sensors Article
Plasma-Modified, Epitaxial Fabricated Graphene on SiC for the Electrochemical Detection of TNT Scott A. Trammell *, Sandra C. Hernández, Rachael L. Myers-Ward, Daniel Zabetakis, David A. Stenger, D. Kurt Gaskill and Scott G. Walton U.S. Naval Research Laboratory, Washington, DC 20375, USA;
[email protected] (S.C.H.);
[email protected] (R.L.M.-W.);
[email protected] (D.Z.);
[email protected] (D.A.S.);
[email protected] (D.K.G.);
[email protected] (S.G.W.) * Correspondence:
[email protected]; Tel.: +1-202-404-6063 Academic Editor: W. Rudolf Seitz Received: 27 May 2016; Accepted: 8 August 2016; Published: 12 August 2016
Abstract: Using square wave voltammetry, we show an increase in the electrochemical detection of trinitrotoluene (TNT) with a working electrode constructed from plasma modified graphene on a SiC surface vs. unmodified graphene. The graphene surface was chemically modified using electron beam generated plasmas produced in oxygen or nitrogen containing backgrounds to introduce oxygen or nitrogen moieties. The use of this chemical modification route enabled enhancement of the electrochemical signal for TNT, with the oxygen treatment showing a more pronounced detection than the nitrogen treatment. For graphene modified with oxygen, the electrochemical response to TNT can be fit to a two-site Langmuir isotherm suggesting different sites on the graphene surface with different affinities for TNT. We estimate a limit of detection for TNT equal to 20 ppb based on the analytical standard S/N ratio of 3. In addition, this approach to sensor fabrication is inherently a high-throughput, high-volume process amenable to industrial applications. High quality epitaxial graphene is easily grown over large area SiC substrates, while plasma processing is a rapid approach to large area substrate processing. This combination facilitates low cost, mass production of sensors. Keywords: plasma modified graphene; epitaxial graphene; electrochemical detection; square wave voltammetry
1. Introduction Being all surface, graphene is an excellent material to sense adsorbates as these can alter the charge carrier concentration, leading to measurable changes in conductivity. Schedin et al. [1], exploited this property by demonstrating detection sensitivity due to single molecule events using graphene flakes. Since large-area graphene is required for realistic applications, graphene formed by the sublimation of Si from SiC (0001) wafer surfaces is attractive for sensor development due to ease of fabrication and obvious manufacturability [2]. Using the epitaxial graphene approach, sensitive molecular sensors for many different analytes, i.e., adsorbates, have been successfully demonstrated by others [3–5]. However, this natural sensitivity to a range of analytes is also a disadvantage and thus sensors should be designed to minimize or discriminate against the effects associated with non-target constituents in the operating ambient conditions [6,7]. Chemical functionalization is a means to achieve this. Square wave voltammetry as an electroanalytical technique appears attractive as a discriminating sensor platform, and, moreover, epitaxial graphene has successfully been used as the working electrode; [8] but the near-perfect crystallinity of epitaxial graphene [9,10] is expected to reduce sensitivity for this approach due to the limited number of electrochemically active (oxidation/reduction) sites [11]. Recently, it was implied that reaction sites can be controllably formed on epitaxial graphene as it was demonstrated that plasma-based, oxygen functionalization of epitaxial
Sensors 2016, 16, 1281; doi:10.3390/s16081281
www.mdpi.com/journal/sensors
Sensors 2016, 16, 1281
2 of 9
graphene enhanced the molecular detection sensitivity for planar device geometries as well as greatly improving sensor response time [12]. For these reasons we have undertaken an investigation using controllably functionalized graphene as the electrode in electroanalytical sensing applications. Carbon nanomaterials have recently received attention in electroanalytical sensing applications with examples in implantable sensors [13], ion sensors for point-of-care analysis [14], and trace analysis of heavy metals [15]. In this work, our target analyte is trinitrotoluene (TNT) as sensitive sensors for this and similar substances are of general interest. Analytical techniques for TNT determination include fluorescence immunoassays [16], fluorescence quenching [17], chromatography [18] and Raman spectrometry [19]. However, electroanalytical techniques have an advantage since the instrumentation is simple and can be miniaturized for field application [20,21]. In addition, nitro-aromatics are well known to give unique signature voltammograms for identification [22]. Previous work on graphene-based electrochemical detection of nitroaromatic substances used a variety of preparation methods utilizing graphene or reduced graphene-oxide as a working electrochemical sensor and reports that limits of detection (LOD) for TNT range from 0.2 ppb to 100 ppb [23–28]. In most cases, achieving low LODs regularly requires an electrochemical accumulation step where an applied voltage at the working electrode is used to accumulate TNT at the surface of the electrode [29]. In addition, the preparation methods for these approaches are often limited in scalability, making practical application on the industrial scale intractable. Limits of detection for other forms of carbon include glassy carbon (LOD = 130 ppb) [30] and screen printed carbon (LOD = 200 ppb) [31]. In this manuscript, we show that plasma modified [32] epitaxial graphene [33] shows improved electrochemical detection of TNT (LOD ≈ 20 ppb) without the need for an accumulation step [29], or preconcentration step that utilizes large volume samples to provide sufficient analyte mass for obtaining a detectable signal [34], or an amplification step in which the TNT signal is enhanced using an electrochemical technique termed redox cycling [35], all to improve the electrochemical response. 2. Materials and Methods 2.1. Growth of Epitaxial Fabricated Graphene on SiC The epitaxial graphene was synthesized by means of Si sublimation from semi-insulating (SI), Si-face, on-axis, 6H-silicon carbide (SiC) substrates [36]. The synthesis took place in a commercial chemical vapor deposition reactor at a temperature of 1540 ◦ C and a pressure of 100 mbar under an Ar ambient. The Ar was used to suppress the sublimation of Si in order to control the thickness of the epitaxial graphene layers. Prior to growth, the substrate was in-situ H2 etched to prepare the SiC surface for epitaxial graphene growth by removing any polishing scratches created during the manufacturing of the SiC substrate and forming bilayer stepped morphology. After growth, the sample was cooled in Ar to 800 ◦ C, at which point the reaction tube was evacuated. The thickness of the epitaxial graphene layers was ≈2 monolayers as determined by X-ray photoelectron spectroscopy (XPS) using a Thermo Scientific K-Alpha spectrometer with a monochromatic Al-Kα source using a 400 µm spot size. Chemical analysis was performed using Avantage. 