Article pubs.acs.

org/JPCC

Single-Walled Carbon NanotubePoly(porphyrin) Hybrid for Volatile Organic Compounds Detection

Tapan Sarkar, Sira Srinives, Santanu Sarkar,, Robert C. Haddon,,, and Ashok Mulchandani*,

Department of Chemical and Environmental Engineering, Department of Chemistry, and Center for Nanoscale Science Engineering, University of California, Riverside, Riverside, California 92521, United States
S Supporting Information *

ABSTRACT: Porphyrins due to their unique and interesting physicochemical properties have been widely investigated as functional materials for chemical sensor fabrication. However, their poor conductivity is a major limitation toward the realization of porphyrin-based eld-eect transistor/chemiresistor sensor. The issue of conductivity can be overcome by exploiting the excellent electrical property of single-walled carbon nanotubes (SWNTs) to make a SWNTs-based hybrid device in which SWNTs would act as a transducer and porphyrin as a sensory layer. The present attempt was to fabricate a SWNTs poly(tetraphenylporphyrin) hybrid through electrochemical route and to evaluate its potential as a low-power chemiresistor sensor for sensing acetone vapor as a model for volatile organic compounds. Functionalization of SWNTs with porphyrin polymer by the electrochemical method resulted in a fuller coverage of SWNTs surface compared to a partial coverage by adsorption and thereby higher sensitivity. SWNTs were coated with poly(tetraphenylporphyrin) of dierent thickness by applying dierent charge density to optimize sensing performance. Dierences in sensing performance were noticed for hybrids fabricated at varying charge densities, and the optimum sensing response was found at 19.65 mC/cm2. The hybrid exhibited a wide dynamic range for acetone vapor sensing from 50 to 230 000 ppm with a limit of detection of 9 ppm. The eld-eect transistor studies showed a negative threshold voltage shift and almost constant transconductance when exposed to air/analyte, indicating electrostatic gating dominated sensing mechanism. Further, the results conrmed a good stability of the device over a period of 180 days. The long-term device stability along with the sensing capability at low analyte concentration with a wide dynamic range and easily scalable fabrication technique signify the potential of SWNTpoly(porphyrin) hybrid for volatile organic compound sensing applications.

Porphyrins are organic macrocyclic compounds with interesting chemical and optical properties as well as good chemical stability.1 Their ability to interact with a variety of chemicals through weak van der Waals forces and coordination interaction with the inner core metal ion2,3 resulting in measurable physical property change make them an attractive class of sensing materials. The sensing capabilities of porphyrin have been explored based on optical47 and mass detection.8,9 However, limited sensitivity, lower dynamic range, and bulky instrument are major limitations of these devices. The reports of work function of metalloporphyrins and its modulation upon adsorption of VOCs are suggestive of the potential of using porphyrin-functionalized chemical eld-eect transistor (ChemFET) as sensor for VOCs. However, eorts to realize ChemFET based on porphyrin lm alone were not successful because the conductivity of porphyrin is not enough for an ecient biasing.10,11 To alleviate this limitation, Anderson et al.10 deposited a layer of gold over the SiO2 and deposited porphyrin as a self-assembled monolayer on the metal layer. While the metal layer was useful to bias the device,
2013 American Chemical Society

INTRODUCTION

its thickness had to be very carefully controlled to avoid the masking of porphyrin work function. Single-walled carbon nanotubes (SWNTs) have unique electrical12,13 and structural14 properties, leading to usability in diverse applications. Its nanometer range dimensions enable fabrication of high-density sensor device within a limited space. In particular, high electrical mobility, which facilitates fabrication of low-power device, and the property of conductance change upon adsorption of analyte gas molecules make SWNTs a promising material for sensor development.1517 However, lack of sensor performance in terms of sensitivity and selectivity is intrinsic in carbon chemistry and limits the use of SWNTs as an individual sensor.18 This problem can be overcome, to a certain degree, through surface modication of SWNT with a suitable molecular recognition system.1921 Surface modication of SWNT with a variety of materials such as polymers,21,22 metals,23 and metal oxides24 has been reported to improve the sensing performance of the
Received: October 3, 2013 Revised: December 13, 2013 Published: December 20, 2013
1602
dx.doi.org/10.1021/jp409851m | J. Phys. Chem. C 2014, 118, 16021610

