemical treatment prior to the attachment process. The electrical properties, spin density, surface energy, and the relationship between spin density and covalent binding with biomolecules are investigated. A successful combination of good electrical conductivity with covalent binding capability would enable the development of a new class of biomaterials for electrochemical biosensors and neural implants.

The apparatus used for plasma deposition includes a vacuum system, a gas reservoir connected to a treatment chamber, and a power supply for delivering pulses in the energy range of 1 kV to 10 kV. The treatment chamber includes an Erlenmeyer flask with a copper electrode fitted around the base on the outside and connected to the negative terminal of the pulsed power supply and a counter electrode around its neck connected to ground. Substrates of glass, quartz, and silicon were placed at the bottom of the flask, inside the electrode area and pumped to a base pressure of 1.5 mPa using a turbomolecular pump. Equal partial pressures of acetylene and nitrogen gases were introduced into the flask and regulated to a total pressure of 30 Pa. Negative voltage pulses of 10 kV were applied to the bottom electrode using a RUP6 pulse generator (GBS Elektronik GmbH, Dresden, Germany) at a frequency of 1.5 kHz pulses of 40 μs duration for 8 min. Figure 1 illustrates schematically a combined plasma deposition and PIII process. During the pulse, the plasma acts as a source of reactive hydrocarbon species produced mainly by collisions between C2H2 molecules and electrons in the plasma. These reactive species form a thin film on the substrate surface that is bombarded by ions, principally nitrogen accelerated toward the growth surface by the applied voltage. This type of discharge is also known as a dielectric barrier discharge,21 and the bombardment process is a type of PIII process. PIII was developed primarily for metal ion implantation22 but is also useful for insulators including polymers.17,23,24

FIG. 1.

A schematic of the combined plasma immersion ion implantation and deposition process. Positive ions are extracted from the plasma by the pulsed bias applied to the copper electrode and periodically accelerated toward the electrode, forming a coating on the substrate surface placed inside the electrode area.

FIG. 1.

A schematic of the combined plasma immersion ion implantation and deposition process. Positive ions are extracted from the plasma by the pulsed bias applied to the copper electrode and periodically accelerated toward the electrode, forming a coating on the substrate surface placed inside the electrode area.

Close modal

Samples were placed in a stainless steel tube and pumped to a pressure of 13 mPa. The tube was heated in a horizontal oven to temperatures in the range of 375-675 °C at a rate of 8 °C/min and maintained for an hour before being cooled to room temperature under vacuum. The temperature was monitored by a thermocouple attached to the end of the tube inside the oven. The average difference between the temperature at the beginning and end of the dwelling period is 10 °C, meaning that the uncertainty in each set temperature is ±5 °C.

Films on silicon wafers were used for contact angle measurement. Contact angles were measured with a Kruss DS10 system using the sessile drop method. Contact angle measurement was conducted after the sample was heat treated and aged for one day. Water and glycerol were used as two liquid probes for polar and dispersive surface energy components using the Owens-Wendt-Rabel-Kaelble method. The results are the average of three separate measurements.

Fourier Transform Infrared Spectroscopy (FTIR) transmission spectra were recorded using an IR-VASE spectrometer (J.A. Woollam Co. Inc., Lincoln, NE, USA) with 500 scans at a resolution of 4 cm−1. To obtain a good signal to noise ratio from the thin film samples, infrared transparent high purity silicon wafers (B-doped p-type, 1-10 Ω cm resistivity, 100 orientation) were used as substrates. The spectrum of the bare silicon substrate was subtracted from the spectrum of the samples to obtain the final absorption spectrum of the film material. All spectra were normalized by multiplication with the ratios of film thicknesses obtained (from spectroscopic ellipsometry) before and after heat treatment.

X-ray photoelectron spectroscopy (XPS) was performed using an AXIS Nova (Kratos Analytical, Manchester, UK) equipped with a monochromated Al Kα X-ray source. Charge neutralisation was not necessary for the acquisition of the XPS data. The Raman spectra were acquired using a Horiba Jobin Yvon HR800 UV and 532 nm laser at a power of 1.6 mW. They were fitted with Lorentzian (at 1350 cm−1) and Breit-Wigner-Fano (BWF) (at 1580 cm−1) functions to obtain information of D and G peaks.