2.2. Plasma Functionalization The epitaxial graphene was chemically modified using electron-beam generated plasmas, which are well-suited for large area plasma processing of atomically thin materials [36]. A detailed description of the processing system and protocols followed here, to introduce and control the relative concentration functional groups on the surface of graphene, can be found in earlier works [37–39]. Briefly, pulsed high-energy electron beams were produced by applying a −2 kV pulse to a linear hollow cathode with a −2 kV square wave for a duration of 1 ms at a duty factor of 10%. The emergent beam passes through a slot (1 × 20 cm2 ) in a ground anode, and terminates at a second grounded anode located further downstream (50 cm). The slotted anode defines the beam cross section and the
Sensors 2016, 16, 1281
3 of 9
volume between these two anodes defines the processing region. The electron beam is magnetically confined to minimize spreading via collisions with the background gas, producing a sheet-like plasma in the processing region. Substrates (graphene) are placed on a processing stage located 2.5 cm from the electron-beam axis. After evacuating the processing reactor to base pressure (~1 × 10−6 Torr), reactive gases, either O2 or N2 depending on the desired functionalities, are introduced at 5% of the total flow rate with argon providing the balance to achieve a target operating pressure. To vary the functional group density on the surface of epitaxial graphene [32], operating pressures were varied from 25 to 90 mTorr by controlling the total flow rate 50 to 180 sccm (standard cubic centimeter per minute). The total plasma processing time was 60 s. 2.3. Surface Characterization Ex-situ surface diagnostics were performed before and immediately after plasma processing to determine the starting material quality and chemistry and the changes resulting from plasma modification. Chemical changes and the resulting bonding characteristics in the graphene due to plasma processing were tracked by XPS. Chemical analysis was performed using Avantage and Unifit softwares. 2.4. Electrochemistry and Cell Design Electrochemical measurements were performed using a potentiostat model #660 from CH Instruments. An electrochemical cell was composed of three electrodes (working, counter (Pt wire), and reference (Ag/AgCl)) all immersed in (or in contact with) an aqueous electrolyte solution. The working electrode was the epitaxial graphene, as received or chemically-modified with an approximately projected area of 1.6 mm2 . Since the current signal generated in electrochemical measurements is proportional to the area of the working electrode, it was advantageous to restrict the active electrode to a well-defined area. When working electrodes are manufactured from rigid substrates, it is practical to form a cell on top of the electrode using an open-ended container and a gasket to prevent leaks. In order to form such a cell and to allow for the electrical connection to the graphene/SiC substrate, a custom cell was designed. The cell contains an integrated gasket, a broad base for clamping, and a cut-out to allow maximum access to the working electrode. The reference and counter electrodes were suspended in the upper funnel-shaped well which also holds the electrolyte. 3. Results 3.1. Plasma Functionalization and Surface Characterization Electron beam generated plasmas produced in O2 /Ar mixtures resulted in the introduction of oxygen functionalities on the surface of graphene. Shown in Figure 1A are the XPS spectra of as-grown and plasma-modified samples, processed at different operating pressures (50–90 mTorr). Following plasma exposures, the presence of oxygen is clear (O1s) and shows a gradual increase in the total oxygen content with increasing operating pressure, going from 3.6% at 50 mTorr to 11.7% at 90 mTorr. A better understanding of the functionalization can be developed by comparing the high-resolution spectra of graphene before and after processing for each plasma processing condition. The results are shown in Figure 1B. The C 1s peaks for untreated EG is composed of three components centered at about 283.6 eV, 284.5 eV and 285.3 eV which correspond to the silicon carbide substrate (Si–C), the epitaxial graphene film (C–C), and the interfacial layer between the SiC and epitaxial graphene, respectively [6]. Here, the graphene peak with sp2 bonding is labeled as EG, the interfacial layer is labeled as IR, and the silicon carbide signal is referred to as SiC. The spectra presented in Figure 1B have been shifted to the intensity of the substrate (Si–C) peak at ≈283.6 eV. Additionally, C1s spectra shows the incorporation of oxygen functionalities which are assigned to carbon bonding involving ethers or alcohols (C–O–C, C–O, or C–OH) and carbonyl bonds (=O) located at ≈286.4 eV and ≈287.1 eV, respectively. Before plasma processing, the O1s scans (Figure 1C) show little to no signal,
Sensors 2016, 16, 1281 Sensors 2016, 16, 1281 Sensors 2016, 16, 1281
4 of 9 4 of 9 4 of 9
signal, indicating no oxygen on the samples. After plasma processing, on the O1s indicating no oxygen presentpresent on the samples. After plasma processing, featuresfeatures on the O1s spectra signal, indicating no oxygen present on the samples. After plasma processing, features on the O1s spectra arise at three different locations corresponding to (Si–O) bonding at ≈534.3 eV, ethers or arise at three different locationslocations corresponding to (Si–O)tobonding at ≈534.3ateV, ethers or ethers alcohols spectra arise at three different corresponding (Si–O) bonding ≈534.3 eV, or alcohols (C–O–C, C–O, or C–OH) at and ≈533.3 eV and carbonyl groups (C=O) at ≈532.2 eV. The O1s de(C–O–C, C–O, or C–OH) at ≈ 533.3 eV carbonyl groups (C=O) at ≈ 532.2 eV. The de-convolved alcohols (C–O–C, C–O,shows or C–OH) at ≈533.3 eV and in carbonyl groups (C=O) at ≈532.2 eV. The deconvolved O1spressure spectra pressure dependence terms theofamount and type of carbonspectra shows dependence in terms of the amount and of type carbon-oxygen bonds formed convolved O1s spectra shows pressure dependence in terms of the amount and type of carbonoxygen bonds formed[32]. in the O2/Ar plasma [32]. in the O /Ar plasma oxygen 2bonds formed in the O2/Ar plasma [32].