The Journal of Physical Chemistry C device toward various analytes. This surface functionalization strategy to enhance sensor performance facilitates the possible development of a SWNTporphyrin device, exploiting both the binding ability of porphyrin with the analyte and the electrical properties of SWNTs. In this device, the binding ability of the porphyrins with dierent analytes contributes toward the improvement of sensor performance, and the high electrical mobility of SWNTs facilitates to make low-power chemiresistive sensor by improving the device conductance. There are reports on porphyrin-functionalized multiwalled carbon nanotube3,25 and single-walled carbon nanotube26 sensors for volatile organic compounds (VOCs). These nanotubeporphyrin devices were fabricated using solvent casting technique, where porphyrin was noncovalently attached on the carbon nanotube surface through interaction without much change in carbon nanotube electronic properties.27 However, this method results in only partial surface coverage of the nanotubes and thus incomplete realization of the full sensing potential.11 A full surface coverage of nanotubes with porphyrin can be achieved through the eletropolymerization of porphyrin. In this work, we report fabrication of SWNTporphyrin hybrid nanostructure through electrochemical polymerization of tetraphenylporphyrin (TPP) on SWNT networks and evaluate the potential of this nanostructure as a chemiresistive sensor for acetone, as a model VOC. The nanostructure was characterized using scanning electron microscopy (SEM), Raman spectroscopy, attenuated total reectance (ATR) infrared spectroscopy (IR), currentvoltage (IV), and eldeect transistor (FET) measurements. The electrochemical process optimization was done through evaluation of the sensing performance of the hybrid fabricated at dierent process conditions. The electrical characterization through eld-eect transistor (FET) measurement of the hybrids prepared at various process conditions enabled understanding the relationship between the fabrication process parameters and the electrical properties of the hybrid and provided an insight about the process optimization. Further, systematic study of sensing performance evaluations revealed a several-fold increase in sensitivity of the hybrid as compared to solvent casting functionalized SWNTs, wide dynamic range, and long-term device stability. FET measurements indicated that the sensing mechanism was governed by charge transfer or dominated by the electrostatic gating eect.

Article

EXPERIMENTAL SECTION Nanosensor Fabrication. SWNTs suspension was prepared by dispersing 0.1 mg of carboxylated SWNTs [P3 SWNT-COOH, 8090% purity from Carbon Solution Inc. (Riverside, CA)] in 10 mL of N,N-dimethylformamide (DMF) (Sigma-Aldrich, Spectral grade) by ultrasonication for 90 min. The suspension was then centrifuge at 31000g for another 90 min to separate the aggregates, and the supernatant was used for device fabrication. Sensor arrays consisting of ve interdigitated microelectrodes were fabricated on highly doped p-type silicon substrate by standard lithographic patterning. An approximately 300 nm thick dielectric layer of SiO2 was rst grown on the Si substrate by low-pressure chemical vapor deposition (LPCVD). The interdigitated electrodes (electrode nger L W: 100 m 5 m; gap between each nger: 3 m; total number of ngers: 20) were then written on the SiO2/Si substrate using photolithography, followed by the deposition of a 20 nm
1603

thick Cr layer and a 180 nm thick Au layer by e-beam evaporation, and nally the electrodes were dened using a standard lift-o processes. SWNTs were aligned across the gaps of four interdigitated Au electrodes by dielectrophoretic (DEP) alignment. A 0.2 L drop of SWNT suspension was placed on top of the interdigitated electrode, and an ac voltage of 3 VPP at a frequency of 4 MHz was applied by a function generator (Wavetek, San Diego, CA) until the desired resistance was achieved. The resistance of the device was adjusted by varying the SWNTs concentration in the suspension and/or the deposition time. After alignment, the electrode was annealed at 300 C for 90 min in a reducing environment (5% H2 in N2) to improve the contact between the Au pads and the SWNTs by removing any possible residual DMF between the Au pads and the SWNTs. Electrochemical Functionalization. Electrochemical functionalization of SWNTs with tetraphenylporphyrin polymer was conducted at room temperature using a threeelectrode electrochemical cell in which the SWNT networks on gold electrodes was the working electrode (WE), the gold electrode without the SWNT networks was the counter electrode (CE), and a silver wire in AgNO3 (Ag/AgNO3 (0.02 M) in acetonitrile) was the reference electrode (RE). Electropolymerization was done in potentiostatic mode at +2 V versus Ag/AgNO3 using a deoxygenated electrolytic solution of 0.2 M tetrabutylammonium perchlorate (TBAP) and 2 mM tetraphenylporphyrin (TPP) in dichloromethane (CH2Cl2) by passing a predetermined xed charge. After electropolymerization, the device was rinsed several times with CH2Cl2 followed by nanopure water and nally dried with nitrogen gas. A custom-made electrochemical cell (Figure S1) was used for electropolymerization. The microfabricated electrodes array chip was carefully designed to t the electrochemical cell so that the electrode area exposed to the electrolyte was constant every time a chip was mounted in the cell. During the electropolymerization process, the polymer growth occurs on the SWNT networks as well as on the gold surface exposed to the electrolyte. The charge density of electropolymerization was computed based on the gold electrode area exposed to the electrolyte as the total surface area of the SWNT networks is negligible with respect to the surface area of the gold electrode exposed to the electrolyte. Characterization. The nanostructures were characterized by scanning electron microscope (SEM), atomic force microscope (AFM), Raman spectrometer, and attenuated total reectance (ATR)-infrared (IR) spectrometer. SEM images were obtained using a Zeiss Leo SUPRA 55 with beam energy of 20 kV. AFM images were taken using a Veeco Innova AFM. Raman spectra were collected in Nicolet Almega XR dispersive Raman microscope with a 0.7 m spot size and 532 nm laser excitation. The ATR-IR spectra were taken using a Thermo Nicolet Nexus 670 FTIR instrument equipped with an ATR sampling accessory. Electrical characterizations were done using a semiconductor parameter analyzer (HP model no. 4155A) through current voltage (IV) measurements and FET measurements to conrm surface functionalization of SWNTs. FET measurements were also performed to understand the eect of the tetraphenylporphyrin polymer thickness on the electrical properties of the hybrids and to elucidate the sensing mechanism. During FET measurements, the gold electrodes served as the source and the drain while the aligned SWNTs or
dx.doi.org/10.1021/jp409851m | J. Phys. Chem. C 2014, 118, 16021610

The Journal of Physical Chemistry C

Article

Figure 1. Schematic representation of reactions involved in electrochemical polymerization of TPP at SWNTs electrode.