Spin density in the carbon thin films was measured by electron paramagnetic resonance (EPR) one day after treatment using an electron paramagnetic resonance spectrometer (Bruker EMX X-band) operating at room temperature with a microwave frequency of 9.8 MHz and a central magnetic field of 3523 G with a sweep width of 200 G. The microwave power was 0.6325 W. The EPR spectra were the average of 10 scans.

Quartz slides (40 mm × 9 mm × 0.9 mm) were used as substrates for plasma deposition to reduce microwave absorption. The EPR spectrum of each sample before and after annealing was compared to calculate the percentage increase in spin density.

Horseradish peroxidase (HRP) is a redox enzyme commonly used in electrochemical biosensors6,25,26 and was chosen to demonstrate the covalent binding of enzymes on the carbon film. HRP was sourced from Sigma Aldrich and prepared in solution with phosphate buffered saline (PBS) (Sigma Aldrich) to obtain a concentration of 50 μg/ml. The HRP solution was incubated with the samples prepared on silicon wafer substrates (1.5 × 1.5 cm2). The samples were mounted at the bottom of wells defined by a Teflon gasket with an O-ring (0.7 mm diameter). After incubation with 100 μl HRP solution for 1 h, each well was washed five times with 200 μl PBS before the HRP assay was conducted.

The nature of HRP attachment on bare silicon wafer and on the coated substrates was further investigated by washing with detergents [Triton and sodium dodecyl sulfate (SDS)]. Samples with HRP were soaked with 200 μl Triton solution (1% in PBS) for 30 min at room temperature and washed again three times with the PBS buffer. Triton is a detergent used to remove physically adsorbed HRP molecules, leaving covalently bound molecules without affecting their activity.27 For SDS testing, samples were soaked in 2% SDS at 70 °C for an hour and washed five times with water to remove unbound molecules. The treatment with SDS is more stringent than with Triton and has been used to prove the covalent binding of biomolecules.28 

HRP assay: A volume of 120 μl of 3,3′,5,5′-tetramethylbenzidine (TMB) (Thermo Scientific) was added to each well and allowed to react with immobilized HRP for 3 min. After that, 20 μl of the reaction solution was transferred into an Eppendorf tube containing 20 μl of H2SO4 2M to stop the reaction. The absorption of this solution at 450 nm was measured using a NanoDrop spectrophotometer (Thermo Scientific). The results presented are the average of 3 replicates.

Silicon wafers (B-doped p-type, 1-10 Ω cm resistivity, 100 orientation) with a 1.0 μm thickness native oxide layer were used as substrates for carbon coating. After carbon deposition and heat treatment, platinum interdigital electrodes were patterned on the carbon layer using optical lithography. The contacts had a total width, w = 0.0221 m and a separation between electrodes, Lgap = 1.6 × 10−5 m. The film thickness, d, was obtained by spectroscopic ellipsometry and subsequent data fitting. Four-probe current-voltage (I-V) measurements were taken from the interdigital electrodes and the resistance, R, of each carbon film was calculated from the slope of its (linear) I-V characteristic. Conductivity, σ, was calculated by

σ=LgapRdw.

The optical properties of the coatings were analyzed using spectroscopic ellipsometry (M-2000D, J.A. Woollam Co. Inc., Lincoln, NE, USA) with three incident angles (60°, 65°, and 70°) in the UV and visible range. A generalized oscillator model was used to fit the Psi and Delta versus wavelength data using the WVASE software to calculate film thickness, refractive index (n), and absorption coefficient (k). Each sample was measured before and after heat treatment for comparison. A Tauc plot29 of Eε2 versus photon energy E is used to determine the optical bandgaps by extrapolation of the linear part of each curve to the X axis, where ε2=2nk and E=hν is the photon energy.