Figure 1. 1. X-ray X-rayphotoelectron photoelectronspectroscopy spectroscopy(XPS) (XPS)of ofunmodified, unmodified,and andoxygen oxygenmodified modifiedgraphene, graphene, Figure depicting the (A) survey spectra; (B) the high resolution C1s spectra and (C) the high resolution O1s Figure 1. X-ray photoelectron spectroscopy (XPS) of unmodified, and oxygen modified graphene, depicting the (A) survey spectra; (B) the high resolution C1s spectra and (C) the high resolution O1s spectra. The black curve represents before functionalization, and red curve after functionalization depicting the (A) survey spectra; (B) the high resolution C1s spectra and (C) the high resolution O1s spectra. The black curve represents before functionalization, and red curve after functionalization in (B). in (B). The black curve represents before functionalization, and red curve after functionalization in (B). spectra.
Similarly, nitrogen functionalities can be introduced in a controlled manner using electron beam Similarly, nitrogen functionalities canmixture be introduced introduced in aa controlled controlled manner using electron beam Similarly, nitrogen functionalities can be in using electron beam generated plasmas produced in an N2/Ar at various operating manner pressures as seen in Figure generated plasmas produced in an N 2 /Ar mixture at various operating pressures as seen in Figure generated plasmas produced in components an N2 /Ar mixture at variousdue operating pressures as seen in Figure 2A. 2A. The assignments of the C1s are challenging to the overlapping binding energies 2A. The assignments of the C1s components are challenging due to the overlapping binding energies The assignments of the C1s components are challenging due to the overlapping binding energies of of nitrogen and oxygen species with those of the interfacial layer. However, based on the combined of nitrogen and oxygen species with those of the interfacial layer. However, based on the combined nitrogen and oxygen species with those of the interfacial layer. However, based on the combined features of the C1s (Figure 2B) and N1s (Figure 2C) high resolution spectra, the identifiable peaks featuresatof of≈283.6 the C1s C1s (Figure 2B) 285.3 and N1s N1s (Figure 2C) high high resolution spectra, the identifiable peaks features the 2B) and 2C) spectra, peaks located eV,(Figure 284.5 eV, eV, (Figure are attributed to resolution the Si–C, C–C sp2, the andidentifiable interfacial layer, 2 2 located at at ≈ ≈283.6 eV, eV, 285.3 eV, are attributed attributed to (>286 the Si–C, Si–C, C–Cattributed sp ,, and located 283.6 eV, 284.5 284.5 eV, at 285.3 eV, binding are to the C–C sp and interfacial interfacial layer, respectively. The peaks located higher energies eV) are to nitrogenlayer, and respectively. The peaks located at higher binding energies (>286 eV) are attributed to nitrogen and respectively. The peaksNotably located C–N at higher energies are attributed nitrogen oxygen functionalities. and binding C–O a ≈286.1 eV,(>286 C=N eV) at ≈287.4 eV, C=O to and N–C=Oand at oxygen functionalities. Notably C–N and C–O a ≈286.1 eV, C=N at ≈287.4 eV, C=O and N–C=O at oxygen functionalities. Notably a≈ 286.1 eV,that C=N at ≈287.4 eV, C=O and N–C=O ≈288.3 eV, and O–C=O at ≈289.4 eV.C–N The and N1s C–O spectra indicates nitrogen functionalities are present ≈288.3 eV,in and O–C=O at ≈289.4 eV.configurations. The indicates that nitrogen functionalities are present at ≈288.3 eV, and O–C=O at ≈289.4 eV.N1s Thespectra N1s spectra indicates that nitrogen functionalities are primarily the amide and pyrolic primarily in the amide and pyrolic configurations. present primarily in the amide and pyrolic configurations.
Figure 2. XPS of unmodified, and nitrogen modified graphene, depicting the (A) survey spectra; (B) 2.resolution XPS of ofunmodified, unmodified, and nitrogen modified graphene, depicting survey spectra; (B) Figure XPS and nitrogen modified graphene, depicting thethe (A)(A) survey spectra; (B) the the high2. C1s spectra and (C) the high resolution N1s spectra. The dotted vertical lines the high resolution C1s spectra and (C) the high resolution N1s spectra. The dotted vertical lines high resolution C1s spectra and (C) the high resolution N1s spectra. The dotted vertical lines represent represent approximate peak positions. represent approximate peak positions. approximate peak positions.