SWNTs-poly(TPP) hybrid acted as the conduction channel. The back gate potential was applied through the highly doped Si substrate. A dielectric layer of 300 nm thick SiO2 was used to separate the back gate from the source drain. A xed potential of 1 V was applied between the source and the drain while the gate voltage was swept from 40 to 40 V. The source-drain current (ID) was recorded at room temperature as a function of the applied gate voltage (VG). The device and the measurement apparatus were carefully grounded to reduce electronic noise. Gas Sensing Studies. An indigenous setup (Figure S2) was constructed to perform the gas sensing studies. During measurement, a bias potential of 1 V was applied across the sensor, and the resistance of the device was recorded with respect to time. For this purpose, a Keithley 2636 dual-channel system source meter was used. A multiplexer connected to the source meter was used for simultaneous measurement of the sensors. An in-house developed LabView program was used to control the process. To interface the sensor with the source meter, it was rst wire-bonded to a chip holder and then connected with the multiplexer through a breadboard. The sensor was covered with a 13 cm3 sealed glass dome with a gas inlet and outlet. Dry air (purity: 99.998%, Airgas Inc., Riverside, CA) was used as carrier gas. Saturated acetone vapor was generated by passing dry air though a bubbler lled with liquid acetone. Dierent concentrations of acetone vapor were obtained by mixing the vapor stream with dry air. A series of
1604

mass ow controllers (Alicate Scientic Inc., Tucson, AZ) were used to control the gas ow rate and to achieve the desired acetone concentration. In all the experiments, sensors were rst exposed to dry air to achieve the baseline followed by the desired concentration of acetone vapor and then back to air, which completed one cycle. Mobility and Carrier Concentration Calculations. The mobility of the hybrid device was calculated assuming the MOSFET linear region model and using eq 1
= gm L2 1 CG VD

(1)

where is the carrier mobility, gm/W is the normalized transconductance of the device (ID/VG), W is the channel width, L is the channel length (3 m), and VD is the drain voltage (1 V).28,29 The gate capacitance (CG) was computed using eq 2
CG = 0dW L Lox

(2)

where 0 is the permittivity of free space, d is the dielectric constant of SiO2, L is the channel length (3 m), and Lox is the thickness of the dielectric layer (300 nm).28 The carrier concentration (n) was computed using eq 3
dx.doi.org/10.1021/jp409851m | J. Phys. Chem. C 2014, 118, 16021610

The Journal of Physical Chemistry C

n= CG VGT eL
(3)

Article

where CG is gate capacitance and determined as above, L is channel length, e is charge of an electron, and VGT is the threshold gate voltage.

RESULTS AND DISCUSSION Verication of SWNTsPoly(TPP) Hybrid Formation. Anodic oxidation of TPP to form an electroactive lm from dichloromethane with tetrabutyl hexauorophosphate on a platinum electrode was reported previously.30 It is well-known that SWNT contains metallic and carbon based impurities, which may strongly inuence its electrochemistry,31,32 and therefore it is necessary to determine the working potential for TPP electropolymerization. Cyclic voltammetry (CV) was conducted at room temperature on spray-coated SWNTs on a gold disk electrode in a deoxygenated solution of 0.2 M tetrabutylammonium perchlorate (TBAP) and 2 mM of tetraphenylporphyrin (TPP) in dichloromethane (CH2Cl2) as electrolyte. The anodic oxidation potential for TPP electropolymerization was found to be around 2 V vs Ag/AgNO3, which is very close to the previously reported value of 1.95 V (Figure S3).30 The CV indicated that the electropolymerization of TPP on SWNTs started when the potential was 2 V (vs Ag/AgNO3). Based on this information, subsequent electropolymerizations were performed at a constant potential of 2 V (vs Ag/AgNO3) to fabricate SWNTpoly(TPP) hybrids. No such anodic oxidation peak was observed when the CV was done in a deoxygenated electrolytic solution of 0.2 M TBAP in CH2Cl2 in the absence of TPP monomer (Figure S3), which indicates no oxidation of SWNT occurred within this potential window. Schematic representation of the reactions involved in the formation of SWNTpoly(TPP) hybrid is shown in Figure 1. During the oxidative electropolymerization, the TPP monomer loses three electrons, one at a time, starting with the formation of radical cation [TPP+] followed by dication [TPP2+] and nally conversion to radical cation of TPP in the form of [(TPP2+)+].30 This reactive radical cation, [(TPP2+)+] then can bind covalently either with the side walls of SWNT via C C bond formation (SWNTaryl bond)33 or reacts with another [(TPP2+)+] through arylaryl30 couplings to produce TPP polymer. Because the radical formation happens at the electrode surface, i.e., SWNTs, it is hypothesized that the covalent attachment of porphyrin with the SWNT will take place rst followed by the additional polymerization of porphyrin on top of the covalently modied SWNT to make an extra layer of porphyrin polymer on the SWNT surface. This hypothesis was supported by SEM (Figures 2a,b) that showed the polymer coating was restricted to SWNT and uniform all over the surface. The formation of TPP polymer on SWNTs was further conrmed by attenuated total reectance (ATR)-IR spectroscopy (Figure 3). The spectrum of SWNTspoly(TPP) showed the characteristic band at about 845 cm1, which is ascribed to the out-of-plane CH waging vibration of the biphenyl group formed due to arylaryl coupling.34 Figure 4 shows a comparison of Raman spectra of SWNTs that are unfunctionalized, noncovalently functionalized with TPP by solution casting (SWNTsTPP), and poly(TPP)coated (SWNTspoly(TPP)). A sharp G (1564 cm1) and G+ (1588 cm1) peak, which is the characteristic signature of SWNTs, along with a small D-peak (1340 cm1) due to the
1605