The plasma deposition condition we used produces carbon films with a thickness in the range of 190-220 nm (data fitted from spectroscopic ellipsometry measurements). Figure 2(c) shows a scanning electron microscope (SEM) image of the as-grown C film surface, which is relatively smooth at the scale of 1 μm. The contact angle measurement of two liquid probes (water and glycerol) [Fig. 2(a)] shows an increase with annealing temperature to 425 °C and a decrease at higher temperatures (except for the water contact angle on the film annealed at 675 °C). The total surface energy [Fig. 2(b)] slightly decreases at 425 °C but increases from 425 to 575 °C and levels off at 575 °C. Although the total surface energies of samples treated between 375 °C and 525 °C do not change substantially, the polar and dispersive components show systematic changes. The initial decrease of the polar component indicates a loss of polar groups, while the increasing dispersive component is consistent with the presentation of a hydrophobic surface to the solution. Polar groups typically contain oxygen and nitrogen such as OH (in C—OH and C—OOH groups) or CO in the COC or NH group. The loss of these groups results in an increase in the relative concentration of hydrophobic structures on the surface that contain the C=C group. FTIR spectroscopy [Figs. 3(a) and 3(b)] shows the reduction in C=O and C=N groups (1850-1650 cm−1) and in O—H and N—H groups (3600-3200 cm−1), when the annealing temperature exceeds 525 °C. The spectrum of the sample annealed at 675 °C suggests that these groups have been totally removed after annealing at this temperature. Saturated aliphatic groups (alkane and alkyl) in the range of 3000-2800 cm−1 have very low absorbance that is not detectable in the spectra from films annealed above 525 °C. This indicates very low levels of sp3 bonds and dominant sp2 bonding within the carbon. There is no peak detected in the range of 2250-2100 cm−1 of the alkyne (CC) group.

FIG. 2.

(a) Contact angle measurement and an inset showing a water droplet on the carbon film. (b) Surface energy as a function of annealing temperature. (c) SEM micrograph of as-grown C film surface (scale bar: 1 μm) deposited on the silicon substrate, showing the smooth surface.

FIG. 2.

(a) Contact angle measurement and an inset showing a water droplet on the carbon film. (b) Surface energy as a function of annealing temperature. (c) SEM micrograph of as-grown C film surface (scale bar: 1 μm) deposited on the silicon substrate, showing the smooth surface.

Close modal
FIG. 3.

FTIR spectra of samples annealed for an hour at the temperatures indicated. (a) shows the change in the number carbon containing groups detected in the range of 1900-1300 cm−1 and (b) shows the reduction of O—H and N—H groups in the range of 3600-3200 cm−1. All spectra were normalized by dividing the raw spectrum by the carbon film thickness.

FIG. 3.

FTIR spectra of samples annealed for an hour at the temperatures indicated. (a) shows the change in the number carbon containing groups detected in the range of 1900-1300 cm−1 and (b) shows the reduction of O—H and N—H groups in the range of 3600-3200 cm−1. All spectra were normalized by dividing the raw spectrum by the carbon film thickness.

Close modal

The XPS survey shown in Fig. 4(a) includes nitrogen, oxygen, and carbon peaks measured on as-grown and 675 °C annealed carbon films. The atomic percentages of these elements measured in all samples are illustrated in Fig. 4(c) and show a continuous trend of oxygen and nitrogen decreasing with annealing temperature. The carbon fraction increases from 82% before annealing to 92% after annealing at 675 °C. These results are consistent with the reduced signal of O—H, N—H, C=O, and C=N groups in the FTIR spectra. Analysis of the C1s region of the XPS spectrum [Fig. 4(d)] reveals a narrowing of the FWHM with increased annealing temperature, consistent with reductions in sp3 C—C bonds and C—O/C=O/C—N bonds within the carbon film and increased size of the regions having sp2 bonding. The reduction in FWHM of the C1s peak is shown clearly in the spectra of Fig. 4(b) in which the high binding energy shoulder (indicated by an arrow) decreases in intensity with increasing annealing temperature.

FIG. 4.

(a) XPS survey spectra from an as-grown and a 675 °C annealed film. (b) C1s spectra from the carbon films annealed for an hour at the temperatures shown. The arrow indicates the shoulder referred to in the text. (c) Percentage of each element as a function of annealing temperature. (d) Full width at half maximum of the C1s peak as a function of annealing temperature.

FIG. 4.

(a) XPS survey spectra from an as-grown and a 675 °C annealed film. (b) C1s spectra from the carbon films annealed for an hour at the temperatures shown. The arrow indicates the shoulder referred to in the text. (c) Percentage of each element as a function of annealing temperature. (d) Full width at half maximum of the C1s peak as a function of annealing temperature.

