Enhanced sensitivity of graphene gas sensors using molecular doping

[0001] This application claims priority from U.S. Provisional Patent Application No. 62/329,780, filed Apr. 29, 2016, entitled “ENHANCED SENSITIVITY OF GRAPHENE GAS SENSORS USING MOLECULAR DOPING”.

[0002] This disclosure relates generally to gas sensors and, more particularly, to enhanced sensitivity of graphene gas sensors using molecular doping.

[0003] Detecting presence of gas molecules is of prominent importance for controlling chemical processes, safety systems, and industrial and medical applications. Despite enormous progress in developing and improving various types of gas sensors, sensors with higher sensitivity, lower sensing limit, and lower cost that can perform at room temperature remain desirable. Graphene is a promising candidate for gas sensing applications due to its unique transport properties, exceptionally high surface-to-volume ratio, and low electrical noise.

[0010] In one aspect, a graphene gas sensor is disclosed. The graphene gas sensor may include a graphene element molecularly doped with nitrogen dioxide (NO₂), a pair of voltage electrodes spaced apart on the graphene element, and a pair of current electrodes at each end of the graphene element. In the graphene gas sensor, the graphene element may be used to detect ammonia (NH₃) based on a conductivity of the graphene element measured using the pair of voltage electrodes and the pair of current electrodes. In the graphene gas sensor, a change in the conductivity upon exposure to NH₃ may be greater for the graphene element doped with NO₂ than an undoped graphene element.

[0011] In any of the disclosed embodiments of graphene gas sensor, the graphene gas sensor may be included in a complementary metal oxide semiconductor (CMOS) device.

[0012] In any of the disclosed embodiments of graphene gas sensor, the graphene element may be molecularly doped with NO₂ using 100 ppm of NO₂ in N₂ at 500 Torr pressure and at room temperature. In any of the disclosed embodiments of graphene gas sensor, the graphene element may be molecularly doped for a duration between 10 minutes and 60 minutes.

[0013] In any of the disclosed embodiments of graphene gas sensor, the conductivity may be measured by applying a current using the pair of current electrodes, measuring a voltage using the pair of voltage electrodes, and determining the conductivity based on the current and the voltage. In any of the disclosed embodiments of graphene gas sensor, the voltage may be measured using a field effect transistor.

[0014] In any of the disclosed embodiments of graphene gas sensor, the graphene element may include a single atomic layer of carbon.

[0015] In any of the disclosed embodiments, the graphene gas sensor may further include a silicon substrate on which the graphene element is situated.

[0016] In a further aspect, a method of detecting ammonia (NH₃) gas is disclosed. The method may include applying a current to a graphene element molecularly doped with nitrogen dioxide (NO₂), measuring a voltage across the graphene element, and exposing the graphene element to NH₃ gas while measuring a change in the voltage. In the method, the change in the voltage may be indicative of the concentration of the NH₃ gas. In the method, the change in voltage may be greater for the graphene element doped with NO₂ than an undoped graphene element.

[0017] In any of the disclosed embodiments of the method, the graphene element may be implemented in a complementary metal oxide semiconductor (CMOS) device.

[0018] In any of the disclosed embodiments, the method may further include molecularly doping the graphene element with NO₂ using 100 ppm of NO₂ in N₂ at 500 Torr pressure and at room temperature. In the method, molecularly doping the graphene element may further include molecularly doping the graphene element for a duration between 10 minutes and 60 minutes.

[0019] In any of the disclosed embodiments, the method may further include determining a conductivity of the graphene element. In the method, determining the conductivity of the graphene element may further include applying the current using a pair of current electrodes spaced apart on the graphene element, measuring a voltage using a pair of voltage electrodes spaced apart on the graphene element, and determining the conductivity based on the current and the voltage.

[0020] In any of the disclosed embodiments of the method, measuring the voltage may further include measuring the voltage using a field effect transistor.

[0021] In any of the disclosed embodiments of the method, the graphene element may include a single atomic layer of carbon.

[0022] In any of the disclosed embodiments of the method, the graphene element may be situated on a silicon substrate.

[0023] In the following description, details are set forth by way of example to facilitate discussion of the disclosed subject matter. It should be apparent to a person of ordinary skill in the field, however, that the disclosed embodiments are exemplary and not exhaustive of all possible embodiments.