Figure 2. SEM image of the same SWNT before (a) and after TPP polymer deposition (b). Insets in (a) and (b) show the nanotube bridging the gold electrode.

Figure 3. ATR-IR spectra of SWNTs (blue/I), SWNTsTPP (black/ II), and SWNTspoly(TPP) (red/III).

presence of carboxyl group on the sidewalls of SWNTs, were evident in the Raman spectra of bare SWNTs.35 A signicant enhancement in ID/IG ratio as compared to SWNTs and an overlapping of G and G+ peak were observed for SWNTs poly(TPP), whereas no signicant change in ID/IG ratio was noticed for noncovalently functionalized SWNTsTPP. This enhancement of ID/IG ratio is attributed to the increase in defects on SWNTs due to the covalent attachment of TPP radical to SWNTs, which creates a new sp3 carbon center in place of an sp2 carbon atom in the SWNT lattice.36,37 On the
dx.doi.org/10.1021/jp409851m | J. Phys. Chem. C 2014, 118, 16021610

The Journal of Physical Chemistry C

Article

Figure 4. Raman spectra (ex = 532 nm) of SWNTs (blue/a; ID/IG = 0.079), SWNTsTPP, solution casted (black/b; ID/IG = 0.083), and SWNTspoly(TPP) (red/c; ID/IG = 0.769).

other hand, the Raman spectra of SWNTs and noncovalently modied SWNTsTPP look alike without any observable change in the ID/IG ratio, which is in agreement with the literature.38 The Raman spectroscopy ndings of the poly(TPP) modifying SWNTs covalently were supported by the observed increase in resistance (few hundred-fold) (Figure 5a) and Ion/Ioff ratio (3-fold) (Figure 5b), attributed to the formation of sp3 carbon centers in SWNTs-poly(TPP), thereby reducing the -electron delocalization of sp2 bonded carbon networks of pristine SWNTs. Electrochemical Process Optimization and Characterization of Hybrid Devices. SWNTspoly(TPP) hybrids were fabricated by applying dierent charge densities to achieve dierent TPP polymer thicknesses on SWNT networks, and the device performance was evaluated on the basis of sensing performance for acetone vapors. Figure S4 shows a typical dynamic response of the hybrid devices prepared at dierent charge densities in terms of normalized resistance change [(R/R0 (%) = 100(R R0)/R0), where R is the resistance of the hybrid exposed to analyte and R0 is the initial baseline resistance before analyte exposure]. In each case, the sensor responded promptly on exposure to acetone reaching a steady state/plateau and reverted toward baseline when the analyte feed was removed. Figure 6 is a plot of the steady state normalized resistance change (from Figure S4) as a function of the charge density applied during electropolymerization. The results showed that for all concentrations of acetone vapor the device response initially increased with increasing charge density, reaching a maximum at 19.56 mC/cm2, and then decreased. This response trend is attributed to an initial increase in the number of porphyrin molecules available for interaction/binding with analyte followed by the polymer coating becoming less conducting due to lower porphyrin conductivity, hindering the charge transfer process. In order to verify this fact, the eect of charge density used for poly(porphyrin) synthesis on normalized change in carrier concentration (n/n0) and normalized change in mobility (/0) were investigated. The n/n0 and /0 values were determined from the FET characteristics curves of each device when exposed to dry air and saturated acetone vapors. A typical transfer characteristic (IDVG) curve of the hybrid device prepared at a charge density of 23.59 mC/cm2 upon exposure to air and saturated acetone vapor is illustrated in Figure S5. Since the device resistance varied widely depending on the charge density applied during the fabrication, transfer character1606

Figure 5. (a) IDVD curve showing the resistance of SWNTs before (blue/(I); 1 k) and after functionalization with poly(TPP) (red/(II); 350 k); (inset) magnied IDVD curve of SWNTspoly(TPP) hybrid. (b) FET curves showing change in Ion/Ioff ratio before (blue/ (I); Ion/Ioff = 1.5) and after (red/(II); Ion/Ioff = 4.7) electrochemical functionalization of SWNTs with poly(TPP).