Close modal

Figure 5 shows Raman spectra and analysis from the carbon films as a function of annealing temperature. Figures 5(a) and 5(b) show Raman spectra from the as-grown and 675 °C annealed films. Visible wavelength Raman spectroscopy is particularly sensitive to the sp2 bonds in the carbon matrix. The main features in the spectra are the D and G bands [at 1350 cm−1 and 1580 cm−1, as indicated in Fig. 5(b)], which were fitted with Lorentzian and BWF peaks, respectively. Figures 5(c)5(e) show the peak width, the G peak position, and the ID/IG ratio as a function of annealing temperature. An increase in the G peak position [Fig. 5(d)], an increase in the ID/IG ratio [Fig. 5(e)], and a decrease in the G peak linewidth [Fig. 5(c)] are consistent with an increase in the crystalline order of sp2 regions in the carbon films as a function of annealing temperature.30 

FIG. 5.

Raman spectra of (a) an as-grown film and (b) a film annealed at 675 °C. (c) Comparison of the peak widths of fitted D and G peaks as a function of annealing temperature. (d) Position of fitted G peak as a function of annealing temperature. (e) Ratios of fitted D and G peak intensities as a function of annealing temperature.

FIG. 5.

Raman spectra of (a) an as-grown film and (b) a film annealed at 675 °C. (c) Comparison of the peak widths of fitted D and G peaks as a function of annealing temperature. (d) Position of fitted G peak as a function of annealing temperature. (e) Ratios of fitted D and G peak intensities as a function of annealing temperature.

Close modal

The refractive index, absorption coefficient, and Tauc plots of carbon coatings before and after annealing can be found in the supplementary material (Figs. 1S and 2S). A decrease in the optical energy gap of the carbon films as a function of annealing temperature is shown in Fig. 6. This is accompanied by an increase in the conductivity of the samples.

FIG. 6.

Dependence of the optical energy gap on annealing temperature.

FIG. 6.

Dependence of the optical energy gap on annealing temperature.

Close modal

The electrical conductivity of the as-grown film was too low to be measurable. Post-deposition annealing at temperatures less than 575 °C also yielded films with conductivity too low to be reliably measured. The current-voltage characteristics measured from the films annealed at 575 °C, 625 °C, and 675 °C can be found in the supplementary material (Fig. 3S). The calculated electrical conductivities of these films were, respectively, 0.25 mS/cm, 1.1 mS/cm, and 56 mS/cm. The increase in electrical conductivity after annealing is attributed mainly to the reduction in the bandgap shown in Fig. 6. This electrical conductivity is comparable with human brain tissue conductivity (0.5-2.4 mS/cm31) and cochlear conductivity (14.2 mS/cm32).

Covalent attachment of proteins to plasma polymer made from acetylene has been attributed to the reaction of radicals embedded in the coating and the protein.17 EPR measurements were used to investigate the change of unpaired electron density in the carbon films as a function of annealing temperature. The spin density increases significantly with annealing and peaks at 2.5 times the density observed in the as-grown film after annealing at 575 °C. However, this ratio decreases after annealing at higher temperatures [Fig. 7(a)], consistent with previous work.18,20 This reduction has been attributed to the joining of neighbouring carbonized clusters to form graphitic sheets which removes unpaired spins.18 

FIG. 7.

Influence of annealing temperature on EPR signal. (a) Spin density. (b) g-factor. (c) Full width at half maximum (FWHM).

FIG. 7.

Influence of annealing temperature on EPR signal. (a) Spin density. (b) g-factor. (c) Full width at half maximum (FWHM).

Close modal

Figure 7(b) shows the proportionality factor (g-factor) detected in the carbon films as a function of annealing temperature. g-factors of 2.003 ± 0.0004 and 2.0045 ± 0.0002 have been assigned to radicals of carbonised clusters33 and cyclohexadienyl radicals,34 respectively, and lie within the range we observe for annealing temperatures less than 575 °C. There is a rapid increase in the g-factor when the annealing temperature exceeds 575 °C, indicating an increase in graphitization. When the material develops local ordered regimes consisting of graphite crystallites, the g-factor of the spins begins to show anisotropy, where the g-factor along the c-axis of the graphite crystallites is much larger than the g-factor of the free electrons. The result is an average g-factor larger than the free electron value of spins localised on small conjugated regions. The g-factor is given by Mrozowski,35 

g=gf+13(g||gf),

where gf is the free electron g-factor and g|| is the much larger g-factor for spins parallel to the graphite c-axis.