3.2. Electrochemistry of TNT at Modified Graphene Modified Graphene Graphene 3.2. Electrochemistry of TNT at Modified The electrochemical response of the reduction of TNT at the graphene working electrode was at the the graphene graphene working technique electrode was The electrochemical responsevoltammetry, of the reduction of TNT at working electrode characterized using square-wave which is a common electroanalytical for a common electroanalytical characterized using square-wave voltammetry, which is a common electroanalytical for analytical applications since it diminishes non-faradaic charging currents that develop attechnique the solution analytical applications applications chargingtechniques currents that that[40]. develop analytical since itemploying diminishes potential non-faradaic charging currents develop the solution electrode interface when sweeping In atsquare-wave electrode interface when employing potential sweeping techniques [40]. In square-wave electrode interface when employing potential sweeping techniques [40]. In square-wave voltammetry, voltammetry, a potential staircase is overlaid on the voltage ramp as the voltage is swept in the a potential staircase is voltage overlaid on the voltage ramp asofthe isoccurs swept in avoltammetry, potential staircase overlaid on the as the voltage is swept in voltage the desired rangeas atthe the desired range at theisworking electrode. The ramp reduction (or oxidation) the analyte desiredapproaches range at the working electrode. Thethe reduction (or ofthe thevoltage analyte occurs as the working electrode. The reduction (or oxidation) of the analyte as approaches its voltage its formal redox potential, potential of oxidation) theoccurs first minimum is the first reduction, voltage approaches its formal redox potential, the potential of the first minimum is the first reduction, named the cathodic peak potential, Epc. The current is then measured at different points of the named the cathodic peak potential, Epc. The current is then measured at different points of the
Sensors 2016, 16, 1281
5 of 9
Sensors 16, 1281 formal2016, redox potential,
5 of 9 the potential of the first minimum is the first reduction, named the cathodic peak potential, Epc . The current is then measured at different points of the potential waveform to potential to minimize charginginterface. at the solution electrode interface. Since faster nonminimize waveform capacitive charging at thecapacitive solution electrode Since non-faradic current decays faradic current decays faster than faradic current [40], a square-wave voltammogram can than faradic current [40], a square-wave voltammogram can be generated by taking the differencebe of generated by taking the difference current betweencharging the points subtracting out detection the capacitive current between the points subtractingofout the capacitive current making the limits charging making detection limits typically better then cyclic voltammetry. In measured this way, typically current better then cyclicthe voltammetry. In this way, the maximum absolute value of the the maximum absolute value the measured current (measured or the net with cathodic peak current, Ipc, is current or the net cathodic peak of current, Ipc , is determined respect to the background determined (measured with respect to the background current taken without TNT) at E pc . current taken without TNT) at Epc . To the square wave voltammograms, To record record the square wave voltammograms, the the modified modified graphene graphene samples samples were were mounted mounted in in aa custom custom built built electrochemical electrochemical cell cell and and background background measurements measurements were were recorded recorded before before the the addition addition of TNT. Representative examples of the square wave voltammetry for the reduction of 10 of TNT. Representative examples of the square wave voltammetry for the reduction of 10 ppm ppm of of TNT at these modified graphene samples are shown in Figure 3. The square wave voltammograms TNT at these modified graphene samples are shown in Figure 3. The square wave voltammograms were saline (PBS, (PBS, pH pH 7.4; were measured measured in in air-saturated air-saturated phosphate phosphate buffered buffered saline 7.4; pH pH did did not not change change with with TNT TNT admixtures) solutions from 0 to −1.0 V vs. Ag/AgCl reference electrode. With unmodified graphene, admixtures) solutions from 0 to −1.0 V vs. Ag/AgCl reference electrode. With unmodified graphene, the reduction of of TNT TNT (E (Epc pc = = −0.8 the first first reduction −0.8VVvs. vs.Ag/AgCl) Ag/AgCl)isisshifted shiftedtotosignificantly significantlymore morenegative negative voltages voltages compared to oxygen functionalized graphene (E pc = −0.58 V vs. Ag/AgCl) and nitrogen functionalized compared to oxygen functionalized graphene (Epc = −0.58 V vs. Ag/AgCl) and nitrogen functionalized graphene V vs. Ag/AgCl) consistent with with a similar reportreport [41]. In[41]. comparison, other carbon graphene (E (Epcpc= =−0.5 −0.5 V vs. Ag/AgCl) consistent a similar In comparison, other electrodes such assuch glassy show the firstthe reduction near −0.575 V0.575 vs. Ag/AgCl [34]. In [34]. square carbon electrodes as carbon glassy carbon show first reduction near − V vs. Ag/AgCl In wave voltammetry, changes in E pc and I pc are a reflection of both thermodynamic and kinetic square wave voltammetry, changes in Epc and Ipc are a reflection of both thermodynamic and kinetic properties ofthe theelectrochemical electrochemical system, hence oxygen functionalized working electrode properties of system, hence thesethese oxygen functionalized working electrode results results indicate that functionalization either increases the active surface area of the graphene or indicate that functionalization either increases the active surface area of the graphene or increases increases the rate of reduction of the TNT or both. The more negative E pc would decrease the the rate of reduction of the TNT or both. The more negative Epc would decrease the resistance to resistance to with interference interference oxygen.with oxygen.
Figure Figure 3. 3. Square Square wave wave voltammetry voltammetry 10 10 ppm ppm of of TNT TNT in in air air saturated saturated phosphate phosphate buffered buffered saline saline (PBS) (PBS) buffer measured at a fabricated epitaxial graphene working electrode on SiC (red), with buffer measured at a fabricated epitaxial graphene working electrode on SiC (red), with oxygen oxygen modification (90 mTorr) mTorr)(blue) (blue)and andwith withnitrogen nitrogenmodification modification mTorr (brown). green trace is modification (90 5050 mTorr (brown). TheThe green trace is the the background signal without TNT. Frequency = 60 Hz and Amplitude = 25 mV. background signal without TNT. Frequency = 60 Hz and Amplitude = 25 mV.
In Figure 4, Ipc and corresponding Epc are shown for a 10 ppm sample of TNT in PBS vs. nitrogen In Figure 4, Ipc and corresponding Epc are shown for a 10 ppm sample of TNT in PBS vs. nitrogen or oxygen functionalized content of the graphene samples. Results for unfunctionalized graphene, or oxygen functionalized content of the graphene samples. Results for unfunctionalized graphene, are given for 0% content. Figure 4A shows that for the nitrogen functionalized samples there is little are given for 0% content. Figure 4A shows that for the nitrogen functionalized samples there is little change to Ipc with increasing content, however, Epc becomes less negative with increasing nitrogen change to I with increasing content, however, Epc becomes less negative with increasing nitrogen content andpcbegins to saturate at 13 atomic % (Figure 4B). For oxygen functionalized samples, the content and begins to saturate at 13 atomic % (Figure 4B). For oxygen functionalized samples, the oxygenated graphene Ipc of TNT is significantly enhanced, being 10-fold larger compared to nitrogen oxygenated graphene I of TNT is significantly enhanced, being 10-fold larger compared to nitrogen functionalized samplespc as can be seen in Figure 4A. Yet, Epc also becomes less negative with increasing functionalized samples as can be seen in Figure 4A. Yet, Epc also becomes less negative with increasing content and saturates earlier, near 4 atomic %, than for the nitrogen functionalized samples. These content and saturates earlier, near 4 atomic %, than for the nitrogen functionalized samples. These results imply the oxygen functionalized samples should have increased TNT detection sensitivity. results imply the oxygen functionalized samples should have increased TNT detection sensitivity.