Figure 6. Sensor responses (R/R0) to acetone vapor exposure at 12.5 (black square), 25 (red circle), and 50% (blue triangle) of saturation versus charge density applied during constant potential electropolymerization (at 2 V) of tetraphenyl porphyrin (TPP) on SWNT networks.

istics of hybrid devices prepared at dierent charge densities were computed in terms of normalized drain current
dx.doi.org/10.1021/jp409851m | J. Phys. Chem. C 2014, 118, 16021610

The Journal of Physical Chemistry C (ID,normalized) from the individual device transfer characteristics and plotted as a function of gate voltage (VG) (Figure S6). The current (ID) measured at 40 V gate potential (VG) was used as the basis for normalizing the source-drain current in the transfer characteristics curves. Figure 7 shows graphs for n/n0 and /0 as a function of charge density. The results showed that there was no

Article

observed negative shift of the threshold gate voltage and correspondingly the carrier concentration without the change in transconductance and the mobility of the FET of SWNTs poly(TPP) hybrid for exposure to acetone compared to dry air indicate that the sensing mechanism is dominated by electrostatic gating eects.39 Structural characterization of the hybrid device fabricated at optimized charge density was done by atomic force microscopy (AFM). AFM image (before and after functionalization with TPP polymer) along with the line scan conrmed an almost uniform coating on SWNT networks and an increased height for each SWNT after coating (Figure S7). The histogram of height distribution as a function of diameter (Figure 8) revealed an increase the average height/diameter of bare SWNTs from 3 to 7 nm after electropolymerization of TPP on SWNT networks.

Figure 8. Histogram of height distribution as a function of diameter/ height of bare and poly(TPP)-coated SWNTs (fabricated at 19.65 mC/cm2 charge density) obtained from AFM analysis. Gaussian tting of the height distribution provides the average diameter of the nanotubes. Figure 7. Normalized change in (a) mobility(/0)and (b) carrier concentration (n/n0) versus charge density applied during electropolymerization. (n/n0) = [(nsaturated vapor of acetone n0)/n0], where n0 is the carrier concentration of the device in air and (/0) = [(saturated vapor of acetone 0)/0], where 0 is the mobility of the device in air.

signicant change in /0 (Figure 7a) with increasing charge density. On the other hand, n/n0 (Figure 7b) rst increased, reaching a maximum at charge density of 19.65 mC/cm2 followed by a decrease, a trend similar to R/R0 versus charge density plot (Figure 6). This behavior conrmed that porphyrin polymer thickness made the hybrid more resistive and restricted the charge transfer process during the interaction with the analyte, thus resulting in the decrease in carrier concentration and the sensing performance of the hybrid when prepared using greater than the optimum charge density (19.65 mC/cm2). The above transfer characteristics curves of SWNTs poly(TPP) transistor also provide an insight into the mechanism of acetone sensing. Acetone is an electron donor; hence, exposure of the p-type SWNTspoly(TPP) (Figure 4b) to acetone will cause a shift in the valence band away from the Fermi level, resulting in a decrease in carrier (hole) concentration (n) and hence a reduction in conductivity and negative shift in threshold voltage (VTH). The experimentally
1607

Evaluation of Sensing Performance of Hybrid Devices. The sensing performance of SWNTpoly(TPP) hybrid networks prepared at optimized charge density was evaluated by exposing to varying acetone concentrations and compared to bare SWNTs and SWNTs modied with TPP by solvent casting (SWNTsTPP) sensors. A higher acetone concentration (12.5, 25, and 50% of saturation) was used for comparison, as the bare SWNT networks sensors were not sensitive to low acetone concentration. As shown in Figures 9a,b, the SWNTspoly(TPP) sensor outperformed the bare SNWTs and SWNTsTPP sensors in sensitivity, response time, and recovery time. The sensitivity of SWNTspoly(TPP) was 4.5-fold higher than bare SWNTs and 3-fold greater than SWNTsTPP (Figure 9b). The average response times (time required to achieve 90% of the maximum response) of bare SWNTs, SWNTsTPP, and SWNTspoly(TPP) devices were 12, 10, and 8 min, respectively (Figure 9a). It is noteworthy to mention that the average time to achieve 50% of the maximum response for SWNTspoly(TPP) hybrid is less than 50 s compared to about a few minutes for bare SWNTs and SWNTsTPP sensors for all acetone concentrations. Similarly, the SWNTspoly(TPP) hybrid sensor recovered faster compared to SWNTs and SWNTsTPP devices. Because of the high sensitivity exhibited by SWNTs poly(TPP) sensor for the higher acetone concentration, the performance of this sensor at lower concentrations of acetone
dx.doi.org/10.1021/jp409851m | J. Phys. Chem. C 2014, 118, 16021610

The Journal of Physical Chemistry C

Article

Figure 9. (a) Real-time responses (R/R0) of nanosensors to acetone vapors at concentrations varying from 12.5 to 50% of saturation: (I) bare SWNTs, (II) SWNTsTPP, and (III) SWNTspoly(TPP). (b) Calibration plots of sensors: () SWNTs, () SWNTs-TPP, () SWNTs-poly(TPP).