[0024] As noted previously, gas sensors of increasingly greater sensitivity, faster response time and portability are desired for many industrial applications. Furthermore, a gas sensor that is coupled with back-end complimentary metal-oxide semiconductor (CMOS) amplification and analysis circuitry also provides advantages from tighter system integration, which may improve functionality and may lower power consumption and cost.

[0025] Current electrochemical sensors may provide suitable response times for certain applications, but may have inherently limited response time due to diffusion processes through an electrolyte. Furthermore, solid state metal oxide semiconductor (MOS) sensors may suffer from high internal resistance, and may be operational at high temperatures to activate the MOS surface. Oxygen from the air may bind to the surface at high temperatures (usually >400 C) to form an active sensing layer at the MOS surface, while the sensing mechanism of MOS sensors may be a redox reaction between the gas analyte and the MOS surface. As a result, MOS sensors may be difficult to integrate in a low-power CMOS package.

[0026] The discovery of graphene and the subsequent progress in nanotechnology and nanomaterials have led to ultra-sensitive room-temperature sensors that can detect individual analyte molecules, such as nitrogen dioxide (NO₂). Certain key factors of graphene as a sensor material include unique transport properties, extremely high surface-to-volume ratio, and low electrical noise. Electronic states of graphene may be affected by adsorbed gas molecules and the charge transfer between graphene and the adsorbed gas molecules can modify carrier concentration without altering mobility.

[0027] However, pristine suspended graphene may not perform particularly well as a gas sensor. The reported sensitivity for graphene gas sensors can be attributed to physical or chemical functionalization of graphene by adsorbates, defects, and the supporting substrate. While such molecular modifications may be determinative for the observed sensitivity of graphene sensors to certain gas molecule analytes, more controllable molecular functionalization methods for graphene may be applied to achieve higher sensitivity. The detection of individual NO₂ molecules using graphene has been observed through a statistical analysis of a prolonged measurement result under strictly defined conditions, rather than a more practical method, such as with a resistive graphene gas sensors, in which the analyte is immediately detected through a change in resistance of the graphene upon exposure to the analyte molecule. Theoretical calculations and recent published scientific results have indicated improvement to the sensitivity of graphene sensors can be achieved by introducing substitutional impurities, which cause localized perturbations in structure and electronic states of graphene and lead to higher binding energy and charge transfer between graphene and the analyte gas molecules.

[0028] Despite extensive information revealed by theoretical studies on the interaction of suspended pristine graphene and gas molecules, the sensing mechanisms for actual fabricated graphene transducers are expected to be more complex than the theoretical models due to the effects of unintentional adsorbates and the supporting substrate. Indeed, the adsorption capacity of graphene for gas molecules predicted by theoretical calculations is smaller than experimental values by almost two orders of magnitude. Compared to substitutional impurities, molecular doping via adsorbed molecules weakly modifies the electronic properties of graphene with minimal effects on molecular structure because of the strong carbon-carbon sp² bonds of graphene.

[0029] As will be described in further detail, molecular doping may be used to functionalize graphene to increase graphene's binding energy and, thereby, sensitivity to specific gas molecules. Enhanced sensitivity of graphene gas sensors using molecular doping may provide high sensitivity, low response time, low-power room temperature operation, and ease of electronic integration, such as with wireless interfaces. Specifically, it was determined that molecular doping of graphene using NO₂ enhances sensitivity of graphene to ammonia (NH₃) gas molecules. It was discovered that the adsorption of NO₂ molecules to graphene may increase sensitivity of graphene to NH₃ molecules by more than an order of magnitude.

[0030] Turning now to the drawings, FIGS. 1A-1B illustrate selected elements of an example of a graphene gas sensor 100 for enhanced sensitivity using molecular doping. A 3-dimensional rendered view is shown as FIG. 1A as a perspective schematic illustration, while a scanning electron microscopy (SEM) image from a top view is shown as FIG. 1B. As shown in FIG. 1A, graphene gas sensor 100 is a resistive sensor with four-probe electrical connections and back-gated graphene field-effect transistors (GFETs). A graphene channel 102 was formed with a length of about 10 μm and a width of about 25 μm between electrodes. It is noted that FIG. 1A is not drawn to scale and is a schematic illustration.