500
Nitrogen modification
400
Oxygen modification
B Epc (V vs Ag/Ag) for TNT Epc (V vs Ag/Ag) for TNT
Ipc, nA for TNT
A
Sensors 3002016, 16, 1281 Sensors 2016, 16, 1281
Ipc, nA for TNT
A
200 500 100
Nitrogen modification
400 0
Oxygen modification 15% 20%
300 0%
5%
10%
B
Surface coverage in percent 200
-1
Nitrogen modification
-0.8
Oxygen modification
6 of 9 6 of 9
-0.6 -0.4 -1 -0.2
Nitrogen modification
-0.8 0
Oxygen modification 20%
-0.6 0%
5% 10% 15% Surface coverage in percent
-0.4
Figure 4. (A) Net cathodic peak current (Ipc) and (B) cathodic peak potential (Epc) vs. surface coverage 100 -0.2 in percent of nitrogen (), and oxygen ( ) with bare graphene at 0%. 0
0
3.3. Trace for Oxygen Modified20% EG 0% Detection 5% of TNT10% 15%
0%
5% 10% 15% Surface coverage in percent
Surface coverage in percent
20%
0.04 0.16
Current, A
0 0.12
B 0
1000
2000 3000 [TNT], ppb
0.08
4000
5000
0.02 0.08 0 0.06
Current, A
A
Current, A
Current, A
We quantified the increased TNT detection sensitivity for oxygen functionalized graphene by acquiring voltammograms various concentrations of TNT Figure 5 shows Figure 4. (A) (A) Net Net cathodic cathodic peak current current (I ) and (B) cathodic peak potential (E pc))in vs.PBS. surface coverage pc Figuresquare-wave 4. peak (Ifor pc) and (B) cathodic peak potential (Epc vs. surface coverage in percent of nitrogen ( ♦ ), and oxygen (2) with bare graphene at 0%. the results of these measurements where Figure 5A displays a larger TNT concentration range (0 to in percent of nitrogen (), and oxygen ( ) with bare graphene at 0%. 5000 ppb) than 5B (0 to 1400 ppb). It is readily apparent that the oxygen functionalized graphene Trace Detection of TNT for over Oxygen Modified yields sensitivities spanning 100-fold. In EG addition, dependence on TNT concentration can be fit 3.3. Trace EG to a two-site Langmuir isotherm (Equation (1)) suggesting different active sites with different TNT graphene by by We quantified the increased TNT detection sensitivity for oxygen functionalized functionalized graphene affinities on the graphene surface. acquiring square-wave voltammograms for various concentrations of TNT in PBS. Figure 5 shows acquiring square-wave voltammograms for various concentrations of TNT in PBS. Figure resultsNet thesemeasurements measurements where Figure displays a larger peak amplitude = a((K a[TNT])/(1 +5A (K5A adisplays [TNT])) b[TNT])/(1 + concentration (Kb[TNT])) the results ofofthese where Figure a+ b((K larger TNTTNT concentration rangerange (0(1) to (0 to 5000 to 1400 ppb). is readily apparentthat thatthe theoxygen oxygenfunctionalized functionalized graphene graphene 5000 ppb)ppb) thanthan 5B (05Bto(01400 ppb). It isIt readily apparent In Equation (1), a is the proportion of current due to the reduction of TNT at sites a, Ka is the yields sensitivities spanning over 100-fold. In addition, addition, dependence dependence on TNT concentration concentration can be fit binding constant of TNT at sites a, b is the proportion of current due to the reduction of TNT at sites isotherm (Equation (Equation (1)) (1)) suggesting suggesting different different active active sites sites with with different different TNT TNT to a two-site Langmuir isotherm b, and Kb is the binding constant of TNT at sites b. The sites with high affinities for TNT saturate with affinities on the graphene surface. increasing TNT concentration and the sites with lower affinities give rise to the large dynamic range. At 40 Net ppbpeak (thepeak lowest TNT=concentration tested) thea [TNT])) S/N ratio is 6. bFrom this, +we of Net amplitude a[TNT])/(1 + (K + b((K [TNT])/(1 (Kestimate b[TNT])) a LOD (1) amplitude a(=Ka((K (1) a [TNT]) / (1 + (Ka [TNT])) + b(Kb [TNT]) / (1 + (Kb [TNT])) 20 ppb based on the analytical standard S/N ratio of 3. In Equation (1), a is the proportion of current due to the reduction of TNT at sites a, Ka is the binding constant of TNT at sites a, b is the proportion of current due to the reduction of TNT at sites A 0.16 B 0.08 b, and Kb is the binding constant of TNT at sites b. The sites with high affinities for TNT saturate with increasing TNT concentration and the sites with lower affinities give rise to the large dynamic range. 0.12 0.06 At 40 ppb (the lowest TNT concentration tested) the S/N ratio is 6. From this, we estimate a LOD of 0.08based on the analytical standard S/N ratio of 3. 0.04 20 ppb
0
200
400
600 800 [TNT], ppb
1000
1200
1400
0.04
Figure Dose response response curves curves for for oxygen oxygen functionalized functionalized epitaxial epitaxial graphene. graphene. (A) (A) Net Net peak peak amplitude, amplitude, Figure 5. 5. Dose 0.04 0.02 Ipc vs. {TNT}, for high ppb range; (B) Net peak amplitude vs. {TNT} for low concentration. The line is a Ipc vs. {TNT}, for high ppb range; (B) Net peak amplitude vs. {TNT} for low concentration. The line is least squares fit of the data to Equation 1 with a = 0.047, b = 3.7, Ka = 0.05 µM−1 ,−1Kb = 1.3 × 10−4−4µM−−11 , a 0least squares fit of the data to Equation 1 with a = 0.047,0b = 3.7, Ka = 0.05 µM , Kb = 1.3 × 10 µM , (K units has1000 been converted to µM−−114000 ) range. 0 2000 3000 0 200 400 600 800 1000 1200 1400 (K units has been converted to µM ) range. 5000 [TNT], ppb
[TNT], ppb