Figure 10. (a) Real-time responses (R/R0) of SWNTspoly(TPP) hybrid (fabricated at optimum charge density) to acetone vapors varying from 50 to 2000 ppm and (b) corresponding calibration curve.

d S=

( ) = R
R R0

dc

k R 0 max (k + c)2

(5)

was evaluated to estimate its full potential. Sensing experiments were performed from 50 to 2000 ppm acetone, corresponding to 5% to 200% of the OSHA (Occupational Safety and Health Administration) PEL (Permissible Exposure Limits) of 1000 ppm (Figure 7). Similar to higher concentration, the sensor exhibited a rapid increase in resistance upon each exposure of acetone concentration and a slow and partial recovery upon removal of the analyte (Figure 10a). The latter can be speeded by raising the temperature to increase the desorption rate. The relationship between the response (normalized resistance change) and acetone concentration (Figure 10b) exhibited the Langmuir-like adsorption isotherm (R2 value of 0.96), where the relationship between the normalized resistance change (R/R0) and the concentration (c) is governed by eq 4
R R c = R0 R 0 max k + c

was calculated to be 0.22% per ppm of acetone vapor. The limit of detection (LOD) was determined from the linear range of calibration data by the equation LOD = 3SD/S, where SD is the standard deviation of the sensor response in dry air, was estimated to be 9 ppm. The long-term storage stability of the SWNTspoly(TPP) sensor was investigated by evaluating the response of the same sensor over a period of 6 months to 50% of saturation acetone vapor and storing in dark conditions under vacuum. The results showed the sensor response declined slightly from the original value of 140% 13 to 118% 5 in the initial 60 days and then remained stable for the next 120 days (Figure S8).

(4)

where (R/R0)max is the maximum change in normalized resistance occurring in the saturation regime, and k is the anity constant. The constants (R/R0)max and k values were found to be 10.62 and 49.15, respectively. The sensitivity (S) of the sensor, computed by using eq 5 at low concentrations (c 0)25
1608

CONCLUSIONS In summary, we have fabricated single-walled carbon nanotubepoly(porphyrin) hybrid through electrochemical route and evaluated the sensing potential. The electrochemical polymerization process is facile, exible, and versatile. It is exible because the porphyrin polymer coating thickness on SWNTs can be varied by modulating of charge density. It is versatile and can be applied to functionalize SWNTs with dierent metalloporphyrins to build sensor arrays.30 Increasing charge density during electropolymerization provided an increased polymer thickness making more amount of porphyrin available to interact with the analyte, which enhanced the
dx.doi.org/10.1021/jp409851m | J. Phys. Chem. C 2014, 118, 16021610

The Journal of Physical Chemistry C sensitivity. However, there was an optimum polymer coating thickness above which the sensor sensitivity decreased. FET analysis revealed that the properties of the hybrids prepared at more than the optimum charge density were dominated by the higher resistance of the hybrid, which restricted the charge transfer process during interaction with the analyte and caused a decrease in charge carrier density that resulted in a decrease in the sensitivity. Further, AFM analysis revealed that the TPP polymer formed a few nanometer thick uniform coating on the SWNTs. Results indicated that the hybrid device fabricated at the optimum condition was about 4-fold more sensitive toward the model VOC acetone as compared to bare and TPP-functionalized SWNTs devices. The sensitivity of sensors, calculated from the sensing responses for the low acetone concentration range of Langmuir like isotherm plot spanning the dynamic range of 50 to 230 000 ppm, was 0.25% per ppm of acetone vapor. The limit of detection was calculated to be 9 ppm acetone. Further, the FET analysis in air and various concentrations of acetone vapor revealed that the sensing mechanism of SWNTspoly(TPP) device was governed by electrostatic gating eects. The hybrid also provided a fairly stable sensing performance until 180 days. The ability of analyte detection well below OSHA PEL level, wide dynamic range, long-term stability, scalable fabrication techniques, and availability of various types of synthetically derived porphyrins, which can provide dierent sensitivity toward various VOCs, make SWNTspoly(porphyrin) hybrid a potential material for further array based sensors development.

Article

S Supporting Information *

ASSOCIATED CONTENT

Schematic representation of the custom-made electrochemical cell, ow diagram of the gas sensing setup, cyclic voltammetry of tetraphenylporphyrin, real-time nanosensor relative responses of the devices prepared at various charge densities to acetone vapor exposure, transfer characteristics of SWNTs poly(TPP) hybrid in the presence of air and saturated vapor of acetone, AFM images of before and after electrodeposition of TPP on SWNTs with line scan, calibration plot of the SWNTspoly(TPP) devices to 50% saturated acetone vapor exposure at various days varying from 0 to 180 days. This material is available free of charge via the Internet at http:// pubs.acs.org.

AUTHOR INFORMATION

Corresponding Author Notes

*E-mail adani@engr.ucr.edu (A.M.). The authors declare no competing nancial interest.