[0031] The graphene used in graphene gas sensor 100 of FIGS. 1A and 1B was synthesized via chemical vapor deposition (CVD). The CVD graphene was transferred to a silicon substrate covered with 300 nm thick SiO₂, which was thermally grown via a poly(methyl methacrylate) PMMA-supported wet-transfer process. The thickness and quality of graphene samples was evaluated with Raman spectroscopy (see FIG. 5), using a neodymium-doped yttrium aluminum garnet (Nd:YAG) laser at 532 nm wavelength under ambient conditions. The undoped graphene samples showed a symmetric 2D peak with full width at half maximum (FWHM) of about 28 cm⁻¹, a 2D peak to G peak intensity ratio (I(2D)/I(G)) greater than about 3, and a negligible ratio of D peak to G peak intensity (I(D)/I(G)), which is indicative of high quality monolayer graphene (e.g. a single atomic layer of carbon atoms). Graphene channel 102 on a Si/SiO₂ substrate was patterned using electron beam lithography and etched using a low-power reactive-ion-etch process in oxygen plasma. Electrical connections to graphene channel 102 were made using electron beam lithography followed by deposition of chromium Cr (5 nm)/gold Au (45 nm) using e-beam evaporation and a lift-off process. During measurements, the gas concentration in the measurement chamber was controlled using mass flow controllers (MFCs) ahead of a mixing manifold, with N₂ as a diluting gas. At the gas sources, concentration of NH₃ and NO₂ gases was 100 parts per million (ppm) in dry air. The experiments described herein were carried out at a constant flow of gas at 500 Torr pressure and at room temperature.

[0032] As shown in FIG. 1A, conductivity measurements of graphene channel 102 were performed using the four-probe electrical connections. For example, a current (I) may be applied and a corresponding voltage (V) may be measured to determine conductivity.

[0033] Referring now to FIGS. 2A and 2B, selected elements of data plots showing sensitivity of graphene gas sensors are illustrated. In FIG. 2A sensitivity of undoped graphene and NO₂-doped graphene sensors to NH₃ gas is shown, while in FIG. 2B sensitivity as a function of NH₃ concentration for undoped and NO₂-doped graphene is illustrated. FIG. 2A shows sensitivity of a graphene gas sensor, such as graphene gas sensor 100 depicted in FIG. 1A, before and after doping with NO₂ under exposure to various concentrations of NH₃, from 2 ppm to 80 ppm NH₃. The sensitivity is defined as the change of electrical conductance (ΔG) normalized to an initial value of the electrical conductance (G₀). The graphene gas sensor was exposed to different concentrations of NH₃ for 15 min., followed by 20 min. of N₂ purge. As shown in FIG. 2b, sensitivity of the graphene gas sensor before doping with NO₂ is 3.7% for 80 ppm NH₃, which is comparable to other prior reports. For concentrations below 8 ppm NH₃, the sensitivity of the graphene gas sensor before doping with NO₂ is less than 1%. The NH₃ sensitivity measurements were repeated after the graphene gas sensor was doped by NO₂. For doping, graphene was exposed to 100 ppm of NO₂ gas for 50 min. at 500 Torr and room temperature, which shifted the Dirac point voltage (V_Dirac) from 18V to 165 V. The response of NO₂-doped graphene gas sensor to various concentrations of NH₃ is depicted in FIG. 2B (Doped Graphene (50 min)). TABLE 1 contains the sensitivity of graphene gas sensors before and after NO₂-doping, which is shown in FIG. 2B.

[0034]

TABLE 1
Conductance sensitivity of graphene gas sensors using undoped
and doped graphene in FIG. 2B.
Graphene	2 ppm	4 ppm	8 Ppm	30 ppm	55 ppm	80 ppm
Type	NH3	NH3	NH3	NH3	NH3	NH3
Undoped	0.1%	0.4%	0.9%	1.9%	2.7%	3.7%
NO2-doped	1.1%	3.4%	6.8%	13.9%	20.2%	25.3%