In Equation (1), a is the proportion of current due to the reduction of TNT at sites a, Ka is the 4. Discussion Figure 5. Doseof response functionalized epitaxialdue graphene. Net peak binding constant TNT atcurves sites a,forb oxygen is the proportion of current to the (A) reduction ofamplitude, TNT at sites b, For carbon electrodes, surface preparation is often critical for good electrochemical response of I pc vs. {TNT}, for high ppb range; (B) Net peak amplitude vs. {TNT} for low concentration. The line is and Kb is the binding constant of TNT at sites b. The sites with high affinities for TNT saturate with analytes Thisfithas been when with comparing basal plane to edge plane graphite, surface a least[42]. squares of the datashown to Equation alower = 0.047, b = 3.7, K a = 0.05 µM−1, Kb = 1.3 × 10−4 µM−1, increasing TNT concentration and the sites1 with affinities give rise to the large dynamic range. modification of glassy carbon, and more recently with different preparations of graphene and −1 ) range. (K units has been converted to µM At 40 ppb (the lowest TNT concentration tested) the S/N ratio is 6. From this, we estimate a LOD graphene oxides [29]. A common theme to describe the phenomenon is that electron transfer between of 20 ppb based on the analytical standard S/N ratio of 3. the analyte and carbon surface occurs at defect sites, or chemical groups from the surface 4. Discussion 4. Discussion For carbon electrodes, surface preparation is often critical for good electrochemical response of analytes [42]. This has beensurface shown preparation when comparing basal plane edgeelectrochemical plane graphite,response surface For carbon electrodes, is often critical forto good modification of This glassy and more withbasal different graphene and of analytes [42]. hascarbon, been shown whenrecently comparing plane preparations to edge planeof graphite, surface graphene oxides [29]. A common theme to describe the phenomenon is that electron transfer between the analyte and carbon surface occurs at defect sites, or chemical groups from the surface
Sensors 2016, 16, 1281
7 of 9
modification of glassy carbon, and more recently with different preparations of graphene and graphene oxides [29]. A common theme to describe the phenomenon is that electron transfer between the analyte and carbon surface occurs at defect sites, or chemical groups from the surface modification. In our case, the initial graphene synthesized on SiC gives a broad signal for TNT reduction at a more negative potential which suggests that graphene without any chemical modification has a slower electron transfer rate requiring a higher over potential to generate the TNT signal. Modification with oxygen and nitrogen both lower the over potential for the reduction of TNT. However, with oxygen modification, a significant increase in signal amplitude is shown. In this case, a two-site Langmuir isotherm can fit the dose response curve and suggests that oxygen modification generates several sites on the graphene surfaces with different affinities of TNT consistent with the range of functional moieties found in the XPS spectra. 5. Conclusions We have shown a 6-fold increase in sensitivity towards the lower detection limit of TNT using chemically-modified graphene on a SiC surface compared to untreated graphene. The graphene surface was chemically modified using electron beam generated plasmas produced in oxygen- and nitrogen-containing backgrounds to introduce oxygen or nitrogen moieties. While both treatments enabled the enhancement of the electrochemical signal for TNT, oxygen-functionalized epitaxial graphene provided a more pronounced response. The oxygen functionalized graphene revealed a two-site Langmuir isotherm response extending over 2 orders of magnitude with a LOD of 20 ppb for TNT in PBS. The approach described here to fabricate graphene sensors sensitive to TNT is attractive for low cost, high-volume industrial production. This is because high quality epitaxial graphene is easily grown over large area SiC substrates and the plasma processing approach to oxygen functionalization is amenable to large area substrate processing. Acknowledgments: This work was funded by the Office of Naval Research. Author Contributions: S.A.T. designed and recorded the electrochemical experiments, analyzed the data, and acted as corresponding author; S.C.H. performed the plasma functionalization and surface charaterization; R.L.MW. performed the growth of epitaxial fabricated graphene on SiC; D.Z. designed and constructed the electrochemical cell; D.A.S. suggested the use of graphene and assisted in editing the manuscript; D.K.G. assisted with the epitaxial graphene synthesis and assisted in editing the manuscript; S.G.W. developed the plasma processing system and helped analyze the results. All authors contributed to the manuscript. Conflicts of Interest: The authors declare no conflict of interest.
Abbreviations The following abbreviations are used in this manuscript: PBS TNT EG
phosphate buffered saline 2,4,6-trinitrotoluene epitaxial graphene
References 1. 2.
3. 4. 5.
Schedin, F.; Geim, A.K.; Morozov, S.V.; Hill, E.W.; Blake, P.; Katsnelson, M.I.; Novoselov, K.S. Detection of individual gas molecules adsorbed on graphene. Nat. Mater. 2007, 6, 652–655. [CrossRef] [PubMed] Nyakiti, L.O.; Wheeler, V.D.; Garces, N.Y.; Myers-Ward, R.L.; Eddy, C.R.; Gaskill, D.K. Enabling graphene-based technologies: Toward wafer-scale production of epitaxial graphene. MRS Bull. 2012, 37, 1149–1157. [CrossRef] Novikov, S.; Lebedeva, N.; Satrapinski, A. Ultrasensitive NO2 Gas Sensor Based on Epitaxial Graphene. J. Sens. 2015, 2015. [CrossRef] Iezhokin, I.; Offermans, P.; Brongersma, S.H.; Giesbers, A.J.M.; Flipse, C.F.J. High sensitive quasi freestanding epitaxial graphene gas sensor on 6H-SiC. Appl. Phys. Lett. 2013, 103. [CrossRef] Pearce, R.; Iakimov, T.; Andersson, M.; Hultman, L.; Spetz, A.L.; Yakimova, R. Epitaxially grown graphene based gas sensors for ultrasensitive NO2 detection. Sens. Actuators B Chem. 2011, 155, 451–455. [CrossRef]
Sensors 2016, 16, 1281
6. 7.