The width of an EPR absorption line is dependent on the interactions of other groups with the major group.19 The broader the peak, the more bonding environments surround the unpaired electron in the material. The full width at half maximum (FWHM) [Fig. 7(c)] shows that the as-grown film has the broadest peak, but this value decreases as the annealing temperature increases up to 575 °C before increasing again at higher temperatures. This increase in peak width has been attributed to interactions between free radicals and conduction electrons in a graphite-like structure.19 The reduction of spin density, the sharp increase of the g-factor, and the increase in FWHM after annealing at temperatures above 575 °C are consistent with the formation of increasingly large sp2 carbon structures.

Figure 8 combines analysis from FTIR, EPR, electrical conductivity, surface energy, and XPS to interpret the phenomena occurring as a function of annealing temperature. Three regions of temperature are separated with dotted lines in Fig. 8. In the low temperature region (less than 375 °C), the density of polar groups as determined from infrared absorbance is highest. Since the spin density is relatively low in this region, we attribute the high polar component of surface energy to the polar groups, especially because the atomic percent of N and O is high and these atoms are present in the polar groups. In the middle region (375-525 °C), the density of the polar groups strongly decreases along with the atomic percent of N and O atoms. At the same time, expulsion of the volatile N and O containing polar groups leads to broken bonds giving rise to unpaired electrons, hence increasing the spin density. The polar surface energy is minimised at the beginning of this region but then increases strongly in correlation with the spin density. In the third region (higher than 525 °C), the contribution of the unpaired electrons to the surface energy becomes dominant and the polar surface energy reaches a maximum. The contribution of the polar groups to the polar component of surface energy becomes very small leading us to conclude that the polar component in this region is the result of the high concentration of unpaired electrons. In this region, as the spin density decreases, the electrical conductivity strongly increases. An important finding for applications is the existence of a region where both electrical conductivity and spin density are high.

FIG. 8.

The results of analytical techniques as a function of annealing temperature. (a) Integrated area under the FTIR features corresponding to polar groups (C=O and C=N) over the range of 1780-1640 cm−1. (b) EPR spin density normalised to the as-grown film and the measured electrical conductivity. (c) Polar component of surface energy determined from contact angle measurements. (d) Atomic percent of elements from XPS analysis. Three regions, I–III, separated by dashed lines are identified.

FIG. 8.

The results of analytical techniques as a function of annealing temperature. (a) Integrated area under the FTIR features corresponding to polar groups (C=O and C=N) over the range of 1780-1640 cm−1. (b) EPR spin density normalised to the as-grown film and the measured electrical conductivity. (c) Polar component of surface energy determined from contact angle measurements. (d) Atomic percent of elements from XPS analysis. Three regions, I–III, separated by dashed lines are identified.

Close modal

As a control, it was found that without incubation with HRP, the bare silicon wafer, the as-grown C films, and the annealed C films all showed no absorbance at 450 nm after reaction with TMB. When as-grown and annealed samples were incubated with HRP and washed with PBS buffer, a high signal corresponding to HRP was found in all samples [Fig. 9(a)]. The signal comes from both covalently attached and physically adsorbed molecules. In order to identify the signal that comes from covalently attached HRP molecules only, we repeated the experiment and washed the samples with Triton detergent before the HRP assay. Figure 9(b) shows that although HRP molecules attach with similar density on the silicon wafer (control) and the carbon films, they can easily be removed from the silicon wafer surface by a detergent wash while strong absorbances were still detected on the carbon films after the same wash. This indicates that most HRP molecules physically adsorb on the silicon wafer, while they are covalently attached to the C films. The activity of the immobilized HRP molecules indicates that they were not denatured when covalently attached to the C films. When comparing as-grown, 625 °C and 675 °C annealed samples, the surface density of covalently attached molecules is approximately the same, while the density of physisorbed molecules varies. We postulate that multiple layers of HRP molecules attach to the carbon film while after detergent washing, only one covalently attached monolayer remains on the surface. This also indicates that spin density of as-grown film is enough to immobilize a monolayer of HRP as a double spin density of annealed film does not increase covalent attached density. Although a minimum density of radicals is required to form a monolayer of covalently bound molecules, the advantage of a higher radical density will be an increased lifetime of the C film, as radicals decay during storage. The decay of radicals has been shown in the carbonized layer formed on polystyrene after PIII treatment.33 

FIG. 9.