[0035] The results in FIGS. 2A and 2B and TABLE 1 show that doping graphene with NO₂ significantly increases the sensitivity of graphene to NH₃. Compared to the undoped graphene gas sensor, the sensitivity of NO₂-doped graphene to 2 ppm NH₃ is increased by 11-fold, and to 80 ppm of NH₃ is increased by 7-fold. The sensitivity of the NO₂-doped graphene to NH₃ is comparable to the reported sensitivity for boron-doped graphene, which is a structural substitutional doping, under continuous UV light illumination. FIGS. 2A and 2B and TABLE 1 show that for improving sensitivity of graphene resistance to adsorption of gas molecules, molecular doping can be as effective as substitutional doping. Moreover, improving the sensitivity by an order of magnitude at low analyte concentrations may indicate an improved detection limit of NO₂-doped graphene to NH₃, which was not directly verified. The detection limit of NO₂-doped graphene gas sensors, such as depicted in FIG. 1A, is estimated to be about 200 ppb NH₃, compared to about 1.4 ppm NH₃ using undoped graphene. To estimate the detection limit, it was assumed that sensitivity is proportional to gas concentration, such that the sensitivity versus concentration curve at low concentrations may be extrapolated to calculate the NH₃ concentration that causes a ratio of ΔG/G₀ Raman peaks comparable to the measurement noise for ΔG/G₀ peaks. In addition, it was observed that the sensitivity of NO₂-doped graphene to NH₃ molecules was proportional to the duration of NO₂ doping.

[0036] The Doped Graphene (20 min.) curve in FIG. 2A shows the sensitivity of the graphene gas sensor when exposed to 100 ppm of NO₂ gas for only 20 min. FIG. 2A also shows, that after NH₃ exposure, the conductance remains nearly stable suggesting that resetting the sensor may involve additional actions such as UV light irradiation, and thermal annealing, which can be performed through Joule heating of the graphene. However, the stability of the adsorbed dopant molecules, NO₂ molecules in this case, during the resetting process might be a concern. To address this concern, graphene may be doped using another molecule that chemisorbs to the graphene and can withstand the resetting process of the sensor. Alternatively, a modified fabrication procedure may be employed. For example, graphene may be doped with NO₂ or another molecule and then transferred to the substrate such that the doped side, which is the exposed side of graphene during the doping process, comes in contact with the substrate. In this arrangement, the NO₂ molecules are trapped between the substrate and the graphene, and the impermeability of graphene to molecules may prevent the NO₂ molecules from escaping during the resetting process of the sensor.

[0037] Referring now to FIGS. 3A-3D, selected elements of data plots showing electronic properties of graphene gas sensors are illustrated. Specifically, FIG. 3A shows drain current (I_d) as a function of back-gate voltage (V_g) for a NO₂-doped GFET for different doping durations. FIG. 3B shows a contour plot of I_d as a function of Vg and doping duration. FIG. 3C shows an I_d−V_g curve of the GFET during recovery. FIG. 3D shows a contour plot of I_d as a function of V_g and recovery duration over 10 days. In FIGS. 3B and 3D, the source-drain electrical bias is 10 mV and V_Dirac is indicated by the dashed line.

[0038] To investigate the mechanism of sensitivity enhancement, GFET devices were fabricated and the effect of NO₂ adsorbates on electronic properties of graphene was examined. With an applied source-drain electrical bias of 10 mV, the drain current (I_d) was measured as a function of back-gate voltage (V_g). FIG. 3A shows the I_d−V_g curve of the GFET before doping (t=0), and after doping for 10 min. (t=10 min.) and 20 min. (t=20 min.). In various embodiments, the doping may be performed for durations as short as 1 min. and as long as 60 min., or longer in some implementations. Similarly, FIG. 3B shows how the I_d−V_g curve evolved over time by exposure to 100 ppm of NO₂ in N₂ at 500 Torr pressure and room temperature. The dashed line in the contour plots of FIGS. 3B and 3D indicates how V_Dirac changed over time. Before NO₂ exposure, the graphene channel was hole-doped with V_Dirac at 18 V. Under ambient conditions, undoped graphene samples become hole-doped due to adsorption of oxygen and water molecules. Exposure to NO₂ gas shifted the Dirac point to higher voltages, which corresponds to enriched concentration of holes.

[0039] As NO₂ exposure continued, V_Dirac became greater than 100 V, beyond the measurement limit. A simple extrapolation suggests that V_Dirac was 165V at the end of the doping process with total duration of 50 min. To determine the stability of NO₂-doped graphene, another GFET was exposed to 100 ppm of NO₂ for 50 min at 500 Torr, and the I_d−V_g curve of the NO₂-doped GFET was characterized immediately after the doping process. Thereafter, the GFET was kept in the same chamber under vacuum at room temperature for 10 days and the I_d−V_g curve was recorded once a day. The results collected in FIGS. 3C and 3D show that V_Dirac decreased from 152V to 74V in 10 days, indicating that NO₂-doped graphene is somewhat unstable. However, the slow recovery rate corresponds to about 0.1V downshift of the Dirac point voltage over 20 min. This indicates that the change in the Dirac point voltage due to recovery can be neglected in the sensitivity measurements in FIG. 2A.