8.
9. 10.
11. 12.
13. 14. 15.
16.
17.
18. 19. 20. 21.
22. 23. 24.
25.
26.
8 of 9
Kong, L.; Enders, A.; Rahman, T.S.; Dowben, P.A. Molecular adsorption on graphene. J. Phys. Condens. Matter 2014, 26. [CrossRef] [PubMed] Giusca, C.E.; Panchal, V.; Munz, M.; Wheeler, V.D.; Nyakiti, L.O.; Myers-Ward, R.L.; Gaskill, D.K.; Kazakova, O. Water affinity to epitaxial graphene: The impact of layer thickness. Adv. Mater. Interfaces 2015, 2. [CrossRef] Daniels, K.M.; Shetu, S.; Staser, J.; Weidner, J.; Williams, C.; Sudarshan, T.S.; Chandrashekhar, M.V.S. Mechanism of Electrochemical Hydrogenation of Epitaxial Graphene. J. Electrochem. Soc. 2015, 162, E37–E42. [CrossRef] Avouris, P.; Dimikratopoulos, C. Graphene: Synthesis and applications. Mater. Today 2012, 86, 86–97. [CrossRef] Fromm, F.; Oliveira, M.H.; Molina-Sanchez, A.; Hundhausen, M.; Lopes, J.M.J.; Riechert, H.; Wirtz, L.; Seyller, T. Contribution of the buffer layer to the Raman spectrum of epitaxial graphene on SiC(0001). New J. Phys. 2013, 15, 043031. [CrossRef] Hernandez, S.C.; Bezares, F.J.; Robinson, J.T.; Caldwell, J.D.; Walton, S.G. Controlling the local chemical reactivity of graphene through spatial functionalization. Carbon 2013, 60, 84–93. [CrossRef] Nagareddy, V.K.; Chan, H.K.; Hernandez, S.C.; Wheeler, V.D.; Nyakiti, L.O.; Myers-Ward, R.L.; Eddy, C.R.; Goss, J.P.; Wright, N.G.; Walton, S.G.; et al. Improved chemical detection and ultra-fast recovery using oxygen functionalized epitaxial graphene sensors. IEEE Sens. J. 2013, 13, 2810–2817. [CrossRef] Wujcik, E.K.; Monty, C.N. Nanotechnology for implantable sensors: Carbon nanotubes and graphene in medicine. Wiley Interdiscip. Rev. Nanomed. Nanobiotechnol. 2013, 5, 233–249. [CrossRef] [PubMed] Wujcik, E.K.; Blasdel, N.J.; Trowbridge, D.; Monty, C.N. Ion sensor for the quantification of sodium in sweat samples. IEEE Sens. J. 2013, 13, 3430–3436. [CrossRef] Zhou, N.; Li, J.; Chen, H.; Liao, C.; Chen, L. A functional graphene oxide-ionic liquid composites-gold nanoparticle sensing platform for ultrasensitive electrochemical detection of Hg2+ . Analyst 2013, 138, 1091–1097. [CrossRef] [PubMed] Charles, P.T.; Adams, A.A.; Howell, P.B.; Trammell, S.A.; Deschamps, J.R.; Kusterbeck, A.W. Fluorescence-based Sensing of 2,4,6-Trinitrotoluene (TNT) Using a Multi-channeled Poly(methyl methacrylate) (PMMA) Microimmunosensor. Sensors 2010, 10, 876–889. [CrossRef] [PubMed] Gao, D.; Wang, Z.; Liu, B.; Ni, L.; Wu, M.; Zhang, Z. Resonance energy transfer-amplifying fluorescence quenching at the surface of silica nanoparticles toward ultrasensitive detection of TNT. Anal. Chem. 2008, 80, 8545–8553. [CrossRef] [PubMed] Walsh, M.E. Determination of nitroaromatic, nitramine, and nitrate ester explosives in soil by gas chromatography and an electron capture detector. Talanta 2001, 54, 427–438. [CrossRef] Liu, M.; Chen, W. Graphene nanosheets-supported Ag nanoparticles for ultrasensitive detection of TNT by surface-enhanced Raman spectroscopy. Biosen. Bioelectron. 2013, 46, 68–73. [CrossRef] [PubMed] Wang, J.; Thongngamdee, S. On-line electrochemical monitoring of (TNT) 2,4,6-trinitrotoluene in natural waters. Anal. Chim. Acta 2003, 485, 139–144. [CrossRef] Ryan, P.; Zabetakis, D.; Stenger, D.A.; Trammell, S.A. Integrating Paper Chromatography with Electrochemical Detection for the Trace Analysis of TNT in Soil. Sensors 2015, 15, 17048–17056. [CrossRef] [PubMed] Galik, M.; O’Mahony, A.M.; Wang, J. Cyclic and square-wave voltammetric signatures of nitro-containing explosives. Electroanalysis 2011, 23, 1193–1204. [CrossRef] Tang, L.; Feng, H.; Cheng, J.; Li, J. Uniform and rich-wrinkled electrophoretic deposited graphene film: A robust electrochemical platform for TNT sensing. Chem. Commun. 2010, 46, 5882–5884. [CrossRef] [PubMed] Tan, S.M.; Chua, C.K.; Pumera, M. Graphenes prepared from multi-walled carbon nanotubes and stacked graphene nanofibers for detection of 2,4,6-trinitrotoluene (TNT) in seawater. Analyst 2013, 138, 1700–1704. [CrossRef] [PubMed] Ong, B.K.; Poh, H.L.; Chua, C.K.; Pumera, M. Graphenes Prepared by hummers, staudenmaier and hofmann methods for analysis of TNT-based nitroaromatic explosives in seawater. Electroanalysis 2012, 24, 2085–2093. [CrossRef] Goh, M.S.; Pumera, M. Graphene-based electrochemical sensor for detection of 2,4,6-trinitrotoluene (TNT) in seawater: The comparison of single-, few-, and multilayer graphene nanoribbons and graphite microparticles. Anal. Bioanal. Chem. 2011, 399, 127–131. [CrossRef] [PubMed]
Sensors 2016, 16, 1281
27. 28.