ACKNOWLEDGMENTS T.S. is grateful to Government of India for his nancial assistance. S.S. acknowledges the Thai Government (Ministry of Science) for his nancial support. REFERENCES
(1) Biesaga, M.; Pyrzynska, K.; Trojanowicz, M. Porphyrins in Analytical Chemistry. A review. Talanta 2000, 51, 209224. (2) Di Natale, C.; Monti, D.; Paolesse, R. Chemical Sensitivity of Porphyrin Assemblies. Mater. Today 2010, 13, 4652. (3) Penza, M.; Rossi, R.; Alvisi, M.; Signore, M. A.; Serra, E.; Paolesse, R.; DAmico, A.; Di Natale, C. Metalloporphyrins-Modified
1609

Carbon Nanotubes Networked Films-Based Chemical Sensors for Enhanced Gas Sensitivity. Sens. Actuators, B 2010, 144, 387394. (4) Lim, S. H.; Feng, L.; Kemling, J. W.; Musto, C. J.; Suslick, K. S. An Optoelectronic Nose for the Detection of Toxic Gases. Nat. Chem. 2009, 1, 562567. (5) Dunbar, A. D. F.; Richardson, T. H.; McNaughton, A. J.; Hutchinson, J.; Hunter, C. A. Investigation of Free Base, Mg, Sn, and Zn Substituted Porphyrin LB Films as Gas Sensors for Organic Analytes. J. Phys. Chem. B 2006, 110, 1664616651. (6) Brittle, S.; Richardson, T. H.; Dunbar, A. D. F.; Turega, S.; Hunter, C. A. Alkylamine Sensing Using Langmuir-Blodgett Films of n-Alkyl-N-phenylamide-Substituted Zinc Porphyrins. J. Phys. Chem. B 2008, 112, 1127811283. (7) Brunink, J. A. J.; Di Natale, C.; Bungaro, F.; Davide, F. A. M.; DAmico, A.; Paolesse, R.; Boschi, T.; Faccio, M.; Ferri, G. The Application of Metalloporphyrins as Coating Material for Quartz Microbalance-Based Chemical Sensors. Anal. Chim. Acta 1996, 325, 5364. (8) Di Natale, C.; Buchholt, K.; Martinelli, E.; Paolesse, R.; Pomarico, G.; DAmico, A.; Lundstrom, I.; Lloyd Spetz, A. Investigation of Quartz Microbalance and ChemFET Transduction of Molecular Recognition Events in a Metalloporphyrin Film. Sens. Actuators, B 2009, 135, 560567. (9) Di Natale, C.; Paolesse, R.; Macagnano, A.; Mantini, A.; Goletti, C.; DAmico, A. Characterization and Design of Porphyrins-Based Broad Selectivity Chemical Sensors for Electronic Nose Applications. Sens. Actuators, B 1998, 52, 162168. (10) Andersson, M.; Holmberg, M.; Lundstrom, I.; Lloyd-Spetz, A.; Martensson, P.; Paolesse, R.; Falconi, C.; Proietti, E.; Di Natale, C.; DAmico, A. Development of a ChemFET Sensor with Molecular Films of Porphyrins as Sensitive Layer. Sens. Actuators, B 2001, 77, 567571. (11) Tanaka, H.; Yajima, T.; Matsumoto, T.; Otsuka, Y.; Ogawa, T. Porphyrin Molecular Nanodevices Wired Using Single-Walled Carbon Nanotubes. Adv. Mater. 2006, 18, 14111415. (12) Bachilo, S. M.; Strano, M. S.; Kittrell, C.; Hauge, R. H.; Smalley, R. E.; Weisman, R. B. Structure-Assigned Optical Spectra of SingleWalled Carbon Nanotubes. Science 2002, 298, 23612366. (13) Baughman, R. H.; Zakhidov, A. A.; de Heer, W. A. Carbon Nanotubes-the Route Toward Applications. Science 2002, 297, 787 792. (14) Wong, E. W.; Sheehan, P. E.; Lieber, C. M. Nanobeam Mechanics: Elasticity, Strength, and Toughness of Nanorods and Nanotubes. Science 1997, 277, 19711975. (15) Kong, J.; Franklin, N. R.; Zhou, C.; Chapline, M. G.; Peng, S.; Cho, K.; Dai, H. Nanotube Molecular Wires as Chemical Sensors. Science 2000, 287, 622625. (16) Collins, P. G.; Bradley, K.; Ishigami, M.; Zettl, A. Extreme Oxygen Sensitivity of Electronic Properties of Carbon Nanotubes. Science 2000, 287, 18011804. (17) Snow, E. S.; Perkins, F. K.; Houser, E. J.; Badescu, S. C.; Reinecke, T. L. Chemical Detection with a Single-Walled Carbon Nanotube Capacitor. Science 2005, 307, 19421945. (18) Zhang, T.; Mubeen, S.; Myung, N. V.; Deshusses, M. A. Recent Progress in Carbon Nanotube-Based Gas Sensors. Nanotechnology 2008, 19, 332001. (19) Kong, J.; Chapline, M. G.; Dai, H. Functionalized Carbon Nanotubes for Molecular Hydrogen Sensors. Adv. Mater. 2001, 13, 13841386. (20) Zhang, T.; Nix, M. A.; Yoo, B. A.; Deshusses, M. A.; Myung, N. Electrochemically Functionalized Single-Walled Carbon Nanotube based Gas Sensor. Electroanalysis 2006, 18, 11531158. (21) Zhang, T.; Mubeen, S.; Bekyarova, E.; Yoo, B. Y.; Haddon, R. C.; Myung, N. V.; Deshusses, M. A. Poly(m-aminobenzene sulfonic acid) Functionalized Single-Walled Carbon Nanotubes based Gas Sensor. Nanotechnology 2007, 18, 165504. (22) Ding, M.; Tang, Y.; Gou, P.; Reber, M. J.; Star, A. Chemical Sensing with Polyaniline Coated Single-Walled Carbon Nanotubes. Adv. Mater. 2011, 23, 536540.
dx.doi.org/10.1021/jp409851m | J. Phys. Chem. C 2014, 118, 16021610