[0040] Referring now to FIGS. 4A and 4B, selected elements of data plots showing electronic properties of graphene gas sensors are illustrated. Specifically, FIG. 4A shows an I_d−V_g curve of an undoped GFET, and FIG. 4B shows an I_d−V_g curve of a NO₂-doped GFET before and after exposure to NH₃. It was observed that adsorption of NO₂ molecules significantly alters the response of the GFET to NH₃ exposure. FIG. 4A shows the I_d−V_g curve of undoped graphene before and after exposure to 100 ppm of NH₃ for 50 min. In FIG. 4A, before exposure to NH₃, V_Dirac was 18 V, with a slight shift to 15.25V after exposure. FIG. 4B, on the other hand, shows the I_d−V_g curve of NO₂-doped graphene before and after 50 min exposure to 100 ppm NH₃. In FIG. 4B, V_Dirac of the GFET was above 100V mainly because NO₂-doping shifted V_Dirac to high voltages. Simple extrapolation shows that V_Dirac of NO₂-doped graphene was 144V before and 98V after exposure to 100 ppm of NH₃. It is noted that the downshift of Dirac voltage after NH₃ exposure cannot be attributed to desorption of NO₂ molecules with a recovery rate of about 0.1V over 20 min.

[0041] Referring now to FIG. 5, selected elements of Raman spectra of graphene gas sensors are illustrated. Raman spectroscopy can be used to probe structural properties and doping of graphene. Specifically, disorder in sp² carbon systems can be monitored from the I(D)/I(G) peak intensity ratio from the Raman spectra. Based on theoretical calculations and experimental observations in both top and bottom gated graphene devices, the G peak stiffens indicating that the FWHM(G) decreases for both electron and hole doping. In comparison, the 2D peak stiffens for hole doping and softens (FWHM is increased) for high concentrations of electron doping. In addition, the I(2D)/I(G) peak intensity ratio decreases for both electron and hole doping. Raman spectra of the graphene gas sensors described herein show that graphene became more hole-doped due to exposure to NO₂ and no disorder was induced by this process. The peak intensity ratio I(D)/I(G) did not change after transfer to the SiO₂/Si substrate and remained similar to the peak intensity ratio I(D)/I(G) of graphene. After doping with NO₂, the G peak upshifted by 3.4 cm⁻¹ and FWHM(G) decreased by 3.3 cm⁻¹. Also after doping with NO₂, the 2D peak FWHM was reduced by 2.6 cm⁻¹, and the peak intensity ratio I(2D)/I(G) decreased by 42%, as shown in FIG. 5.

[0042] Theoretical calculations have shown that NO₂ and NH₃ molecules are physisorbed onto pristine suspended graphene. According to these studies, charge transfer between the adsorbate and graphene is due to the relative position of the density of states (DOS) of the highest occupied molecular orbital (HOMO) and the lowest unoccupied molecular orbital (LUMO) of adsorbate with respect to the Fermi level of graphene. Another mechanism is hybridization of HOMO and LUMO of adsorbate with graphene orbitals. Theoretical calculations have shown that doping graphene with substitutional dopant atoms induces localized perturbations in the structure and the electronic states near dopant atoms, which provides energetically favorable adsorption sites for gas molecules. Since NO₂ molecules are physisorbed onto graphene, no structural disorder is expected to occur for the weak physisorption interaction. On the other hand, several theoretical studies have shown that NO₂ adsorption alters electronic states of graphene near the Dirac point. It is believed that the perturbation in the electronic states of graphene that makes graphene more hole-doped and electron-deficient increases the binding affinity and amount of charge transfer for the adsorption of electron-donating NH₃ molecules. Similar behavior is expected to occur for adsorption of other electron-donating, closed-shell molecules like carbon monoxide (CO) that weakly physisorb to graphene with long molecule-graphene equilibrium distances comparable to that of NH₃-graphene. Therefore, the electronic structure of the adsorbate and adsorbent atoms is scarcely perturbed by the adsorption process and the interaction of the molecule with graphene is a small charge transfer upon adsorption. It is noted that the theoretical calculations on graphene sensors are most relevant for the ideal case of a pristine suspended graphene sample. In the graphene gas sensors described herein, the direct interaction of gas molecules with graphene is suppressed by any residues of organic and water molecules. Therefore, it is contemplated that the sensitivity of graphene to other electron-donating analyte molecules may be enhanced using NO₂-doping.