29. 30.
31. 32.
33.
34.
35. 36.
37. 38.
39.
40. 41.
42.
9 of 9
Chen, T.W.; Sheng, Z.H.; Wang, K.; Wang, F.B.; Xia, X.H. Determination of explosives using electrochemically reduced graphene. Chem. Asian J. 2011, 6, 1210–1216. [CrossRef] [PubMed] Zhang, R.; Sun, C.L.; Lu, Y.J.; Chen, W. Graphene nanoribbon-supported PtPd concave nanocubes for electrochemical detection of TNT with high sensitivity and selectivity. Anal. Chem. 2015, 87, 12262–12269. [CrossRef] [PubMed] O’Mahony, A.M.; Wang, J. Nanomaterial-based electrochemical detection of explosives: A review of recent developments. Anal. Methods 2013, 5, 4296–4309. [CrossRef] Trammell, S.A.; Zeinali, M.; Melde, B.J.; Charles, P.T.; Velez, F.L.; Dinderman, M.A.; Kusterbeck, A.; Markowitz, M.A. Nanoporous organosilicas as preconcentration materials for the electrochemical detection of trinitrotoluene. Anal. Chem. 2008, 80, 4627–4633. [CrossRef] [PubMed] Wang, J.; Lu, F.; MacDonald, D.; Lu, J.M.; Ozsoz, M.E.S.; Rogers, K.R. Screen-printed voltammetric sensor for TNT. Talanta 1998, 46, 1405–1412. [CrossRef] Hernández, S.C.; Wheeler, V.D.; Osofsky, M.S.; Jernigan, G.G.; Nagareddy, V.K.; Nath, A.; Lock, E.H.; Nyakiti, L.O.; Myers-Ward, R.L.; Sridhara, K.; et al. Plasma-based chemical modification of epitaxial graphene with oxygen functionalities. Surf. Coat. Technol. 2014, 241, 8–12. [CrossRef] Jernigan, G.G.; VanMil, B.L.; Tedesco, J.L.; Tischler, J.G.; Glaser, E.R.; Davidson, A., 3rd; Campbell, P.M.; Gaskill, D.K. Comparison of epitaxial graphene on Si-face and C-face 4H SiC formed by ultrahigh vacuum and RF furnace production. Nano Lett. 2009, 9, 2605–2609. [CrossRef] [PubMed] Trammell, S.A.; Melde, B.J.; Zabetakis, D.; Deschamps, J.R.; Dinderman, M.A.; Johnson, B.J.; Kusterbeck, A.W. Electrochemical detection of TNT with in-line pre-concentration using imprinted diethylbenzene-bridged periodic mesoporous organosilicas. Sens. Actuators B Chem. 2011, 155, 737–744. [CrossRef] Trammell, S.A.; Velez, F.; Charles, P.T.; Kusterbeck, A. Electrochemical Detection of 2,4,6-Trinitrotoluene Using Interdigitated Array Electrodes. Anal. Lett. 2008, 41, 2634–2645. [CrossRef] Walton, S.G.; Boris, D.R.; Hernández, S.C.; Lock, E.H.; Petrova, T.B.; Petrov, G.M.; Fernsler, R.F. Electron Beam Generated Plasmas for Ultra Low Te Processing. ECS J. Solid State Sci. Technol. 2015, 4, N5033–N5040. [CrossRef] Baraket, M.; Stine, R.; Lee, W.K.; Robinson, J.T.; Tamanaha, C.R.; Sheehan, P.E.; Walton, S.G. Aminated graphene for DNA attachment produced via plasma functionalization. App. Phys. Lett. 2012, 100. [CrossRef] Walton, S.G.; Hernández, S.C.; Baraket, M.; Wheeler, V.D.; Nyakiti, L.O.; Myers-Ward, R.L.; Eddy, C.R.; Gaskill, D.K. Plasma-based chemical modification of graphene. Mater. Sci. Forum 2012, 657, 717–720. [CrossRef] Foley, B.M.; Hernández, S.C.; Duda, J.C.; Robinson, J.T.; Walton, S.G.; Hopkins, P.E. Modifying surface energy of graphene via plasma-based chemical functionalization to tune thermal and electrical transport at metal interfaces. Nano Lett. 2015, 15, 4876–4882. [CrossRef] [PubMed] Bard, A.J.; Faulkner, L.R. Electrochemical Methods: Fundamentals and Applications; Wiley: Hoboken, NJ, USA, 2001. Chen, T.-W.; Xu, J.-Y.; Sheng, Z.-H.; Wang, K.; Wang, F.-B.; Liang, T.-M.; Xia, X.-H. Enhanced electrocatalytic activity of nitrogen-doped graphene for the reduction of nitro explosives. Electrochem. Commun. 2012, 16, 30–33. [CrossRef] McCreery, R.L.; Cline, K.K. Carbon Electrodes. In Laboratory Techniques in Electroanalytical Chemistry, 2nd ed.; Kissinger, P.T., Heineman, W.R., Eds.; Marcel Dekker, Inc.: New York, NY, USA, 1996; pp. 293–330. © 2016 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC-BY) license (http://creativecommons.org/licenses/by/4.0/).