Comparison of absorbance using the HRP assay on bare silicon wafer (denoted as Si) and C coated Si wafer annealed at 675 °C (denoted as Pl): (a) at different annealing temperatures and (b) after washing the samples with Triton. Comparison of XPS analysis on Si wafer and C coated Si annealed at 675 °C before and after SDS wash: (c) percentage of elements, (d) percentage of nitrogen, (e) N1s peak detected on silicon wafer, and (f) N1s peak detected on the C coating. These results confirm that the relative density of covalently attached molecules on the carbon is largely retained after both Triton and SDS washing processes.

FIG. 9.

Comparison of absorbance using the HRP assay on bare silicon wafer (denoted as Si) and C coated Si wafer annealed at 675 °C (denoted as Pl): (a) at different annealing temperatures and (b) after washing the samples with Triton. Comparison of XPS analysis on Si wafer and C coated Si annealed at 675 °C before and after SDS wash: (c) percentage of elements, (d) percentage of nitrogen, (e) N1s peak detected on silicon wafer, and (f) N1s peak detected on the C coating. These results confirm that the relative density of covalently attached molecules on the carbon is largely retained after both Triton and SDS washing processes.

Close modal

The results from XPS analysis [Figs. 9(c)9(f)] provide further confirmation of covalent attachment of HRP on the C coatings. Figure 9(d) shows the increasing nitrogen percentage on silicon wafer and C film after incubated with HRP. After the stringent wash with SDS, nitrogen on silicon wafer is mostly removed [Fig. 9(e)] while it only slightly reduces on the C film [Fig. 9(f)].

In summary, we have demonstrated the formation of conductive carbon films containing free radicals from acetylene and nitrogen plasma. These films are amorphous and contain some sp3 hybridised bonds, while the majority are sp2 hybridised bonds. In the as-grown film, electrons are confined to unconnected clusters, resulting in very high electrical resistivity. After annealing, C=O, C=N, O—H, and N—H groups (Fig. 3) are reduced and elemental oxygen, nitrogen, and hydrogen are also removed from the carbon structure, creating unpaired electrons. Annealing at higher temperatures causes the carbon clusters to enlarge and merge, increasing the electrical conductivity while preserving a high density of unpaired electrons. Our demonstration of covalent binding with HRP via reactions with the unpaired electrons suggests a broad range of biomolecules including proteins, enzymes, antibodies, oligonucleotides, and cells could also be bound to the surface following previous work.16,36,37 The ability to covalently bind cell adhesive proteins such as fibronectin and tropoelastin to the surface of an implant will increase its biocompatibility.38–40 The need to anneal the film at a temperature up to 675 °C is not a drawback for many applications as the annealing step could be carried out prior to device fabrication. Most commonly used substrates such as platinum, titanium, and stainless steel for medical implants or metals, graphene, glassy carbon, or silicon wafer to make transducers in biosensing are compatible with this annealing process. These films therefore have potential application as conducting electrodes or neural probes on which neural friendly biomolecules are covalently bound.

We have reported the formation of carbon films, deposited from an acetylene/nitrogen plasma, with the essential characteristics for electrochemical biosensor electrodes and neural implants, namely, electrical conductivity, high polar surface energy, and a high density of covalent binding sites associated with unpaired electrons. Annealing under vacuum allows property modification to proceed in three regions. In the first region, the surface properties are dominated by polar groups. In the second region, unpaired electrons make a dominant contribution to the polar component of surface energy as the polar groups are expelled. In the third region, unpaired electrons recombine, while electrical conductivity rapidly increases. In this region, the conductivity and the covalent binding can be tailored for applications. In all regions, enzyme molecules were covalently attached without loss of enzyme activity and were largely retained after washing with detergent. The combination of these characteristics makes these carbon films suitable for electrochemical biosensors and neural implants with the additional benefit of a simple process for covalently bonding biomolecules.

See the supplementary material for details on the refractive index and absorption coefficient from ellipsometric spectroscopy data, Tauc plots at different annealing temperatures, and I-V measurements on conducting C films.