[0043] The above discussions suggest that the mechanism of sensitivity enhancement for NO₂-doped graphene is comparable to that based on the electric-field effect, where the effect of adsorption of gas molecules on the charge carrier density of graphene can be modified by modulating the Fermi level using the back-gate voltage. The effectiveness of electric field doping for enhancing sensitivity can be compared to that of molecular doping by evaluating the ratio of change of conductance sensitivity ΔS=Δ(ΔG/G₀) to the change of charge carrier concentration Δn, which is given in Equation 1 below.

[0044] $\begin{array}{cc}\frac{\Delta S}{\Delta n}=\frac{qt}{{\mathrm{\varepsilon \varepsilon }}_0}\Delta S/\Delta \left(V_g-V_{\mathrm{Dirac}}\right)& \mathrm{Equation}\left(1\right)\end{array}$
In Equation 1:

• • [0045] q is the electron charge;
  • [0046] t is the SiO₂ thickness;
  • [0047] ∈ is the dielectric constant of SiO₂; and
  • [0048] ∈₀ is the dielectric constant of vacuum permittivity.

[0049] Based on reported values, a graphene ΔS/Δn is about 101 nm² for 2 min. exposure to 550 ppm of NH₃. As measured, V_g is fixed and Δ(Vg−V_Dirac)=−ΔV_Dirac. Assuming that sensitivity linearly depends on gas concentration and exposure time (see FIG. 2a), ΔS/Δn for the NO₂-doped graphene gas sensor in a similar measurement condition is estimated to be 347 nm². This suggests that the local effects of adsorbed NO₂ molecules on graphene electronic structure may be more effective than the modulation of the Fermi level via the electric-field effect. It is noted that different ΔS/Δn values may be partly due to the difference in thickness of graphene samples, which were monolayer in the NO₂-doped graphene gas sensor and few-layer in the back-gated graphene gas sensors, respectively. On the other hand, since the ΔS/Δn values are rather comparable, the mechanism of enhancing sensitivity using NO₂-doping of graphene may be comparable to that of electric-field doping of graphene. This also suggests that similar enhancement is expected to occur for the sensitivity of NO₂-doped graphene to other electron-donating molecules such as CO, among others. It is noted that for back-gated graphene, sensors obtaining sensitivity values comparable to molecular-doped graphene sensors utilized prohibitively large back-gate voltages that may not be compatible with by many electronic devices and circuits. Many common electronic devices are based on CMOS technology, where the maximum voltage is relatively small, around a few volts.

[0050] Using molecular doping, sensitive gas sensors based on graphene and other two-dimensional materials can be developed that can be integrated into CMOS-based electronic devices, which is an important step towards realization of commercial graphene gas sensors.

[0051] In summary, the sensitivity of a graphene gas sensor to a gas molecule may be significantly enhanced using molecular doping, which may be as effective as substitutional doping and more effective than electric-field doping. In particular, the room temperature sensitivity of NO₂-doped graphene to NH₃ was measured to be comparable to the sensitivity of graphene doped with substitutional boron atoms and superior to that of undoped graphene by an order of magnitude. The detection limit for NO₂-doped graphene gas sensors was estimated to be about 200 ppb, which may be improved with extended exposure to NO₂, compared to a detection limit of about 1.4 ppm for undoped graphene. While the stability analysis of NO₂-doped graphene sensors indicates that the doping method may not be completely stable, molecular doping is nevertheless a candidate technique for sensitivity improvement by enhancing the initial carrier concentration. The high levels of doping described herein using molecular doping may be difficult to obtain via the electric field effect in real applications due to restrictions on power consumption and maximum supply voltage, especially for the case of CMOS compatible integrated devices. Electrical characterization and Raman spectroscopy results indicated that the observed sensitivity enhancement was due to localized hole doping of graphene via adsorption of NO₂ molecules.

[0052] The above disclosed subject matter is to be considered illustrative, and not restrictive, and the appended claims are intended to cover all such modifications, enhancements, and other embodiments which fall within the true spirit and scope of the present disclosure. Thus, to the maximum extent allowed by law, the scope of the present disclosure is to be determined by the broadest permissible interpretation of the following claims and their equivalents, and shall not be restricted or limited by the foregoing detailed description.