The Journal of Physical Chemistry C

(23) Penza, M.; Rossi, R.; Alvisi, M.; Serra, E. Metal-Modified and Vertically Aligned Carbon Nanotube Sensors Array for Landfill Gas Monitoring Applications. Nanotechnology 2010, 21, 105501. (24) Wei, B.-Y.; Hsu, M.-C.; Su, P.-G.; Lin, H.-M.; Wu, R.-J.; Lai, H.J. A Novel SnO2 Gas Sensor Doped with Carbon Nanotubes Operating at Room Temperature. Sens. Actuators, B 2004, 101, 8189. (25) Penza, M.; Alvisi, M.; Rossi, R.; Serra, E.; Paolesse, R.; DAmico, A.; Natale, C. D. Carbon Nanotube Films as a Platform to Transduce Molecular Recognition Events in Metalloporphyrins. Nanotechnology 2011, 22, 125502. (26) Shirsat, M. D.; Sarkar, T.; Kakoullis, J.; Myung, N. V.; Konnanath, B.; Spanias, A.; Mulchandani, A. Porphyrin-Functionalized Single-Walled Carbon Nanotube Chemiresistive Sensor Arrays for VOCs. J. Phys. Chem. C 2012, 116, 384550. (27) Rahman, G. M. A.; Guldi, D. M.; Campidelli, S.; Prato, M. Electronically Interacting Single Wall Carbon NanotubePorphyrin Nanohybrids. J. Mater. Chem. 2006, 16, 6265. (28) Snow, E. S.; Campbell, P. M.; Ancona, M. G.; Novak, J. P. HighMobility Carbon-Nanotube Thin-Film Transistors on a Polymeric Substrate. Appl. Phys. Lett. 2005, 86, 033105. (29) Martel, R.; Schmidt, T.; Shea, H. R.; Hertel, T.; Avouris, P. Single- and Multi-Wall Carbon Nanotube Field-Effect Transistors. Appl. Phys. Lett. 1998, 73, 24472449. (30) Paul-Roth, C.; Rault-Berthelot, J.; Simonneaux, G.; Poriel, C.; Abdalilah, M.; Letessier, J. Electroactive Films of Poly(Tetraphenylporphyrins) with Reduced Bandgap. J. Electroanal. Chem. 2006, 597, 1927. (31) Pumera, M. Volatmmetry of Carbon Nanotubes and Graphene: Excitement, Disappointment, and Reality. Chem. Rec. 2012, 12, 201 213. (32) Pumera, M.; Ambrosi, A.; Chng, E. L. K. Impurities in Graphene and Carbon Nanotubes and Their Influence on the Redox Properties. Chem. Sci. 2012, 3, 33473355. (33) Strano, M. S.; Dyke, C. A.; Usrey, M. L.; Barone, P. W.; Allen, M. J.; Shan, H.; Kittrell, C.; Hauge, R. H.; Tour, J. M.; Smalley, R. E. Electronic Structure Control of Single-Walled Carbon Nanotube Functionalization. Science 2003, 301, 15191522. (34) Halls, M. D.; Tripp, C. P.; Schlegel, H. B. Structure and Infrared (IR) Assignments for the OLED Material: N,N-diphenyl-N,N-bis(1naphthyl)-1,1-biphenyl-4,4-diamine (NPB). Phys. Chem. Chem. Phys. 2001, 3, 21312136. (35) Jorio, A.; Pimenta, M. A.; Filho, A. G. S.; Saito, R.; Dresselhaus, G.; Dresselhaus, M. S. Characterizing Carbon Nanotube Samples with Resonance Raman Scattering. New J. Phys. 2003, 5, 139.1139.17. (36) Bekyarova, E.; Itkis, M. E.; Ramesh, P.; Berger, C.; Sprinkle, M.; de Heer, W. A.; Haddon, R. C. Chemical Modification of Epitaxial Graphene: Spontaneous Grafting of Aryl Groups. J. Am. Chem. Soc. 2009, 131, 13361337. (37) Sarkar, S.; Bekyarova, E.; Haddon, R. C. Reversible Grafting of -Naphthylmethyl Radicals to Epitaxial Graphene. Angew. Chem. Int. Ed. 2012, 51, 49014904. (38) Satake, A.; Miyajima, Y.; Kobuke, Y. PorphyrinCarbon Nanotube Composites Formed by Noncovalent Polymer Wrapping. Chem. Mater. 2005, 17, 716724. (39) Heller, I.; Janssens, A. M.; Mannik, J.; Minot, E. D.; Lemay, S. G.; Dekker, C. Identifying the Mechanism of Biosensing with Carbon Nanotube Transistors. Nano Lett. 2008, 8, 591595.

Article

1610

dx.doi.org/10.1021/jp409851m | J. Phys. Chem. C 2014, 118, 16021610