The authors acknowledge the financial support for this work from the Australian Research Council through the provision of Grant No. DP170102086. We thank the Mark Wainwright Analytical Centre at the University of New South Wales and in particular Dr. Donald Thomas for assistance with the EPR analysis. We also thank Dr. Cenk Kocer for giving permission to use the tube furnace. This work was performed in part at the Australian National Fabrication Facility (ANFF), a company established under the National Collaborative Research Infrastructure Strategy, through the La Trobe University Centre for Materials and Surface Science. Special thanks to Ms. Neha Singh from J.A. Woollam Co., Inc. for assisting with the ellipsometric spectroscopy model. The authors acknowledge the facilities and technical assistance of the Sydney Analytical and Australian Microscopy & Microanalysis Research Facility at the Australian Centre for Microscopy & Microanalysis at the University of Sydney.

1.
R. A.
Green
,
N. H.
Lovell
,
G. G.
Wallace
, and
L. A.
Poole-Warren
,
Biomaterials
29
,
3393
(
2008
).
2.
T. D.
Kozai
,
A. S.
Jaquins-Gerstl
,
A. L.
Vazquez
,
A. C.
Michael
, and
X. T.
Cui
,
ACS Chem. Neurosci.
6
,
48
(
2015
).
3.
Y.
Zhong
,
X.
Yu
,
R.
Gilbert
, and
R. V.
Bellamkonda
,
J. Rehabil. Res. Dev.
38
,
627
(
2001
).
4.
W.
He
,
G. C.
McConnell
, and
R. V.
Bellamkonda
,
J. Neural. Eng.
3
,
316
(
2006
).
5.
X.
Cui
,
J.
Wiler
,
M.
Dzaman
,
R. A.
Altschuler
, and
D. C.
Martin
,
Biomaterials
24
,
777
(
2003
).
6.
Y.
Okawa
,
N.
Yokoyama
,
Y.
Sakai
, and
F.
Shiba
,
Anal. Chem. Res.
5
,
1
(
2015
).
7.
S.
Huang
,
S.
Lu
,
C.
Huang
,
J.
Sheng
,
L.
Zhang
,
W.
Su
, and
Q.
Xiao
,
Sensors Actuators B Chem.
224
,
22
(
2016
).
8.
N.
Aydemir
,
J.
Malmstrom
, and
J.
Travas-Sejdic
,
Phys. Chem. Chem. Phys.
18
,
8264
(
2016
).
9.
J.-M.
Moon
,
N.
Thapliyal
,
K. K.
Hussain
,
R. N.
Goyal
, and
Y.-B.
Shim
,
Biosens. Bioelectron.
102
,
540
(
2018
).
10.
R.
Green
and
M. R.
Abidian
,
Adv. Mater.
27
,
7620
(
2015
).
11.
X.
Luo
and
J. J.
Davis
,
Chem. Soc. Rev.
42
,
5944
(
2013
).
12.
M.
Bilek
and
D.
McKenzie
,
Biophys. Rev.
2
,
55
(
2010
).
13.
B. R.
Coad
,
M.
Jasieniak
,
S. S.
Griesser
, and
H. J.
Griesser
,
Surf. Coat. Technol.
233
,
169
(
2013
).
14.
C. T. H.
Tran
,
N.
Nosworthy
,
M. M. M.
Bilek
, and
D. R.
McKenzie
,
Biomass. Bioenergy
81
,
234
(
2015
).
15.
E.
Kosobrodova
,
A.
Mohamed
,
Y.
Su
,
A.
Kondyurin
,
C. G.
dos Remedios
,
D. R.
McKenzie
, and
M. M. M.
Bilek
,
Mater. Sci. Eng. C
35
,
434
(
2014
).
16.
C. T. H.
Tran
,
M.
Craggs
,
L. M.
Smith
,
K.
Stanley
,
A.
Kondyurin
,
M. M.
Bilek
, and
D. R.
McKenzie
,
RSC Adv.
6
,
83328
(
2016
).
17.
Y.
Yin
,
K.
Fisher
,
N. J.
Nosworthy
,
D.
Bax
,
S.
Rubanov
,
B.
Gong
,
A. S.
Weiss
,
D. R.
McKenzie
, and
M. M. M.
Bilek
,
Plasma Process. Polym.
6
,
658
(
2009
).
18.
D. J.
Miller
and
D. R.
McKenzie
,
Thin Solid Films
108
,
257
(
1983
).
19.
D. E. G.
Austen
,
D. J. E.
Ingram
, and
J. G.
Tapley
,
Trans. Faraday Soc.
54
,
400
(
1958
).
20.
K.
Fabisiak
,
S.
Orzeszko
,
F.
Rozpo̵ch
, and
J.
Szatkowski
,
J. Non-Cryst. Solids
99
,
12
(
1988
).
21.
U.
Kogelschatz
,
Plasma Chem. Plasma Process.
23
,
1
(
2003
).
22.
J.
Tendys
,
I. J.
Donnelly
,
M. J.
Kenny
, and
J. T. A.
Pollock
,
Appl. Phys. Lett.
53
,
2143
(
1988
).
23.
R. C.
Powles
,
D. R.
McKenzie
,
N.
Fujisawa
, and
D. G.
McCulloch
,
Diam. Relat. Mater.
14
,
1577
(
2005
).
24.
A.
Anders
,
Surf. Coat. Technol.
156
,
3
(
2002
).
25.
I.
Taurino
,
G.
Sanzò
,
R.
Antiochia
,
C.
Tortolini
,
F.
Mazzei
,
G.
Favero
,
G.
De Micheli
, and
S.
Carrara
,
Trends Anal. Chem.
79
,
151
(
2016
).
26.
W.
Zhang
and
G.
Li
,
Anal. Sci.
20
,
603
(
2004
).
27.
N. J.
Nosworthy
,
A.
Kondyurin
,
M. M. M.
Bilek
, and
D. R.
McKenzie
,
Enzyme Microb. Technol.
54
,
20
(
2014
).
28.
A.
Kondyurin
,
N. J.
Nosworthy
, and
M. M. M.
Bilek
,
Langmuir
27
,
6138
(
2011
).
30.
A. C.
Ferrari
and
J.
Robertson
,
Philos. Trans. R. Soc. Lond. A Math. Phys. Eng. Sci.
362
,
2477
(
2004
).
31.
M.
Akhtari
,
D.
Emin
,
B. M.
Ellingson
,
D.
Woodworth
,
A.
Frew
, and
G. W.
Mathern
,
J. Appl. Phys.
119
,
064701
(
2016
).
32.
P.
Tran
,
P.
Wong
,
A.
Sue
,
Q.
Li
, and
P.
Carter
,
35th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC)
(IEEE,
2013
), p.
5291
.
33.
E. A.
Kosobrodova
,
A. V.
Kondyurin
,
K.
Fisher
,
W.
Moeller
,
D. R.
McKenzie
, and
M. M. M.
Bilek
,
Nucl. Instrum. Methods Phys. Res. B
280
,
26
(
2012
).
34.
M.
Kuzuya
,
A.
Noguchi
,
H.
Ito
,
S. I.
Kondo
, and
N.
Noda
,
J. Polym. Sci. A Polym. Chem.
29
,
1
(
1991
).
36.
M. M. M.
Bilek
,
Appl. Surf. Sci.
310
,
3
(
2014
).
37.
Y.
Yin
,
M. M. M.
Bilek
, and
D. R.
McKenzie
,
Plasma Process. Polym.
7
,
708
(
2010
).
38.
Y.
Yin
,
S. G.
Wise
,
N. J.
Nosworthy
,
A.
Waterhouse
,
D. V.
Bax
,
H.
Youssef
,
M. J.
Byrom
,
M. M.
Bilek
,
D. R.
McKenzie
,
A. S.
Weiss
, and
M. K.
Ng
,
Biomaterials
30
,
1675
(
2009
).
39.
A.
Waterhouse
,
Y.
Yin
,
S. G.
Wise
,
D. V.
Bax
,
D. R.
McKenzie
,
M. M.
Bilek
,
A. S.
Weiss
, and
M. K.
Ng
,
Biomaterials
31
,
8332
(
2010
).
40.
W.
Chrzanowski
,
A.
Kondyurin
,
J. H.
Lee
,
M. S.
Lord
,
M. M.
Bilek
, and
H. W.
Kim
,
J. Mater. Sci. Mater. Med.
23
,
2203
(
2012
).

Supplementary Material